From 29ef9455578c37f5581609171bb1abc7d0410162 Mon Sep 17 00:00:00 2001 From: unknown <jcole@tetra.spaceapes.com> Date: Wed, 25 Jul 2001 15:47:31 -0500 Subject: [PATCH] New Chapter 1 a la DocTOC. Sorry for such a big patch. :) --- Docs/manual.texi | 75593 +++++++++++++++++++++++---------------------- 1 file changed, 37837 insertions(+), 37756 deletions(-) diff --git a/Docs/manual.texi b/Docs/manual.texi index bb158c60f9..8b5052aa19 100644 --- a/Docs/manual.texi +++ b/Docs/manual.texi @@ -111,10 +111,7 @@ distribution for that version. @menu * Introduction:: General Information about @strong{MySQL} -* Questions:: @strong{MySQL} mailing lists and how to ask questions or report errors (bugs) -* Licensing and Support:: @strong{MySQL} licensing and support * Installing:: Installing @strong{MySQL} -* Compatibility:: How standards-compatible is @strong{MySQL}? * Privilege system:: The @strong{MySQL} access privilege system * Reference:: @strong{MySQL} language reference * Table types:: @strong{MySQL} table types @@ -134,7 +131,6 @@ distribution for that version. * Common problems:: Solving some common problems with @strong{MySQL} * Log files:: * Clients:: @strong{MySQL} client tools and APIs -* Comparisons:: How does @strong{MySQL} compare with other databases? * MySQL internals:: @strong{MySQL} internals * Environment variables:: @strong{MySQL} environment variables * Users:: Some @strong{MySQL} users @@ -142,8 +138,6 @@ distribution for that version. * Contrib:: Contributed programs * Credits:: Contributors to @strong{MySQL} * News:: @strong{MySQL} change history -* Bugs:: Known errors and design deficiencies in @strong{MySQL} -* TODO:: List of things we want to add to @strong{MySQL} in the future (The TODO) * Porting:: Comments on porting to other systems * Regexp:: Description of @strong{MySQL} regular expression syntax * Unireg:: What is Unireg? @@ -157,20 +151,31 @@ distribution for that version. General Information About MySQL +* MySQL and MySQL AB:: +* MySQL Information Sources:: +* Licensing and Support:: +* Compatibility:: +* Comparisons:: +* TODO:: + +MySQL, MySQL AB, and Open Source + * What-is:: What is @strong{MySQL}? -* What is MySQL AB:: +* What is MySQL AB:: What is @strong{MySQL AB}? * Manual-info:: About this manual +* Manual conventions:: Conventions used in this manual * History:: History of @strong{MySQL} -* MySQL-Books:: Books about MySQL * Features:: The main features of @strong{MySQL} * Stability:: How stable is @strong{MySQL}? +* Table size:: * Year 2000 compliance:: Year 2000 compliance -* General-SQL:: General SQL information and tutorials -* Useful Links:: Useful @strong{MySQL}-related links -About This Manual +MySQL Information Sources -* Manual conventions:: Conventions used in this manual +* MySQL-Books:: +* General-SQL:: +* Useful Links:: +* Questions:: MySQL Mailing Lists @@ -211,6 +216,61 @@ Types of Commercial Support * Telephone support:: Telephone support * Table handler support:: Support for other table handlers +How Standards-compatible Is MySQL? + +* Extensions to ANSI:: @strong{MySQL} extensions to ANSI SQL92 +* Differences from ANSI:: @strong{MySQL} differences compared to ANSI SQL92 +* ANSI mode:: Running @strong{MySQL} in ANSI mode +* Missing functions:: Functionality missing from @strong{MySQL} +* Standards:: What standards does @strong{MySQL} follow? +* Commit-rollback:: How to cope without @code{COMMIT}-@code{ROLLBACK} +* Bugs:: + +Functionality Missing from MySQL + +* Missing Sub-selects:: Sub-selects +* Missing SELECT INTO TABLE:: @code{SELECT INTO TABLE} +* Missing Transactions:: Transactions +* Missing Triggers:: Triggers +* Missing Foreign Keys:: Foreign Keys +* Broken Foreign KEY:: +* Missing Views:: Views +* Missing comments:: @samp{--} as the start of a comment + +Foreign Keys + +* Broken Foreign KEY:: Reasons NOT to use foreign keys constraints + +How MySQL Compares to Other Databases + +* Compare mSQL:: How @strong{MySQL} compares to @code{mSQL} +* Protocol differences:: +* Compare PostgreSQL:: How @strong{MySQL} compares with PostgreSQL +* MySQL-PostgreSQL features:: + +How MySQL Compares to @code{mSQL} + +* Using mSQL tools:: How to convert @code{mSQL} tools for @strong{MySQL} + +How @code{mSQL} and MySQL Client/Server Communications Protocols Differ + +* Syntax differences:: + +How MySQL Compares to PostgreSQL + +* MySQL-PostgreSQL goals:: MySQL and PostgreSQL development strategies + +Featurevise Comparison of MySQL and PostgreSQL + +* MySQL-PostgreSQL benchmarks:: + +MySQL and the future (The TODO) + +* TODO MySQL 4.0:: Things that should be in Version 4.0 +* TODO future:: Things that must be done in the near future +* TODO sometime:: Things that have to be done sometime +* TODO unplanned:: Some things we don't have any plans to do + Installing MySQL * Getting MySQL:: How to get @strong{MySQL} @@ -331,29 +391,6 @@ Upgrading/Downgrading MySQL * Upgrading-from-3.20:: Upgrading from a 3.20 version to 3.21 * Upgrading-to-arch:: Upgrading to another architecture -How Standards-compatible Is MySQL? - -* Extensions to ANSI:: @strong{MySQL} extensions to ANSI SQL92 -* ANSI mode:: Running @strong{MySQL} in ANSI mode -* Differences from ANSI:: @strong{MySQL} differences compared to ANSI SQL92 -* Missing functions:: Functionality missing from @strong{MySQL} -* Standards:: What standards does @strong{MySQL} follow? -* Commit-rollback:: How to cope without @code{COMMIT}-@code{ROLLBACK} - -Functionality Missing from MySQL - -* Missing Sub-selects:: Sub-selects -* Missing SELECT INTO TABLE:: @code{SELECT INTO TABLE} -* Missing Transactions:: Transactions -* Missing Triggers:: Triggers -* Missing Foreign Keys:: Foreign Keys -* Missing Views:: Views -* Missing comments:: @samp{--} as the start of a comment - -Foreign Keys - -* Broken Foreign KEY:: Reasons NOT to use foreign keys constraints - The MySQL Access Privilege System * General security:: General security @@ -619,7 +656,6 @@ Queries from Twin Project MySQL Server Functions * Languages:: What languages are supported by @strong{MySQL}? -* Table size:: How big @strong{MySQL} tables can be What Languages Are Supported by MySQL? @@ -919,23 +955,6 @@ MySQL PHP API * PHP problems:: Common problems with MySQL and PHP -How MySQL Compares to Other Databases - -* Compare mSQL:: How @strong{MySQL} compares to @code{mSQL} -* Compare PostgreSQL:: How @strong{MySQL} compares with PostgreSQL - -How MySQL Compares to @code{mSQL} - -* Using mSQL tools:: How to convert @code{mSQL} tools for @strong{MySQL} -* Protocol differences:: How @code{mSQL} and @strong{MySQL} client/server communications protocols differ -* Syntax differences:: How @code{mSQL} 2.0 SQL syntax differs from @strong{MySQL} - -How MySQL Compares to PostgreSQL - -* MySQL-PostgreSQL goals:: MySQL and PostgreSQL development strategies -* MySQL-PostgreSQL features:: Featurevise Comparison of MySQL and PostgreSQL -* MySQL-PostgreSQL benchmarks:: Benchmarking MySQL and PostgreSQL - MySQL Internals * MySQL threads:: MySQL threads @@ -1111,13 +1130,6 @@ Changes in release 3.19.x * News-3.19.4:: Changes in release 3.19.4 * News-3.19.3:: Changes in release 3.19.3 -MySQL and the future (The TODO) - -* TODO MySQL 4.0:: Things that should be in Version 4.0 -* TODO future:: Things that must be done in the near future -* TODO sometime:: Things that have to be done sometime -* TODO unplanned:: Some things we don't have any plans to do - Comments on porting to other systems * Debugging server:: Debugging a @strong{MySQL} server @@ -1144,22 +1156,9 @@ Debugging a MySQL server @cindex online location of manual @cindex manual, online location -@node Introduction, Questions, Top, Top +@node Introduction, Installing, Top, Top @chapter General Information About MySQL -@menu -* What-is:: What is @strong{MySQL}? -* What is MySQL AB:: -* Manual-info:: About this manual -* History:: History of @strong{MySQL} -* MySQL-Books:: Books about MySQL -* Features:: The main features of @strong{MySQL} -* Stability:: How stable is @strong{MySQL}? -* Year 2000 compliance:: Year 2000 compliance -* General-SQL:: General SQL information and tutorials -* Useful Links:: Useful @strong{MySQL}-related links -@end menu - This is the @strong{MySQL} reference manual; it documents @strong{MySQL} Version @value{mysql_version}. As @strong{MySQL} is work in progress, the manual gets updated frequently. There is a very good chance that @@ -1246,10 +1245,36 @@ or relational database concepts. If you want general information about SQL, see @ref{General-SQL}. For books that focus more specifically on @strong{MySQL}, see @ref{MySQL-Books}. + +@menu +* MySQL and MySQL AB:: +* MySQL Information Sources:: +* Licensing and Support:: +* Compatibility:: +* Comparisons:: +* TODO:: +@end menu + +@node MySQL and MySQL AB, MySQL Information Sources, Introduction, Introduction +@section MySQL, MySQL AB, and Open Source + +@menu +* What-is:: What is @strong{MySQL}? +* What is MySQL AB:: What is @strong{MySQL AB}? +* Manual-info:: About this manual +* Manual conventions:: Conventions used in this manual +* History:: History of @strong{MySQL} +* Features:: The main features of @strong{MySQL} +* Stability:: How stable is @strong{MySQL}? +* Table size:: +* Year 2000 compliance:: Year 2000 compliance +@end menu + +@node What-is, What is MySQL AB, MySQL and MySQL AB, MySQL and MySQL AB +@subsection What Is MySQL + @cindex MySQL, defined @cindex MySQL, introduction -@node What-is, What is MySQL AB, Introduction, Introduction -@section What Is MySQL @strong{MySQL}, the most popular Open Source SQL database, is provided by @strong{MySQL AB}. @strong{MySQL AB} is a commercial company that @@ -1335,9 +1360,10 @@ language already supports @strong{MySQL}. The official way to pronounce @strong{MySQL} is ``My Ess Que Ell'' (not MY-SEQUEL). But we try to avoid correcting people who say MY-SEQUEL. +@node What is MySQL AB, Manual-info, What-is, MySQL and MySQL AB +@subsection What Is MySQL AB + @cindex MySQL AB, defined -@node What is MySQL AB, Manual-info, What-is, Introduction -@section What Is MySQL AB @strong{MySQL AB} is the Swedish company owned and run by the @strong{MySQL} founders and main developers. We are dedicated to developing @strong{MySQL} @@ -1446,12 +1472,8 @@ Are a virtual company, networking with others. Work against software patents. @end itemize -@node Manual-info, History, What is MySQL AB, Introduction -@section About This Manual - -@menu -* Manual conventions:: Conventions used in this manual -@end menu +@node Manual-info, Manual conventions, What is MySQL AB, MySQL and MySQL AB +@subsection About This Manual This manual is currently available in Texinfo, plain text, Info, HTML, PostScript, and PDF versions. The primary document is the Texinfo file. @@ -1467,7 +1489,7 @@ This manual is written and maintained by David Axmark, Michael (Monty) Widenius, Jeremy Cole, and Paul DuBois. For other contributors, see @ref{Credits}. -@node Manual conventions, , Manual-info, Manual-info +@node Manual conventions, History, Manual-info, MySQL and MySQL AB @subsection Conventions Used in This Manual This manual uses certain typographical conventions: @@ -1591,12 +1613,13 @@ alternatives are listed within braces (@samp{@{} and @samp{@}}): @{DESCRIBE | DESC@} tbl_name @{col_name | wild@} @end example +@node History, Features, Manual conventions, MySQL and MySQL AB +@subsection History of MySQL + @cindex MySQL history @cindex history of MySQL @cindex MySQL name @cindex My, derivation -@node History, MySQL-Books, Manual-info, Introduction -@section History of MySQL We once started out with the intention of using @code{mSQL} to connect to our tables using our own fast low-level (ISAM) routines. However, after some @@ -1611,363 +1634,9 @@ directory and a large number of our libraries and tools have had the prefix is also named My. Which of the two gave its name to @strong{MySQL} is still a mystery, even for us. -@node MySQL-Books, Features, History, Introduction -@section Books About MySQL - -@cindex books, about MySQL -@cindex manuals, about MySQL - -While this manual is still the right place for up to date technical -information, its primary goal is to contain everything there is to know -about @strong{MySQL}. It is sometimes nice to have a bound book to read -in bed or while you travel. Here is a list of books about @strong{MySQL} and -related subjects (in English). - -By purchasing a book through these hyperlinks provided herein, you are -contributing to the development of @strong{MySQL}. - -@emph{MySQL} -@multitable @columnfractions .3 .7 -@item Available @tab @uref{http://service.bfast.com/bfast/click?bfmid=2181&sourceid=34233559&bfpid=0735709211&bfmtype=book, Barnes and Noble} -@item Publisher @tab New Riders -@item Author @tab Paul DuBois -@item Pub Date @tab 1st Edition December 1999 -@item ISBN @tab 0735709211 -@item Pages @tab 800 -@item Price @tab $49.99 US -@item Downloadable examples @tab - @uref{http://www.kitebird.com/mysql-book/, @code{samp_db} distribution} -@item Errata @tab -@uref{http://www.kitebird.com/mysql-book/errata.html, are available here} -@end multitable - -Foreword by Michael ``Monty'' Widenius, @strong{MySQL} Moderator. -@* - -In @emph{MySQL}, Paul DuBois provides you with a comprehensive guide to -one of the most popular relational database systems. Paul has -contributed to the online documentation for @strong{MySQL} and is an -active member of the @strong{MySQL} community. The principal @strong{MySQL} -developer, Monty Widenius, and a network of his fellow developers -reviewed the manuscript, and provided Paul with the kind of insight -no one else could supply. -@* - -Instead of merely giving you a general overview of @strong{MySQL}, Paul -teaches you how to make the most of its capabilities. Through two -sample database applications that run throughout the book, he -gives you solutions to problems you're sure to face. He helps you -integrate @strong{MySQL} efficiently with third-party tools, such as PHP -and Perl, enabling you to generate dynamic Web pages through -database queries. He teaches you to write programs that access -@strong{MySQL} databases, and also provides a comprehensive set of -references to column types, operators, functions, SQL syntax, -@strong{MySQL} programming, C API, Perl @code{DBI}, and PHP API. -@emph{MySQL} simply gives you the kind of information you won't find -anywhere else. -@* - -If you use @strong{MySQL}, this book provides you with: -@itemize @bullet -@item -An introduction to @strong{MySQL} and SQL. -@item -Coverage of @strong{MySQL}'s data types and how to use them. -@item -Thorough treatment of how to write client programs in C. -@item -A guide to using the Perl @code{DBI} and PHP APIs for developing -command-line and Web-based applications. -@item -Tips on administrative issues such as user accounts, backup, -crash recovery, and security. -@item -Help in choosing an ISP for @strong{MySQL} access. -@item -A comprehensive reference for @strong{MySQL}'s data types, operators, -functions, and SQL statements and utilities. -@item -Complete reference guides for @strong{MySQL}'s C API, the Perl @code{DBI} API, -and PHP's @strong{MySQL}-related functions. -@end itemize -@* - -@emph{MySQL & mSQL} -@multitable @columnfractions .3 .7 -@item Available @tab @uref{http://service.bfast.com/bfast/click?bfmid=2181&sourceid=34233559&bfpid=1565924347&bfmtype=book, Barnes and Noble} -@item Publisher @tab O'Reilly -@item Authors @tab Randy Jay Yarger, George Reese & Tim King -@item Pub Date @tab 1st Edition July 1999 -@item ISBN @tab 1-56592-434-7, Order Number: 4347 -@item Pages @tab 506 -@item Price @tab $34.95 -@end multitable - -This book teaches you how to use @strong{MySQL} and @code{mSQL}, two popular -and robust database products that support key subsets of SQL on both Linux -and Unix systems. Anyone who knows basic C, Java, Perl, or Python can -write a program to interact with a database, either as a stand-alone -application or through a Web page. This book takes you through the -whole process, from installation and configuration to programming -interfaces and basic administration. Includes plenty of tutorial -material. -@* - -@emph{Sams' Teach Yourself MySQL in 21 Days} -@multitable @columnfractions .3 .7 -@item Available @tab @uref{http://service.bfast.com/bfast/click?bfmid=2181&sourceid=34233559&bfpid=0672319144&bfmtype=book, Barnes and Noble} -@item Publisher @tab Sams -@item Authors @tab Mark Maslakowski and Tony Butcher -@item Pub Date @tab June 2000 -@item ISBN @tab 0672319144 -@item Pages @tab 650 -@item Price @tab $39.99 -@end multitable - -Sams' @emph{Teach Yourself MySQL in 21 Days} is for intermediate Linux users -who want to move into databases. A large share of the audience is Web -developers who need a database to store large amounts of information that -can be retrieved via the Web. - -Sams' @emph{Teach Yourself MySQL in 21 Days} is a practical, step-by-step -tutorial. The reader will learn to design and employ this open source -database technology into his or her Web site using practical, hands-on -examples to follow. -@* - -@emph{E-Commerce Solutions with MySQL} -@multitable @columnfractions .3 .7 -@item Available @tab @uref{http://service.bfast.com/bfast/click?bfmid=2181&sourceid=34233559&bfpid=0761524452&bfmtype=book, Barnes and Noble} -@item Publisher @tab Prima Communications, Inc. -@item Authors @tab N/A -@item Pub Date @tab January 2000 -@item ISBN @tab 0761524452 -@item Pages @tab 500 -@item Price @tab $39.99 -@end multitable - -No description available. -@* - -@emph{MySQL and PHP from Scratch} -@multitable @columnfractions .3 .7 -@item Available @tab @uref{http://service.bfast.com/bfast/click?bfmid=2181&sourceid=34233559&bfpid=0789724405&bfmtype=book, Barnes and Noble} -@item Publisher @tab Que -@item Authors @tab N/A -@item Pub Date @tab September 2000 -@item ISBN @tab 0789724405 -@item Pages @tab 550 -@item Price @tab $34.99 -@end multitable - -This book puts together information on installing, setting up, and -troubleshooting Apache, @strong{MySQL}, PHP3, and IMP into one complete -volume. You also learn how each piece is part of a whole by learning, -step-by-step, how to create a web-based e-mail system. Learn to run -the equivalent of Active Server Pages (ASP) using PHP3, set up an -e-commerce site using a database and the Apache web server, and create -a data entry system (such as sales, product quality tracking, customer -preferences, etc) that no installation in the PC. -@* - -@emph{Professional MySQL Programming} -@multitable @columnfractions .3 .7 -@item Available @tab @uref{http://shop.barnesandnoble.com/bookSearch/isbnInquiry.asp?isbn=1861005164, Barnes and Noble} -@item Publisher @tab Wrox Press, Inc. -@item Authors @tab N/A -@item Pub Date @tab Late 2001 -@item ISBN @tab 1861005164 -@item Pages @tab 1000 -@item Price @tab $49.99 -@end multitable - -No description available. -@* - -@emph{Professional Linux Programming} -@multitable @columnfractions .3 .7 -@item Available @tab @uref{http://service.bfast.com/bfast/click?bfmid=2181&sourceid=34233559&bfpid=1861003013&bfmtype=book, Barnes and Noble} -@item Publisher @tab Wrox Press, Inc. -@item Authors @tab N/A -@item Pub Date @tab September 2000 -@item ISBN @tab 1861003013 -@item Pages @tab 1155 -@item Price @tab $47.99 -@end multitable - -In this follow-up to the best-selling @emph{Beginning Linux Programming}, -you will learn from the authors' real-world knowledge and experience of -developing software for Linux; you'll be taken through the development -of a sample 'DVD Store' application, with 'theme' chapters addressing -different aspects of its implementation. Meanwhile, individual -``take-a-break'' chapters cover important topics that go beyond the -bounds of the central theme. All focus on the practical aspects of -programming, showing how crucial it is to choose the right tools for -the job, use them as they should be used, and get things right first -time. -@* - -@emph{PHP and MySQL Web Development} -@multitable @columnfractions .3 .7 -@item Available @tab @uref{http://service.bfast.com/bfast/click?bfmid=2181&sourceid=34233559&bfpid=0672317842&bfmtype=book, Barnes and Noble} -@item Publisher @tab Sams -@item Authors @tab Luke Welling, Laura Thomson -@item Pub Date @tab March 2001 -@item ISBN @tab 0672317842 -@item Pages @tab 700 -@item Price @tab $49.99 -@end multitable - -@emph{PHP and MySQL Web Development} introduces you to the advantages -of implementing both @strong{MySQL} and PHP. These advantages are detailed -through the provision of both statistics and several case studies. A -practical web application is developed throughout the book, providing -you with the tools necessary to implement a functional online -database. Each function is developed separately, allowing you the -choice to incorporate only those parts that you would like to -implement. Programming concepts of the PHP language are highlighted, -including functions which tie @strong{MySQL} support into a PHP script and -advanced topics regarding table manipulation. -@* - -@strong{Books recommended by the MySQL Developers} - -@emph{SQL-99 Complete, Really} -@multitable @columnfractions .3 .7 -@item Available @tab @uref{http://service.bfast.com/bfast/click?bfmid=2181&sourceid=34233559&bfpid=0879305681&bfmtype=book, Barnes and Noble} -@item Publisher @tab CMP Books -@item Authors @tab Peter Gulutzan, Trudy Pelzer -@item Pub Date @tab April 1999 -@item ISBN @tab 0879305681 -@item Pages @tab 1104 -@item Price @tab $55.96 -@end multitable - -This book contains complete descriptions of the new standards for -syntax, data structures, and retrieval processes of SQL databases. As -an example-based reference manual, it includes all of the CLI -functions, information, schema tables, and status codes, as well as a -working SQL database provided on the companion disk. -@* - -@emph{C, A reference manual} -@multitable @columnfractions .3 .7 -@item Available @tab @uref{http://service.bfast.com/bfast/click?bfmid=2181&sourceid=34233559&bfpid=0133262243&bfmtype=book, Barnes and Noble} -@item Publisher @tab Prentice Hall -@item Authors @tab Samuel P. Harbison, Guy L. Steele -@item Pub Date @tab September 1994 -@item ISBN @tab 0133262243 -@item Pages @tab 480 -@item Price @tab $35.99 -@end multitable - -A new and improved revision of the bestselling C language -reference. This manual introduces the notion of "Clean C", writing C -code that can be compiled as a C++ program, C programming style that -emphasizes correctness, portability, maintainability, and -incorporates the ISO C Amendment 1 (1994) which specifies new -facilities for writing portable, international programs in C. -@* - -@emph{C++ for Real Programmers} -@multitable @columnfractions .3 .7 -@item Available @tab @uref{http://service.bfast.com/bfast/click?bfmid=2181&sourceid=34233559&bfpid=0120499428&bfmtype=book, Barnes and Noble} -@item Publisher @tab Academic Press, Incorporated -@item Authors @tab Jeff Alger, Jim Keogh -@item Pub Date @tab February 1998 -@item ISBN @tab 0120499428 -@item Pages @tab 388 -@item Price @tab $39.95 -@end multitable - -@emph{C++ For Real Programmers} bridges the gap between C++ as described -in beginner and intermediate-level books and C++ as it is practiced by -experts. Numerous valuable techniques are described, organized into -three simple themes: indirection, class hierarchies, and memory -management. It also provides in-depth coverage of template creation, -exception handling, pointers and optimization techniques. The focus of -the book is on ANSI C++ and, as such, is compiler independent. - -@emph{C++ For Real Programmers} is a revision of -@emph{Secrets of the C++ Masters} and includes a new appendix comparing C++ -with Java. The book comes with a 3.5" disk for Windows with source code. -@* - -@emph{Algorithms in C} -@multitable @columnfractions .3 .7 -@item Available @tab @uref{http://service.bfast.com/bfast/click?bfmid=2181&sourceid=34233559&bfpid=0201514257&bfmtype=book, Barnes and Noble} -@item Publisher @tab Addison Wesley Longman, Inc. -@item Authors @tab Robert Sedgewick -@item Pub Date @tab April 1990 -@item ISBN @tab 0201514257 -@item Pages @tab 648 -@item Price @tab $45.75 -@end multitable - -@emph{Algorithms in C} describes a variety of algorithms in a number of -areas of interest, including: sorting, searching, string-processing, and -geometric, graph and mathematical algorithms. The book emphasizes -fundamental techniques, providing readers with the tools to confidently -implement, run, and debug useful algorithms. -@* - -@emph{Multithreaded Programming with Pthreads} -@multitable @columnfractions .3 .7 -@item Available @tab @uref{http://service.bfast.com/bfast/click?bfmid=2181&sourceid=34233559&bfpid=0136807291&bfmtype=book, Barnes and Noble} -@item Publisher @tab Prentice Hall -@item Authors @tab Bil Lewis, Daniel J. Berg -@item Pub Date @tab October 1997 -@item ISBN @tab 0136807291 -@item Pages @tab 432 -@item Price @tab $34.95 -@end multitable - -Based on the best-selling @emph{Threads Primer}, -@emph{Multithreaded Programming with Pthreads} gives you a solid -understanding of Posix threads: what they are, how they work, when to use -them, and how to optimize them. It retains the clarity and humor of -@emph{Threads Primer}, but includes expanded comparisons to Win32 and OS/2 -implementations. Code examples tested on all of the major UNIX platforms -are featured along with detailed explanations of how and why they use threads. -@* - -@emph{Programming the PERL DBI: Database Programming with PERL} -@multitable @columnfractions .3 .7 -@item Available @tab @uref{http://service.bfast.com/bfast/click?bfmid=2181&sourceid=34233559&bfpid=1565926994&bfmtype=book, Barnes and Noble} -@item Publisher @tab O'Reilly & Associates, Incorporated -@item Authors @tab Alligator Descartes, Tim Bunce -@item Pub Date @tab February 2000 -@item ISBN @tab 1565926994 -@item Pages @tab 400 -@item Price @tab $27.96 -@end multitable - -@emph{Programming the Perl DBI} is coauthored by Alligator Descartes, one -of the most active members of the DBI community, and by Tim Bunce, the -inventor of DBI. For the uninitiated, the book explains the architecture -of DBI and shows you how to write DBI-based programs. For the experienced -DBI dabbler, this book explains DBI's nuances and the peculiarities of each -individual DBD. - -The book includes: -@itemize @bullet -@item -An introduction to DBI and its design. -@item -How to construct queries and bind parameters. -@item -Working with database, driver, and statement handles. -@item -Debugging techniques. -@item -Coverage of each existing DBD. -@item -A complete reference to DBI. -@end itemize -@* -@node Features, Stability, MySQL-Books, Introduction -@section The Main Features of MySQL +@node Features, Stability, History, MySQL and MySQL AB +@subsection The Main Features of MySQL @cindex main features of MySQL @cindex features of MySQL @@ -2130,8 +1799,8 @@ can be used to determine how the optimizer resolves a query. @end itemize @cindex stability -@node Stability, Year 2000 compliance, Features, Introduction -@section How Stable Is MySQL? +@node Stability, Table size, Features, MySQL and MySQL AB +@subsection How Stable Is MySQL? This section addresses the questions ``How stable is @strong{MySQL}?'' and ``Can I depend on @strong{MySQL} in this project?'' We will try to clarify @@ -2146,8 +1815,6 @@ new users who made queries in a manner different than our own. Each new release has had fewer portability problems than the previous one (even though each has had many new features). -@c FIX We've been stable for quite a while now. :) (jcole) - Each release of @strong{MySQL} has been usable, and there have been problems only when users start to use code from the ``gray zones.'' Naturally, outside users don't know what the gray zones are; this section attempts to indicate @@ -2290,11 +1957,68 @@ Text search seems to work, but is still not widely used. Bugs are usually fixed right away with a patch; for serious bugs, there is almost always a new release. +@node Table size, Year 2000 compliance, Stability, MySQL and MySQL AB +@subsection How Big Can MySQL Tables Be? + +@cindex tables, maximum size +@cindex size of tables +@cindex operating systems, file size limits +@cindex limits, file size +@cindex files, size limits + +@strong{MySQL} Version 3.22 has a 4G limit on table size. With the new +@code{MyISAM} in @strong{MySQL} Version 3.23 the maximum table size is +pushed up to 8 million terabytes (2 ^ 63 bytes). + +Note, however, that operating systems have their own file size +limits. Here are some examples: + +@multitable @columnfractions .5 .5 +@item @strong{Operating System} @tab @strong{File Size Limit} +@item Linux-Intel 32 bit @tab 2G, 4G or more, depends on Linux version +@item Linux-Alpha @tab 8T (?) +@item Solaris 2.5.1 @tab 2G (possible 4G with patch) +@item Solaris 2.6 @tab 4G +@item Solaris 2.7 Intel @tab 4G +@item Solaris 2.7 ULTRA-SPARC @tab 8T (?) +@end multitable + +On Linux 2.2 you can get bigger tables than 2G by using the LFS patch for +the ext2 file system. On Linux 2.4 there exists also patches for ReiserFS +to get support for big files. + +This means that the table size for @strong{MySQL} is normally limited by +the operating system. + +By default, @strong{MySQL} tables have a maximum size of about 4G. You can +check the maximum table size for a table with the @code{SHOW TABLE STATUS} +command or with the @code{myisamchk -dv table_name}. +@xref{SHOW}. + +If you need bigger tables than 4G (and your operating system supports +this), you should set the @code{AVG_ROW_LENGTH} and @code{MAX_ROWS} +parameter when you create your table. @xref{CREATE TABLE}. You can +also set these later with @code{ALTER TABLE}. @xref{ALTER TABLE}. + +If your big table is going to be read-only, you could use +@code{myisampack} to merge and compress many tables to one. +@code{myisampack} usually compresses a table by at least 50%, so you can +have, in effect, much bigger tables. @xref{myisampack, , +@code{myisampack}}. + +You can go around the operating system file limit for @code{MyISAM} data +files by using the @code{RAID} option. @xref{CREATE TABLE}. + +Another solution can be the included MERGE library, which allows you to +handle a collection of identical tables as one. @xref{MERGE, MERGE +tables}. + +@node Year 2000 compliance, , Table size, MySQL and MySQL AB +@subsection Year 2000 Compliance + @cindex Year 2000 compliance @cindex compliance, Y2K @cindex date functions, Y2K compliance -@node Year 2000 compliance, General-SQL, Stability, Introduction -@section Year 2000 Compliance @strong{MySQL} itself has no problems with Year 2000 (Y2K) compliance: @@ -2388,10 +2112,373 @@ provide unambiguous input. See @ref{Y2K issues} for @strong{MySQL}'s rules for dealing with ambiguous date input data (data containing 2-digit year values). +@node MySQL Information Sources, Licensing and Support, MySQL and MySQL AB, Introduction +@section MySQL Information Sources + @cindex manuals, about @strong{MySQL} @cindex books, about @strong{MySQL} -@node General-SQL, Useful Links, Year 2000 compliance, Introduction -@section General SQL Information and Tutorials + +@menu +* MySQL-Books:: +* General-SQL:: +* Useful Links:: +* Questions:: +@end menu + +@node MySQL-Books, General-SQL, MySQL Information Sources, MySQL Information Sources +@subsection Books About MySQL + +While this manual is still the right place for up to date technical +information, its primary goal is to contain everything there is to know +about @strong{MySQL}. It is sometimes nice to have a bound book to read +in bed or while you travel. Here is a list of books about @strong{MySQL} and +related subjects (in English). + +By purchasing a book through these hyperlinks provided herein, you are +contributing to the development of @strong{MySQL}. + +@emph{MySQL} +@multitable @columnfractions .3 .7 +@item Available @tab @uref{http://service.bfast.com/bfast/click?bfmid=2181&sourceid=34233559&bfpid=0735709211&bfmtype=book, Barnes and Noble} +@item Publisher @tab New Riders +@item Author @tab Paul DuBois +@item Pub Date @tab 1st Edition December 1999 +@item ISBN @tab 0735709211 +@item Pages @tab 800 +@item Price @tab $49.99 US +@item Downloadable examples @tab + @uref{http://www.kitebird.com/mysql-book/, @code{samp_db} distribution} +@item Errata @tab +@uref{http://www.kitebird.com/mysql-book/errata.html, are available here} +@end multitable + +Foreword by Michael ``Monty'' Widenius, @strong{MySQL} Moderator. +@* + +In @emph{MySQL}, Paul DuBois provides you with a comprehensive guide to +one of the most popular relational database systems. Paul has +contributed to the online documentation for @strong{MySQL} and is an +active member of the @strong{MySQL} community. The principal @strong{MySQL} +developer, Monty Widenius, and a network of his fellow developers +reviewed the manuscript, and provided Paul with the kind of insight +no one else could supply. +@* + +Instead of merely giving you a general overview of @strong{MySQL}, Paul +teaches you how to make the most of its capabilities. Through two +sample database applications that run throughout the book, he +gives you solutions to problems you're sure to face. He helps you +integrate @strong{MySQL} efficiently with third-party tools, such as PHP +and Perl, enabling you to generate dynamic Web pages through +database queries. He teaches you to write programs that access +@strong{MySQL} databases, and also provides a comprehensive set of +references to column types, operators, functions, SQL syntax, +@strong{MySQL} programming, C API, Perl @code{DBI}, and PHP API. +@emph{MySQL} simply gives you the kind of information you won't find +anywhere else. +@* + +If you use @strong{MySQL}, this book provides you with: +@itemize @bullet +@item +An introduction to @strong{MySQL} and SQL. +@item +Coverage of @strong{MySQL}'s data types and how to use them. +@item +Thorough treatment of how to write client programs in C. +@item +A guide to using the Perl @code{DBI} and PHP APIs for developing +command-line and Web-based applications. +@item +Tips on administrative issues such as user accounts, backup, +crash recovery, and security. +@item +Help in choosing an ISP for @strong{MySQL} access. +@item +A comprehensive reference for @strong{MySQL}'s data types, operators, +functions, and SQL statements and utilities. +@item +Complete reference guides for @strong{MySQL}'s C API, the Perl @code{DBI} API, +and PHP's @strong{MySQL}-related functions. +@end itemize +@* + +@emph{MySQL & mSQL} +@multitable @columnfractions .3 .7 +@item Available @tab @uref{http://service.bfast.com/bfast/click?bfmid=2181&sourceid=34233559&bfpid=1565924347&bfmtype=book, Barnes and Noble} +@item Publisher @tab O'Reilly +@item Authors @tab Randy Jay Yarger, George Reese & Tim King +@item Pub Date @tab 1st Edition July 1999 +@item ISBN @tab 1-56592-434-7, Order Number: 4347 +@item Pages @tab 506 +@item Price @tab $34.95 +@end multitable + +This book teaches you how to use @strong{MySQL} and @code{mSQL}, two popular +and robust database products that support key subsets of SQL on both Linux +and Unix systems. Anyone who knows basic C, Java, Perl, or Python can +write a program to interact with a database, either as a stand-alone +application or through a Web page. This book takes you through the +whole process, from installation and configuration to programming +interfaces and basic administration. Includes plenty of tutorial +material. +@* + +@emph{Sams' Teach Yourself MySQL in 21 Days} +@multitable @columnfractions .3 .7 +@item Available @tab @uref{http://service.bfast.com/bfast/click?bfmid=2181&sourceid=34233559&bfpid=0672319144&bfmtype=book, Barnes and Noble} +@item Publisher @tab Sams +@item Authors @tab Mark Maslakowski and Tony Butcher +@item Pub Date @tab June 2000 +@item ISBN @tab 0672319144 +@item Pages @tab 650 +@item Price @tab $39.99 +@end multitable + +Sams' @emph{Teach Yourself MySQL in 21 Days} is for intermediate Linux users +who want to move into databases. A large share of the audience is Web +developers who need a database to store large amounts of information that +can be retrieved via the Web. + +Sams' @emph{Teach Yourself MySQL in 21 Days} is a practical, step-by-step +tutorial. The reader will learn to design and employ this open source +database technology into his or her Web site using practical, hands-on +examples to follow. +@* + +@emph{E-Commerce Solutions with MySQL} +@multitable @columnfractions .3 .7 +@item Available @tab @uref{http://service.bfast.com/bfast/click?bfmid=2181&sourceid=34233559&bfpid=0761524452&bfmtype=book, Barnes and Noble} +@item Publisher @tab Prima Communications, Inc. +@item Authors @tab N/A +@item Pub Date @tab January 2000 +@item ISBN @tab 0761524452 +@item Pages @tab 500 +@item Price @tab $39.99 +@end multitable + +No description available. +@* + +@emph{MySQL and PHP from Scratch} +@multitable @columnfractions .3 .7 +@item Available @tab @uref{http://service.bfast.com/bfast/click?bfmid=2181&sourceid=34233559&bfpid=0789724405&bfmtype=book, Barnes and Noble} +@item Publisher @tab Que +@item Authors @tab N/A +@item Pub Date @tab September 2000 +@item ISBN @tab 0789724405 +@item Pages @tab 550 +@item Price @tab $34.99 +@end multitable + +This book puts together information on installing, setting up, and +troubleshooting Apache, @strong{MySQL}, PHP3, and IMP into one complete +volume. You also learn how each piece is part of a whole by learning, +step-by-step, how to create a web-based e-mail system. Learn to run +the equivalent of Active Server Pages (ASP) using PHP3, set up an +e-commerce site using a database and the Apache web server, and create +a data entry system (such as sales, product quality tracking, customer +preferences, etc) that no installation in the PC. +@* + +@emph{Professional MySQL Programming} +@multitable @columnfractions .3 .7 +@item Available @tab @uref{http://shop.barnesandnoble.com/bookSearch/isbnInquiry.asp?isbn=1861005164, Barnes and Noble} +@item Publisher @tab Wrox Press, Inc. +@item Authors @tab N/A +@item Pub Date @tab Late 2001 +@item ISBN @tab 1861005164 +@item Pages @tab 1000 +@item Price @tab $49.99 +@end multitable + +No description available. +@* + +@emph{Professional Linux Programming} +@multitable @columnfractions .3 .7 +@item Available @tab @uref{http://service.bfast.com/bfast/click?bfmid=2181&sourceid=34233559&bfpid=1861003013&bfmtype=book, Barnes and Noble} +@item Publisher @tab Wrox Press, Inc. +@item Authors @tab N/A +@item Pub Date @tab September 2000 +@item ISBN @tab 1861003013 +@item Pages @tab 1155 +@item Price @tab $47.99 +@end multitable + +In this follow-up to the best-selling @emph{Beginning Linux Programming}, +you will learn from the authors' real-world knowledge and experience of +developing software for Linux; you'll be taken through the development +of a sample 'DVD Store' application, with 'theme' chapters addressing +different aspects of its implementation. Meanwhile, individual +``take-a-break'' chapters cover important topics that go beyond the +bounds of the central theme. All focus on the practical aspects of +programming, showing how crucial it is to choose the right tools for +the job, use them as they should be used, and get things right first +time. +@* + +@emph{PHP and MySQL Web Development} +@multitable @columnfractions .3 .7 +@item Available @tab @uref{http://service.bfast.com/bfast/click?bfmid=2181&sourceid=34233559&bfpid=0672317842&bfmtype=book, Barnes and Noble} +@item Publisher @tab Sams +@item Authors @tab Luke Welling, Laura Thomson +@item Pub Date @tab March 2001 +@item ISBN @tab 0672317842 +@item Pages @tab 700 +@item Price @tab $49.99 +@end multitable + +@emph{PHP and MySQL Web Development} introduces you to the advantages +of implementing both @strong{MySQL} and PHP. These advantages are detailed +through the provision of both statistics and several case studies. A +practical web application is developed throughout the book, providing +you with the tools necessary to implement a functional online +database. Each function is developed separately, allowing you the +choice to incorporate only those parts that you would like to +implement. Programming concepts of the PHP language are highlighted, +including functions which tie @strong{MySQL} support into a PHP script and +advanced topics regarding table manipulation. +@* + +@strong{Books recommended by the MySQL Developers} + +@emph{SQL-99 Complete, Really} +@multitable @columnfractions .3 .7 +@item Available @tab @uref{http://service.bfast.com/bfast/click?bfmid=2181&sourceid=34233559&bfpid=0879305681&bfmtype=book, Barnes and Noble} +@item Publisher @tab CMP Books +@item Authors @tab Peter Gulutzan, Trudy Pelzer +@item Pub Date @tab April 1999 +@item ISBN @tab 0879305681 +@item Pages @tab 1104 +@item Price @tab $55.96 +@end multitable + +This book contains complete descriptions of the new standards for +syntax, data structures, and retrieval processes of SQL databases. As +an example-based reference manual, it includes all of the CLI +functions, information, schema tables, and status codes, as well as a +working SQL database provided on the companion disk. +@* + +@emph{C, A reference manual} +@multitable @columnfractions .3 .7 +@item Available @tab @uref{http://service.bfast.com/bfast/click?bfmid=2181&sourceid=34233559&bfpid=0133262243&bfmtype=book, Barnes and Noble} +@item Publisher @tab Prentice Hall +@item Authors @tab Samuel P. Harbison, Guy L. Steele +@item Pub Date @tab September 1994 +@item ISBN @tab 0133262243 +@item Pages @tab 480 +@item Price @tab $35.99 +@end multitable + +A new and improved revision of the bestselling C language +reference. This manual introduces the notion of "Clean C", writing C +code that can be compiled as a C++ program, C programming style that +emphasizes correctness, portability, maintainability, and +incorporates the ISO C Amendment 1 (1994) which specifies new +facilities for writing portable, international programs in C. +@* + +@emph{C++ for Real Programmers} +@multitable @columnfractions .3 .7 +@item Available @tab @uref{http://service.bfast.com/bfast/click?bfmid=2181&sourceid=34233559&bfpid=0120499428&bfmtype=book, Barnes and Noble} +@item Publisher @tab Academic Press, Incorporated +@item Authors @tab Jeff Alger, Jim Keogh +@item Pub Date @tab February 1998 +@item ISBN @tab 0120499428 +@item Pages @tab 388 +@item Price @tab $39.95 +@end multitable + +@emph{C++ For Real Programmers} bridges the gap between C++ as described +in beginner and intermediate-level books and C++ as it is practiced by +experts. Numerous valuable techniques are described, organized into +three simple themes: indirection, class hierarchies, and memory +management. It also provides in-depth coverage of template creation, +exception handling, pointers and optimization techniques. The focus of +the book is on ANSI C++ and, as such, is compiler independent. + +@emph{C++ For Real Programmers} is a revision of +@emph{Secrets of the C++ Masters} and includes a new appendix comparing C++ +with Java. The book comes with a 3.5" disk for Windows with source code. +@* + +@emph{Algorithms in C} +@multitable @columnfractions .3 .7 +@item Available @tab @uref{http://service.bfast.com/bfast/click?bfmid=2181&sourceid=34233559&bfpid=0201514257&bfmtype=book, Barnes and Noble} +@item Publisher @tab Addison Wesley Longman, Inc. +@item Authors @tab Robert Sedgewick +@item Pub Date @tab April 1990 +@item ISBN @tab 0201514257 +@item Pages @tab 648 +@item Price @tab $45.75 +@end multitable + +@emph{Algorithms in C} describes a variety of algorithms in a number of +areas of interest, including: sorting, searching, string-processing, and +geometric, graph and mathematical algorithms. The book emphasizes +fundamental techniques, providing readers with the tools to confidently +implement, run, and debug useful algorithms. +@* + +@emph{Multithreaded Programming with Pthreads} +@multitable @columnfractions .3 .7 +@item Available @tab @uref{http://service.bfast.com/bfast/click?bfmid=2181&sourceid=34233559&bfpid=0136807291&bfmtype=book, Barnes and Noble} +@item Publisher @tab Prentice Hall +@item Authors @tab Bil Lewis, Daniel J. Berg +@item Pub Date @tab October 1997 +@item ISBN @tab 0136807291 +@item Pages @tab 432 +@item Price @tab $34.95 +@end multitable + +Based on the best-selling @emph{Threads Primer}, +@emph{Multithreaded Programming with Pthreads} gives you a solid +understanding of Posix threads: what they are, how they work, when to use +them, and how to optimize them. It retains the clarity and humor of +@emph{Threads Primer}, but includes expanded comparisons to Win32 and OS/2 +implementations. Code examples tested on all of the major UNIX platforms +are featured along with detailed explanations of how and why they use threads. +@* + +@emph{Programming the PERL DBI: Database Programming with PERL} +@multitable @columnfractions .3 .7 +@item Available @tab @uref{http://service.bfast.com/bfast/click?bfmid=2181&sourceid=34233559&bfpid=1565926994&bfmtype=book, Barnes and Noble} +@item Publisher @tab O'Reilly & Associates, Incorporated +@item Authors @tab Alligator Descartes, Tim Bunce +@item Pub Date @tab February 2000 +@item ISBN @tab 1565926994 +@item Pages @tab 400 +@item Price @tab $27.96 +@end multitable + +@emph{Programming the Perl DBI} is coauthored by Alligator Descartes, one +of the most active members of the DBI community, and by Tim Bunce, the +inventor of DBI. For the uninitiated, the book explains the architecture +of DBI and shows you how to write DBI-based programs. For the experienced +DBI dabbler, this book explains DBI's nuances and the peculiarities of each +individual DBD. + +The book includes: +@itemize @bullet +@item +An introduction to DBI and its design. +@item +How to construct queries and bind parameters. +@item +Working with database, driver, and statement handles. +@item +Debugging techniques. +@item +Coverage of each existing DBD. +@item +A complete reference to DBI. +@end itemize +@* + +@node General-SQL, Useful Links, MySQL-Books, MySQL Information Sources +@subsection General SQL Information and Tutorials The following book has been recommended by several people on the @strong{MySQL} mailing list: @@ -2419,8 +2506,8 @@ Alameda, CA USA A SQL tutorial is available on the net at http://w3.one.net/~jhoffman/sqltut.htm -@node Useful Links, , General-SQL, Introduction -@section Useful MySQL-related Links +@node Useful Links, Questions, General-SQL, MySQL Information Sources +@subsection Useful MySQL-related Links Apart from the following links, you can find and download a lot of @strong{MySQL} programs, tools and APIs from the @@ -3007,10 +3094,11 @@ require that you show a @strong{MySQL} logo somewhere if you wish your site to be added. It is okay to have it on a ``used tools'' page or something similar. +@node Questions, , Useful Links, MySQL Information Sources +@subsection MySQL Mailing Lists + @cindex reporting, errors @cindex MySQL mailing lists -@node Questions, Licensing and Support, Introduction, Top -@chapter MySQL Mailing Lists @menu * Mailing-list:: The @strong{MySQL} mailing lists @@ -3019,11 +3107,11 @@ something similar. * Answering questions:: Guidelines for answering questions on the mailing list @end menu -This chapter introduces you to the @strong{MySQL} mailing lists, and gives +This section introduces you to the @strong{MySQL} mailing lists, and gives some guidelines as to how to use them. @node Mailing-list, Asking questions, Questions, Questions -@section The MySQL Mailing Lists +@subsubsection The MySQL Mailing Lists @cindex mailing lists @cindex email lists @@ -3167,11 +3255,12 @@ Email @code{subscribe mysql-br your@@email.address} to this list. Email @code{subscribe mysql your@@email.address} to this list. @end table +@node Asking questions, Bug reports, Mailing-list, Questions +@subsubsection Asking Questions or Reporting Bugs + @cindex net etiquette @cindex mailing lists, archive location @cindex searching, MySQL webpages -@node Asking questions, Bug reports, Mailing-list, Questions -@section Asking Questions or Reporting Bugs Before posting a bug report or question, please do the following: @@ -3200,6 +3289,9 @@ local @strong{MySQL} expert. If you still can't find an answer to your question, go ahead and read the next section about how to send mail to @email{mysql@@lists.mysql.com}. +@node Bug reports, Answering questions, Asking questions, Questions +@subsubsection How to Report Bugs or Problems + @cindex bugs, reporting @cindex reporting, bugs @cindex problems, reporting @@ -3207,8 +3299,6 @@ question, go ahead and read the next section about how to send mail to @cindex @code{mysqlbug} script @cindex creating, bug reports @cindex scripts, @code{mysqlbug} -@node Bug reports, Answering questions, Asking questions, Questions -@section How to Report Bugs or Problems Writing a good bug report takes patience, but doing it right the first time saves time for us and for you. A good bug report containing a full @@ -3506,12 +3596,13 @@ it is considered good etiquette to summarize the answers and send the summary to the mailing list so that others may have the benefit of responses you received that helped you solve your problem! +@node Answering questions, , Bug reports, Questions +@subsubsection Guidelines for Answering Question on the Mailing List + @cindex net etiquette @cindex questions, answering @cindex answering questions, etiquette @cindex mailing lists, guidelines -@node Answering questions, , Bug reports, Questions -@section Guidelines for Answering Question on the Mailing List If you consider your answer to have broad interest, you may want to post it to the mailing list instead of replying directly to the individual who @@ -3525,10 +3616,11 @@ obliged to quote the entire original message. Please don't post mail messages from your browser with HTML mode turned on! Many users don't read mail with a browser! +@node Licensing and Support, Compatibility, MySQL Information Sources, Introduction +@section MySQL Licensing and Support + @cindex licensing terms @cindex support terms -@node Licensing and Support, Installing, Questions, Top -@chapter MySQL Licensing and Support @menu * Licensing policy:: @strong{MySQL} licensing policy @@ -3538,7 +3630,7 @@ Many users don't read mail with a browser! * Support:: Types of commercial support @end menu -This chapter describes @strong{MySQL} support and licensing +This section describes @strong{MySQL} support and licensing arrangements: @itemize @bullet @@ -3555,12 +3647,13 @@ arrangements: @item Commercial licensing costs @end itemize +@node Licensing policy, Copyright, Licensing and Support, Licensing and Support +@subsection MySQL Licensing Policy + @cindex licensing policy @cindex technical support, licensing @cindex support, licensing @cindex General Public License, MySQL -@node Licensing policy, Copyright, Licensing and Support, Licensing and Support -@section MySQL Licensing Policy The formal terms of the GPL license can be found at @ref{GPL license}. Basically, our licensing policy and interpretation of the GPL is as follows: @@ -3640,9 +3733,10 @@ is to use the license form on @strong{MySQL}'s secure server at @uref{https://order.mysql.com/}. Other forms of payment are discussed in @ref{Payment information}. -@cindex copyrights @node Copyright, Licensing examples, Licensing policy, Licensing and Support -@section Copyrights Used by MySQL +@subsection Copyrights Used by MySQL + +@cindex copyrights @menu * Copyright changes:: Possible future copyright changes @@ -3686,13 +3780,13 @@ want to help secure its development by purchasing licenses or a support contract. @xref{Support}. @node Copyright changes, , Copyright, Copyright -@subsection Copyright Changes +@subsubsection Copyright Changes Version 3.22 of @strong{MySQL} is still using a more strict license. See the documentation for that version for more information. @node Licensing examples, Cost, Copyright, Licensing and Support -@section Example Licensing Situations +@subsection Example Licensing Situations @menu * Products that use MySQL:: Selling products that use @strong{MySQL} @@ -3713,7 +3807,7 @@ Note that a single @strong{MySQL} license covers any number of CPUs and number of clients that connect to the server in any way. @node Products that use MySQL, ISP, Licensing examples, Licensing examples -@subsection Selling Products that use MySQL +@subsubsection Selling Products that use MySQL To determine whether or not you need a @strong{MySQL} license when selling your application, you should ask whether the proper functioning @@ -3757,7 +3851,7 @@ don't need a license. @end itemize @node ISP, Web server, Products that use MySQL, Licensing examples -@subsection ISP MySQL Services +@subsubsection ISP MySQL Services @cindex ISP services @cindex services, ISP @@ -3782,10 +3876,11 @@ it should give its customers at least read access to the source of the @strong{MySQL} installation so that its customer can verify that it is patched correctly. +@node Web server, , ISP, Licensing examples +@subsubsection Running a Web Server Using MySQL + @cindex web server, running @cindex running, a web server -@node Web server, , ISP, Licensing examples -@subsection Running a Web Server Using MySQL If you use @strong{MySQL} in conjunction with a Web server on Unix, you don't have to pay for a license. @@ -3795,12 +3890,13 @@ This is true even if you run a commercial Web server that uses version yourself. However, in this case we would like you to purchase @strong{MySQL} support, because @strong{MySQL} is helping your enterprise. +@node Cost, Support, Licensing examples, Licensing and Support +@subsection MySQL Licensing and Support Costs + @cindex costs, licensing and support @cindex licensing costs @cindex support costs @cindex prices, licensing and support -@node Cost, Support, Licensing examples, Licensing and Support -@section MySQL Licensing and Support Costs @menu * Payment information:: Payment information @@ -3876,13 +3972,15 @@ Note that as our sales staff is very busy, it may take some time until your request is handled. Our support staff does however always answer promptly to support questions! -@cindex payment information @node Payment information, Contact information, Cost, Cost -@subsection Payment information +@subsubsection Payment information + +@cindex payment information Currently we can take SWIFT payments, checks, or credit cards. Payment should be made to: + @example Postgirot Bank AB 105 06 STOCKHOLM, SWEDEN @@ -3922,15 +4020,17 @@ us'' in the comment field. You can also mail a message to @email{sales@@mysql.com} (@strong{not} @code{mysql@@lists.mysql.com}!) with your company information and ask us to bill you. +@node Contact information, , Payment information, Cost +@subsubsection Contact Information + @cindex contact information @cindex licensing, contact information -@cindex partnering, with MySQL +@cindex advertising, contact information +@cindex employment, contact information +@cindex partnering with MySQL @cindex employment with MySQL @cindex jobs at MySQL -@node Contact information, , Payment information, Cost -@subsection Contact Information - For commercial licensing, please contact the @strong{MySQL} licensing team. The much preferred method is by e-mail to @email{licensing@@mysql.com}. Fax is also possible but handling of @@ -3962,11 +4062,12 @@ For general information inquires, please send e-mail to For questions or comments about the workings or content of the Web site, please send e-mail to @email{webmaster@@mysql.com}. +@node Support, , Cost, Licensing and Support +@subsection Types of Commercial Support + @cindex support, types @cindex types, of support @cindex commercial support, types -@node Support, , Cost, Licensing and Support -@section Types of Commercial Support @menu * Basic email support:: Basic email support @@ -3999,10 +4100,11 @@ if you have login support (2000 EURO), you can expect us to work up to 10 hours to help you with things like this. @end itemize +@node Basic email support, Extended email support, Support, Support +@subsubsection Basic E-mail Support + @cindex email, technical support @cindex technical support, by email -@node Basic email support, Extended email support, Support, Support -@subsection Basic E-mail Support Basic e-mail support is a very inexpensive support option and should be thought of more as a way to support our development of @strong{MySQL} @@ -4071,9 +4173,10 @@ If you want us to help optimize your system, you must upgrade to a higher level of support. @end itemize -@cindex extended email support @node Extended email support, Login support, Basic email support, Support -@subsection Extended E-mail Support +@subsubsection Extended E-mail Support + +@cindex extended email support Extended e-mail support includes everything in basic e-mail support with these additions: @@ -4117,9 +4220,10 @@ TODO List. @xref{TODO}. This will ensure that the features you really need will be implemented sooner than they might be otherwise. @end itemize -@cindex login support @node Login support, Extended login support, Extended email support, Support -@subsection Login Support +@subsubsection Login Support + +@cindex login support Login support includes everything in extended e-mail support with these additions: @@ -4160,7 +4264,7 @@ solving the problem. @end itemize @node Extended login support, Telephone support, Login support, Support -@subsection Extended Login Support +@subsubsection Extended Login Support Extended login support includes everything in login support with these additions: @@ -4197,7 +4301,7 @@ rate that is at least 20 % less than our standard rates. @end itemize @node Telephone support, Table handler support, Extended login support, Support -@subsection Telephone Support +@subsubsection Telephone Support Telephone support includes everything in extended login support with these additions: @@ -4212,10 +4316,11 @@ For non critical problem, you can request a @strong{MySQL} developer to phone back within 48 hours to discuss @code{MySQL} related issues. @end itemize +@node Table handler support, , Telephone support, Support +@subsubsection Support for other table handlers + @cindex support, BDB Tables @cindex support, InnoDB Tables -@node Table handler support, , Telephone support, Support -@subsection Support for other table handlers To get support for @code{BDB} tables, @code{InnoDB} tables you have to pay an additional 30% on the standard support price for each of @@ -4231,37182 +4336,37225 @@ timely manner, we can't guarantee a quick solution for any problems you can get with the different table handlers. We will however do our best to help you get the problem solved. +@node Compatibility, Comparisons, Licensing and Support, Introduction +@section How Standards-compatible Is MySQL? -@node Installing, Compatibility, Licensing and Support, Top -@chapter Installing MySQL -@cindex installing, overview +@cindex compatibility, with ANSI SQL +@cindex standards compatibility +@cindex extensions, to ANSI SQL +@cindex ANSI SQL92, extensions to @menu -* Getting MySQL:: How to get @strong{MySQL} -* Which OS:: Operating systems supported by @strong{MySQL} -* Which version:: Which @strong{MySQL} version to use -* Many versions:: How and when updates are released -* Installation layouts:: Installation layouts -* Installing binary:: Installing a @strong{MySQL} binary distribution -* Installing source:: Installing a @strong{MySQL} source distribution -* Installing source tree:: Installing @strong{MySQL} from development source tree -* Compilation problems:: Problems compiling? -* MIT-pthreads:: MIT-pthreads notes -* Perl support:: Perl installation comments -* Source install system issues:: System-specific issues -* Windows:: Windows notes -* OS/2:: OS/2 notes -* MySQL binaries:: MySQL binaries -* Post-installation:: Post-installation setup and testing -* Installing many servers:: Installing many servers on the same machine -* Upgrade:: Upgrading/Downgrading MySQL +* Extensions to ANSI:: @strong{MySQL} extensions to ANSI SQL92 +* Differences from ANSI:: @strong{MySQL} differences compared to ANSI SQL92 +* ANSI mode:: Running @strong{MySQL} in ANSI mode +* Missing functions:: Functionality missing from @strong{MySQL} +* Standards:: What standards does @strong{MySQL} follow? +* Commit-rollback:: How to cope without @code{COMMIT}-@code{ROLLBACK} +* Bugs:: @end menu -This chapter describes how to obtain and install @strong{MySQL}: - -@itemize @bullet -@item -For a list of sites from which you can obtain @strong{MySQL}, see -@ref{Getting MySQL, , Getting @strong{MySQL}}. +This section describes how @strong{MySQL} relates to the ANSI SQL standards. +@strong{MySQL} has many extensions to the ANSI SQL standards, and here you +will find out what they are, and how to use them. You will also find +information about functionality missing from @strong{MySQL}, and how to work +around some differences. -@item -To see which platforms are supported, see @ref{Which OS}. Please note that -not all supported system are equally good for running @strong{MySQL} on them. -On some it is much more robust and efficient than others - see @ref{Which OS} -for details. +@node Extensions to ANSI, Differences from ANSI, Compatibility, Compatibility +@subsection MySQL Extensions to ANSI SQL92 -@item -Several versions of @strong{MySQL} are available in both binary and -source distributions. We also provide public access to our current -source tree for those who want to see our most recent developments and -help us test new code. To determine which version and type of -distribution you should use, see @ref{Which version}. When in doubt, -use the binary distribution. +@strong{MySQL} includes some extensions that you probably will not find in +other SQL databases. Be warned that if you use them, your code will not be +portable to other SQL servers. In some cases, you can write code that +includes @strong{MySQL} extensions, but is still portable, by using comments +of the form @code{/*! ... */}. In this case, @strong{MySQL} will parse and +execute the code within the comment as it would any other @strong{MySQL} +statement, but other SQL servers will ignore the extensions. For example: -@item -Installation instructions for binary and source distributions are described -in @ref{Installing binary} and @ref{Installing source}. Each set of -instructions includes a section on system-specific problems you may run -into. +@example +SELECT /*! STRAIGHT_JOIN */ col_name FROM table1,table2 WHERE ... +@end example -@item -For post-installation procedures, see @ref{Post-installation}. These -procedures apply whether you install @strong{MySQL} using a binary or -source distribution. -@end itemize +If you add a version number after the @code{'!'}, the syntax will only be +executed if the @strong{MySQL} version is equal to or newer than the used +version number: -@cindex downloading -@cindex MySQL version -@cindex version, latest -@cindex getting MySQL -@cindex mirror sites -@cindex URLS for downloading MySQL -@node Getting MySQL, Which OS, Installing, Installing -@section How to Get MySQL +@example +CREATE /*!32302 TEMPORARY */ TABLE (a int); +@end example -Check the @uref{http://www.mysql.com/, @strong{MySQL} home page} for -information about the current version and for downloading instructions. +The above means that if you have Version 3.23.02 or newer, then @strong{MySQL} +will use the @code{TEMPORARY} keyword. -Our main download mirror is located at: +@strong{MySQL} extensions are listed below: -@uref{http://download.sourceforge.net/mirrors/mysql/} +@itemize @bullet +@item +The field types @code{MEDIUMINT}, @code{SET}, @code{ENUM}, and the +different @code{BLOB} and @code{TEXT} types. -If you are interested in becoming a @strong{MySQL} mirror site, you may -anonymously rsync with: @code{rsync://download.sourceforge.net/mysql/}. Please -send e-mail to @email{webmaster@@mysql.com} notifying us of your mirror to be -added to the list below. +@item +The field attributes @code{AUTO_INCREMENT}, @code{BINARY}, @code{NULL}, +@code{UNSIGNED}, and @code{ZEROFILL}. -If you have problems downloading from our main site, try using one of the -mirrors listed below. +@item +All string comparisons are case insensitive by default, with sort +ordering determined by the current character set (ISO-8859-1 Latin1 by +default). If you don't like this, you should declare your columns with +the @code{BINARY} attribute or use the @code{BINARY} cast, which causes +comparisons to be done according to the ASCII order used on the +@strong{MySQL} server host. -Please report bad or out-of-date mirrors to @email{webmaster@@mysql.com}. +@item +@strong{MySQL} maps each database to a directory under the @strong{MySQL} +data directory, and tables within a database to filenames in the database +directory. -@c START_OF_MIRROR_LISTING +This has a few implications: -@strong{Europe:} +@cindex database names, case sensitivity +@cindex table names, case sensitivity +@cindex case sensitivity, of database names +@cindex case sensitivity, of table names -@itemize @bullet +@itemize @minus @item -@c EMAIL: sl@iuinfo.tuwien.ac.at (Tony Sprinzl) -@image{Flags/austria} Austria [Univ. of Technology/Vienna] @ -@uref{http://gd.tuwien.ac.at/db/mysql/, WWW} -@uref{ftp://gd.tuwien.ac.at/db/mysql/, FTP} - -@c @item -@c Not ok 20000919; Old site (Matt) -@c EMAIL: delian@naturella.com (Delian Delchev) -@c @image{Flags/bulgaria} Bulgaria [Naturella] @ -@c @uref{http://archive.nat.bg/pub/mirror/mysql/, WWW} -@c @uref{ftp://ftp.ntrl.net/pub/mirror/mysql/, FTP} +Database names and table names are case sensitive in @strong{MySQL} on +operating systems that have case-sensitive filenames (like most Unix +systems). @xref{Name case sensitivity}. @item -@c EMAIL: salle@online.bg (Admin) -@image{Flags/bulgaria} Bulgaria [online.bg/Sofia] @ -@uref{http://mysql.online.bg/, WWW} -@uref{ftp://mysql.online.bg/, FTP} - -@c @item -@c Added: 990614 -@c Not ok 20000919; Out of date (Matt) -@c EMAIL: vuksan@veus.hr (Vladimir Vuksan) -@c @image{Flags/croatia} Croatia [HULK] @ -@c @uref{http://ftp.linux.hr/pub/mysql/, WWW} -@c @uref{ftp://ftp.linux.hr/pub/mysql/, FTP} +Database, table, index, column, or alias names may begin with a digit +(but may not consist solely of digits). @item -@c Added: 990614 -@c EMAIL: kas@informatics.muni.cz (Jan Kasprzak) -@image{Flags/czech-republic} Czech Republic [Masaryk University in Brno] @ -@uref{http://mysql.linux.cz/index.html, WWW} -@uref{ftp://ftp.fi.muni.cz/pub/mysql/, FTP} +You can use standard system commands to backup, rename, move, delete, and copy +tables. For example, to rename a table, rename the @file{.MYD}, @file{.MYI}, +and @file{.frm} files to which the table corresponds. +@end itemize @item -@c Added: 990920 -@c EMAIL: <radek@sopik.cz> (Radek Libovicky) -@image{Flags/czech-republic} Czech Republic [www.sopik.cz] @ -@uref{http://www.mysql.cz/, WWW} +In SQL statements, you can access tables from different databases +with the @code{db_name.tbl_name} syntax. Some SQL servers provide +the same functionality but call this @code{User space}. +@strong{MySQL} doesn't support tablespaces as in: +@code{create table ralph.my_table...IN my_tablespace}. @item -@c Added: 000418 -@c EMAIL: <feela@ipex.cz> (Ondrej Feela Filip) -@image{Flags/czech-republic} Czech Republic [www.gin.cz] @ -@uref{http://mysql.gin.cz/, WWW} -@uref{ftp://ftp.gin.cz/pub/MIRRORS/www.mysql.com/, FTP} +@code{LIKE} is allowed on numeric columns. @item -@c removed 991020 (no DNS entry). New name 991026. Added 991121 -@c Statistics at http://mirror.borsen.dk/ -@c EMAIL: mirrorman@borsen.dk (Michael Kyed) -@image{Flags/denmark} Denmark [Borsen] @ -@uref{ http://mysql.borsen.dk/, WWW} +Use of @code{INTO OUTFILE} and @code{STRAIGHT_JOIN} in a @code{SELECT} +statement. @xref{SELECT, , @code{SELECT}}. @item -@c EMAIL: mkp@socsci.auc.dk (Martin Kasper Petersen) -@image{Flags/denmark} Denmark [SunSITE] @ -@uref{http://SunSITE.auc.dk/mysql/, WWW} -@uref{ftp://SunSITE.auc.dk/pub/databases/mysql/, FTP} - -@c @item -@c Tõnu does not work there anymore :-) 990720 -@c EMAIL: tonu@tradenet.ee (Tõnu Samuel) -@c @image{Flags/estonia} Estonia [Tradenet] @ -@c @uref{http://mysql.tradenet.ee, WWW} +The @code{SQL_SMALL_RESULT} option in a @code{SELECT} statement. @item -@c EMAIL: tonu@spam.ee (Tõnu Samuel) -@image{Flags/estonia} Estonia [OKinteractive] @ -@uref{http://mysql.mirror.ok.ee, WWW} +@code{EXPLAIN SELECT} to get a description on how tables are joined. -@c @item -@c Changed 990531 -@c Not ok 20000919; Old site (Matt) -@c EMAIL: Steeve.Devergne@minet.net (Steeve Devergne) -@c @image{Flags/france} France [minet] @ -@c @uref{http://www.minet.net/devel/mysql/, WWW} +@item +Use of index names, indexes on a prefix of a field, and use of +@code{INDEX} or @code{KEY} in a @code{CREATE TABLE} +statement. @xref{CREATE TABLE, , @code{CREATE TABLE}}. @item -@c EMAIL: alex@mtesa.net (Alexandre Dupouy) -@image{Flags/france} France [mtesa.net] @ -@uref{http://mysql.mtesa.net/, WWW} +Use of @code{TEMPORARY} or @code{IF NOT EXISTS} with @code{CREATE TABLE}. @item -@c EMAIL: mrjack@fastorama.com (MrJacK) -@image{Flags/france} France [fastorama.com, Chatenois] @ -@uref{http://mysql.fastorama.com/, WWW} -@uref{ftp://ftp.fastorama.com/unix/mysql/, FTP} +Use of @code{COUNT(DISTINCT list)} where 'list' is more than one element. -@c @item -@c Not ok 20000919; Old site (Matt) -@c EMAIL: Jaakko.Hyvatti@eunet.fi -@c @image{Flags/finland} Finland [EUnet] @ -@c @uref{http://mysql.eunet.fi/, WWW} +@item +Use of @code{CHANGE col_name}, @code{DROP col_name}, or @code{DROP +INDEX}, @code{IGNORE} or @code{RENAME} in an @code{ALTER TABLE} +statement. @xref{ALTER TABLE, , @code{ALTER TABLE}}. -@c @item -@c Added 990829 -@c Not ok 20000919; Non-existent (Matt) -@c EMAIL: tomi.hakala@clinet.fi (Tomi Hakala) -@c @image{Flags/finland} Finland [clinet] @ -@c @uref{ftp://ftp.clinet.fi/mirrors/ftp.mysql.org/pub/mysql/, FTP} +@item +Use of @code{RENAME TABLE}. @xref{RENAME TABLE, , @code{RENAME TABLE}}. @item -@c Added 20001019 -@c EMAIL: d@d.tj (Dominique L . Bouix) -@image{Flags/finland} Finland [tonnikala.net] @ -@uref{http://mysql.tonnikala.org/, WWW} +Use of multiple @code{ADD}, @code{ALTER}, @code{DROP}, or @code{CHANGE} +clauses in an @code{ALTER TABLE} statement. @item -@c Added 981208 -@c EMAIL: noel@uni-bonn.de (Noel Koethe) -@image{Flags/germany} Germany [Kernelnotes.de, Bonn] @ -@uref{http://www.kernelnotes.de/mysql/, WWW} -@uref{ftp://ftp.kernelnotes.de/pub/mirror/mysql.org/, FTP} +Use of @code{DROP TABLE} with the keywords @code{IF EXISTS}. @item -@c EMAIL: th@rz.fh-wolfenbuettel.de (Thorsten Ludewig) -@image{Flags/germany} Germany [Wolfenbuettel] @ -@uref{http://www.fh-wolfenbuettel.de/ftp/pub/database/mysql/, WWW} -@uref{ftp://ftp.fh-wolfenbuettel.de/pub/database/mysql/, FTP} +You can drop multiple tables with a single @code{DROP TABLE} statement. -@c @item -@c Ok 980114. Removed 981208 (down > 3 days) ok 981214 -@c Not ok 20000919; Non-existent (Matt) -@c EMAIL: straub@gks.de (Hans-Peter Straub) -@c @image{Flags/germany} Germany [Staufen] @ -@c @uref{http://mysql.staufen.de/, WWW} +@item +The @code{LIMIT} clause of the @code{DELETE} statement. -@c @item -@c Added 990614 -@c Not ok 20000919; Non-existent (Matt) -@c EMAIL: thomas.rohde@ecrc.de (Thomas Rohde) -@c @image{Flags/germany} Germany [Cable & Wireless] @ -@c @uref{ftp://ftp.ecrc.net/pub/database/mysql/, FTP} +@item +The @code{DELAYED} clause of the @code{INSERT} and @code{REPLACE} +statements. @item -@c Added 981208 -@c EMAIL: christias@noc.ntua.gr (Panagiotis Christias) -@image{Flags/greece} Greece [NTUA, Athens] @ -@uref{http://www.ntua.gr/mysql/, WWW} -@uref{ftp://ftp.ntua.gr/pub/databases/mysql/, FTP} +The @code{LOW_PRIORITY} clause of the @code{INSERT}, @code{REPLACE}, +@code{DELETE}, and @code{UPDATE} statements. +@cindex Oracle compatibility +@cindex compatibility, with Oracle @item -@c EMAIL: torlasz@xenia.sote.hu (Laszlo L. Tornoc) -@image{Flags/hungary} Hungary [Xenia] @ -@uref{http://mysql.sote.hu/, WWW} -@uref{ftp://xenia.sote.hu/pub/mirrors/www.mysql.com/, FTP} +Use of @code{LOAD DATA INFILE}. In many cases, this syntax is compatible with +Oracle's @code{LOAD DATA INFILE}. @xref{LOAD DATA, , @code{LOAD DATA}}. @item -@c EMAIL: hendlein@tiszanet.hu (Hendlein Peter) -@image{Flags/hungary} Hungary [TiszaneT] @ -@uref{http://mysql.tiszanet.hu/, WWW} -@uref{ftp://mysql.tiszanet.hu/pub/mirrors/mysql/, FTP} +The @code{ANALYZE TABLE}, @code{CHECK TABLE}, @code{OPTIMIZE TABLE}, and +@code{REPAIR TABLE} statements. @item -@c EMAIL: mirrors@gm.is (Tomas Edwardsson) -@image{Flags/iceland} Iceland [GM] @ -@uref{http://mysql.gm.is/, WWW} -@uref{ftp://ftp.gm.is/pub/mysql/, FTP} +The @code{SHOW} statement. +@xref{SHOW, , @code{SHOW}}. -@c @item -@c Out of date 990906 -@c EMAIL: bourbon@netvision.net.il (Zeev Suraski) -@c @image{Flags/israel} Israel [Netvision] @ -@c @uref{http://mysql.netvision.net.il/, WWW} +@item +Strings may be enclosed by either @samp{"} or @samp{'}, not just by @samp{'}. -@c @item -@c Not working 99.03.06 -@c EMAIL: maruzz@matrice.it (Giovanni Maruzzelli) -@c @image{Flags/italy} Italy [Matrice] @ -@c @uref{http://www.matrice.it/risorse/mysql/, WWW} +@item +Use of the escape @samp{\} character. @item -@c EMAIL: gabriele@feelinglinux.com (Giansante Gabriele) -@image{Flags/italy} Italy [feelinglinux.com] @ -@uref{http://mysql.feelinglinux.com/, WWW} +The @code{SET OPTION} statement. @xref{SET OPTION, , @code{SET OPTION}}. @item -@c EMAIL: irena@yacc.it -@image{Flags/italy} Italy [Teta Srl] @ -@uref{http://www.teta.it/mysql/, WWW} +You don't need to name all selected columns in the @code{GROUP BY} part. +This gives better performance for some very specific, but quite normal +queries. +@xref{Group by functions}. @item -@c EMAIL: rudy@comm2000.it (Rudy) -@image{Flags/italy} Italy [tzone.it] @ -@uref{http://mysql.tzone.it/, WWW} +One can specify @code{ASC} and @code{DESC} with @code{GROUP BY}. -@c @item -@c Added 991121 -@c Not ok 20000919; Old site (Matt) -@c EMAIL: nick@iol.ie (Nick Hilliard) -@c @image{Flags/ireland} Ireland [Ireland On-Line/Dublin] @ -@c @uref{http://mysql.iol.ie, WWW} -@c @uref{ftp://ftp.iol.ie/pub/mysql, FTP} +@item +To make it easier for users who come from other SQL environments, +@strong{MySQL} supports aliases for many functions. For example, all +string functions support both ANSI SQL syntax and ODBC syntax. @item -@c Added 20001031 -@c EMAIL: dave@esat.net (Dave Rynne) -@image{Flags/ireland} Ireland [Esat Net] @ -@uref{http://ftp.esat.net/mirrors/download.sourceforge.net/pub/mirrors/mysql/, WWW} -@uref{ftp://ftp.esat.net/mirrors/download.sourceforge.net/pub/mirrors/mysql/, FTP} +@strong{MySQL} understands the @code{||} and @code{&&} operators to mean +logical OR and AND, as in the C programming language. In @strong{MySQL}, +@code{||} and @code{OR} are synonyms, as are @code{&&} and @code{AND}. +Because of this nice syntax, @strong{MySQL} doesn't support +the ANSI SQL @code{||} operator for string concatenation; use +@code{CONCAT()} instead. Because @code{CONCAT()} takes any number +of arguments, it's easy to convert use of the @code{||} operator to +@strong{MySQL}. @item -@c Added 20010524 -@c EMAIL: arvids@parks.lv (Arvids) -@image{Flags/latvia} Latvia [linux.lv] @ -@uref{ftp://ftp.linux.lv/pub/software/mysql/, FTP} +@code{CREATE DATABASE} or @code{DROP DATABASE}. +@xref{CREATE DATABASE, , @code{CREATE DATABASE}}. +@cindex PostgreSQL compatibility +@cindex compatibility, with PostgreSQL @item -@c Added 20001125 -@c EMAIL: mleicher@silverpoint.nl (Marcel Leicher) -@image{Flags/netherlands} Netherlands [Silverpoint] @ -@uref{http://mysql.silverpoint.nl/, WWW} +The @code{%} operator is a synonym for @code{MOD()}. That is, +@code{N % M} is equivalent to @code{MOD(N,M)}. @code{%} is supported +for C programmers and for compatibility with PostgreSQL. @item -@c EMAIL: mysql@widexs.nl (Wouter de Jong) -@image{Flags/netherlands} Netherlands [Widexs BV] @ -@uref{http://mysql.widexs.nl/, WWW} -@uref{ftp://mysql.widexs.nl/pub/mysql/, FTP} +The @code{=}, @code{<>}, @code{<=} ,@code{<}, @code{>=},@code{>}, +@code{<<}, @code{>>}, @code{<=>}, @code{AND}, @code{OR}, or @code{LIKE} +operators may be used in column comparisons to the left of the +@code{FROM} in @code{SELECT} statements. For example: + +@example +mysql> SELECT col1=1 AND col2=2 FROM tbl_name; +@end example @item -@c EMAIL: peter@proserve.nl (ProServe - Peter Batenburg) -@image{Flags/netherlands} Netherlands [ProServe] @ -@uref{http://mysql.proserve.nl/, WWW} +The @code{LAST_INSERT_ID()} function. +@xref{mysql_insert_id, , @code{mysql_insert_id()}}. @item -@c EMAIL: W.Sylwestrzak@icm.edu.pl (Wojtek Sylwestrzak) -@c mirroring nightly at 05:25 -@image{Flags/poland} Poland [Sunsite] @ -@uref{http://sunsite.icm.edu.pl/mysql/, WWW} -@uref{ftp://sunsite.icm.edu.pl/pub/unix/mysql/, FTP} +The @code{REGEXP} and @code{NOT REGEXP} extended regular expression +operators. @item -@c EMAIL: admin@net.ncservice.com.pl (Marian Witkowski) -@image{Flags/poland} Poland [ncservice.com/Gdansk] @ -@uref{http://mysql.service.net.pl/, WWW} +@code{CONCAT()} or @code{CHAR()} with one argument or more than two +arguments. (In @strong{MySQL}, these functions can take any number of +arguments.) -@c @item -@c EMAIL: melo@co.telenet.pt (Pedro Melo) -@c Temp out of service (email from Pedro) -@c @image{Flags/portugal} Portugal [IP] @ -@c @uref{http://mysql.ip.pt, WWW} +@item The @code{BIT_COUNT()}, @code{CASE}, @code{ELT()}, +@code{FROM_DAYS()}, @code{FORMAT()}, @code{IF()}, @code{PASSWORD()}, +@code{ENCRYPT()}, @code{md5()}, @code{ENCODE()}, @code{DECODE()}, +@code{PERIOD_ADD()}, @code{PERIOD_DIFF()}, @code{TO_DAYS()}, or +@code{WEEKDAY()} functions. -@c @item -@c Not ok 20000919; Non-existent (Matt) -@c EMAIL: Equipa de suporte do Leirianet <support@leirianet.pt> -@c @image{Flags/portugal} Portugal [lerianet] @ -@c @uref{http://mysql.leirianet.pt, WWW} -@c @uref{ftp://ftp.leirianet.pt/pub/mysql/,FTP} +@item +Use of @code{TRIM()} to trim substrings. ANSI SQL only supports removal +of single characters. @item -@c Added 20001031 -@c bofh@netc.pt (Bruno Rodrigues) -@image{Flags/portugal} Portugal [Netc] @ -@uref{http://ftp.netc.pt/pub/mysql/, WWW} -@uref{ftp://ftp.netc.pt/pub/mysql/, FTP} +The @code{GROUP BY} functions @code{STD()}, @code{BIT_OR()}, and +@code{BIT_AND()}. @item -@c EMAIL: keeper@roedu.net (Mihai RUSU) -@image{Flags/romania} Romania [roedu.net/Bucharest] @ -@uref{ftp://ftp.roedu.net/pub/mirrors/ftp.mysql.com/, FTP} +Use of @code{REPLACE} instead of @code{DELETE} + @code{INSERT}. +@xref{REPLACE, , @code{REPLACE}}. @item -@c EMAIL: kuzmin@dn.ru (Roma Kuzmin) -@image{Flags/russia} Russia [DirectNet] @ -@uref{http://mysql.directnet.ru/, WWW} -@uref{ftp://ftp.dn.ru/pub/MySQL/, FTP} +The @code{FLUSH flush_option} statement. -@c @item -@c down 990113 -@c EMAIL: nikkic@cityline.ru (Nikki Chumakov) -@c @image{Flags/russia} Russia [Cityline] @ -@c @uref{ftp://mysql.cityline.ru/pub/mysql, FTP} -@c @uref{http://mysql.cityline.ru, WWW} -@c EMAIL: bar@izhcom.ru (Alexander I Barkov) +@item +The possibility to set variables in a statement with @code{:=}: +@example +SELECT @@a:=SUM(total),@@b=COUNT(*),@@a/@@b AS avg FROM test_table; +SELECT @@t1:=(@@t2:=1)+@@t3:=4,@@t1,@@t2,@@t3; +@end example -@c @item -@c Not ok 20000919; Non-existent (Matt) -@c @image{Flags/russia} Russia [IZHCOM] @ -@c @uref{http://mysql.udm.net/, WWW} -@c @uref{ftp://ftp.izhcom.ru/pub/mysql/,FTP} +@end itemize -@item -@c Added 990507 -@c EMAIL: demon@gpad.ac.ru (Dima Sivachenko) -@image{Flags/russia} Russia [Scientific Center/Chernogolovka] @ -@uref{ftp://ftp.chg.ru/pub/databases/mysql/, FTP} +@node Differences from ANSI, ANSI mode, Extensions to ANSI, Compatibility +@subsection MySQL Differences Compared to ANSI SQL92 -@c @item -@c Not ok 20000919; Old site (Matt) -@c EMAIL: sebi@dnttm.ro (Sebastian DEAC) -@c @image{Flags/romania} Romania [Timisoara] @ -@c @uref{http://download.dnttm.ro/mysql, WWW} -@c @uref{ftp://ftp.dnttm.ro/pub/mysql, FTP} - -@c @item -@c Not ok 20001123; Non-existent (Matt) -@c EMAIL: tim@lbi.ro (Bogdan Surdu) -@c @image{Flags/romania} -@c Romania [Bucharest] @ -@c @uref{http://www.lbi.ro/MySQL/, WWW} -@c @uref{ftp://ftp.lbi.ro/mirrors/ftp.tcx.se, FTP} - -@c @item -@c Removed 20000521 because there is no mirror here. -@c EMAIL: jips@masterd.es (Juan Ignacio Pérez Sacristán) -@c @image{Flags/spain} Spain [MasterD] -@c @uref{http://mysql.masterd.es, WWW} - -@c @item -@c Not ok 20000919; Old site (Matt) -@c EMAIL: Patrik.Karen@sdi.slu.se (Patrik Karen) -@c ftp -> remove old files -@c @image{Flags/sweden} Sweden [Sunet] @ -@c @uref{http://ftp.sunet.se/pub/unix/databases/relational/mysql/, WWW} -@c @uref{ftp://ftp.sunet.se/pub/unix/databases/relational/mysql/, FTP} +We try to make @strong{MySQL} follow the ANSI SQL standard and the +ODBC SQL standard, but in some cases @strong{MySQL} does some things +differently: +@itemize @bullet @item -@c EMAIL: archive@sunsite.cnlab-switch.ch (Thomas Lenggenhager) -@image{Flags/switzerland} Switzerland [Sunsite] @ -@uref{http://sunsite.cnlab-switch.ch/ftp/mirror/mysql/, WWW} -@uref{ftp://sunsite.cnlab-switch.ch/mirror/mysql/, FTP} - -@c @item -@c Not ok 20000919; Non-existent (Matt) -@c simon@oyster.co.uk (Simon Gornall) -@c @image{Flags/great-britain} UK [Oyster/UK] @ -@c @uref{ftp://ftp.oyster.co.uk/pub/mysql, FTP} - +@code{--} is only a comment if followed by a white space. @xref{Missing +comments}. @item -@c EMAIL: gareth@omnipotent.net (Gareth Watts) -@image{Flags/great-britain} UK [Omnipotent/UK] @ -@uref{http://mysql.omnipotent.net/, WWW} -@uref{ftp://mysql.omnipotent.net/, FTP} - +For @code{VARCHAR} columns, trailing spaces are removed when the value is +stored. @xref{Bugs}. @item -@c EMAIL: keet@mordor.plig.net (Christiaan Keet) -@image{Flags/great-britain} UK [PLiG/UK] @ -@uref{http://ftp.plig.org/pub/mysql/, WWW} -@uref{ftp://ftp.plig.org/pub/mysql/, FTP} - +In some cases, @code{CHAR} columns are silently changed to @code{VARCHAR} +columns. @xref{Silent column changes}. @item -@c EMAIL: sean@telekon.co.uk (Sean Gibson) -@image{Flags/great-britain} UK [Telekon Internet/UK] @ -@uref{ftp://ftp.telekon.co.uk/pub/mysql/, FTP} +Privileges for a table are not automatically revoked when you delete a +table. You must explicitly issue a @code{REVOKE} to revoke privileges for +a table. @xref{GRANT, , @code{GRANT}}. +@item +@code{NULL AND FALSE} will evaluate to @code{NULL} and not to @code{FALSE}. +This is because we don't think it's good to have to evaluate a lot of +extra conditions in this case. +@end itemize -@c @item -@c lance@uklinux.net (Lance) -@c @image{Flags/great-britain} UK [uklinux.net] @ -@c @uref{http://mirror.uklinux.net/mysql/, WWW} -@c @uref{ftp://mirror.uklinux.net/pub/mysql/, FTP} +@node ANSI mode, Missing functions, Differences from ANSI, Compatibility +@subsection Running MySQL in ANSI Mode -@c @item -@c unknown -@c @image{Flags/great-britain} UK [MicroMuse] @ -@c @uref{ftp://ftp.micromuse.co.uk/pub/packages/unix/databases/mysql/, FTP} +@cindex running, ANSI mode +@cindex ANSI mode, running -@c @item -@c Not ok 20000919; Old site (Matt) -@c lmjm@icparc.ic.ac.uk (Lee McLoughlin) -@c @image{Flags/great-britain} UK [SunSITE] @ -@c @uref{http://sunsite.org.uk/packages/mysql/, WWW} -@c @uref{ftp://sunsite.org.uk/packages/mysql/, FTP} +If you start @code{mysqld} with the @code{--ansi} option, the following behavior +of @strong{MySQL} changes: +@itemize @bullet @item -@c sander@paco.net (Alexander Ivanov) -@image{Flags/ukraine} Ukraine [PACO] @ -@uref{http://mysql.paco.net.ua, WWW} -@uref{ftp://mysql.paco.net.ua/, FTP} - +@code{||} is string concatenation instead of @code{OR}. @item -@c mizi@alkar.net (Alexander Ryumshin) -@image{Flags/ukraine} Ukraine [ISP Alkar Teleport/Dnepropetrovsk] @ -@uref{http://mysql.dp.ua/, WWW} - +You can have any number of spaces between a function name and the @samp{(}. +This forces all function names to be treated as reserved words. +@item +@samp{"} will be an identifier quote character (like the @strong{MySQL} +@samp{`} quote character) and not a string quote character. +@item +@code{REAL} will be a synonym for @code{FLOAT} instead of a synonym of +@code{DOUBLE}. +@item +The default transaction isolation level is @code{SERIALIZABLE}. +@xref{SET TRANSACTION}. @end itemize -@strong{North America:} +@node Missing functions, Standards, ANSI mode, Compatibility +@subsection Functionality Missing from MySQL -@itemize @bullet -@c @item -@c Not ok 990101 (only to 981007) -@c EMAIL: sysop@polarcom.com (Seamus Venasse) -@c @image{Flags/canada} Canada [Polaris Computing] @ -@c @uref{http://mysql.polaris.ca/, WWW} +@cindex missing functionality +@cindex functionality, missing -@item -@c Ok 980109 -@c EMAIL: wojtek@tryc.on.ca (Wojtjeck Tryc) -@image{Flags/canada} Canada [Tryc] @ -@uref{http://web.tryc.on.ca/mysql/, WWW} +The following functionality is missing in the current version of +@strong{MySQL}. For a prioritized list indicating when new extensions +may be added to @strong{MySQL}, you should consult +@uref{http://www.mysql.com/documentation/manual.php?section=TODO, the +online @strong{MySQL} TODO list}. That is the latest version of the TODO +list in this manual. @xref{TODO}. -@c @item -@c not updated 990218. Added again 990918 -@c Not ok 20000919; Non-existent (Matt) -@c EMAIL: rhooper@cyberus.ca (Roy Hooper) -@c @image{Flags/canada} Canada [Cyberus] @ -@c @uref{http://mysql.cyberus.ca/, WWW} -@c @uref{ftp://mysql.cyberus.ca/, FTP} +@menu +* Missing Sub-selects:: Sub-selects +* Missing SELECT INTO TABLE:: @code{SELECT INTO TABLE} +* Missing Transactions:: Transactions +* Missing Triggers:: Triggers +* Missing Foreign Keys:: Foreign Keys +* Broken Foreign KEY:: +* Missing Views:: Views +* Missing comments:: @samp{--} as the start of a comment +@end menu -@item -@c EMAIL: mleber@he.net (Mike Leber) -@c Added 980312 -@image{Flags/usa} USA [Hurricane Electric/San Jose] @ -@uref{http://mysql.he.net/, WWW} +@node Missing Sub-selects, Missing SELECT INTO TABLE, Missing functions, Missing functions +@subsubsection Sub-selects -@item -@c EMAIL: ask@valueclick.com (Ask Bjoern Hansen) -@c Added 20000925 -@image{Flags/usa} USA [ValueClick, Los Angeles CA] @ -@uref{http://mysql.valueclick.com/, WWW} -@uref{ftp://mysql.valueclick.com/pub/mysql/Downloads/, FTP} +@cindex sub-selects -@c @item -@c Not ok 20000919; Non-existent (Matt) -@c EMAIL: meltzer@icsnet.com (Jeffrey Meltzer) -@c Added 000108 -@c @image{Flags/usa} USA [Meltzer/New York State] @ -@c @uref{ftp://ftp.meltzer.org/pub/mysql/, FTP} +The following will not yet work in @strong{MySQL}: -@c @item -@c No such directory 990830 -@c EMAIL: tps@users.buoy.com (Tim Sailer) -@c @image{Flags/usa} USA [Buoy/New York] @ -@c @uref{http://www.buoy.com/mysql/, WWW} +@example +SELECT * FROM table1 WHERE id IN (SELECT id FROM table2); +SELECT * FROM table1 WHERE id NOT IN (SELECT id FROM table2); +SELECT * FROM table1 WHERE NOT EXISTS (SELECT id FROM table2 where table1.id=table2.id); +@end example -@c @item -@c EMAIL: db@hpnc.com (Douglas Bowyer) -@c Added 980107, removed 981124 because of 'file not found' -@c @image{Flags/usa} USA [Hypernet Communications/Dallas] @ -@c @uref{http://epsilon.hpnc.com/mysql, WWW} +However, in many cases you can rewrite the query without a sub-select: -@c @item @c ********************************** -@c Not updated 980106 -@c EX: twh@iquest.net (Thomas Holt) who no longer works there -@c @image{Flags/usa} USA [IQuest/Indiana] @ -@c @uref{http://mirrors.iquest.net/mysql/, WWW} +@example +SELECT table1.* FROM table1,table2 WHERE table1.id=table2.id; +SELECT table1.* FROM table1 LEFT JOIN table2 ON table1.id=table2.id where table2.id IS NULL +@end example -@c @item @c ********************************** -@c Only a partial mirror so we exclude it from the list -@c EX: lindberg@id.wustl.edu (Fred Lindberg) -@c @image{Flags/usa} USA [Washington University/St. Louis] @ -@c @uref{ftp://ftp.id.wustl.edu/pub/database/mysql/, FTP} +For more complicated subqueries you can often create temporary tables +to hold the subquery. In some cases, however this option will not +work. The most frequently encountered of these cases arises with +@code{DELETE} statements, for which standard SQL does not support joins +(except in sub-selects). For this situation there are two options +available until subqueries are supported by @strong{MySQL}. -@c removed 991111 -> no answer -@c @item -@c EMAIL: andrew@netcasting.net (Andrew Sawyers) -@c @image{Flags/usa} USA [Netcasting/West Coast] @ -@c @uref{ftp://ftp.netcasting.net/pub/mysql/, FTP} +The first option is to use a procedural programming language (such as +Perl or PHP) to submit a @code{SELECT} query to obtain the primary keys +for the records to be deleted, and then use these values to construct +the @code{DELETE} statement (@code{DELETE FROM ... WHERE ... IN (key1, +key2, ...)}). -@c @item -@c No mirror! 980809 David -@c EMAIL: savages@savages.com (Shaun Savage) -@c @image{Flags/usa} USA [Savages/Oregon] @ -@c @uref{http://mysql.savages.com, WWW} +The second option is to use interactive SQL to contruct a set of +@code{DELETE} statements automatically, using the @strong{MySQL} +extension @code{CONCAT()} (in lieu of the standard @code{||} operator). +For example: -@c @item -@c Not ok 20000919; Non-existent (Matt) -@c EMAIL: paul@gina.net (Paul Vining) -@c mirrors ftp.sunet.se -@c @image{Flags/usa} USA [Gina net/Florida] @ -@c @uref{http://www.gina.net/mysql/, WWW} +@example +SELECT CONCAT('DELETE FROM tab1 WHERE pkid = ', tab1.pkid, ';') + FROM tab1, tab2 + WHERE tab1.col1 = tab2.col2; +@end example -@c Out of date 2000-01-08 (Not updated since 1999-10) -@c @item -@c EMAIL: wswanson@pingzero.net (Wylie Swanson) -@c mirrors mysql.org -@c @image{Flags/usa} USA [pingzero/Los Angeles] @ -@c @uref{http://mysql.pingzero.net/, WWW} +You can place this query in a script file and redirect input from it to +the @code{mysql} command-line interpreter, piping its output back to a +second instance of the interpreter: -@item -@c EMAIL: ftpkeeper@mirror.sit.wisc.edu -@image{Flags/usa} USA [Wisconsin University/Wisconsin] @ -@uref{http://mirror.sit.wisc.edu/mysql/, WWW} -@uref{ftp://mirror.sit.wisc.edu/mirrors/mysql/, FTP} +@example +prompt> mysql --skip-column-names mydb < myscript.sql | mysql mydb +@end example -@c @item -@c Not ok 20000919; Out of date (Matt) -@c EMAIL: ftp-admin@digex.net -@c @image{Flags/usa} USA [DIGEX] @ -@c @uref{ftp://ftp.digex.net/pub/packages/database/mysql/, FTP} +@strong{MySQL} only supports @code{INSERT ... SELECT ...} and +@code{REPLACE ... SELECT ...} Independent sub-selects will probably +be available in Version 4.0. You can now use the function @code{IN()} in +other contexts, however. -@item -@c EMAIL: andrew.sawyers@thelinuxstore.com -@image{Flags/usa} USA [LinuxWired/Scottsdale, AZ] @ -@uref{http://mysql.linuxwired.net/, WWW} -@uref{ftp://ftp.linuxwired.net/pub/mirrors/mysql/, FTP} +@node Missing SELECT INTO TABLE, Missing Transactions, Missing Sub-selects, Missing functions +@subsubsection @code{SELECT INTO TABLE} -@c @item -@c EMAIL: dan@surfsouth.com (Dan Muntz) -@c @image{Flags/usa} USA [Venoma.Org/Valdosta, GA] @ -@c @uref{http://mysql.venoma.org/, WWW} +@findex SELECT INTO TABLE -@item -@c EMAIL: hkind@adgrafix.com (Hans Kind) -@image{Flags/usa} USA [adgrafix.com/Boston, MA] @ -@uref{http://mysql.adgrafix.com/, WWW} +@strong{MySQL} doesn't yet support the Oracle SQL extension: +@code{SELECT ... INTO TABLE ...}. @strong{MySQL} supports instead the +ANSI SQL syntax @code{INSERT INTO ... SELECT ...}, which is basically +the same thing. @xref{INSERT SELECT}. -@end itemize +@example +INSERT INTO tblTemp2 (fldID) SELECT tblTemp1.fldOrder_ID FROM tblTemp1 WHERE +tblTemp1.fldOrder_ID > 100; +@end example -@strong{South America:} +Alternatively, you can use @code{SELECT INTO OUTFILE...} or @code{CREATE +TABLE ... SELECT} to solve your problem. -@itemize @bullet +@node Missing Transactions, Missing Triggers, Missing SELECT INTO TABLE, Missing functions +@subsubsection Transactions -@item -@c Added 20001102 -@c EMAIL: nico@bannerlandia.com (Nicolas Moldavsky) -@image{Flags/argentina} Argentina [bannerlandia.com] @ -@uref{http://mysql.bannerlandia.com.ar/, WWW} -@uref{ftp://mysql.bannerlandia.com.ar/mirrors/mysql/, FTP} +@cindex transactions, support -@c @item -@c Not ok 20000919; Non-existent (Matt) -@c EMAIL: gaiser@matrix.com.br (Roberto Gaiser) -@c @image{Flags/brazil} Brazil [Matrix] @ -@c @uref{http://mysql.matrix.com.br, WWW} +As @strong{MySQL} does nowadays support transactions, the following +discussion is only valid if you are only using the non-transaction-safe +table types. @xref{COMMIT}. -@item -@c jpabuyer@vision.cl -@image{Flags/chile} Chile [Vision] @ -@uref{http://mysql.vision.cl/, WWW} +The question is often asked, by the curious and the critical, ``Why is +@strong{MySQL} not a transactional database?'' or ``Why does @strong{MySQL} +not support transactions?'' -@item -@c EMAIL: bartschm@psi.com (Marcelo Bartsch) -@image{Flags/chile} Chile [PSINet] @ -@uref{http://mysql.psinet.cl/, WWW} -@uref{ftp://ftp.psinet.cl/pub/database/mysql/, FTP} +@strong{MySQL} has made a conscious decision to support another paradigm +for data integrity, ``atomic operations.'' It is our thinking and +experience that atomic operations offer equal or even better integrity +with much better performance. We, nonetheless, appreciate and understand +the transactional database paradigm and plan, within the next few releases, +to introduce transaction-safe tables on a per table basis. We will be +giving our users the possibility to decide if they need the speed of +atomic operations or if they need to use transactional features in their +applications. -@item -@c EMAIL: jpabuyer@tecnoera.com (Juan Pablo Abuyeres) -@image{Flags/chile} Chile [Tecnoera] @ -@uref{http://mysql.tecnoera.com/, WWW} +How does one use the features of @strong{MySQL} to maintain rigorous integrity +and how do these features compare with the transactional paradigm? -@c @item -@c Removed 990730 -@c @c EMAIL: dan@amerikanclaris.com (Danilo Lotina F.) -@c @image{Flags/chile} Chile [Amerikanclaris] @ -@c @uref{http://www.labs.amerikanclaris.cl/mysql, WWW} -@c @uref{ftp://ftp.amerikanclaris.cl/pub/mysql, FTP} -@end itemize +First, in the transactional paradigm, if your applications are written +in a way that is dependent on the calling of ``rollback'' instead of +``commit'' in critical situations, then transactions are more +convenient. Moreover, transactions ensure that unfinished updates or +corrupting activities are not committed to the database; the server is +given the opportunity to do an automatic rollback and your database is +saved. -@strong{Asia:} +@strong{MySQL}, in almost all cases, allows you to solve for potential +problems by including simple checks before updates and by running simple +scripts that check the databases for inconsistencies and automatically +repair or warn if such occurs. Note that just by using the +@strong{MySQL} log or even adding one extra log, one can normally fix +tables perfectly with no data integrity loss. -@itemize @bullet -@item -@c EMAIL: mirnshi@263.net (Meng Lingbo) -@image{Flags/china} China [Freecode] @ -@uref{http://www.freecode.net.cn/mirror/mysql/, WWW} +Moreover, fatal transactional updates can be rewritten to be +atomic. In fact,we will go so far as to say that all integrity problems +that transactions solve can be done with @code{LOCK TABLES} or atomic updates, +ensuring that you never will get an automatic abort from the database, +which is a common problem with transactional databases. -@item -@c EMAIL: yusun@atwell.co.jp (yu sun) -@image{Flags/china} China [linuxforum.net] @ -@uref{http://www2.linuxforum.net/mirror/mysql/, WWW} +Not even transactions can prevent all loss if the server goes down. In +such cases even a transactional system can lose data. The difference +between different systems lies in just how small the time-lap is where +they could lose data. No system is 100% secure, only ``secure +enough.'' Even Oracle, reputed to be the safest of transactional +databases, is reported to sometimes lose data in such situations. -@item -@c EMAIL: Vincent_Fong@innovator.com.hk (Vincent Fong) -@image{Flags/china} China [ISL/Hong Kong] @ -@uref{http://mysql.islnet.net, WWW} +To be safe with @strong{MySQL}, you only need to have backups and have +the update logging turned on. With this you can recover from any +situation that you could with any transactional database. It is, of +course, always good to have backups, independent of which database you +use. -@item -@c EMAIL: xcyber@yahoo.com (xcyber) -@image{Flags/china} China [xcyber.org/Hong Kong] @ -@uref{http://mysql.xcyber.org/, WWW} +The transactional paradigm has its benefits and its drawbacks. Many +users and application developers depend on the ease with which they can +code around problems where an abort appears to be, or is necessary, and they +may have to do a little more work with @strong{MySQL} to either think +differently or write more. If you are new to the atomic operations +paradigm, or more familiar or more comfortable with transactions, do not +jump to the conclusion that @strong{MySQL} has not addressed these +issues. Reliability and integrity are foremost in our minds. Recent +estimates indicate that there are more than 1,000,000 @code{mysqld} servers +currently running, many of which are in production environments. We +hear very, very seldom from our users that they have lost any data, and +in almost all of those cases user error is involved. This is, in our +opinion, the best proof of @strong{MySQL}'s stability and reliability. -@c @item -@c Not ok 20010330; Non-existent! (Matt) -@c EMAIL: marquischan@hotmail.com (Marquis Chan) -@c @image{Flags/china} China [TraLand.com/Hong Kong] @ -@c @uref{http://www.traland.com/mysql/, WWW} +Lastly, in situations where integrity is of highest importance, +@strong{MySQL}'s current features allow for transaction-level or better +reliability and integrity. If you lock tables with @code{LOCK TABLES}, all +updates will stall until any integrity checks are made. If you only obtain +a read lock (as opposed to a write lock), then reads and inserts are +still allowed to happen. The new inserted records will not be seen by +any of the clients that have a @code{READ} lock until they release their read +locks. With @code{INSERT DELAYED} you can queue inserts into a local queue, +until the locks are released, without having the client wait for the insert +to complete. @xref{INSERT DELAYED}. -@c @item -@c Not ok 20000919; Non-existent (Matt) -@c EMAIL: george@netfirm.net (Hongsheng Zhu) -@c @image{Flags/china} China [Netfirm] @ -@c @uref{http://mysql.netfirm.net, WWW} +``Atomic,'' in the sense that we mean it, is nothing magical. It only means +that you can be sure that while each specific update is running, no other +user can interfere with it, and there will never be an automatic +rollback (which can happen on transaction based systems if you are not +very careful). @strong{MySQL} also guarantees that there will not be +any dirty reads. You can find some example of how to write atomic updates +in the commit-rollback section. @xref{Commit-rollback}. -@c @item -@c Not ok 20000919; Old site (Matt) -@c EMAIL: ahmlhs@nmsvr.chosun.com (Ho-sun Lee) -@c @image{Flags/south-korea} South Korea [KREONet] @ -@c @uref{http://linux.kreonet.re.kr/mysql/, WWW} +We have thought quite a bit about integrity and performance, and we +believe that our atomic operations paradigm allows for both high +reliability and extremely high performance, on the order of three to +five times the speed of the fastest and most optimally tuned of +transactional databases. We didn't leave out transactions because they +are hard to do. The main reason we went with atomic operations as +opposed to transactions is that by doing this we could apply many speed +optimizations that would not otherwise have been possible. -@item -@c EMAIL: jasper@webiiz.com (Kang, Tae-jin) -@image{Flags/south-korea} South Korea [Webiiz] @ -@uref{http://mysql.webiiz.com/, WWW} +Many of our users who have speed foremost in their minds are not at all +concerned about transactions. For them transactions are not an +issue. For those of our users who are concerned with or have wondered +about transactions vis-a-vis @strong{MySQL}, there is a ``@strong{MySQL} +way'' as we have outlined above. For those where safety is more +important than speed, we recommend them to use the @code{BDB}, +or @code{InnoDB} tables for all their critical data. @xref{Table types}. -@item -@c EMAIL: hollywar@holywar.net (Oh Junseon) -@image{Flags/south-korea} South Korea [PanworldNet] @ -@uref{http://mysql.holywar.net/, WWW} +One final note: We are currently working on a safe replication schema +that we believe to be better than any commercial replication system we +know of. This system will work most reliably under the atomic +operations, non-transactional, paradigm. Stay tuned. -@c @item -@c ftp -> remove old files -@c EX: ahmlhs@nmsvr.chosun.com (Ho-sun Lee) -@c @image{Flags/south-korea} South Korea [KREONet] @ -@c @uref{ftp://linux.kreonet.re.kr/pub/tools/db/mysql/, FTP} +@node Missing Triggers, Missing Foreign Keys, Missing Transactions, Missing functions +@subsubsection Stored Procedures and Triggers -@item -@c Ok 980805 -@c EMAIL: takeshi@SoftAgency.co.jp -@image{Flags/japan} Japan [Soft Agency] @ -@uref{http://www.softagency.co.jp/MySQL, WWW} +@cindex stored procedures and triggers, defined +@cindex procedures, stored +@cindex triggers, stored -@item -@c EMAIL: takafumi@u-aizu.ac.jp (Takafumi Hayashi) -@image{Flags/japan} Japan [u-aizu.ac.jp/Aizu] @ -@uref{ftp://ftp.u-aizu.ac.jp/ftp/pub/dbms/mysql/mysql.com, FTP} -@c @item -@c Ok 980109 Removed 990730 -@c EMAIL: satoshi@HappySize.co.jp (Satoshi TATSUOKA) -@c @image{Flags/japan} Japan [HappySize] @ -@c @uref{http://www.happysize.co.jp/mysql/, WWW} -@c @uref{ftp://ftp.happysize.co.jp/pub/mysql/, FTP} +A stored procedure is a set of SQL commands that can be compiled and stored +in the server. Once this has been done, clients don't need to keep reissuing +the entire query but can refer to the stored procedure. This provides better +performance because the query has to be parsed only once, and less information +needs to be sent between the server and the client. You can also raise the +conceptual level by having libraries of functions in the server. -@c @item -@c Ok 981204 -@c Not ok 20000919; Old site (Matt) -@c EMAIL: hiroyuki@nucba.ac.jp (hiroyuki kurimoto) -@c @image{Flags/japan} Japan [Nagoya Syouka University] @ -@c @uref{http://mirror.nucba.ac.jp/mirror/mysql, WWW} -@c @uref{ftp://mirror.nucba.ac.jp/mirror/mysql, FTP} +A trigger is a stored procedure that is invoked when a particular event +occurs. For example, you can install a stored procedure that is triggered +each time a record is deleted from a transaction table and that automatically +deletes the corresponding customer from a customer table when all his +transactions are deleted. -@c @item -@c Removed 990308 -@c EMAIL: terence@com5.net (Terence Chan) -@c @image{Flags/singapore} Singapore [Com5 Productions] @ -@c @uref{http://mysql.com5.net, WWW} -@c @uref{ftp://ftp.com5.net/pub/mysql, FTP} +The planned update language will be able to +handle stored procedures, but without triggers. Triggers usually slow +down everything, even queries for which they are not needed. -@item -@c EMAIL: csy@hjc.edu.sg -@image{Flags/singapore} Singapore [HJC] @ -@uref{http://mysql.hjc.edu.sg, WWW} -@uref{ftp://ftp.hjc.edu.sg/mysql, FTP} +To see when @strong{MySQL} might get stored procedures, see @ref{TODO}. -@c @item -@c Not ok 20000919; Old site (Matt) -@c 991118: Removed because a user complained about that the page contains -@c nothing about MySQL. 991119: Added again because it is a mirror again -@c EMAIL: dean@ht.net.tw (Dean Lin) -@c @image{Flags/taiwan} Taiwan [HT] @ -@c @uref{http://mysql.ht.net.tw, WWW} +@node Missing Foreign Keys, Broken Foreign KEY, Missing Triggers, Missing functions +@subsubsection Foreign Keys -@item -@c EMAIL: linda@ttn.com.tw (Linda Hu) -@image{Flags/taiwan} Taiwan [TTN] @ -@uref{http://mysql.ttn.net, WWW} - -@c @item -@c Ok 980321 No connect -> removed 990730 -@c EMAIL: tby@ccca.nctu.edu.tw (Bao-Yi Tuang) -@c @image{Flags/taiwan} Taiwan [NCTU] @ -@c @uref{http://mysql.taconet.com.tw, WWW} - -@c @item -@c Out of date 990905 (~2 months) -@c @item @c ********************************** -@c Error 980106 -@c EX: WolfySu@acer.net (Wolfy Su) -@c @image{Flags/taiwan} Taiwan [Acer] @ -@c @uref{http://mysql.acer.net/, WWW} - -@c @item @c ********************************** -@c files to delete -@c EX: service@wownet.net -@c @image{Flags/taiwan} Taiwan [Wownet] @ -@c @uref{ftp://ftp.wownet.net/mysql/, FTP} +@cindex foreign keys +@cindex keys, foreign -@c @item @c ********************************** -@c No contact 980106 -@c EX: serge@oneway.net -@c @image{Flags/taiwan} Taiwan [Oneway] @ -@c @uref{ftp://ftp.oneway.com.tw/pub/mysql/, FTP} +Note that foreign keys in SQL are not used to join tables, but are used +mostly for checking referential integrity (foreign key constraints). If +you want to get results from multiple tables from a @code{SELECT} +statement, you do this by joining tables: -@item -@c Added 20001031 -@c EMAIL: ijliao@php.nctu.edu.tw (Ying-Chieh Liao) -@image{Flags/taiwan} Taiwan [nctu.edu/HsinChu] @ -@uref{http://mysql.nctu.edu.tw/, WWW} +@example +SELECT * from table1,table2 where table1.id = table2.id; +@end example -@end itemize +@xref{JOIN, , @code{JOIN}}. @xref{example-Foreign keys}. -@strong{Australia:} +The @code{FOREIGN KEY} syntax in @strong{MySQL} exists only for compatibility +with other SQL vendors' @code{CREATE TABLE} commands; it doesn't do +anything. The @code{FOREIGN KEY} syntax without @code{ON DELETE ...} is +mostly used for documentation purposes. Some ODBC applications may use this +to produce automatic @code{WHERE} clauses, but this is usually easy to +override. @code{FOREIGN KEY} is sometimes used as a constraint check, but +this check is unnecessary in practice if rows are inserted into the tables in +the right order. @strong{MySQL} only supports these clauses because some +applications require them to exist (regardless of whether or not they +work). -@itemize @bullet -@item -@c Added 980610 -@c EMAIL: jason@dstc.edu.au (Jason Andrade) -@image{Flags/australia} Australia [AARNet/Queensland] @ -@uref{http://mysql.mirror.aarnet.edu.au/, WWW} -@uref{ftp://mysql.mirror.aarnet.edu.au/, FTP} +In @strong{MySQL}, you can work around the problem of @code{ON DELETE +...} not being implemented by adding the appropriate @code{DELETE} statement to +an application when you delete records from a table that has a foreign key. +In practice this is as quick (in some cases quicker) and much more portable +than using foreign keys. -@c @item -@c Added 980805. Removed 000102 'no such directory' -@c EMAIL: sdd@ntccc.tas.gov.au (Scott Donovan) -@c @image{Flags/australia} Australia [Tas] @ -@c @uref{http://ftp.tas.gov.au/mysql, WWW} -@c @uref{ftp://ftp.tas.gov.au/pub/mysql, FTP} +In the near future we will extend the @code{FOREIGN KEY} implementation so +that at least the information will be saved in the table specification file +and may be retrieved by @code{mysqldump} and ODBC. At a later stage we will +implement the foreign key constraints for application that can't easily be +coded to avoid them. -@c @item -@c Ok 980623 -@c Not ok 20000919; Old site (Matt) -@c EMAIL: samh@bluep.com (Sam Hadzajlic) -@c @image{Flags/australia} Australia [Blue Planet/Melbourne] @ -@c @uref{http://mysql.bluep.com/, WWW} -@c removed because ftp was not working 990729 & 30 -@c @uref{ftp://mysql.bluep.com/pub/mirror1/mysql/, FTP} +@menu +* Broken Foreign KEY:: Reasons NOT to use foreign keys constraints +@end menu -@c @item -@c Added 990531 -@c EMAIL: gavin@itworks.com.au (Gavin Cameron) -@c @image{Flags/australia} Australia [ITworks Consulting/Victoria] @ -@c @uref{http://mysql.itworks.com.au, WWW} +@node Broken Foreign KEY, Missing Views, Missing Foreign Keys, Missing functions +@subsubsection Reasons NOT to Use Foreign Keys constraints -@c @item -@c 980610 Only the toplevel dir! -@c EMAIL: lucifer@maths.uq.edu.au (David Conran) -@c @image{Flags/australia} Australia FTP @ -@c @uref{ftp://ftp.sage-au.org.au/pub/database/mysql, [Sage]} -@end itemize +@cindex foreign keys, reasons not to use -@strong{Africa:} +There are so many problems with foreign key constraints that we don't +know where to start: @itemize @bullet - @item -@c EMAIL: ftp-admin@mweb.com (Warren Baker) -@image{Flags/south-africa1} South-Africa [Mweb] @ -@uref{http://www.mysql.mweb.co.za/, WWW} +Foreign key constraints make life very complicated, because the foreign +key definitions must be stored in a database and implementing them would +destroy the whole ``nice approach'' of using files that can be moved, +copied, and removed. @item -@c Ok 981010 -@c EMAIL: oskar@is.co.za (Oskar Pearson) -@image{Flags/south-africa1} South Africa [The Internet Solution/Johannesburg] @ -@uref{ftp://ftp.is.co.za/linux/mysql/, FTP} - -@end itemize - -@c END_OF_MIRROR_LISTING - - -@cindex operating systems, supported -@cindex native thread support -@cindex thread support -@cindex process support -@cindex support, for operating systems -@node Which OS, Which version, Getting MySQL, Installing -@section Operating Systems Supported by MySQL - -We use GNU Autoconf, so it is possible to port @strong{MySQL} to all modern -systems with working Posix threads and a C++ compiler. (To compile only the -client code, a C++ compiler is required but not threads.) We use and develop -the software ourselves primarily on Sun Solaris (Versions 2.5 - 2.7) and -SuSE Linux Version 7.x. - -Note that for many operating systems, the native thread support works only -in the latest versions. @strong{MySQL} has been reported to compile -successfully on the following operating system/thread package combinations: +The speed impact is terrible for @code{INSERT} and @code{UPDATE} +statements, and in this case almost all @code{FOREIGN KEY} constraint +checks are useless because you usually insert records in the right +tables in the right order, anyway. -@itemize @bullet -@item -AIX 4.x with native threads. @xref{IBM-AIX}. -@item -Amiga. -@item -BSDI 2.x with the included MIT-pthreads package. @xref{BSDI}. -@item -BSDI 3.0, 3.1 and 4.x with native threads. @xref{BSDI}. -@item -DEC Unix 4.x with native threads. @xref{Alpha-DEC-UNIX}. -@item -FreeBSD 2.x with the included MIT-pthreads package. @xref{FreeBSD}. -@item -FreeBSD 3.x and 4.x with native threads. @xref{FreeBSD}. -@item -HP-UX 10.20 with the included MIT-pthreads package. @xref{HP-UX 10.20}. @item -HP-UX 11.x with the native threads. @xref{HP-UX 11.x}. -@item -Linux 2.0+ with LinuxThreads 0.7.1+ or @code{glibc} 2.0.7+. @xref{Linux}. -@item -Mac OS X Server. @xref{Mac OS X}. -@item -NetBSD 1.3/1.4 Intel and NetBSD 1.3 Alpha (Requires GNU make). @xref{NetBSD}. -@item -OpenBSD > 2.5 with native therads. OpenBSD < 2.5 with the included -MIT-pthreads package. @xref{OpenBSD}. -@item -OS/2 Warp 3, FixPack 29 and OS/2 Warp 4, FixPack 4. @xref{OS/2}. -@item -SGI Irix 6.x with native threads. @xref{SGI-Irix}. -@item -Solaris 2.5 and above with native threads on SPARC and x86. @xref{Solaris}. -@item -SunOS 4.x with the included MIT-pthreads package. @xref{Solaris}. -@item -SCO OpenServer with a recent port of the FSU Pthreads package. @xref{SCO}. -@item -SCO UnixWare 7.0.1. @xref{SCO Unixware}. -@item -Tru64 Unix -@item -Win95, Win98, NT, and Win2000. @xref{Windows}. -@end itemize - -Note that not all platforms are suited equally well for running -@strong{MySQL}. How well a certain platform is suited for a high-load -mission critical @strong{MySQL} server is determined by the following -factors: +There is also a need to hold locks on many more tables when updating one +table, because the side effects can cascade through the entire database. It's +MUCH faster to delete records from one table first and subsequently delete +them from the other tables. -@itemize -@item -General stability of the thread library. A platform may have excellent -reputation otherwise, but if the thread library is unstable in the code -that is called by @strong{MySQL}, even if -everything else is perfect, @strong{MySQL} will be only as stable as the -thread library. -@item -The ability of the kernel and/or thread library to take advantage of -@strong{SMP} on -multi-processor systems. In other words, when a process creates a thread, it -should be possible for that thread to run on a different CPU than the original -process. -@item -The ability of the kernel and/or the thread library to run many threads which -acquire/release a mutex over a short critical region frequently without -excessive context switches. In other words, if the implementation of -@code{pthread_mutex_lock()} is too anxious to yield CPU, this will hurt -@strong{MySQL} tremendously. If this issue -is not taken care of, adding extra CPUs will actually make @strong{MySQL} -slower. -@item -General file system stability/performance. @item -Ability of the file system to deal with large files at all and deal with them -efficiently, if your tables are big. +You can no longer restore a table by doing a full delete from the table +and then restoring all records (from a new source or from a backup). + @item -Our level of expertise here at @strong{MySQL AB} with the platform. If we know -a platform well, we introduce platform-specific optimizations/fixes enabled at -compile time. We can also provide advice on configuring your system optimally -for @strong{MySQL}. +If you use foreign key constraints you can't dump and restore tables +unless you do so in a very specific order. + @item -The amount of testing of similar configurations we have done internally. +It's very easy to do ``allowed'' circular definitions that make the +tables impossible to re-create each table with a single create statement, +even if the definition works and is usable. + @item -The number of users that have successfully run @strong{MySQL} on that -platform in similar configurations. If this number is high, the chances of -hitting some platform-specific surprise are much smaller. +It's very easy to overlook @code{FOREIGN KEY ... ON DELETE} rules when +one codes an application. It's not unusual that one loses a lot of +important information just because a wrong or misused @code{ON DELETE} rule. @end itemize -Based on the above criteria, the best platforms for running -@strong{MySQL} at this point are x86 with SuSE Linux 7.1, 2.4 kernel and -ReiserFS (or any similar Linux distribution) and Sparc with Solaris 2.7 -or 2.8. FreeBSD comes third, but we really hope it will join the top -club once the thread library is improved. We also hope that at some -point we will be able to include all other platforms on which -@strong{MySQL} compiles, runs ok, but not quite with the same level of -stability and performance, into the top category. This will require some -effort on our part in cooperation with the developers of the OS/library -components @strong{MySQL} depends upon. If you are interested in making -one of those components better, are in a position to influence their -development, and need more detailed instructions on what @strong{MySQL} -needs to run better, send an e-mail to -@email{internals@@lists.mysql.com}. - -Please note that the comparison above is not to say that one OS is better or -worse than the other in general. We are talking about choosing a particular OS -for a dedicated purpose - running @strong{MySQL}, and compare platforms in that -regard only. With this in mind, the result of this comparison -would be different if we included more issues into it. And in some cases, -the reason one OS is better than the other could simply be that we have put -forth more effort into testing on and optimizing for that particular platform. -We are just stating our observations to help you make a -decision on which platform to use @strong{MySQL} on in your setup. +The only nice aspect of @code{FOREIGN KEY} is that it gives ODBC and some +other client programs the ability to see how a table is connected and to use +this to show connection diagrams and to help in building applications. -@cindex MySQL binary distribution -@cindex MySQL source distribution -@cindex release numbers -@cindex version, choosing -@cindex choosing, a MySQL version -@node Which version, Many versions, Which OS, Installing -@section Which MySQL Version to Use +@strong{MySQL} will soon store @code{FOREIGN KEY} definitions so that a +client can ask for and receive an answer about how the original +connection was made. The current @file{.frm} file format does not have +any place for it. At a later stage we will implement the foreign key +constraints for application that can't easily be coded to avoid them. -The first decision to make is whether you want to use the latest development -release or the last stable release: +@node Missing Views, Missing comments, Broken Foreign KEY, Missing functions +@subsubsection Views -@itemize @bullet -@item -Normally, if you are beginning to use @strong{MySQL} for the first time -or trying to port it to some system for which there is no binary -distribution, we recommend going with the stable release (currently -Version @value{mysql_version}. Note that all @strong{MySQL} releases are -checked with the @strong{MySQL} benchmarks and an extensive test suite -before each release. +@cindex views -@item -Otherwise, if you are running an old system and want to upgrade, but -don't want to take chances with a non-seamless upgrade, you should -upgrade to the latest in the same branch you are using (where only the -last version number is newer than yours). We have tried to fix only -fatal bugs and make small, relatively safe changes to that version. -@end itemize +@strong{MySQL} doesn't yet support views, but we plan to implement these +to about 4.1. -The second decision to make is whether you want to use a source -distribution or a binary distribution. In most cases you should probably -use a binary distribution, if one exists for your platform, as this -generally will be easier to install than a source distribution. +Views are mostly useful for letting users access a set of relations as one +table (in read-only mode). Many SQL databases don't allow one to update +any rows in a view, but you have to do the updates in the separate tables. -In the following cases you probably will be better off with a source -installation: +As @strong{MySQL} is mostly used in applications and on web system where +the application writer has full control on the database usage, most of +our users haven't regarded views to be very important. (At least no one +has been interested enough in this to be prepared to finance the +implementation of views). -@itemize @bullet -@item -If you want to install @strong{MySQL} at some explicit location. (The standard -binary distributions are ``ready to run'' at any place, but you may want -to get even more flexibility). +One doesn't need views in @strong{MySQL} to restrict access to columns +as @strong{MySQL} has a very sophisticated privilege +system. @xref{Privilege system}. -@item -To be able to satisfy different user requirements, we are providing two -different binary versions; One compiled with the non-transactional table -handlers, (a small, fast binary), and one configured with the most -important extended options like transaction-safe tables. Both versions -are compiled from the same source distribution. All native @code{MySQL} -clients can connect to both @strong{MySQL} versions. +@node Missing comments, , Missing Views, Missing functions +@subsubsection @samp{--} as the Start of a Comment -The extended @strong{MySQL} binary distribution is marked with the -@code{-max} suffix and is configured with the same options as -@code{mysqld-max}. @xref{mysqld-max, , @code{mysqld-max}}. +@cindex comments, starting +@cindex starting, comments -If you want to use the @code{MySQL-Max} RPM, you must first -install the standard @code{MySQL} RPM. +Some other SQL databases use @samp{--} to start comments. @strong{MySQL} +has @samp{#} as the start comment character, even if the @code{mysql} +command-line tool removes all lines that start with @samp{--}. +You can also use the C comment style @code{/* this is a comment */} with +@strong{MySQL}. +@xref{Comments}. -@item -If you want to configure @code{mysqld} with some extra feature that are NOT in -the standard binary distributions. Here is a list of the most common -extra options that you may want to use: +@strong{MySQL} Version 3.23.3 and above supports the @samp{--} comment style +only if the comment is followed by a space. This is because this +degenerate comment style has caused many problems with automatically +generated SQL queries that have used something like the following code, +where we automatically insert the value of the payment for +@code{!payment!}: -@itemize @bullet -@item --with-berkeley-db -@item --with-innodb -@item --with-raid -@item --with-libwrap -@item --with-named-z-lib (This is done for some of the binaries) -@item --with-debug[=full] -@end itemize +@example +UPDATE tbl_name SET credit=credit-!payment! +@end example -@item -The default binary distribution is normally compiled with support -for all characters sets and should work on a variety of processors from -the same processor family. +What do you think will happen when the value of @code{payment} is negative? -If you want a faster @strong{MySQL} server you may want to recompile it -with support for only the character sets you need, use a better compiler -(like @code{pgcc}) or use compiler options that are better optimized for your -processor. +Because @code{1--1} is legal in SQL, we think it is terrible that +@samp{--} means start comment. -@item -If you have found a bug and reported it to the @strong{MySQL} -development team you will probably receive a patch that you need to apply to -the source distribution to get the bug fixed. +In @strong{MySQL} Version 3.23 you can, however, use: +@code{1-- This is a comment} -@item -If you want to read (and/or modify) the C and C++ code that makes up -@strong{MySQL}, you should get a source distribution. The source code is -always the ultimate manual. Source distributions also contain more -tests and examples than binary distributions. -@end itemize +The following discussion only concerns you if you are running a @strong{MySQL} +version earlier than Version 3.23: -@cindex naming, releases of MySQL -@cindex releases, naming scheme +If you have a SQL program in a text file that contains @samp{--} comments +you should use: -The @strong{MySQL} naming scheme uses release numbers that consist of three -numbers and a suffix. For example, a release name like -@code{mysql-3.21.17-beta} is interpreted like this: +@example +shell> replace " --" " #" < text-file-with-funny-comments.sql \ + | mysql database +@end example -@itemize @bullet -@item -The first number (@code{3}) describes the file format. All Version 3 -releases have the same file format. +instead of the usual: -@item -The second number (@code{21}) is the release level. Normally there are two to -choose from. One is the release/stable branch (currently @code{23}) and the -other is the development branch (currently @code{4.0}). Normally both are -stable, but the development version may have quirks, missing documentation on -new features, or may fail to compile on some systems. +@example +shell> mysql database < text-file-with-funny-comments.sql +@end example -@item -The third number (@code{17}) is the version number within the -release level. This is incremented for each new distribution. Usually you -want the latest version for the release level you have chosen. +You can also edit the command file ``in place'' to change the @samp{--} +comments to @samp{#} comments: -@item -The suffix (@code{beta}) indicates the stability level of the release. -The possible suffixes are: +@example +shell> replace " --" " #" -- text-file-with-funny-comments.sql +@end example -@itemize @minus -@item -@code{alpha} indicates that the release contains some large section of -new code that hasn't been 100% tested. Known bugs (usually there are none) -should be documented in the News section. @xref{News}. There are also new -commands and extensions in most alpha releases. Active development that -may involve major code changes can occur on an alpha release, but everything -will be tested before doing a release. There should be no known bugs in any -@strong{MySQL} release. +Change them back with this command: -@item -@code{beta} means that all new code has been tested. No major new -features that could cause corruption on old code are added. There should -be no known bugs. A version changes from alpha to beta when there -haven't been any reported fatal bugs within an alpha version for at least -a month and we don't plan to add any features that could make any old command -more unreliable. +@example +shell> replace " #" " --" -- text-file-with-funny-comments.sql +@end example -@item -@code{gamma} is a beta that has been around a while and seems to work fine. -Only minor fixes are added. This is what many other companies call a release. +@node Standards, Commit-rollback, Missing functions, Compatibility +@subsection What Standards Does MySQL Follow? -@item -If there is no suffix, it means that the version has been run for a -while at many different sites with no reports of bugs other than -platform-specific bugs. Only critical bug fixes are applied to the -release. This is what we call a stable release. -@end itemize -@end itemize +Entry level SQL92. ODBC levels 0-2. -All versions of @strong{MySQL} are run through our standard tests and -benchmarks to ensure that they are relatively safe to use. Because the -standard tests are extended over time to check for all previously found bugs, -the test suite keeps getting better. +@node Commit-rollback, Bugs, Standards, Compatibility +@subsection How to Cope Without @code{COMMIT}/@code{ROLLBACK} -@cindex releases, testing -@cindex testing, of MySQL releases +@findex COMMIT +@findex ROLLBACK +@cindex transaction-safe tables +@cindex tables, updating +@cindex updating, tables +@cindex @code{BDB} tables +@cindex @code{InnoDB} tables -Note that all releases have been tested at least with: +The following mostly applies only for @code{ISAM}, @code{MyISAM}, and +@code{HEAP} tables. If you only use transaction-safe tables (@code{BDB}, +or @code{InnoDB} tables) in an an update, you can do +@code{COMMIT} and @code{ROLLBACK} also with @strong{MySQL}. +@xref{COMMIT}. -@table @asis -@item An internal test suite -This is part of a production system for a customer. It has many tables with -hundreds of megabytes of data. +The problem with handling @code{COMMIT}-@code{ROLLBACK} efficiently with +the above table types would require a completely different table layout +than @strong{MySQL} uses today. The table type would also need extra +threads that do automatic cleanups on the tables, and the disk usage +would be much higher. This would make these table types about 2-4 times +slower than they are today. -@item The @strong{MySQL} benchmark suite -This runs a range of common queries. It is also a test to see whether the -latest batch of optimizations actually made the code faster. -@xref{Benchmarks}. +For the moment, we prefer implementing the SQL server language (something +like stored procedures). With this you would very seldom really need +@code{COMMIT}-@code{ROLLBACK.} This would also give much better performance. -@item The @code{crash-me} test -This tries to determine what features the database supports and what its -capabilities and limitations are. @xref{Benchmarks}. -@end table +Loops that need transactions normally can be coded with the help of +@code{LOCK TABLES}, and you don't need cursors when you can update records +on the fly. -Another test is that we use the newest @strong{MySQL} version in our internal -production environment, on at least one machine. We have more than 100 -gigabytes of data to work with. +We at TcX had a greater need for a real fast database than a 100% +general database. Whenever we find a way to implement these features without +any speed loss, we will probably do it. For the moment, there are many more +important things to do. Check the TODO for how we prioritize things at +the moment. (Customers with higher levels of support can alter this, so +things may be reprioritized.) -@cindex releases, updating -@cindex updating, releases of MySQL -@node Many versions, Installation layouts, Which version, Installing -@section How and When Updates Are Released +The current problem is actually @code{ROLLBACK}. Without +@code{ROLLBACK}, you can do any kind of @code{COMMIT} action with +@code{LOCK TABLES}. To support @code{ROLLBACK} with the above table +types, @strong{MySQL} would have to be changed to store all old records +that were updated and revert everything back to the starting point if +@code{ROLLBACK} was issued. For simple cases, this isn't that hard to do +(the current @code{isamlog} could be used for this purpose), but it +would be much more difficult to implement @code{ROLLBACK} for +@code{ALTER/DROP/CREATE TABLE}. -@strong{MySQL} is evolving quite rapidly here at @strong{MySQL AB} and we -want to share this with other @strong{MySQL} users. We try to make a release -when we have very useful features that others seem to have a need for. +To avoid using @code{ROLLBACK}, you can use the following strategy: -We also try to help out users who request features that are easy to -implement. We take note of what our licensed users want to have, and -we especially take note of what our extended e-mail supported customers -want and try to help them out. +@enumerate +@item +Use @code{LOCK TABLES ...} to lock all the tables you want to access. +@item +Test conditions. +@item +Update if everything is okay. +@item +Use @code{UNLOCK TABLES} to release your locks. +@end enumerate -No one has to download a new release. The News section will tell you if -the new release has something you really want. @xref{News}. +This is usually a much faster method than using transactions with possible +@code{ROLLBACK}s, although not always. The only situation this solution +doesn't handle is when someone kills the threads in the middle of an +update. In this case, all locks will be released but some of the updates may +not have been executed. -We use the following policy when updating @strong{MySQL}: +You can also use functions to update records in a single operation. +You can get a very efficient application by using the following techniques: @itemize @bullet -@item -For each minor update, the last number in the version string is incremented. -When there are major new features or minor incompatibilities with previous -versions, the second number in the version string is incremented. When the -file format changes, the first number is increased. - -@item -Stable tested releases are meant to appear about 1-2 times a year, but -if small bugs are found, a release with only bug fixes will be released. +@item Modify fields relative to their current value. +@item Update only those fields that actually have changed. +@end itemize -@item -Working releases are meant to appear about every 1-8 weeks. +For example, when we are doing updates to some customer information, we +update only the customer data that has changed and test only that none of +the changed data, or data that depend on the changed data, has changed +compared to the original row. The test for changed data is done with the +@code{WHERE} clause in the @code{UPDATE} statement. If the record wasn't +updated, we give the client a message: "Some of the data you have changed +have been changed by another user". Then we show the old row versus the new +row in a window, so the user can decide which version of the customer record +he should use. -@item -Binary distributions for some platforms will be made by us for major releases. -Other people may make binary distributions for other systems but probably -less frequently. +This gives us something that is similar to column locking but is actually +even better, because we only update some of the columns, using values that +are relative to their current values. This means that typical @code{UPDATE} +statements look something like these: -@item -We usually make patches available as soon as we have located and fixed -small bugs. +@example +UPDATE tablename SET pay_back=pay_back+'relative change'; -@item -For non-critical but annoying bugs, we will make patches available if they -are sent to us. Otherwise we will combine many of them into a larger -patch. +UPDATE customer + SET + customer_date='current_date', + address='new address', + phone='new phone', + money_he_owes_us=money_he_owes_us+'new_money' + WHERE + customer_id=id AND address='old address' AND phone='old phone'; +@end example -@item -If there is, by any chance, a fatal bug in a release we will make a new -release as soon as possible. We would like other companies to do this, -too. -@end itemize +As you can see, this is very efficient and works even if another client has +changed the values in the @code{pay_back} or @code{money_he_owes_us} columns. -The current stable release is Version 3.23; We have already moved active -development to Version 4.0. Bugs will still be fixed in the stable version. -We don't believe in a complete freeze, as this also leaves out bug fixes -and things that ``must be done.'' ``Somewhat frozen'' means that we may -add small things that ``almost surely will not affect anything that's -already working.'' +@findex mysql_insert_id() +@findex LAST_INSERT_ID() +In many cases, users have wanted @code{ROLLBACK} and/or @code{LOCK +TABLES} for the purpose of managing unique identifiers for some tables. This +can be handled much more efficiently by using an @code{AUTO_INCREMENT} column +and either the SQL function @code{LAST_INSERT_ID()} or the C API function +@code{mysql_insert_id()}. @xref{mysql_insert_id, , @code{mysql_insert_id()}}. -@node Installation layouts, Installing binary, Many versions, Installing -@section Installation Layouts +@cindex rows, locking +At @strong{MySQL AB}, we have never had any need for row-level locking +because we have always been able to code around it. Some cases really need +row locking, but they are very few. If you want row-level locking, you +can use a flag column in the table and do something like this: -@cindex installation layouts -@cindex layout of installation -@cindex directory structure, default -@cindex default installation location +@example +UPDATE tbl_name SET row_flag=1 WHERE id=ID; +@end example -This section describes the default layout of the directories created by -installing binary and source distributions. +@strong{MySQL} returns 1 for the number of affected rows if the row was +found and @code{row_flag} wasn't already 1 in the original row. -A binary distribution is installed by unpacking it at the installation -location you choose (typically @file{/usr/local/mysql}) and creates the -following directories in that location: +You can think of it as @strong{MySQL} changed the above query to: -@multitable @columnfractions .3 .7 -@item @strong{Directory} @tab @strong{Contents of directory} -@item @file{bin} @tab Client programs and the @code{mysqld} server -@item @file{data} @tab Log files, databases -@item @file{include} @tab Include (header) files -@item @file{lib} @tab Libraries -@item @file{scripts} @tab @code{mysql_install_db} -@item @file{share/mysql} @tab Error message files -@item @file{sql-bench} @tab Benchmarks -@end multitable +@example +UPDATE tbl_name SET row_flag=1 WHERE id=ID and row_flag <> 1; +@end example -A source distribution is installed after you configure and compile it. By -default, the installation step installs files under @file{/usr/local}, in the -following subdirectories: +@node Bugs, , Commit-rollback, Compatibility +@subsection Known errors and design deficiencies in MySQL -@multitable @columnfractions .3 .7 -@item @strong{Directory} @tab @strong{Contents of directory} -@item @file{bin} @tab Client programs and scripts -@item @file{include/mysql} @tab Include (header) files -@item @file{info} @tab Documentation in Info format -@item @file{lib/mysql} @tab Libraries -@item @file{libexec} @tab The @code{mysqld} server -@item @file{share/mysql} @tab Error message files -@item @file{sql-bench} @tab Benchmarks and @code{crash-me} test -@item @file{var} @tab Databases and log files -@end multitable +@cindex bugs, known +@cindex errors, known +@cindex design, issues +@cindex known errors -Within an installation directory, the layout of a source installation differs -from that of a binary installation in the following ways: +The following problems are known and have a very high priority to get +fixed: @itemize @bullet @item -The @code{mysqld} server is installed in the @file{libexec} -directory rather than in the @file{bin} directory. +@code{ANALYZE TABLE} on a BDB table may in some case make the table +unusable until one has restarted @code{mysqld}. When this happens you will +see errors like the following in the @strong{MySQL} error file: -@item -The data directory is @file{var} rather than @file{data}. +@example +001207 22:07:56 bdb: log_flush: LSN past current end-of-log +@end example @item -@code{mysql_install_db} is installed in the @file{/usr/local/bin} directory -rather than in @file{/usr/local/mysql/scripts}. +Don't execute @code{ALTER TABLE} on a @code{BDB} table on which you are +running not completed multi-statement transactions. (The transaction +will probably be ignored). @item -The header file and library directories are @file{include/mysql} and -@file{lib/mysql} rather than @file{include} and @file{lib}. -@end itemize +@code{ANALYZE TABLE}, @code{OPTIMIZE TABLE} and @code{REPAIR TABLE} may +cause problems on tables for which you are using @code{INSERT DELAYED}. -You can create your own binary installation from a compiled source -distribution by executing the script @file{scripts/make_binary_distribution}. +@item +Doing a @code{LOCK TABLE ..} and @code{FLUSH TABLES ..} doesn't +guarantee that there isn't a half-finished transaction in progress on the +table. -@cindex installing, binary distribution -@cindex binary distributions, installing -@node Installing binary, Installing source, Installation layouts, Installing -@section Installing a MySQL Binary Distribution +@item +BDB tables are a bit slow to open. If you have many BDB tables in a +database, it will take a long time to use the @code{mysql} client on the +database if you are not using the @code{-A} option or if you are using +@code{rehash}. This is especially notable when you have a big table +cache. -@menu -* Linux-RPM:: Linux RPM files -* Building clients:: Building client programs -* Binary install system issues:: System-specific issues -@end menu +@item +Th current replication protocol cannot deal with @code{LOAD DATA INFILE} +and line terminator characters of more than 1 character. +@end itemize -You need the following tools to install a @strong{MySQL} binary distribution: +The following problems are known and will be fixed in due time: @itemize @bullet @item -GNU @code{gunzip} to uncompress the distribution. - +For the moment @code{MATCH} only works with @code{SELECT} statements. @item -A reasonable @code{tar} to unpack the distribution. GNU @code{tar} is -known to work. Sun @code{tar} is known to have problems. -@end itemize +When using @code{SET CHARACTER SET}, one can't use translated +characters in database, table and column names. +@item +@code{DELETE FROM merge_table} used without a @code{WHERE} +will only clear the mapping for the table, not delete everything in the +mapped tables +@item +You cannot build in another directory when using +MIT-pthreads. Because this requires changes to MIT-pthreads, we are not +likely to fix this. +@item +@code{BLOB} values can't ``reliably'' be used in @code{GROUP BY} or +@code{ORDER BY} or @code{DISTINCT}. Only the first @code{max_sort_length} +bytes (default 1024) are used when comparing @code{BLOB}bs in these cases. +This can be changed with the @code{-O max_sort_length} option to +@code{mysqld}. A workaround for most cases is to use a substring: +@code{SELECT DISTINCT LEFT(blob,2048) FROM tbl_name}. +@item +Calculation is done with @code{BIGINT} or @code{DOUBLE} (both are +normally 64 bits long). It depends on the function which precision one +gets. The general rule is that bit functions are done with @code{BIGINT} +precision, @code{IF}, and @code{ELT()} with @code{BIGINT} or @code{DOUBLE} +precision and the rest with @code{DOUBLE} precision. One should try to +avoid using bigger unsigned long long values than 63 bits +(9223372036854775807) for anything else than bit fields! +@item +All string columns, except @code{BLOB} and @code{TEXT} columns, automatically +have all trailing spaces removed when retrieved. For @code{CHAR} types this +is okay, and may be regarded as a feature according to ANSI SQL92. The bug is +that in @strong{MySQL}, @code{VARCHAR} columns are treated the same way. +@item +You can only have up to 255 @code{ENUM} and @code{SET} columns in one table. +@item +@code{safe_mysqld} re-directs all messages from @code{mysqld} to the +@code{mysqld} log. One problem with this is that if you execute +@code{mysqladmin refresh} to close and reopen the log, +@code{stdout} and @code{stderr} are still redirected to the old log. +If you use @code{--log} extensively, you should edit @code{safe_mysqld} to +log to @file{'hostname'.err} instead of @file{'hostname'.log} so you can +easily reclaim the space for the old log by deleting the old one and +executing @code{mysqladmin refresh}. +@item +In the @code{UPDATE} statement, columns are updated from left to right. +If you refer to an updated column, you will get the updated value instead of the +original value. For example: +@example +mysql> UPDATE tbl_name SET KEY=KEY+1,KEY=KEY+1; +@end example +will update @code{KEY} with @code{2} instead of with @code{1}. +@item +You can't use temporary tables more than once in the same query. +For example, the following doesn't work. -@cindex RPM, defined -@cindex RedHat Package Manager -An alternative installation method under Linux is to use RPM (RedHat Package -Manager) distributions. @xref{Linux-RPM}. +@example +select * from temporary_table, temporary_table as t2; +@end example -@c texi2html fails to split chapters if I use strong for all of this. -If you run into problems, @strong{PLEASE ALWAYS USE} @code{mysqlbug} when -posting questions to @email{mysql@@lists.mysql.com}. Even if the problem -isn't a bug, @code{mysqlbug} gathers system information that will help others -solve your problem. By not using @code{mysqlbug}, you lessen the likelihood -of getting a solution to your problem! You will find @code{mysqlbug} in the -@file{bin} directory after you unpack the distribution. @xref{Bug reports}. +@item +@code{RENAME} doesn't work with @code{TEMPORARY} tables. -@cindex commands, for binary distribution -The basic commands you must execute to install and use a @strong{MySQL} -binary distribution are: +@item +The optimizer may handle @code{DISTINCT} differently if you are using +'hidden' columns in a join or not. In a join, hidden columns are +counted as part of the result (even if they are not shown) while in +normal queries hidden columns doesn't participate in the @code{DISTINCT} +comparison. We will probably change this in the future to never compare +the hidden columns when executing @code{DISTINCT} + +An example of this is: @example -shell> groupadd mysql -shell> useradd -g mysql mysql -shell> cd /usr/local -shell> gunzip < /path/to/mysql-VERSION-OS.tar.gz | tar xvf - -shell> ln -s mysql-VERSION-OS mysql -shell> cd mysql -shell> scripts/mysql_install_db -shell> chown -R root /usr/local/mysql -shell> chown -R mysql /usr/local/mysql/data -shell> chgrp -R mysql /usr/local/mysql -shell> chown -R root /usr/local/mysql/bin -shell> bin/safe_mysqld --user=mysql & +SELECT DISTINCT mp3id FROM band_downloads WHERE userid = 9 ORDER BY id +DESC; + +and + +SELECT DISTINCT band_downloads.mp3id, FROM band_downloads,band_mp3 +WHERE band_downloads.userid = 9 AND band_mp3.id = band_downloads.mp3id +ORDER BY band_downloads.id DESC; @end example -@cindex adding, new users -@cindex new users, adding -@cindex users, adding +In the second case you may in @strong{MySQL} 3.23.x get two identical rows +in the result set (because the hidden 'id' column may differ). -You can add new users using the @code{bin/mysql_setpermission} script if -you install the @code{DBI} and @code{Msql-Mysql-modules} Perl modules. +Note that the this only happens for queries where you don't have the +ORDER BY columns in the result, something that is you are not allowed +to do in ANSI SQL. -A more detailed description follows. +@item +Because @strong{MySQL} allows you to work with table types that doesn't +support transactions (and thus can't @code{rollback} data) some things +behaves a little different in @strong{MySQL} than in other SQL servers: +(This is just to ensure that @strong{MySQL} never need to do a rollback +for a SQL command). This may be a little awkward at times as column +Because @strong{MySQL} allows you to work with table types that don't +support transactions, and thus can't @code{rollback} data, some things +behave a little differently in @strong{MySQL} than in other SQL servers. +This is just to ensure that @strong{MySQL} never need to do a rollback +for a SQL command. This may be a little awkward at times as column +values must be checked in the application, but this will actually give +you a nice speed increase as it allows @strong{MySQL} to do some +optimizations that otherwise would be very hard to do. -To install a binary distribution, follow the steps below, then proceed -to @ref{Post-installation}, for post-installation setup and testing: +If you set a column to an incorrect value, @strong{MySQL} will, instead of +doing a rollback, store the @code{best possible value} in the column: -@enumerate +@itemize @bullet @item -Pick the directory under which you want to unpack the distribution, and move -into it. In the example below, we unpack the distribution under -@file{/usr/local} and create a directory @file{/usr/local/mysql} into which -@strong{MySQL} is installed. (The following instructions therefore assume -you have permission to create files in @file{/usr/local}. If that directory -is protected, you will need to perform the installation as @code{root}.) +If you try to store a value outside the range in a numerical column, +@strong{MySQL} will instead store the smallest or biggest possible value in +the column. +@item +If you try to store a string that doesn't start with a number into a +numerical column, @strong{MySQL} will store 0 into it. +@item +If you try to store @code{NULL} into a column that doesn't take +@code{NULL} values, @strong{MySQL} will store 0 or @code{''} (empty +string) in it instead. (This behavior can, however, be changed with the +-DDONT_USE_DEFAULT_FIELDS compile option). +@item +@strong{MySQL} allows you to store some wrong date values into +@code{DATE} and @code{DATETIME} columns. (Like 2000-02-31 or 2000-02-00). +If the date is totally wrong, @strong{MySQL} will store the special +0000-00-00 date value in the column. +@item +If you set an @code{enum} to an unsupported value, it will be set to +the error value 'empty string', with numeric value 0. +@end itemize @item -Obtain a distribution file from one of the sites listed in -@ref{Getting MySQL, , Getting @strong{MySQL}}. +If you execute a @code{PROCEDURE} on a query that returns an empty set, +in some cases the @code{PROCEDURE} will not transform the columns. +@item +Creation of a table of type @code{MERGE} doesn't check if the underlying +tables are of compatible types. +@item +@strong{MySQL} can't yet handle @code{NaN}, @code{-Inf} and @code{Inf} +values in double. Using these will cause problems when trying to export +and import data. We should as an intermediate solution change @code{NaN} to +@code{NULL} (if possible) and @code{-Inf} and @code{Inf} to the +Minimum respective maximum possible @code{double} value. +@item +@code{LIMIT} on negative numbers are treated as big positive numbers. +@end itemize -@strong{MySQL} binary distributions are provided as compressed @code{tar} -archives and have names like @file{mysql-VERSION-OS.tar.gz}, where -@code{VERSION} is a number (for example, @code{3.21.15}), and @code{OS} -indicates the type of operating system for which the distribution is intended -(for example, @code{pc-linux-gnu-i586}). +The following are known bugs in earlier versions of @strong{MySQL}: +@itemize @bullet @item -If you see a binary distribution marked with the @code{-max} prefix, this -means that the binary has support for transaction-safe tables and other -features. @xref{mysqld-max, , @code{mysqld-max}}. Note that all binaries -are built from the same @strong{MySQL} source distribution. +You can get a hung thread if you do a @code{DROP TABLE} on a table that is +one among many tables that is locked with @code{LOCK TABLES}. @item -Add a user and group for @code{mysqld} to run as: - -@example -shell> groupadd mysql -shell> useradd -g mysql mysql -@end example - -These commands add the @code{mysql} group and the @code{mysql} user. The -syntax for @code{useradd} and @code{groupadd} may differ slightly on different -versions of Unix. They may also be called @code{adduser} and @code{addgroup}. -You may wish to call the user and group something else instead of @code{mysql}. +In the following case you can get a core dump: +@enumerate +@item +Delayed insert handler has pending inserts to a table. +@item +@code{LOCK table} with @code{WRITE} +@item +@code{FLUSH TABLES} +@end enumerate @item -Change into the intended installation directory: +Before @strong{MySQL} Version 3.23.2 an @code{UPDATE} that updated a key with +a @code{WHERE} on the same key may have failed because the key was used to +search for records and the same row may have been found multiple times: @example -shell> cd /usr/local +UPDATE tbl_name SET KEY=KEY+1 WHERE KEY > 100; @end example -@item -Unpack the distribution and create the installation directory: +A workaround is to use: @example -shell> gunzip < /path/to/mysql-VERSION-OS.tar.gz | tar xvf - -shell> ln -s mysql-VERSION-OS mysql +mysql> UPDATE tbl_name SET KEY=KEY+1 WHERE KEY+0 > 100; @end example -The first command creates a directory named @file{mysql-VERSION-OS}. The -second command makes a symbolic link to that directory. This lets you refer -more easily to the installation directory as @file{/usr/local/mysql}. - +This will work because @strong{MySQL} will not use index on expressions in +the @code{WHERE} clause. @item -Change into the installation directory: +Before @strong{MySQL} Version 3.23, all numeric types where treated as fixed-point +fields. That means you had to specify how many decimals a floating-point +field shall have. All results were returned with the correct number of +decimals. +@end itemize -@example -shell> cd mysql -@end example +For platform-specific bugs, see the sections about compiling and porting. -You will find several files and subdirectories in the @code{mysql} directory. -The most important for installation purposes are the @file{bin} and -@file{scripts} subdirectories. +@node Comparisons, TODO, Compatibility, Introduction +@section How MySQL Compares to Other Databases -@table @file -@item bin -@tindex PATH environment variable -@tindex environment variable, PATH -This directory contains client programs and the server -You should add the full pathname of this directory to your -@code{PATH} environment variable so that your shell finds the @strong{MySQL} -programs properly. @xref{Environment variables}. +@cindex databases, MySQL vs. others +@cindex comparisons, MySQL vs. others -@item scripts -This directory contains the @code{mysql_install_db} script used to initialize -the @code{mysql} database containing the grant tables that store the server -access permissions. -@end table +@menu +* Compare mSQL:: How @strong{MySQL} compares to @code{mSQL} +* Protocol differences:: +* Compare PostgreSQL:: How @strong{MySQL} compares with PostgreSQL +* MySQL-PostgreSQL features:: +@end menu -@item -If you would like to use @code{mysqlaccess} and have the @strong{MySQL} -distribution in some non-standard place, you must change the location where -@code{mysqlaccess} expects to find the @code{mysql} client. Edit the -@file{bin/mysqlaccess} script at approximately line 18. Search for a line -that looks like this: +This section compares @strong{MySQL} to other popular databases. -@example -$MYSQL = '/usr/local/bin/mysql'; # path to mysql executable -@end example +This section has been written by the @strong{MySQL} developers, so it +should be read with that in mind. There are no factual errors contained +in this section that we know of. If you find something which you believe +to be an error, please contact us about it at @email{docs@@mysql.com}. -Change the path to reflect the location where @code{mysql} actually is -stored on your system. If you do not do this, you will get a @code{Broken -pipe} error when you run @code{mysqlaccess}. +For a list of all supported limits, functions, and types, see the +@code{crash-me} Web page at +@uref{http://www.mysql.com/information/crash-me.php}. -@item -Create the @strong{MySQL} grant tables (necessary only if you haven't -installed @strong{MySQL} before): -@example -shell> scripts/mysql_install_db -@end example +@node Compare mSQL, Protocol differences, Comparisons, Comparisons +@subsection How MySQL Compares to @code{mSQL} -Note that @strong{MySQL} versions older than Version 3.22.10 started the -@strong{MySQL} server when you run @code{mysql_install_db}. This is no -longer true! +@table @strong +@item Performance -@item -Change ownership of binaries to @code{root} and ownership of the data -directory to the user that you will run @code{mysqld} as: +For a true comparison of speed, consult the growing @strong{MySQL} benchmark +suite. @xref{Benchmarks}. -@example -shell> chown -R root /usr/local/mysql -shell> chown -R mysql /usr/local/mysql/data -shell> chgrp -R mysql /usr/local/mysql -@end example - -The first command changes the @code{owner} attribute of the files to the -@code{root} user, the second one changes the @code{owner} attribute of the -data directory to the @code{mysql} user, and the third one changes the -@code{group} attribute to the @code{mysql} group. +Because there is no thread creation overhead, a small parser, few features, and +simple security, @code{mSQL} should be quicker at: +@itemize @bullet @item -If you want to install support for the Perl @code{DBI}/@code{DBD} interface, -see @ref{Perl support}. - +Tests that perform repeated connects and disconnects, running a very simple +query during each connection. @item -If you would like @strong{MySQL} to start automatically when you boot your -machine, you can copy @code{support-files/mysql.server} to the location where -your system has its startup files. More information can be found in the -@code{support-files/mysql.server} script itself and in -@ref{Automatic start}. - -@end enumerate - -After everything has been unpacked and installed, you should initialize -and test your distribution. - -You can start the @strong{MySQL} server with the following command: - -@example -shell> bin/safe_mysqld --user=mysql & -@end example - -@xref{safe_mysqld, , @code{safe_mysqld}}. - -@xref{Post-installation}. - -@cindex RPM file -@cindex RedHat Package Manager -@c This node name is special -@node Linux-RPM, Building clients, Installing binary, Installing binary -@subsection Linux RPM Notes - -The recommended way to install @strong{MySQL} on Linux is by using an RPM -file. The @strong{MySQL} RPMs are currently being built on a RedHat Version -6.2 system but should work on other versions of Linux that support @code{rpm} -and use @code{glibc}. +@code{INSERT} operations into very simple tables with few columns and keys. +@item +@code{CREATE TABLE} and @code{DROP TABLE}. +@item +@code{SELECT} on something that isn't an index. (A table scan is very +easy.) +@end itemize -If you have problems with an RPM file, for example, if you receive the error -``@code{Sorry, the host 'xxxx' could not be looked up}'', see -@ref{Binary notes-Linux}. +Because these operations are so simple, it is hard to be better at them when +you have a higher startup overhead. After the connection is established, +@strong{MySQL} should perform much better. -The RPM files you may want to use are: +On the other hand, @strong{MySQL} is much faster than @code{mSQL} (and +most other SQL implementations) on the following: @itemize @bullet -@item @code{MySQL-VERSION.i386.rpm} - -The @strong{MySQL} server. You will need this unless you only want to -connect to a @strong{MySQL} server running on another machine. - -@item @code{MySQL-client-VERSION.i386.rpm} +@item +Complex @code{SELECT} operations. +@item +Retrieving large results (@strong{MySQL} has a better, faster, and safer +protocol). +@item +Tables with variable-length strings, because @strong{MySQL} has more efficient +handling and can have indexes on @code{VARCHAR} columns. +@item +Handling tables with many columns. +@item +Handling tables with large record lengths. +@item +@code{SELECT} with many expressions. +@item +@code{SELECT} on large tables. +@item +Handling many connections at the same time. @strong{MySQL} is fully +multi-threaded. Each connection has its own thread, which means that +no thread has to wait for another (unless a thread is modifying +a table another thread wants to access). In @code{mSQL}, once one connection +is established, all others must wait until the first has finished, regardless +of whether the connection is running a query that is short or long. When the +first connection terminates, the next can be served, while all the others wait +again, etc. +@item +Joins. +@code{mSQL} can become pathologically slow if you change the order of tables +in a @code{SELECT}. In the benchmark suite, a time more than 15000 times +slower than @strong{MySQL} was seen. This is due to @code{mSQL}'s lack of a +join optimizer to order tables in the optimal order. However, if you put the +tables in exactly the right order in @code{mSQL}2 and the @code{WHERE} is +simple and uses index columns, the join will be relatively fast! +@xref{Benchmarks}. +@item +@code{ORDER BY} and @code{GROUP BY}. +@item +@code{DISTINCT}. +@item +Using @code{TEXT} or @code{BLOB} columns. +@end itemize -The standard @strong{MySQL} client programs. You probably always want to -install this package. +@item SQL Features -@item @code{MySQL-bench-VERSION.i386.rpm} +@itemize @bullet +@item @code{GROUP BY} and @code{HAVING}. +@code{mSQL} does not support @code{GROUP BY} at all. +@strong{MySQL} supports a full @code{GROUP BY} with both @code{HAVING} and +the following functions: @code{COUNT()}, @code{AVG()}, @code{MIN()}, +@code{MAX()}, @code{SUM()}, and @code{STD()}. @code{COUNT(*)} is optimized to +return very quickly if the @code{SELECT} retrieves from one table, no other +columns are retrieved, and there is no @code{WHERE} clause. @code{MIN()} and +@code{MAX()} may take string arguments. -Tests and benchmarks. Requires Perl and msql-mysql-modules RPMs. +@item @code{INSERT} and @code{UPDATE} with calculations. +@strong{MySQL} can do calculations in an @code{INSERT} or @code{UPDATE}. +For example: +@example +mysql> UPDATE SET x=x*10+y WHERE x<20; +@end example -@item @code{MySQL-devel-VERSION.i386.rpm} +@item Aliasing. +@strong{MySQL} has column aliasing. -Libraries and include files needed if you want to compile other -@strong{MySQL} clients, such as the Perl modules. +@item Qualifying column names. +In @strong{MySQL}, if a column name is unique among the tables used in a +query, you do not have to use the full qualifier. -@item @code{MySQL-VERSION.src.rpm} +@item @code{SELECT} with functions. +@strong{MySQL} has many functions (too many to list here; see @ref{Functions}). -This contains the source code for all of the above packages. It can also -be used to try to build RPMs for other architectures (for example, Alpha -or SPARC). @end itemize -To see all files in an RPM package, run: -@example -shell> rpm -qpl MySQL-VERSION.i386.rpm -@end example +@item Disk Space Efficiency +That is, how small can you make your tables? -To perform a standard minimal installation, run: +@strong{MySQL} has very precise types, so you can create tables that take +very little space. An example of a useful @strong{MySQL} datatype is the +@code{MEDIUMINT} that is 3 bytes long. If you have 100,000,000 records, +saving even one byte per record is very important. -@example -shell> rpm -i MySQL-VERSION.i386.rpm MySQL-client-VERSION.i386.rpm -@end example +@code{mSQL2} has a more limited set of column types, so it is +more difficult to get small tables. +@item Stability +This is harder to judge objectively. For a discussion of @strong{MySQL} +stability, see @ref{Stability}. -To install just the client package, run: +We have no experience with @code{mSQL} stability, so we cannot say +anything about that. -@example -shell> rpm -i MySQL-client-VERSION.i386.rpm -@end example +@item Price +Another important issue is the license. @strong{MySQL} has a +more flexible license than @code{mSQL}, and is also less expensive than +@code{mSQL}. Whichever product you choose to use, remember to at least +consider paying for a license or e-mail support. (You are required to get +a license if you include @strong{MySQL} with a product that you sell, +of course.) -The RPM places data in @file{/var/lib/mysql}. The RPM also creates the -appropriate entries in @file{/etc/rc.d/} to start the server automatically -at boot time. (This means that if you have performed a previous -installation, you may want to make a copy of your previously installed -@strong{MySQL} startup file if you made any changes to it, so you don't lose -your changes.) +@item Perl Interfaces +@strong{MySQL} has basically the same interfaces to Perl as @code{mSQL} with +some added features. -After installing the RPM file(s), the @code{mysqld} daemon should be running -and you should now be able to start using @strong{MySQL}. -@xref{Post-installation}. +@item JDBC (Java) +@strong{MySQL} currently has a lot of different JDBC drivers: +@itemize @bullet +@item +The mm driver: A type 4 JDBC driver by Mark Matthews +@email{mmatthew@@ecn.purdue.edu}. This is released under the LGPL. +@item +The Resin driver. This is a commercial JDBC driver released under open +source. @uref{http://www.caucho.com/projects/jdbc-mysql/index.xtp} +@item +The gwe driver: A Java interface by GWE technologies (not supported anymore). +@item +The jms driver: An improved gwe driver by Xiaokun Kelvin ZHU +@email{X.Zhu@@brad.ac.uk} (not supported anymore). +@item +The twz driver: A type 4 JDBC driver by Terrence W. Zellers +@email{zellert@@voicenet.com}. This is commercial but is free for private +and educational use (not supported anymore). +@end itemize -If something goes wrong, you can find more information in the binary -installation chapter. @xref{Installing binary}. +The recommended driver is the mm driver. The Resin driver may also be +good (at least the benchmarks looks good), but we haven't received that much +information about this yet. -@cindex client programs, building -@cindex linking -@cindex building, client programs -@cindex programs, client -@node Building clients, Binary install system issues, Linux-RPM, Installing binary -@subsection Building Client Programs +We know that @code{mSQL} has a JDBC driver, but we have too little experience +with it to compare. -If you compile @strong{MySQL} clients that you've written yourself or that -you obtain from a third party, they must be linked using the -@code{-lmysqlclient -lz} option on the link command. You may also need to -specify a @code{-L} option to tell the linker where to find the library. For -example, if the library is installed in @file{/usr/local/mysql/lib}, use -@code{-L/usr/local/mysql/lib -lmysqlclient -lz} on the link command. +@item Rate of Development +@strong{MySQL} has a very small team of developers, but we are quite +used to coding C and C++ very rapidly. Because threads, functions, +@code{GROUP BY}, and so on are still not implemented in @code{mSQL}, it +has a lot of catching up to do. To get some perspective on this, you +can view the @code{mSQL} @file{HISTORY} file for the last year and +compare it with the News section of the @strong{MySQL} Reference Manual +(@pxref{News}). It should be pretty obvious which one has developed +most rapidly. -For clients that use @strong{MySQL} header files, you may need to specify a -@code{-I} option when you compile them (for example, -@code{-I/usr/local/mysql/include}), so the compiler can find the header -files. +@item Utility Programs +Both @code{mSQL} and @strong{MySQL} have many interesting third-party +tools. Because it is very easy to port upward (from @code{mSQL} to +@strong{MySQL}), almost all the interesting applications that are available for +@code{mSQL} are also available for @strong{MySQL}. -@node Binary install system issues, , Building clients, Installing binary -@subsection System-specific Issues +@strong{MySQL} comes with a simple @code{msql2mysql} program that fixes +differences in spelling between @code{mSQL} and @strong{MySQL} for the +most-used C API functions. +For example, it changes instances of @code{msqlConnect()} to +@code{mysql_connect()}. Converting a client program from @code{mSQL} to +@strong{MySQL} usually takes a couple of minutes. +@end table @menu -* Binary notes-Linux:: Linux notes for binary distribution -* Binary notes-HP-UX:: HP-UX notes for binary distribution +* Using mSQL tools:: How to convert @code{mSQL} tools for @strong{MySQL} @end menu -The following sections indicate some of the issues that have been observed -on particular systems when installing @strong{MySQL} from a binary -distribution or from RPM files. - -@cindex binary distributions, on Linux -@cindex Linux, binary distribution -@node Binary notes-Linux, Binary notes-HP-UX, Binary install system issues, Binary install system issues -@subsubsection Linux Notes for Binary Distributions - -@strong{MySQL} needs at least Linux Version 2.0. - -The binary release is linked with @code{-static}, which means you do not -normally need to worry about which version of the system libraries you -have. You need not install LinuxThreads, either. A program linked with -@code{-static} is slightly bigger than a dynamically linked program but -also slightly faster (3-5%). One problem, however, is that you can't use -user-definable functions (UDFs) with a statically linked program. If -you are going to write or use UDF functions (this is something only for -C or C++ programmers), you must compile @strong{MySQL} yourself, using -dynamic linking. - -If you are using a @code{libc}-based system (instead of a @code{glibc2} -system), you will probably get some problems with hostname resolving and -@code{getpwnam()} with the binary release. (This is because @code{glibc} -unfortunately depends on some external libraries to resolve hostnames -and @code{getpwent()}, even when compiled with @code{-static}). In this -case you probably get the following error message when you run -@code{mysql_install_db}: +@node Using mSQL tools, , Compare mSQL, Compare mSQL +@subsubsection How to Convert @code{mSQL} Tools for MySQL -@example -Sorry, the host 'xxxx' could not be looked up -@end example +@cindex MySQL tools, conversion +@cindex converting, tools +@cindex tools, converting -or the following error when you try to run @code{mysqld} with the @code{--user} -option: +According to our experience, it would just take a few hours to convert tools +such as @code{msql-tcl} and @code{msqljava} that use the +@code{mSQL} C API so that they work with the @strong{MySQL} C API. -@example -getpwnam: No such file or directory -@end example +The conversion procedure is: -You can solve this problem in one of the following ways: +@enumerate +@item +Run the shell script @code{msql2mysql} on the source. This requires the +@code{replace} program, which is distributed with @strong{MySQL}. +@item +Compile. +@item +Fix all compiler errors. +@end enumerate +Differences between the @code{mSQL} C API and the @strong{MySQL} C API are: @itemize @bullet @item -Get a @strong{MySQL} source distribution (an RPM or the @code{tar.gz} -distribution) and install this instead. +@strong{MySQL} uses a @code{MYSQL} structure as a connection type (@code{mSQL} +uses an @code{int}). @item -Execute @code{mysql_install_db --force}; This will not execute the -@code{resolveip} test in @code{mysql_install_db}. The downside is that -you can't use host names in the grant tables; you must use IP numbers -instead (except for @code{localhost}). If you are using an old @strong{MySQL} -release that doesn't support @code{--force}, you have to remove the -@code{resolveip} test in @code{mysql_install} with an editor. +@code{mysql_connect()} takes a pointer to a @code{MYSQL} structure as a +parameter. It is easy to define one globally or to use @code{malloc()} to get +one. +@code{mysql_connect()} also takes two parameters for specifying the user and +password. You may set these to @code{NULL, NULL} for default use. @item -Start @code{mysqld} with @code{su} instead of using @code{--user}. +@code{mysql_error()} takes the @code{MYSQL} structure as a parameter. Just add +the parameter to your old @code{msql_error()} code if you are porting old code. +@item +@strong{MySQL} returns an error number and a text error message for all +errors. @code{mSQL} returns only a text error message. +@item +Some incompatibilities exist as a result of @strong{MySQL} supporting +multiple connections to the server from the same process. @end itemize -The Linux-Intel binary and RPM releases of @strong{MySQL} are configured -for the highest possible speed. We are always trying to use the fastest -stable compiler available. - -@strong{MySQL} Perl support requires Version Perl 5.004_03 or newer. - -On some Linux 2.2 versions, you may get the error @code{Resource -temporarily unavailable} when you do a lot of new connections to a -@code{mysqld} server over TCP/IP. - -The problem is that Linux has a delay between when you close a TCP/IP -socket and until this is actually freed by the system. As there is only -room for a finite number of TCP/IP slots, you will get the above error if -you try to do too many new TCP/IP connections during a small time, like -when you run the @strong{MySQL} @file{test-connect} benchmark over -TCP/IP. +@cindex communications protocols +@cindex mSQL vs. MySQL +@node Protocol differences, Compare PostgreSQL, Compare mSQL, Comparisons +@subsection How @code{mSQL} and MySQL Client/Server Communications Protocols Differ -We have mailed about this problem a couple of times to different Linux -mailing lists but have never been able to resolve this properly. +There are enough differences that it is impossible (or at least not easy) +to support both. -The only known 'fix' to this problem is to use persistent connections in -your clients or use sockets, if you are running the database server -and clients on the same machine. We hope that the @code{Linux 2.4} -kernel will fix this problem in the future. +The most significant ways in which the @strong{MySQL} protocol differs +from the @code{mSQL} protocol are listed below: -@cindex HP-UX, binary distribution -@cindex binary distributions, on HP-UX -@node Binary notes-HP-UX, , Binary notes-Linux, Binary install system issues -@subsubsection HP-UX Notes for Binary Distributions +@itemize @bullet +@item +A message buffer may contain many result rows. +@item +The message buffers are dynamically enlarged if the query or the +result is bigger than the current buffer, up to a configurable server and +client limit. +@item +All packets are numbered to catch duplicated or missing packets. +@item +All column values are sent in ASCII. The lengths of columns and rows are sent +in packed binary coding (1, 2, or 3 bytes). +@item +@strong{MySQL} can read in the result unbuffered (without having to store the +full set in the client). +@item +If a single read/write takes more than 30 seconds, the server closes +the connection. +@item +If a connection is idle for 8 hours, the server closes the connection. +@end itemize -Some of the binary distributions of @strong{MySQL} for HP-UX is -distributed as an HP depot file and as a tar file. To use the depot -file you must be running at least HP-UX 10.x to have access to HP's -software depot tools. +@menu +* Syntax differences:: +@end menu -The HP version of @strong{MySQL} was compiled on an HP 9000/8xx server -under HP-UX 10.20, and uses MIT-pthreads. It is known to work well under -this configuration. @strong{MySQL} Version 3.22.26 and newer can also be -built with HP's native thread package. +@node Syntax differences, , Protocol differences, Protocol differences +@subsubsection How @code{mSQL} 2.0 SQL Syntax Differs from MySQL -Other configurations that may work: +@noindent +@strong{Column types} +@table @code +@item @strong{MySQL} +Has the following additional types (among others; +@pxref{CREATE TABLE, , @code{CREATE TABLE}}): @itemize @bullet @item -HP 9000/7xx running HP-UX 10.20+ +@code{ENUM} type for one of a set of strings. @item -HP 9000/8xx running HP-UX 10.30 +@code{SET} type for many of a set of strings. +@item +@code{BIGINT} type for 64-bit integers. @end itemize - -The following configurations almost definitely won't work: - +@item +@strong{MySQL} also supports +the following additional type attributes: @itemize @bullet @item -HP 9000/7xx or 8xx running HP-UX 10.x where x < 2 +@code{UNSIGNED} option for integer columns. @item -HP 9000/7xx or 8xx running HP-UX 9.x +@code{ZEROFILL} option for integer columns. +@item +@code{AUTO_INCREMENT} option for integer columns that are a +@code{PRIMARY KEY}. +@xref{mysql_insert_id, , @code{mysql_insert_id()}}. +@item +@code{DEFAULT} value for all columns. @end itemize +@item mSQL2 +@code{mSQL} column types correspond to the @strong{MySQL} types shown below: +@multitable @columnfractions .15 .85 +@item @code{mSQL} @strong{type} @tab @strong{Corresponding @strong{MySQL} type} +@item @code{CHAR(len)} @tab @code{CHAR(len)} +@item @code{TEXT(len)} @tab @code{TEXT(len)}. @code{len} is the maximal length. +And @code{LIKE} works. +@item @code{INT} @tab @code{INT}. With many more options! +@item @code{REAL} @tab @code{REAL}. Or @code{FLOAT}. Both 4- and 8-byte versions are available. +@item @code{UINT} @tab @code{INT UNSIGNED} +@item @code{DATE} @tab @code{DATE}. Uses ANSI SQL format rather than @code{mSQL}'s own format. +@item @code{TIME} @tab @code{TIME} +@item @code{MONEY} @tab @code{DECIMAL(12,2)}. A fixed-point value with two decimals. +@end multitable +@end table -To install the distribution, use one of the commands below, where -@code{/path/to/depot} is the full pathname of the depot file: +@noindent +@strong{Index Creation} -@itemize @bullet -@item -To install everything, including the server, client and development tools: +@table @code +@item @strong{MySQL} +Indexes may be specified at table creation time with the @code{CREATE TABLE} +statement. +@item mSQL +Indexes must be created after the table has been created, with separate +@code{CREATE INDEX} statements. +@end table -@example -shell> /usr/sbin/swinstall -s /path/to/depot mysql.full -@end example +@noindent +@strong{To Insert a Unique Identifier into a Table} -@item -To install only the server: +@table @code +@item @strong{MySQL} +Use @code{AUTO_INCREMENT} as a column type +specifier. +@xref{mysql_insert_id, , @code{mysql_insert_id()}}. +@item mSQL +Create a @code{SEQUENCE} on a table and select the @code{_seq} column. +@end table -@example -shell> /usr/sbin/swinstall -s /path/to/depot mysql.server -@end example +@noindent +@strong{To Obtain a Unique Identifier for a Row} -@item -To install only the client package: +@table @code +@item @strong{MySQL} +Add a @code{PRIMARY KEY} or @code{UNIQUE} key to the table and use this. +New in Version 3.23.11: If the @code{PRIMARY} or @code{UNIQUE} key consists of only one +column and this is of type integer, one can also refer to it as +@code{_rowid}. +@item mSQL +Use the @code{_rowid} column. Observe that @code{_rowid} may change over time +depending on many factors. +@end table -@example -shell> /usr/sbin/swinstall -s /path/to/depot mysql.client -@end example +@noindent +@strong{To Get the Time a Column Was Last Modified} -@item -To install only the development tools: +@table @code +@item @strong{MySQL} +Add a @code{TIMESTAMP} column to the table. This column is automatically set +to the current date and time for @code{INSERT} or @code{UPDATE} statements if +you don't give the column a value or if you give it a @code{NULL} value. -@example -shell> /usr/sbin/swinstall -s /path/to/depot mysql.developer -@end example -@end itemize +@item mSQL +Use the @code{_timestamp} column. +@end table -The depot places binaries and libraries in @file{/opt/mysql} and data in -@file{/var/opt/mysql}. The depot also creates the appropriate entries in -@file{/etc/init.d} and @file{/etc/rc2.d} to start the server automatically -at boot time. Obviously, this entails being @code{root} to install. +@noindent +@strong{@code{NULL} Value Comparisons} -To install the HP-UX tar.gz distribution, you must have a copy of GNU -@code{tar}. +@table @code +@item @strong{MySQL} +@strong{MySQL} follows +ANSI SQL, and a comparison with @code{NULL} is always @code{NULL}. +@item mSQL +In @code{mSQL}, @code{NULL = NULL} is TRUE. You +must change @code{=NULL} to @code{IS NULL} and @code{<>NULL} to +@code{IS NOT NULL} when porting old code from @code{mSQL} to @strong{MySQL}. +@end table -@cindex installing, source distribution -@cindex source distribution, installing -@cindex installation overview -@node Installing source, Installing source tree, Installing binary, Installing -@section Installing a MySQL Source Distribution +@noindent +@strong{String Comparisons} -Before you proceed with the source installation, check first to see if our -binary is available for your platform and if it will work for you. We -put in a lot of effort into making sure that our binaries are built with the -best possible options. +@table @code +@item @strong{MySQL} +Normally, string comparisons are performed in case-independent fashion with +the sort order determined by the current character set (ISO-8859-1 Latin1 by +default). If you don't like this, declare your columns with the +@code{BINARY} attribute, which causes comparisons to be done according to the +ASCII order used on the @strong{MySQL} server host. +@item mSQL +All string comparisons are performed in case-sensitive fashion with +sorting in ASCII order. +@end table -You need the following tools to build and install @strong{MySQL} from source: +@noindent +@strong{Case-insensitive Searching} -@itemize @bullet -@item -GNU @code{gunzip} to uncompress the distribution. +@table @code +@item @strong{MySQL} +@code{LIKE} is a case-insensitive or case-sensitive operator, depending on +the columns involved. If possible, @strong{MySQL} uses indexes if the +@code{LIKE} argument doesn't start with a wild-card character. +@item mSQL +Use @code{CLIKE}. +@end table -@item -A reasonable @code{tar} to unpack the distribution. GNU @code{tar} is -known to work. Sun @code{tar} is known to have problems. +@noindent +@strong{Handling of Trailing Spaces} -@item -A working ANSI C++ compiler. @code{gcc} >= 2.95.2, @code{egcs} >= 1.0.2 -or @code{egcs 2.91.66}, SGI C++, and SunPro C++ are some of the -compilers that are known to work. @code{libg++} is not needed when -using @code{gcc}. @code{gcc} 2.7.x has a bug that makes it impossible -to compile some perfectly legal C++ files, such as -@file{sql/sql_base.cc}. If you only have @code{gcc} 2.7.x, you must -upgrade your @code{gcc} to be able to compile @strong{MySQL}. @code{gcc} -2.8.1 is also known to have problems on some platforms so it should be -avoided if there exists a new compiler for the platform.. +@table @code +@item @strong{MySQL} +Strips all spaces at the end of @code{CHAR} and @code{VARCHAR} +columns. Use a @code{TEXT} column if this behavior is not desired. +@item mSQL +Retains trailing space. +@end table -@code{gcc} >= 2.95.2 is recommended when compiling @strong{MySQL} -Version 3.23.x. +@noindent +@strong{@code{WHERE} Clauses} -@item -A good @code{make} program. GNU @code{make} is always recommended and is -sometimes required. If you have problems, we recommend trying GNU -@code{make} 3.75 or newer. -@end itemize +@table @code +@item @strong{MySQL} +@strong{MySQL} correctly prioritizes everything (@code{AND} is evaluated +before @code{OR}). To get @code{mSQL} behavior in @strong{MySQL}, use +parentheses (as shown in an example below). +@item mSQL +Evaluates everything from left to right. This means that some logical +calculations with more than three arguments cannot be expressed in any +way. It also means you must change some queries when you upgrade to +@strong{MySQL}. You do this easily by adding parentheses. Suppose you +have the following @code{mSQL} query: +@example +mysql> SELECT * FROM table WHERE a=1 AND b=2 OR a=3 AND b=4; +@end example +To make @strong{MySQL} evaluate this the way that @code{mSQL} would, +you must add parentheses: +@example +mysql> SELECT * FROM table WHERE (a=1 AND (b=2 OR (a=3 AND (b=4)))); +@end example +@end table -If you are using a recent version of @strong{gcc}, recent enough to understand -@code{-fno-exceptions} option, it is @strong{VERY IMPORTANT} that you use -it. Otherwise, you may compile a binary that crashes randomly. We also -recommend that you use @code{-felide-contructors} and @code{-fno-rtti} along -with @code{-fno-exceptions}. When in doubt, do the following: +@noindent +@strong{Access Control} -@example +@table @code +@item @strong{MySQL} +Has tables to store grant (permission) options per user, host, and +database. @xref{Privileges}. +@item mSQL +Has a file @file{mSQL.acl} in which you can grant read/write privileges for +users. +@item +@end table -CFLAGS="-O3" CXX=gcc CXXFLAGS="-O3 -felide-constructors -fno-exceptions -fno-rtti" ./configure --prefix=/usr/local/mysql --enable-assembler --with-mysqld-ldflags=-all-static +@node Compare PostgreSQL, MySQL-PostgreSQL features, Protocol differences, Comparisons +@subsection How MySQL Compares to PostgreSQL -@end example +@cindex PostgreSQL vs. MySQL, overview -On most systems this will give you a fast and stable binary. +When reading the following, please note that both products are +continually evolving. We at @strong{MySQL AB} and the PostgreSQL +developers are both working on making our respective database as good as +possible, so we are both a serious choice to any commercial database. -@c texi2html fails to split chapters if I use strong for all of this. -If you run into problems, @strong{PLEASE ALWAYS USE @code{mysqlbug}} when -posting questions to @email{mysql@@lists.mysql.com}. Even if the problem -isn't a bug, @code{mysqlbug} gathers system information that will help others -solve your problem. By not using @code{mysqlbug}, you lessen the likelihood -of getting a solution to your problem! You will find @code{mysqlbug} in the -@file{scripts} directory after you unpack the distribution. -@xref{Bug reports}. +The following comparison is made by us at MySQL AB. We have tried to be +as accurate and fair as possible, but because we don't have a full +knowledge of all PostgreSQL features while we know MySQL througly, we +may have got some things wrong. We will however correct these when they +come to our attention. + +We would first like to note that @code{PostgreSQL} and @strong{MySQL} +are both widely used products, but with different design goals, even if +we are both striving to be ANSI SQL compatible. This means that for +some applications @strong{MySQL} is more suitable and for others +@code{PostgreSQL} is more suitable. When choosing which database to +use, you should first check if the database's feature set satisfies your +application. If you need speed, @strong{MySQL} is probably your best +choice. If you need some of the extra features that only @code{PostgreSQL} +can offer, you should use @code{PostgreSQL}. +@cindex PostgreSQL/MySQL, strategies @menu -* Quick install:: Quick installation overview -* Applying patches:: Applying patches -* configure options:: Typical @code{configure} options +* MySQL-PostgreSQL goals:: MySQL and PostgreSQL development strategies @end menu -@node Quick install, Applying patches, Installing source, Installing source -@subsection Quick Installation Overview +@node MySQL-PostgreSQL goals, , Compare PostgreSQL, Compare PostgreSQL +@subsubsection MySQL and PostgreSQL development strategies -The basic commands you must execute to install a @strong{MySQL} source -distribution are: +When adding things to MySQL we take pride to do an optimal, definite +solution. The code should be so good that we shouldn't have any need to +change it in the foreseeable future. We also do not like to sacrifice +speed for features but instead will do our utmost to find a solution +that will give maximal throughput. This means that development will take +a little longer, but the end result will be well worth this. This kind +of development is only possible because all server code are checked by +one of a few (currently two) persons before it's included in the +@strong{MySQL} server. -@example -shell> groupadd mysql -shell> useradd -g mysql mysql -shell> gunzip < mysql-VERSION.tar.gz | tar -xvf - -shell> cd mysql-VERSION -shell> ./configure --prefix=/usr/local/mysql -shell> make -shell> make install -shell> scripts/mysql_install_db -shell> chown -R root /usr/local/mysql -shell> chown -R mysql /usr/local/mysql/var -shell> chgrp -R mysql /usr/local/mysql -shell> cp support-files/my-medium.cnf /etc/my.cnf -shell> /usr/local/mysql/bin/safe_mysqld --user=mysql & -@end example +We at MySQL AB believe in frequent releases to be able to push out new +features quickly to our users. Because of this we do a new small release +about every 3 weeks, which a major branch every year. All releases are +throughly tested with our testing tools on a lot of different platforms. -If you want have support for InnoDB tables, you should edit the -@code{/etc/my.cnf} file and remove the @code{#} character before the -parameters that starts with @code{innodb_...}. @xref{Option -files}. @xref{InnoDB start}. +PostgreSQL is based on a kernel with lots of contributors. In this setup +it makes sense to prioritize adding a lot of new features, instead of +implementing them optimally, because one can always optimize things +later if there arises a need for this. -If you start from a source RPM, then do the following: +Another big difference between @strong{MySQL} and PostgreSQL is that +nearly all of the code in the MySQL server are coded by developers that +are employed by MySQL AB and are still working on the server code. The +exceptions are the transaction engines and the regexp library. -@example -shell> rpm --rebuild MySQL-VERSION.src.rpm -@end example +This is in sharp contrast to the PostgreSQL code where the majority of +the code is coded by a big group of people with different backgrounds. +It was only recently that the PostgreSQL developers announced that they +current developer group had finally had time to take a look at all +the code in the current PostgreSQL release. -This will make a binary RPM that you can install. +Both of the above development methods has it's own merits and drawbacks. +We here at @strong{MySQL AB} think of course that our model is better +because our model gives better code consistence, more optimal and +reusable code and, in our opinion, fewer bugs. Because we are the +authors of the @strong{MySQL} server code we are better able to +coordinate new features and releases. -You can add new users using the @code{bin/mysql_setpermission} script if -you install the @code{DBI} and @code{Msql-Mysql-modules} Perl modules. +@cindex PostgreSQL/MySQL, features +@node MySQL-PostgreSQL features, , Compare PostgreSQL, Comparisons +@subsection Featurevise Comparison of MySQL and PostgreSQL -A more detailed description follows. +On the @uref{http://www.mysql.com/information/crash-me.php, crash-me} +page you can find a list of those database constructs and limits that +one can detect automatically with a program. Note however that a lot of +the numerical limits may be changed with startup options for respective +database. The above web page is however extremely useful when you want to +ensure that your applications works with many different databases or +when you want to convert your application from one datbase to another. -To install a source distribution, follow the steps below, then proceed -to @ref{Post-installation}, for post-installation initialization and testing: +@strong{MySQL} offers the following advantages over PostgreSQL: -@enumerate +@itemize @bullet @item -Pick the directory under which you want to unpack the distribution, and move -into it. - +@code{MySQL} is generally much faster than PostgreSQL. +@xref{MySQL-PostgreSQL benchmarks}. @item -Obtain a distribution file from one of the sites listed in -@ref{Getting MySQL, , Getting @strong{MySQL}}. - +Because @strong{MySQL} has a much larger user base than PostgreSQL the +code is more tested and has historically been more stable than +PostgreSQL. @strong{MySQL} is the much more used in production +environments than PostgreSQL, mostly thanks to that @strong{MySQL AB}, +former TCX DataKonsult AB, has provided top quality commercial support +for @strong{MySQL} from the day it was released, whereas until recently +PostgreSQL was unsupported. @item -If you are interested in using Berkeley DB tables with MySQL, you -will need to obtain a patched version of the Berkeley DB source -code. Please read the chapter on Berkeley DB tables before -proceeding. @xref{BDB}. - -@strong{MySQL} source distributions are provided as compressed @code{tar} -archives and have names like @file{mysql-VERSION.tar.gz}, where -@code{VERSION} is a number like @value{mysql_version}. - +@strong{MySQL} works on more platforms than PostgreSQL. @xref{Which OS}. @item -Add a user and group for @code{mysqld} to run as: - -@example -shell> groupadd mysql -shell> useradd -g mysql mysql -@end example - -These commands add the @code{mysql} group, and the @code{mysql} user. The -syntax for @code{useradd} and @code{groupadd} may differ slightly on different -versions of Unix. They may also be called @code{adduser} and @code{addgroup}. -You may wish to call the user and group something else instead of @code{mysql}. - +@strong{MySQL} works better on Windows; @strong{MySQL} is running as a +native windows application (a service on NT/Win2000/WinXP), while +PostgreSQL is run under the cygwin emulation. We have heard that +PostgreSQL is not yet that stable on windows but we haven't been able to +verify this ourselves. @item -Unpack the distribution into the current directory: -@example -shell> gunzip < /path/to/mysql-VERSION.tar.gz | tar xvf - -@end example - -This command creates a directory named @file{mysql-VERSION}. - +@strong{MySQL} has more API to other languages and is supported by more +programs than PostgreSQL. @xref{Contrib}. @item -Change into the top-level directory of the unpacked distribution: - -@example -shell> cd mysql-VERSION -@end example +@strong{MySQL} works on 24/7 heavy duty systems. In most circumstances +you never have to run any cleanups on @code{MySQL}. PostgreSQL doesn't +yet support 24/7 systems because you have have to run @code{vacuum()} +once in a while to reclaim space from @code{UPDATE} and @code{DELETE} +commands and to perform statistics analyzes that are critical to get +good performance with PostgreSQL. Vacuum is also needed after adding +a lot of new rows to a table. On a busy system with lots of changes +vacuum must be run very frequently, in the worst cases even many times a +day. During the @code{vacuum()} run, which may take hours if the +database is big, the database is from a production standpoint +practically dead. The PostgreSQL team has fixing this on their TODO, +but we assume that this is not an easy thing to fix permanently. +@item +A working, tested replication feature used by sites like +@uref{http://finance.yahoo.com, Yahoo finance}, +@uref{http://www.mobile.de/,mobile.de} and +@uref{http://www.slashdot.org,Slashdot}. +@item +Included in the @strong{MySQL} distribution is included two different +testing suits (@file{mysql-test-run} and +@uref{http://www.mysql.com/information/crash-me.php,crash-me}) and a +benchmark suite. The test system is actively updated with code to test +each new feature and almost all repeatable bugs that comes to our +attention. We test @strong{MySQL} with these on a lot of platforms +before every release. These tests are more sofisticated than anything +have seen from PostgreSQL and ensures that the @strong{MySQL} code keeps +at a high standard. +@item +There are far moore books in print on @strong{MySQL} than on PostgreSQL. +O'Reilly, Sams, Que, and New Riders are all major publishers with books +about MySQL. All @strong{MySQL} features is also documented in the +@strong{MySQL} on-line manual because when a feature is implemented, the +@strong{MySQL} developers are required to document it before it's +included in the source. +@item +@strong{MySQL} has supports more of the standard ODBC functions than +@code{PostgreSQL}. +@item +@strong{MySQL} has a much more sophisticated @code{ALTER TABLE}. +@item +@strong{MySQL} has support for tables without transactions for +applications that need all speed they can get. The tables may be memory +based,@code{HEAP} tables or disk based @code{MyISAM}. @xref{Table types}. +@item +@strong{MySQL} has support for 3 different table handles that support +transactions (@code{BDB} and @code{InnoDB}). Because +every transaction engine performs differently under different +conditions, this gives the application writer more options to find an +optimal solution for his/her setup. @xref{Table types}. +@item +@code{MERGE} tables gives you a unique way to instantly make a view over +a set of identical tables and use these as one. This is perfectly for +systems where you have log files that you order for example by month. +@xref{MERGE}. +@item +The option to compress read-only tables, but still have direct access to +the rows in the table, gives you better performance by minimizing disk +reads. This is very useful when you are archiving +things.@xref{myisampack}. +@item +@strong{MySQL} has internal support for text search. @xref{Fulltext Search}. +@item +You can access many databases from the same connection (depending of course +on your privileges). +@item +@strong{MySQL} is coded from the start with multi-threading while +PostgreSQL uses processes. Because context switching and access to +common storage areas is much faster between threads, than are separate +processes, this gives @strong{MySQL} a big speed advantage in multi-user +applications and also makes it easier for @strong{MySQL} to take full +advantage of symmetric multiprocessor systems (SMP). +@item +@strong{MySQL} has a much more sophisticated privilege system than +PostgreSQL. While PostgreSQL only supports @code{INSERT}, +@code{SELECT}, @code{update/delete} grants per user on a database or a +table @strong{MySQL} allows you to define a full set of different +privileges on database, table and columns level. @strong{MySQL} also allows +you to specify the privilege on host+user combinations. @xref{GRANT}. +@item +@strong{MySQL} supports a compressed server/client protocol which +improves performance over slow links. +@item +@strong{MySQL} employs the table handler concept and is the only +relational database we know of built around this concept. This allows +different low level table types to be swapped into the SQL engine, each +table type optimized for a different performance characteristics. +@item +All @code{MySQL} table types (except @strong{InnoDB}) are implemented as +files (ie: one table per file), which makes it really easy to backup, +move, delete and even symlink databases and tables when the server is +down. +@item +Tools to repair and optimize @strong{MyISAM} tables (the most common +@strong{MySQL} table type). A repair tool is only needed when a +physical corruption of a data file happens, usually from a hardware +failure. It allows a majority of the data to be recovered. +@item +Upgrading @strong{MySQL} is painless. When you are upgrading @strong{MySQL}, +you don't need to dump/restore your data, as you have to do with most +PostgreSQL upgrades. +@end itemize -Note that currently you must configure and build @strong{MySQL} from -this top-level directory. You can not build it in a different -directory. +Drawbacks with @strong{MySQL} compared to PostgreSQL: +@itemize @bullet @item -Configure the release and compile everything: +The transaction support in @strong{MySQL} is not yet as well tested as +PostgreSQL's system. +@item +Because @strong{MySQL} uses threads, which are still a moving target on +many OS, one must either use binaries from +@uref{http://www.mysql.com/downloads} or carefully follow our +instructions on +@uref{http://www.mysql.com/doc/I/n/Installing_source.html} to get an +optimal binary that works in all cases. +@item +Table locking, as used by the non-transactional @code{MyISAM} tables, is +in many cases faster than page locks, row locks or versioning. The +drawback however is that if one doesn't take into account how table +locks works, a single long-running query can block a table for updates +for a long time. This can usable be avoided when designing the +application. If not, one can always switch the trouble table to use one +of the transactional table types. @xref{Table locking}. +@item +With UDF (user defined functions) one can extend @strong{MySQL} with +both normal SQL functions and aggregates, but this is not as easy or as +flexible as in PostgreSQL. @xref{Adding functions}. +@item +Updates and deletes that goes over multiple tables is harder to do in +@strong{MySQL}. (Will be fixed in @strong{MySQL} 4.0 with multi-table +@code{DELETE} and multi-table @code{UPDATE} and in @strong{MySQL} 4.1 +with @code{SUB-SELECT}) +@end itemize -@example -shell> ./configure --prefix=/usr/local/mysql -shell> make -@end example +PostgreSQL offers currently the following advantages over @strong{MySQL}: -When you run @code{configure}, you might want to specify some options. -Run @code{./configure --help} for a list of options. -@ref{configure options, , @code{configure} options}, discusses some of the -more useful options. +Note that because we know the @strong{MySQL} road map, we have included +in the following table the version when @strong{MySQL} should support +this feature. Unfortunately we couldn't do this for previous comparison, +because we don't know the PostgreSQL roadmap. -If @code{configure} fails, and you are going to send mail to -@email{mysql@@lists.mysql.com} to ask for assistance, please include any -lines from @file{config.log} that you think can help solve the problem. Also -include the last couple of lines of output from @code{configure} if -@code{configure} aborts. Post the bug report using the @code{mysqlbug} -script. @xref{Bug reports}. +@multitable @columnfractions .70 .30 +@item @strong{Feature} @tab @strong{MySQL version} +@item Subselects @tab 4.1 +@item Foreign keys @tab 4.0 and 4.1 +@item Views. @tab 4.2 +@item Stored procedures in multiple languages @tab 4.1 +@item Extensible type system. @tab Not planed +@item Unions @tab 4.0. +@item Full join. @tab 4.0 or 4.1. +@item Triggers. @tab 4.1 +@item Constrainst @tab 4.1 +@item Cursors @tab 4.1 or 4.2 +@item Extensible index types like R-trees @tab R-trees are planned to 4.2 +@item Inherited tables @tab Not planned +@end multitable -If the compile fails, see @ref{Compilation problems}, for help with -a number of common problems. +Other reasons to use PostgreSQL: +@itemize @bullet @item -Install everything: - -@example -shell> make install -@end example +Standard usage is in PostgreSQL closer to ANSI SQL in some cases. +@item +One can get speed up PostgreSQL by coding things as stored procedures. +@item +Bigger team of developers that contributes to the server. +@end itemize -You might need to run this command as @code{root}. +Drawbacks with PostgreSQL compared to @strong{MySQL}: +@itemize @bullet @item -Create the @strong{MySQL} grant tables (necessary only if you haven't -installed @strong{MySQL} before): +@code{Vaccum()} makes PostgreSQL hard to use in a 24/7 environment. +@item +Only transactional tables. +@item +Much slower insert/delete/update. +@end itemize -@example -shell> scripts/mysql_install_db -@end example +For a complete list of drawbacks, you should also examine the first table +in this section. -Note that @strong{MySQL} versions older than Version 3.22.10 started the -@strong{MySQL} server when you run @code{mysql_install_db}. This is no -longer true! +@menu +* MySQL-PostgreSQL benchmarks:: +@end menu -@item -Change ownership of binaries to @code{root} and ownership of the data -directory to the user that you will run @code{mysqld} as: +@node MySQL-PostgreSQL benchmarks, , MySQL-PostgreSQL features, MySQL-PostgreSQL features +@subsubsection Benchmarking MySQL and PostgreSQL -@example -shell> chown -R root /usr/local/mysql -shell> chown -R mysql /usr/local/mysql/var -shell> chgrp -R mysql /usr/local/mysql -@end example +@cindex PostgreSQL vs. MySQL, benchmarks -The first command changes the @code{owner} attribute of the files to the -@code{root} user, the second one changes the @code{owner} attribute of the -data directory to the @code{mysql} user, and the third one changes the -@code{group} attribute to the @code{mysql} group. +The only open source benchmark, that we know of, that can be used to +benchmark @strong{MySQL} and PostgreSQL (and other databases) is our +own. It can be found at: +@uref{http://www.mysql.com/information/benchmarks.html}. -@item -If you want to install support for the Perl @code{DBI}/@code{DBD} interface, -see @ref{Perl support}. +We have many times asked the PostgreSQL developers and some PostgreSQL +users to help us extend this benchmark to make the definitive benchmark +for databases, but unfortunately we haven't got any feedback for this. -@item -If you would like @strong{MySQL} to start automatically when you boot your -machine, you can copy @code{support-files/mysql.server} to the location where -your system has its startup files. More information can be found in the -@code{support-files/mysql.server} script itself and in -@ref{Automatic start}. -@end enumerate - -After everything has been installed, you should initialize and test your -distribution: - -@example -shell> /usr/local/mysql/bin/safe_mysqld --user=mysql & -@end example - -If that command fails immediately with @code{mysqld daemon ended} then you can -find some information in the file @file{mysql-data-directory/'hostname'.err}. -The likely reason is that you already have another @code{mysqld} server -running. @xref{Multiple servers}. +We, the @strong{MySQL} developers, have because of this spent a lot of +hours to get maximum performance from PostgreSQL for the benchmarks, but +because we don't know PostgreSQL intimately we are sure that there are +things that we have missed. We have on the benchmark page documented +exactly how we did run the benchmark so that it should be easy for +anyone to repeat and verify our results. -@xref{Post-installation}. +The benchmarks are usually run with and without the @code{--fast} +option. When run with @code{--fast} we are trying to use every trick +the server can do to get the code to execute as fast as possible. +The idea is that the normal run should show how the server would work in +a default setup and the @code{--fast} run shows how the server would do +if the application developer would use extensions in the server to make +his application run faster. -@cindex patches, applying -@cindex applying, patches -@node Applying patches, configure options, Quick install, Installing source -@subsection Applying Patches +When running with PostgreSQL and @code{--fast} we do a @code{vacuum()} +between after every major table update and drop table to make the database +in perfect shape for the following selects. The time for vacuum() is +measured separately. -Sometimes patches appear on the mailing list or are placed in the -@uref{http://www.mysql.com/Downloads/Patches, patches area} of the -@strong{MySQL} Web site. +When running with PostgreSQL 7.1.1 we could however not run with +@code{--fast} because during the insert test, the postmaster (the +PostgreSQL deamon) died and the database was so corrupted that it was +impossible to restart postmaster. (The details about the machine we run +the benchmark can be found on the benchmark page). After this happened +twice, we decided to postpone the @code{--fast} test until next +PostgreSQL release. -To apply a patch from the mailing list, save the message in which the patch -appears in a file, change into the top-level directory of your @strong{MySQL} -source tree, and run these commands: +Before going to the other benchmarks we know of, We would like to give +some background to benchmarks: -@example -shell> patch -p1 < patch-file-name -shell> rm config.cache -shell> make clean -@end example +It's very easy to write a test that shows ANY database to be best +database in the world, by just restricting the test to something the +database is very good at and not test anything that the database is not +good at; If one after this publish the result with a single figure +things is even easier. -Patches from the FTP site are distributed as plain text files or as files -compressed with @code{gzip}. Apply a plain patch as shown above for -mailing list patches. To apply a compressed patch, change into the -top-level directory of your @strong{MySQL} source tree and run these -commands: +This would be like we would measure the speed of @strong{MySQL} compared +to PostgreSQL by looking at the summary time of the MySQL benchmarks on +our web page. Based on this @strong{MySQL} would be more than 40 times +faster than PostgreSQL, something that is of course not true. We could +make things even worse by just taking the test where PostgreSQL performs +worst and claim that @strong{MySQL} is more than 2000 times faster than +PostgreSQL. -@example -shell> gunzip < patch-file-name.gz | patch -p1 -shell> rm config.cache -shell> make clean -@end example +The case is that @strong{MySQL} does a lot of optimizations that +PostgreSQL doesn't do and the other way around. An SQL optimizer is a +very complex thing and a company could spend years on just making the +optimizer faster and faster. -After applying a patch, follow the instructions for a normal source install, -beginning with the @code{./configure} step. After running the @code{make -install} step, restart your @strong{MySQL} server. +When looking at the benchmark results you should look for things that +you do in your application and just use these results to decide which +database would be best suited for your application. The benchmark +results also shows things a particular database is not good at and should +give you a notion about things to avoid and what you may have to do in +other ways. -You may need to bring down any currently running server before you run -@code{make install}. (Use @code{mysqladmin shutdown} to do this.) Some -systems do not allow you to install a new version of a program if it replaces -the version that is currently executing. +We know of two benchmark tests that claims that PostgreSQL performers +better than @strong{MySQL}. These both where multi-user tests, a test +that we here at @strong{MySQL AB} haven't had time to write and include in +the benchmark suite, mainly because it's a big task to do this in a +manner that is fair against all databases. -@cindex @code{configure} script -@cindex options, configure -@cindex configuration options -@findex without-server option -@cindex log files -@cindex files, log -@node configure options, , Applying patches, Installing source -@subsection Typical @code{configure} Options +One is the benchmark paid for by +@uref{http://www.greatbridge.com/about/press.php?content_id=4,Great +Bridge}. -The @code{configure} script gives you a great deal of control over how -you configure your @strong{MySQL} distribution. Typically you do this -using options on the @code{configure} command line. You can also affect -@code{configure} using certain environment variables. @xref{Environment -variables}. For a list of options supported by @code{configure}, run -this command: +This is the worst benchmark we have ever seen anyone ever conduct. This +was not only tuned to only test what PostgreSQL is absolutely best at, +it was also totally unfair against every other database involved in the +test. -@example -shell> ./configure --help -@end example +@strong{NOTE}: We know that not even some of the main PostgreSQL +developers did like the way Great Bridge conducted the benchmark, so we +don't blame them for the way the benchmark was made. -Some of the more commonly-used @code{configure} options are described below: +This benchmark has been condemned in a lot of postings and newsgroups so +we will here just shortly repeat some things that where wrong with it. @itemize @bullet @item -To compile just the @strong{MySQL} client libraries and client programs and -not the server, use the @code{--without-server} option: - -@example -shell> ./configure --without-server -@end example - -@c FIX incorrect.. - -If you don't have a C++ compiler, @code{mysql} will not compile (it is the -one client program that requires C++). In this case, -you can remove the code in @code{configure} that tests for the C++ compiler -and then run @code{./configure} with the @code{--without-server} option. The -compile step will still try to build @code{mysql}, but you can ignore any -warnings about @file{mysql.cc}. (If @code{make} stops, try @code{make -k} -to tell it to continue with the rest of the build even if errors occur.) - +The tests where run with an expensive commercial tool, that makes it +impossible for an open source company like us to verify the benchmarks, +or even check how the benchmark where really done. The tool is not even +a true benchmark tool, but a application/setup testing tool. To refer +this as STANDARD benchmark tool is to stretch the truth a long way. @item -If you don't want your log files and database directories located under -@file{/usr/local/var}, use a @code{configure} command, something like one -of these: - -@example -shell> ./configure --prefix=/usr/local/mysql -shell> ./configure --prefix=/usr/local \ - --localstatedir=/usr/local/mysql/data -@end example - -The first command changes the installation prefix so that everything is -installed under @file{/usr/local/mysql} rather than the default of -@file{/usr/local}. The second command preserves the default installation -prefix, but overrides the default location for database directories -(normally @file{/usr/local/var}) and changes it to -@code{/usr/local/mysql/data}. - -@cindex changing socket location -@cindex socket location, changing -@cindex mysql.sock, changing location of +Great Bridge admitted that they had optimized the PostgreSQL database +(with vacuum() before the test) and tuned the startup for the tests, +something they hadn't done for any of the other databases involved. To +say "This process optimizes indexes and frees up disk space a bit. The +optimized indexes boost performance by some margin". Our benchmarks +clearly indicates that the difference in running a lot of selects on a +database with and without vacuum() can easily differ by a factor of 10. @item -If you are using Unix and you want the @strong{MySQL} socket located somewhere -other than the default location (normally in the directory @file{/tmp} or -@file{/var/run}) use a @code{configure} command like this: - -@example -shell> ./configure --with-unix-socket-path=/usr/local/mysql/tmp/mysql.sock -@end example +The test results where also strange; The AS3AP test documentation +mentions that the test does: -Note that the given file must be an absolute pathname! +"selections, simple joins, projections, aggregates, one-tuple updates, +and bulk updates" -@cindex compiling statically -@cindex statically, compiling +PostgreSQL is good at doing selects and joins (especially after a +vacuum()), but doesn't perform as well on inserts/updates; The +benchmarks seem to indicate that only SELECTs where done (or very few +updates) . This could easily explain they good results for PostgreSQL in +this test. The bad results for MySQL will be obvious a bit down in this +document. @item -If you want to compile statically linked programs (for example, to make a -binary distribution, to get more speed, or to work around problems with some -RedHat Linux distributions), run @code{configure} like this: - -@example -shell> ./configure --with-client-ldflags=-all-static \ - --with-mysqld-ldflags=-all-static -@end example +They did run the so called benchmark from a Windows machine against a +Linux machine over ODBC, a setup that no normal database user would ever +do when running a heavy multi-user application. This tested more the +ODBC driver and the Windows protocol used between the clients than the +database itself. +@item +When running the database against Oracle and MS-SQL (Great Bridge has +indirectly indicated that the databases they used in the test), +they didn't use the native protocol but instead ODBC. Anyone that has +ever used Oracle, knows that all real application uses the native +interface instead of ODBC. Doing a test through ODBC and claiming that +the results had anything to do with using the database for real can't +be regarded as fair play. They should have done two tests with and +without ODBC to provide the right facts (after having got experts to tune +all involved databases of course). +@item +They refer to the TPC-C tests, but doesn't anywhere mention that the +tests they did where not a true TPC-C test and they where not even +allowed to call it a TPC-C test. A TPC-C test can only be conducted by +the rules approved by the @uref{http://www.tpc.org,TPC-council}. Great +Bridge didn't do that. By doing this they have both violated the TPC +trademark and miscredited their own benchmarks. The rules set by the +TPC-council are very strict to ensure that no one can produce false +results or make unprovable statements. Apparently Great Bridge wasn't +interested in doing this. +@item +After the first test, we contacted Great Bridge and mentioned to them +some of the obvious mistakes they had done with @strong{MySQL}; Running +with a debug version of our ODBC driver, running on a Linux system that +wasn't optimized for threads, using an old MySQL version when there was +a recommended newer one available, not starting @strong{MySQL} with the +right options for heavy multi-user use (the default installation of +MySQL is tuned for minimal resource use). Great Bridge did run a new +test, with our optimized ODBC driver and with better startup options for +MySQL, but refused to either use our updated glibc library or our +standard binary (used by 80% of our users), which was statically linked +with a fixed glibc library. -@tindex CC environment variable -@tindex environment variable, CC -@tindex CXX environment variable -@tindex environment variable, CXX -@cindex gcc -@cindex C++ complier, gcc -@cindex compiler, C++ gcc +According to what we know, Great Bridge did nothing to ensure that the +other databases where setup correctly to run good in their test +environment. We are sure however that they didn't contact Oracle or +Microsoft to ask for their advice in this matter ;) @item -If you are using @code{gcc} and don't have @code{libg++} or @code{libstdc++} -installed, you can tell @code{configure} to use @code{gcc} as your C++ -compiler: +The benchmark was paid for by Great Bridge, and they decided to publish +only partial chosen results (instead of publishing it all). +@end itemize -@example -shell> CC=gcc CXX=gcc ./configure -@end example +Tim Perdue, a long time PostgreSQL fan and a reluctant MySQL user +published a comparison on +@uref{http://www.phpbuilder.com/columns/tim20001112.php3,phpbuider}. -When you use @code{gcc} as your C++ compiler, it will not attempt to link in -@code{libg++} or @code{libstdc++}. +When we got aware of the comparison, we phoned Tim Perdue about this +because there was a lot of strange things in his results. For example, +he claimed that MySQL had a problem with five users in his tests, when we +know that there are users with similar machines as his that are using +MySQL with 2000 simultaneous connections doing 400 queries per second (In +this case the limit was the web bandwidth, not the database). -Here is some common environment variables to set depending on -the compiler you are using: +It sounded like he was using a Linux kernel that either had some +problems with many threads (Linux kernels before 2.4 had a problem with +this but we have documented how to fix this and Tim should be aware of +this problem). The other possible problem could have been an old glibc +library and that Tim didn't use a MySQL binary from our site, which is +linked with a corrected glibc library, but had compiled a version of his +own with. In any of the above cases, the symptom would have been exactly +what Tim had measured. -@tindex CXXFLAGS environment variable -@tindex environment variable, CXXFLAGS -@multitable @columnfractions .20 .80 -@item gcc 2.7.2.1 @tab -CC=gcc CXX=gcc CXXFLAGS="-O3 -felide-constructors" -@item egcs 1.0.3a @tab -CC=gcc CXX=gcc CXXFLAGS="-O3 -felide-constructors -fno-exceptions -fno-rtti" -@item gcc 2.95.2 @tab -CFLAGS="-O3 -mpentiumpro" CXX=gcc CXXFLAGS="-O3 -mpentiumpro -felide-constructors -fno-exceptions -fno-rtti" -@item pgcc 2.90.29 or newer @tab -CFLAGS="-O3 -mpentiumpro -mstack-align-double" CXX=gcc CXXFLAGS="-O3 -mpentiumpro -mstack-align-double -felide-constructors -fno-exceptions -fno-rtti" -@end multitable +We asked Tim if we could get access to his data so that we could repeat +the benchmark and if he could check the MySQL version on the machine to +find out what was wrong and he promised to come back to us about this. +He has not done that yet. -In most cases you can get a reasonably optimal @strong{MySQL} binary by -using the options from the above and adding the following options to the -configure line: +Because of this we can't put any trust in this benchmark either :( -@example ---prefix=/usr/local/mysql --enable-assembler --with-mysqld-ldflags=-all-static -@end example +Conclusion: -The full configure line would in other words be something like the -following for all recent gcc versions: +The only benchmarks that exist today that anyone can download and run +against @strong{MySQL}and PostgreSQL is the MySQL benchmarks. We here +at @strong{MySQL} believe that open source databases should be tested +with open source tools! This is the only way to ensure that no one +does tests that nobody can reproduce and use this to claim that a +database is better than another. Without knowing all the facts it's +impossible to answer the claims of the tester. -@example -CFLAGS="-O3 -mpentiumpro" CXX=gcc CXXFLAGS="-O3 -mpentiumpro -felide-constructors -fno-exceptions -fno-rtti" ./configure --prefix=/usr/local/mysql --enable-assembler --with-mysqld-ldflags=-all-static -@end example +The thing we find strange is that every test we have seen about +PostgreSQL, that is impossible to reproduce, claims that PostgreSQL is +better in most cases while our tests, which anyone can reproduce, +clearly shows otherwise. With this we don't want to say that PostgreSQL +isn't good at many things (It is!) We would just like to see a fair test +where they are very good so that we could get some friendly competition +going! -The binaries we provide on the @strong{MySQL} Web site at -@uref{http://www.mysql.com} are all compiled with full optimization and -should be perfect for most users. @xref{MySQL binaries}. There are some -things you can tweak to make an even faster binary, but this is only for -advanced users. @xref{Compile and link options}. +For more information about our benchmarks suite see @xref{MySQL +Benchmarks}. -If the build fails and produces errors about your compiler or linker not -being able to create the shared library @file{libmysqlclient.so.#} (@samp{#} -is a version number), you can work around this problem by giving the -@code{--disable-shared} option to @code{configure}. In this case, -@code{configure} will not build a shared @code{libmysqlclient.so.#} library. +We are working on an even better benchmark suite, including much better +documentation of what the individual tests really do and how to add more +tests to the suite. -@cindex default values, suppression -@cindex suppression, default values -@item -You can configure @strong{MySQL} not to use @code{DEFAULT} column values for -non-@code{NULL} columns (that is, columns that are not allowed to be -@code{NULL}). This causes @code{INSERT} statements to generate an error -unless you explicitly specify values for all columns that require a -non-@code{NULL} value. To suppress use of default values, run -@code{configure} like this: +@node TODO, , Comparisons, Introduction +@section MySQL and the future (The TODO) -@tindex CXXFLAGS environment variable -@tindex environment variable, CXXFLAGS -@example -shell> CXXFLAGS=-DDONT_USE_DEFAULT_FIELDS ./configure -@end example +@cindex ToDo list for MySQL -@cindex character sets -@findex configure option, --with-charset -@findex configure option, --with-extra-charset -@item -By default, @strong{MySQL} uses the ISO-8859-1 (Latin1) character set. To -change the default set, use the @code{--with-charset} option: -@example -shell> ./configure --with-charset=CHARSET -@end example -@code{CHARSET} may be one of @code{big5}, @code{cp1251}, @code{cp1257}, -@code{czech}, @code{danish}, @code{dec8}, @code{dos}, @code{euc_kr}, -@code{gb2312}, @code{gbk}, @code{german1}, @code{hebrew}, @code{hp8}, -@code{hungarian}, @code{koi8_ru}, @code{koi8_ukr}, @code{latin1}, -@code{latin2}, @code{sjis}, @code{swe7}, @code{tis620}, @code{ujis}, -@code{usa7}, or @code{win1251ukr}. -@xref{Character sets}. +@menu +* TODO MySQL 4.0:: Things that should be in Version 4.0 +* TODO future:: Things that must be done in the near future +* TODO sometime:: Things that have to be done sometime +* TODO unplanned:: Some things we don't have any plans to do +@end menu -If you want to convert characters between the server and the client, -you should take a look at the @code{SET OPTION CHARACTER SET} command. -@xref{SET OPTION, , @code{SET OPTION}}. +This appendix lists the features that we plan to implement in @strong{MySQL}. -@cindex @code{myisamchk} -@strong{Warning:} If you change character sets after having created any -tables, you will have to run @code{myisamchk -r -q} on every table. Your -indexes may be sorted incorrectly otherwise. (This can happen if you -install @strong{MySQL}, create some tables, then reconfigure -@strong{MySQL} to use a different character set and reinstall it.) +Everything in this list is approximately in the order it will be done. If you +want to affect the priority order, please register a license or support us and +tell us what you want to have done more quickly. @xref{Licensing and Support}. -With the option @code{--with-extra-charset=LIST} you can define -which additional character sets should be incompiled in the server. +The plan is that we in the future will support the full ANSI SQL99 +standard, but with a lot of useful extensions. The challenge is to do +this without sacrifying the speed or compromise the code. -Here @code{LIST} is either a list of character set separated with space, -@code{complex} to include all characters that can't be dynamically loaded -or @code{all} to include all character sets into the binaries. +@node TODO MySQL 4.0, TODO future, TODO, TODO +@subsection Things that should be in 4.0 -@item -To configure @strong{MySQL} with debugging code, use the @code{--with-debug} -option: -@example -shell> ./configure --with-debug -@end example -This causes a safe memory allocator to be included that can find some errors -and that provides output about what is happening. -@xref{Debugging server}. +We plan to make @strong{MySQL} Version 4.0 a ``quick'' release where we only +add some new stuff to enable others to help us with developing new features +into Version 4.1. The @strong{MySQL} 4.0 version should only take us about +a month to make after which we want to stabilize it and start working on +Version 4.1. Version 4.0 should have the following new features: -@item -If your client programs are using threads, you need to also compile a -thread-safe version of the @strong{MySQL} client library with the -@code{--with-thread-safe-client} configure options. This will create a -@code{libmysqlclient_r} library with which you should link your threaded -applications. @xref{Thread-safe clients}. +The news section for 4.0 includes a list of the features we have already +implemented in the 4.0 tree. @xref{News-4.0.x}. +@itemize @bullet @item -Options that pertain to particular systems can be found in the -system-specific sections later in this chapter. @xref{Source install -system issues}. -@end itemize - -@node Installing source tree, Compilation problems, Installing source, Installing -@section Installing from the Development Source Tree - -@cindex development source tree -@cindex BitKeeper tree -@cindex cvs tree +New table definition file format (@code{.frm} files) This will enable us +to not run out of bits when adding more table options. One will still +be able to use the old @code{.frm} file format with 4.0. All newly created +tables will, however, use the new format. -@strong{CAUTION:} You should read this section only if you are interested -in helping us test our new code. If you just want to get @strong{MySQL} up -and running on your system, you should use a standard release distribution -(either a source or binary distribution will do). +The new file format will enable us to add new column types, more options +for keys and @code{FOREIGN KEY} support. +@item +@code{mysqld} as a library. This will have the same interface as the +standard @strong{MySQL} client (with an extra function to just set up +startup parameters) but will be faster (no TCP/IP or socket overhead), +smaller and much easier to use for embedded products. -To obtain our most recent development source tree, use these instructions: +One will be able to define at link time if one wants to use the +client/server model or a stand-alone application just by defining which +library to link with. -@enumerate +The @code{mysqld} will support all standard @strong{MySQL} features and +one can use it in a threaded client to run different queries in each +thread. @item -Download @strong{BitKeeper} from -@uref{http://www.bitmover.com/cgi-bin/download.cgi}. You will need -@strong{Bitkeeper} 2.0 or newer to access our repository. - +Replication should work with @code{RAND()} and user variables @code{@@var}. @item -Follow the instructions to install it. - +Online backup with very low performance penalty. The online backup will +make it easy to add a new replication slave without taking down the +master. @item -After @strong{BitKeeper} is installed, use this command if you want to clone -the @strong{MySQL} 3.23 branch: - -@example -shell> bk clone bk://work.mysql.com:7000 mysql -@end example - -To clone the 4.0 branch, use this command instead: - -@example -shell> bk clone bk://work.mysql.com:7001 mysql-4.0 -@end example - -The initial download of the source tree may take a while, depending on the -speed of your connection; be patient. - +@code{DELETE FROM table_name} will return the number of deleted rows. For +fast execution one should use @code{TRUNCATE table_name}. @item -You will need GNU @code{autoconf}, @code{automake}, @code{libtool}, and -@code{m4} to run the next set of commands. -If you get some strange error during this stage, check that you really -have @code{libtool} installed! - -@example -shell> cd mysql -shell> bk -r edit -shell> aclocal; autoheader; autoconf; automake; -shell> ./configure # Add your favorite options here -shell> make -@end example - -A collection of our standard configure scripts is located in the -@file{BUILD/} subdirectory. If you are lazy, you can use -@file{BUILD/compile-pentium-debug}. To compile on a different architecture, -modify the script removing flags that are Pentium-specific. - +Allow @code{DELETE} on @code{MyISAM} tables to use the record cache. +To do this, we need to update the threads record cache when we update +the @code{.MYD} file. @item -When the build is done, run @code{make install}. Be careful with this -on a production machine; the command may overwrite your live release -installation. If you have another installation of @strong{MySQL}, we -recommand that you run @code{./configure} with different values for the -@code{prefix}, @code{tcp-port}, and @code{unix-socket-path} options than -those used for your production server. - +Better replication. @item -Play hard with your new installation and try to make the new features -crash. Start by running @code{make test}. @xref{MySQL test suite}. - +More functions for full-text search. +@xref{Fulltext Features to Appear in MySQL 4.0}. @item -If you have gotten to the @code{make} stage and the distribution does -not compile, please report it to @email{bugs@@lists.mysql.com}. If you -have installed the latest versions of the required GNU tools, and they -crash trying to process our configuration files, please report that also. -However, if you execute @code{aclocal} and get a @code{command not found} -error or a similar problem, do not report it. Instead, make sure all -the necessary tools are installed and that your @code{PATH} variable is -set correctly so your shell can find them. - +Character set casts and syntax for handling multiple character sets. @item -After the initial @code{bk clone} operation to get the source tree, you -should run @code{bk pull} periodically to get the updates. - +Allow users to change startup options without taking down the server. @item -You can examine the change history for the tree with all the diffs by using -@code{bk sccstool}. If you see some funny diffs or code that you have a -question about, do not hesitate to send e-mail to -@email{internals@@lists.mysql.com}. Also, if you think you have a better idea -on how to do something, send an email to the same address with a patch. -@code{bk diffs} will produce a patch for you after you have made changes -to the source. If you do not have the time to code your idea, just send -a description. - +Help for all commands from the client. @item -@strong{BitKeeper} has a nice help utility that you can access via -@code{bk helptool}. - -@end enumerate - -@cindex compiling, problems -@cindex problems, compiling -@cindex reconfiguring -@cindex @code{config.cache} file -@cindex files, @code{config.cache} -@node Compilation problems, MIT-pthreads, Installing source tree, Installing -@section Problems Compiling? - -All @strong{MySQL} programs compile cleanly for us with no warnings on -Solaris using @code{gcc}. On other systems, warnings may occur due to -differences in system include files. See @ref{MIT-pthreads} for warnings -that may occur when using MIT-pthreads. For other problems, check the list -below. - -The solution to many problems involves reconfiguring. If you do need to -reconfigure, take note of the following: - -@cindex running @code{configure} after prior invocation -@cindex @code{configure}, running after prior invocation -@cindex reconfiguring -@cindex config.cache -@itemize @bullet +Secure connections (with SSL). @item -If @code{configure} is run after it already has been run, it may use -information that was gathered during its previous invocation. This -information is stored in @file{config.cache}. When @code{configure} starts -up, it looks for that file and reads its contents if it exists, on the -assumption that the information is still correct. That assumption is invalid -when you reconfigure. - +Extend the optimizer to be able to optimize some @code{ORDER BY key_name DESC} +queries. @item -Each time you run @code{configure}, you must run @code{make} again -to recompile. However, you may want to remove old object files from previous -builds first, because they were compiled using different configuration options. +@code{SHOW COLUMNS FROM table_name} (used by @code{mysql} client to allow +expansions of column names) should not open the table, but only the +definition file. This will require less memory and be much faster. +@item +New key cache +@item +When using @code{SET CHARACTER SET} we should translate the whole query +at once and not only strings. This will enable users to use the translated +characters in database, table and column names. +@item +Add a portable interface over @code{gethostbyaddr_r()} so that we can change +@code{ip_to_hostname()} to not block other threads while doing DNS lookups. +@item +Add @code{record_in_range()} method to @code{MERGE} tables to be +able to choose the right index when there is many to choose from. We should +also extend the info interface to get the key distribution for each index, +of @code{analyze} is run on all sub tables. +@item +@code{SET SQL_DEFAULT_TABLE_TYPE=[MyISAM | INNODB | BDB | HEAP]}. @end itemize -To prevent old configuration information or object files from being used, -run these commands before rerunning @code{configure}: - -@example -shell> rm config.cache -shell> make clean -@end example - -Alternatively, you can run @code{make distclean}. - -The list below describes some of the problems compiling @strong{MySQL} -that have been found to occur most often: +@node TODO future, TODO sometime, TODO MySQL 4.0, TODO +@subsection Things that must be done in the real near future @itemize @bullet @item -@cindex @code{cc1plus} problems -@cindex @code{fatal signal 11} -@cindex @code{sql_yacc.cc} problems -@cindex internal compiler errors -@cindex virtual memory, problems while compiling -@cindex configure option, --with-low-memory -If you get errors when compiling @file{sql_yacc.cc}, such as the ones shown -below, you have probably run out of memory or swap space: - -@example -Internal compiler error: program cc1plus got fatal signal 11 - or -Out of virtual memory - or -Virtual memory exhausted -@end example - -The problem is that @code{gcc} requires huge amounts of memory to compile -@file{sql_yacc.cc} with inline functions. Try running @code{configure} with -the @code{--with-low-memory} option: - -@example -shell> ./configure --with-low-memory -@end example - -This option causes @code{-fno-inline} to be added to the compile line if you -are using @code{gcc} and @code{-O0} if you are using something else. You -should try the @code{--with-low-memory} option even if you have so much -memory and swap space that you think you can't possibly have run out. This -problem has been observed to occur even on systems with generous hardware -configurations, and the @code{--with-low-memory} option usually fixes it. - +Fail safe replication. @item -By default, @code{configure} picks @code{c++} as the compiler name and -GNU @code{c++} links with @code{-lg++}. If you are using @code{gcc}, -that behavior can cause problems during configuration such as this: - -@cindex C++ compiler cannot create executables -@example -configure: error: installation or configuration problem: -C++ compiler cannot create executables. -@end example - -@tindex CXX environment variable -@tindex Environment variable, CXX -You might also observe problems during compilation related to -@code{g++}, @code{libg++}, or @code{libstdc++}. - -One cause of these problems is that you may not have @code{g++}, or you may -have @code{g++} but not @code{libg++}, or @code{libstdc++}. Take a look at -the @file{config.log} file. It should contain the exact reason why your c++ -compiler didn't work! To work around these problems, you can use @code{gcc} -as your C++ compiler. Try setting the environment variable @code{CXX} to -@code{"gcc -O3"}. For example: - -@tindex CXX environment variable -@tindex Environment variables, CXX -@example -shell> CXX="gcc -O3" ./configure -@end example - -This works because @code{gcc} compiles C++ sources as well as @code{g++} -does, but does not link in @code{libg++} or @code{libstdc++} by default. - -Another way to fix these problems, of course, is to install @code{g++}, -@code{libg++} and @code{libstdc++}. - +Subqueries. +@code{select id from t where grp in (select grp from g where u > 100)} @item -If your compile fails with errors, such as any of the following, -you must upgrade your version of @code{make} to GNU @code{make}: - +Derived tables. @example -making all in mit-pthreads -make: Fatal error in reader: Makefile, line 18: -Badly formed macro assignment - or -make: file `Makefile' line 18: Must be a separator (: - or -pthread.h: No such file or directory +select a.col1, b.col2 from (select max(col1) as col1 from root_table ) a, +other_table b where a.col1=b.col1 @end example -@cindex Solaris troubleshooting -@cindex FreeBSD troubleshooting -@cindex troubleshooting, Solaris -@cindex troubleshooting, FreeBSD -Solaris and FreeBSD are known to have troublesome @code{make} programs. - -GNU @code{make} Version 3.75 is known to work. - -@tindex CC environment variable -@tindex environment variable, CC -@tindex CXX environment variable -@tindex environment variable, CXX -@tindex CFLAGS environment variable -@tindex environment variable, CFLAGS -@tindex CXXFLAGS environment variable -@tindex environment variable, CXXFLAGS +This could be done by automatically creating temporary tables for the +derived tables for the duration of the query. @item -If you want to define flags to be used by your C or C++ compilers, do so by -adding the flags to the @code{CFLAGS} and @code{CXXFLAGS} environment -variables. You can also specify the compiler names this way using @code{CC} -and @code{CXX}. For example: - -@example -shell> CC=gcc -shell> CFLAGS=-O3 -shell> CXX=gcc -shell> CXXFLAGS=-O3 -shell> export CC CFLAGS CXX CXXFLAGS -@end example - -See @ref{MySQL binaries}, for a list of flag definitions that have been found -to be useful on various systems. - +Add @code{PREPARE} of statements and sending of parameters to @code{mysqld}. @item -If you get an error message like this, -you need to upgrade your @code{gcc} compiler: - -@example -client/libmysql.c:273: parse error before `__attribute__' -@end example - -@code{gcc} 2.8.1 is known to work, but we recommend using @code{gcc} 2.95.2 or -@code{egcs} 1.0.3a instead. - +Extend the server/client protocol to support warnings. @item -If you get errors such as those shown below when compiling @code{mysqld}, -@code{configure} didn't correctly detect the type of the last argument to -@code{accept()}, @code{getsockname()}, or @code{getpeername()}: - -@example -cxx: Error: mysqld.cc, line 645: In this statement, the referenced - type of the pointer value "&length" is "unsigned long", which - is not compatible with "int". -new_sock = accept(sock, (struct sockaddr *)&cAddr, &length); -@end example - -To fix this, edit the @file{config.h} file (which is generated by -@code{configure}). Look for these lines: - -@example -/* Define as the base type of the last arg to accept */ -#define SOCKET_SIZE_TYPE XXX -@end example - -Change @code{XXX} to @code{size_t} or @code{int}, depending on your -operating system. (Note that you will have to do this each time you run -@code{configure}, because @code{configure} regenerates @file{config.h}.) - +Add options to the server/protocol protocol to get progress notes +for long running commands. @item -The @file{sql_yacc.cc} file is generated from @file{sql_yacc.yy}. Normally -the build process doesn't need to create @file{sql_yacc.cc}, because -@strong{MySQL} comes with an already-generated copy. However, if you do need -to re-create it, you might encounter this error: - -@example -"sql_yacc.yy", line xxx fatal: default action causes potential... -@end example - -This is a sign that your version of @code{yacc} is deficient. -You probably need to install @code{bison} (the GNU version of @code{yacc}) -and use that instead. - +Add database and real table name (in case of alias) to the MYSQL_FIELD +structure. @item -If you need to debug @code{mysqld} or a @strong{MySQL} client, run -@code{configure} with the @code{--with-debug} option, then recompile and -link your clients with the new client library. @xref{Debugging client}. -@end itemize - -@cindex MIT-pthreads -@cindex thread support, non-native -@node MIT-pthreads, Perl support, Compilation problems, Installing -@section MIT-pthreads Notes - -This section describes some of the issues involved in using MIT-pthreads. - -Note that on Linux you should NOT use MIT-pthreads but install LinuxThreads! -@xref{Linux}. - -If your system does not provide native thread support, you will need to -build @strong{MySQL} using the MIT-pthreads package. This includes -older FreeBSD systems, SunOS 4.x, Solaris 2.4 and earlier, and some others. -@xref{Which OS}. - +Don't allow more than a defined number of threads to run MyISAM recover +at the same time. +@item +Change @code{INSERT ... SELECT} to optionally use concurrent inserts. +@item +Implement @code{RENAME DATABASE}. To make this safe for all table handlers, +it should work as follows: @itemize @bullet @item -On most systems, you can force MIT-pthreads to be used by running -@code{configure} with the @code{--with-mit-threads} option: - -@example -shell> ./configure --with-mit-threads -@end example - -Building in a non-source directory is not supported when using -MIT-pthreads, because we want to minimize our changes to this code. - +Create the new database. @item -The checks that determine whether or not to use MIT-pthreads occur only -during the part of the configuration process that deals with the server -code. If you have configured the distribution using @code{--without-server} -to build only the client code, clients will not know whether or not -MIT-pthreads is being used and will use Unix socket connections by default. -Because Unix sockets do not work under MIT-pthreads, this means you will need -to use @code{-h} or @code{--host} when you run client programs. - +For every table do a rename of the table to another database, as +we do with the @code{RENAME} command. @item -When @strong{MySQL} is compiled using MIT-pthreads, system locking is -disabled by default for performance reasons. You can tell the server to use -system locking with the @code{--use-locking} option. - +Drop the old database. +@end itemize @item -Sometimes the pthread @code{bind()} command fails to bind to a socket without -any error message (at least on Solaris). The result is that all connections -to the server fail. For example: - -@example -shell> mysqladmin version -mysqladmin: connect to server at '' failed; -error: 'Can't connect to mysql server on localhost (146)' -@end example - -The solution to this is to kill the @code{mysqld} server and restart it. -This has only happened to us when we have forced the server down and done -a restart immediately. - +Return the original field types() when doing @code{SELECT MIN(column) +... GROUP BY}. @item -With MIT-pthreads, the @code{sleep()} system call isn't interruptible with -@code{SIGINT} (break). This is only noticeable when you run -@code{mysqladmin --sleep}. You must wait for the @code{sleep()} call to -terminate before the interrupt is served and the process stops. - +Multiple result sets. @item -When linking, you may receive warning messages like these (at least on -Solaris); they can be ignored: - -@example -ld: warning: symbol `_iob' has differing sizes: - (file /my/local/pthreads/lib/libpthread.a(findfp.o) value=0x4; -file /usr/lib/libc.so value=0x140); - /my/local/pthreads/lib/libpthread.a(findfp.o) definition taken -ld: warning: symbol `__iob' has differing sizes: - (file /my/local/pthreads/lib/libpthread.a(findfp.o) value=0x4; -file /usr/lib/libc.so value=0x140); - /my/local/pthreads/lib/libpthread.a(findfp.o) definition taken -@end example - +Change the protocol to allow binary transfer of values. To do this +efficiently, we need to add an API to allow binding of variables. @item -Some other warnings also can be ignored: - -@example -implicit declaration of function `int strtoll(...)' -implicit declaration of function `int strtoul(...)' -@end example - +Make it possible to specify @code{long_query_time} with a granularity +in microseconds. @item -We haven't gotten @code{readline} to work with MIT-pthreads. (This isn't -needed, but may be interesting for someone.) -@end itemize - -@cindex Perl, installing -@cindex installing, Perl -@node Perl support, Source install system issues, MIT-pthreads, Installing -@section Perl Installation Comments - -@menu -* Perl installation:: Installing Perl on Unix -* ActiveState Perl:: Installing ActiveState Perl on Windows -* Windows Perl:: Installing the @strong{MySQL} Perl distribution on Windows -* Perl support problems:: Problems using the Perl @code{DBI}/@code{DBD} interface -@end menu - -@node Perl installation, ActiveState Perl, Perl support, Perl support -@subsection Installing Perl on Unix - -Perl support for @strong{MySQL} is provided by means of the -@code{DBI}/@code{DBD} client interface. @xref{Perl}. The Perl -@code{DBD}/@code{DBI} client code requires Perl Version 5.004 or later. The -interface @strong{will not work} if you have an older version of Perl. - -@strong{MySQL} Perl support also requires that you've installed -@strong{MySQL} client programming support. If you installed @strong{MySQL} -from RPM files, client programs are in the client RPM, but client programming -support is in the developer RPM. Make sure you've installed the latter RPM. - -As of Version 3.22.8, Perl support is distributed separately from the main -@strong{MySQL} distribution. If you want to install Perl support, the files -you will need can be obtained from -@uref{http://www.mysql.com/Downloads/Contrib/}. - -The Perl distributions are provided as compressed @code{tar} archives and -have names like @file{MODULE-VERSION.tar.gz}, where @code{MODULE} is the -module name and @code{VERSION} is the version number. You should get the -@code{Data-Dumper}, @code{DBI}, and @code{Msql-Mysql-modules} distributions -and install them in that order. The installation procedure is shown below. -The example shown is for the @code{Data-Dumper} module, but the procedure is -the same for all three distributions: - -@enumerate +Add a configurable prompt to the @code{mysql} command line client, with +options like database in use, time and date... @item -Unpack the distribution into the current directory: -@example -shell> gunzip < Data-Dumper-VERSION.tar.gz | tar xvf - -@end example -This command creates a directory named @file{Data-Dumper-VERSION}. - +Add range checking to @code{MERGE} tables. @item -Change into the top-level directory of the unpacked distribution: -@example -shell> cd Data-Dumper-VERSION -@end example - +Link the @code{myisampack} code into the server. @item -Build the distribution and compile everything: -@example -shell> perl Makefile.PL -shell> make -shell> make test -shell> make install -@end example -@end enumerate - -The @code{make test} command is important because it verifies that the -module is working. Note that when you run that command during the -@code{Msql-Mysql-modules} installation to exercise the interface code, the -@strong{MySQL} server must be running or the test will fail. - -It is a good idea to rebuild and reinstall the @code{Msql-Mysql-modules} -distribution whenever you install a new release of @strong{MySQL}, -particularly if you notice symptoms such as all your @code{DBI} scripts -dumping core after you upgrade @strong{MySQL}. - -If you don't have the right to install Perl modules in the system directory -or if you to install local Perl modules, the following reference may help -you: - -@example -@uref{http://www.iserver.com/support/contrib/perl5/modules.html} -@end example - -Look under the heading -@code{Installing New Modules that Require Locally Installed Modules}. - -@node ActiveState Perl, Windows Perl, Perl installation, Perl support -@subsection Installing ActiveState Perl on Windows -@cindex installing, Perl on Windows -@cindex Perl, installing on Windows -@cindex ActiveState Perl - -To install the @strong{MySQL} @code{DBD} module with ActiveState Perl on -Windows, you should do the following: - -@itemize @bullet -@item -Get ActiveState Perl from -@uref{http://www.activestate.com/Products/ActivePerl/index.html} -and install it. - +Port of @strong{MySQL} to BeOS. @item -Open a DOS shell. - -@item -If required, set the HTTP_proxy variable. For example, you might try: - -@example -set HTTP_proxy=my.proxy.com:3128 -@end example - +Port of the @strong{MySQL} clients to LynxOS. @item -Start the PPM program: - -@example -C:\> c:\perl\bin\ppm.pl -@end example - +Add a temporary key buffer cache during @code{INSERT/DELETE/UPDATE} so that we +can gracefully recover if the index file gets full. @item -If you have not already done so, install @code{DBI}: - -@example -ppm> install DBI -@end example - +If you perform an @code{ALTER TABLE} on a table that is symlinked to another +disk, create temporary tables on this disk. @item -If this succeeds, run the following command: - -@example -install ftp://ftp.de.uu.net/pub/CPAN/authors/id/JWIED/DBD-mysql-1.2212.x86.ppd -@end example -@end itemize - -The above should work at least with ActiveState Perl Version 5.6. +Implement a @code{DATE/DATETIME} type that handles time zone information +properly, so that dealing with dates in different time zones is easier. +@item +FreeBSD and MIT-pthreads; Do sleeping threads take CPU? +@item +Check if locked threads take any CPU. +@item +Fix configure so that one can compile all libraries (like @code{MyISAM}) +without threads. +@item +Add an option to periodically flush key pages for tables with delayed keys +if they haven't been used in a while. +@item +Allow join on key parts (optimization issue). +@item +@code{INSERT SQL_CONCURRENT} and @code{mysqld --concurrent-insert} to do +a concurrent insert at the end of the file if the file is read-locked. +@item +Remember @code{FOREIGN} key definitions in the @file{.frm} file. +@item +Cascading @code{DELETE} +@item +Server side cursors. +@item +Check if @code{lockd} works with modern Linux kernels; If not, we have +to fix @code{lockd}! To test this, start @code{mysqld} with +@code{--enable-locking} and run the different fork* test suits. They shouldn't +give any errors if @code{lockd} works. +@item +Allow SQL variables in @code{LIMIT}, like in @code{LIMIT @@a,@@b}. +@item +Allow update of variables in @code{UPDATE} statements. For example: +@code{UPDATE TABLE foo SET @@a=a+b,a=@@a, b=@@a+c} +@item +Change when user variables are updated so that one can use them with +@code{GROUP BY}, as in the following example: +@code{SELECT id, @@a:=count(*), sum(sum_col)/@@a FROM table_name GROUP BY id}. +@item +Don't add automatic @code{DEFAULT} values to columns. Give an error when using +an @code{INSERT} that doesn't contain a column that doesn't have a +@code{DEFAULT}. +@item +Caching of queries and results. This should be done as a separated +module that examines each query and if this is query is in the cache +the cached result should be returned. When one updates a table one +should remove as few queries as possible from the cache. +This should give a big speed bost on machines with much RAM where +queries are often repeated (like WWW applications). +One idea would be to only cache queries of type: +@code{SELECT CACHED ...} +@item +Fix @file{libmysql.c} to allow two @code{mysql_query()} commands in a row +without reading results or give a nice error message when one does this. +@item +Optimize @code{BIT} type to take 1 bit (now @code{BIT} takes 1 char). +@item +Check why MIT-pthreads @code{ctime()} doesn't work on some FreeBSD systems. +@item +Add an @code{IMAGE} option to @code{LOAD DATA INFILE} to not update +@code{TIMESTAMP} and @code{AUTO_INCREMENT} fields. +@item +Added @code{LOAD DATE INFILE.. UPDATE} syntax. +@itemize @bullet +@item +For tables with primary keys, if the data contains the primary key, +entries matching that primary key are updated from the remainder of the +columns. However, columns MISSING from the incoming data feed are not +touched. +@item +For tables tables with primary keys that are missing some part of the key +in the incoming data stream, or that have no primary key, the feed is +treated as a @code{LOAD DATA INFILE ... REPLACE INTO} is now. +@end itemize +@item +Make @code{LOAD DATA INFILE} understand syntax like: +@example +LOAD DATA INFILE 'file_name.txt' INTO TABLE tbl_name +TEXT_FIELDS (text_field1, text_field2, text_field3) +SET table_field1=concatenate(text_field1, text_field2), table_field3=23 +IGNORE text_field3 -If you can't get the above to work, you should instead install the -@strong{MyODBC} driver and connect to @strong{MySQL} server through -ODBC: +This can be used to skip over extra columns in the text file, or update columns +based on expressions of the read data... +@end example +@item +@code{LOAD DATA INFILE 'file_name' INTO TABLE 'table_name' ERRORS TO err_table_name} +This would cause any errors and warnings to be logged into the err_table_name +table. That table would have a structure like: @example -use DBI; -$dbh= DBI->connect("DBI:ODBC:$dsn","$user","$password") || - die "Got error $DBI::errstr when connecting to $dsn\n"; +line_number - line number in data file +error_message - the error/warning message +and maybe +data_line - the line from the data file @end example +@item +Add true @code{VARCHAR} support (There is already support for this in MyISAM). +@item +Automatic output from @code{mysql} to netscape. +@item +@code{LOCK DATABASES}. (with various options) +@item +Change sort to allocate memory in ``hunks'' to get better memory utilization. +@item +@code{DECIMAL} and @code{NUMERIC} types can't read exponential numbers; +@code{Field_decimal::store(const char *from,uint len)} must be recoded +to fix this. +@item +Fix @code{mysql.cc} to do fewer @code{malloc()} calls when hashing field +names. +@item +Functions: +ADD_TO_SET(value,set) and REMOVE_FROM_SET(value,set) +@item +Add use of @code{t1 JOIN t2 ON ...} and @code{t1 JOIN t2 USING ...} +Currently, you can only use this syntax with @code{LEFT JOIN}. +@item +Add full support for @code{unsigned long long} type. +@item +Many more variables for @code{show status}. Counts for: +@code{INSERT}/@code{DELETE}/@code{UPDATE} statements. Records reads and +updated. Selects on 1 table and selects with joins. Mean number of +tables in select. Number of @code{ORDER BY} and @code{GROUP BY} queries. +@item +If you abort @code{mysql} in the middle of a query, you should open +another connection and kill the old running query. +Alternatively, an attempt should be made to detect this in the server. +@item +Add a handler interface for table information so you can use it as a system +table. This would be a bit slow if you requested information about all tables, +but very flexible. @code{SHOW INFO FROM tbl_name} for basic table information +should be implemented. +@item +Add support for UNICODE. +@item +@code{NATURAL JOIN} and @code{UNION JOIN} +@item +Allow @code{select a from crash_me left join crash_me2 using (a)}; In this +case a is assumed to come from the crash_me table. +@item +Fix that @code{ON} and @code{USING} works with the @code{JOIN} and +@code{INNER JOIN} join types. +@item +Oracle like @code{CONNECT BY PRIOR ...} to search hierarchy structures. +@item +@code{RENAME DATABASE} +@item +@code{mysqladmin copy database new-database}. -- Requires COPY command to be +added to @code{mysqld} +@item +Processlist should show number of queries/thread. +@item +@code{SHOW HOSTS} for printing information about the hostname cache. +@item +@code{DELETE} and @code{REPLACE} options to the @code{UPDATE} statement +(this will delete rows when one gets a duplicate key error while updating). +@item +Change the format of @code{DATETIME} to store fractions of seconds. +@item +Add all missing ANSI92 and ODBC 3.0 types. +@item +Change table names from empty strings to @code{NULL} for calculated columns. +@item +Don't use 'Item_copy_string' on numerical values to avoid +number->string->number conversion in case of: +@code{SELECT COUNT(*)*(id+0) FROM table_name GROUP BY id} +@item +Make it possible to use the new GNU regexp library instead of the current +one (The GNU library should be much faster than the old one). +@item +Change that @code{ALTER TABLE} doesn't abort clients that executes +@code{INSERT DELAYED}. +@item +Fix that when columns referenced in an @code{UPDATE} clause contains the old +values before the update started. +@item +@code{myisamchk}, @code{REPAIR} and @code{OPTIMIZE TABLE} should be able +to handle cases where the data and/or index files are symbolic links. +@item +Add simulation of @code{pread()}/@code{pwrite()} on Windows to enable +concurrent inserts. +@item +A logfile analyzer that could parsed out information about which tables +are hit most often, how often multi-table joins are executed, etc. It +should help users identify areas or table design that could be optimized +to execute much more efficient queries. +@item +Add @code{SUM(DISTINCT)} +@item +Add @code{ANY()},@code{EVERY()} and @code{SOME()} group functions. In +ANSI SQL these only works on boolean columns, but we can extend these to +work on any columns/expressions by applying: value == 0 -> FALSE and +value <> 0 -> TRUE. +@item +Fix that the type for @code{MAX(column)} is the same as the column type. +@example +create table t1 (a DATE); +insert into t1 values (now()); +create table t2 select max(a) from t1; +show columns from t2; +@end example +@item +Come up with a nice syntax for a statement that will @code{UPDATE} the row +if it exists and @code{INSERT} a new row if the row didn't exist. +(Like @code{REPLACE} works with @code{INSERT} / @code{DELETE}) +@end itemize -@node Windows Perl, Perl support problems, ActiveState Perl, Perl support -@subsection Installing the MySQL Perl Distribution on Windows - -The @strong{MySQL} Perl distribution contains @code{DBI}, -@code{DBD:MySQL} and @code{DBD:ODBC}. +@node TODO sometime, TODO unplanned, TODO future, TODO +@subsection Things that have to be done sometime @itemize @bullet -@item -Get the Perl distribution for Windows from -@uref{http://www.mysql.com/download.html}. - @item -Unzip the distribution in @code{C:} so that you get a @file{C:\PERL} directory. - +Implement function: @code{get_changed_tables(timeout,table1,table2,...)} @item -Add the directory @file{C:\PERL\BIN} to your path. - +Atomic multi-table updates, eg @code{update items,month set +items.price=month.price where items.id=month.id;}; @item -Add the directory @file{C:\PERL\BIN\MSWIN32-x86-thread} or -@file{C:\PERL\BIN\MSWIN32-x86} to your path. +Change reading through tables to use memmap when possible. Now only +compressed tables use memmap. +@item +Add a new privilege @strong{'Show_priv'} for @code{SHOW} commands. +@item +Make the automatic timestamp code nicer. Add timestamps to the update +log with @code{SET TIMESTAMP=#;} +@item +Use read/write mutex in some places to get more speed. +@item +Full foreign key support. One probably wants to implement a procedural +language first. +@item +Simple views (first on one table, later on any expression). +@item +Automatically close some tables if a table, temporary table or temporary files +gets error 23 (not enough open files). +@item +When one finds a field=#, change all occurrences of field to #. Now this +is only done for some simple cases. +@item +Change all const expressions with calculated expressions if possible. +@item +Optimize key = expression. At the moment only key = field or key = +constant are optimized. +@item +Join some of the copy functions for nicer code. +@item +Change @file{sql_yacc.yy} to an inline parser to reduce its size and get +better error messages (5 days). +@item +Change the parser to use only one rule per different number of arguments +in function. +@item +Use of full calculation names in the order part. (For ACCESS97) +@item +@code{UNION}, @code{MINUS}, @code{INTERSECT} and @code{FULL OUTER JOIN}. +(Currently only @code{LEFT OUTER JOIN} is supported) +@item +Allow @code{UNIQUE} on fields that can be @code{NULL}. +@item +@code{SQL_OPTION MAX_SELECT_TIME=#} to put a time limit on a query. +@item +Make the update log to a database. +@item +Negative @code{LIMIT} to retrieve data from the end. +@item +Alarm around client connect/read/write functions. +@item +Please note the changes to @code{safe_mysqld}: according to FSSTND (which +Debian tries to follow) PID files should go into @file{/var/run/<progname>.pid} +and log files into @file{/var/log}. It would be nice if you could put the +"DATADIR" in the first declaration of "pidfile" and "log", so the +placement of these files can be changed with a single statement. +@item +Allow a client to request logging. +@item +Add use of @code{zlib()} for @code{gzip}-ed files to @code{LOAD DATA INFILE}. +@item +Fix sorting and grouping of @code{BLOB} columns (partly solved now). +@item +Stored procedures. This is currently not regarded to be very +important as stored procedures are not very standardized yet. +Another problem is that true stored procedures make it much harder for +the optimizer and in many cases the result is slower than before +We will, on the other hand, add a simple (atomic) update language that +can be used to write loops and such in the @strong{MySQL} server. +@item +Change to use semaphores when counting threads. One should first implement +a semaphore library to MIT-pthreads. +@item +Don't assign a new @code{AUTO_INCREMENT} value when one sets a column to 0. +Use @code{NULL} instead. +@item +Add full support for @code{JOIN} with parentheses. +@item +As an alternative for one thread / connection manage a pool of threads +to handle the queries. +@item +Allow one to get more than one lock with @code{GET_LOCK}. When doing this, +one must also handle the possible deadlocks this change will introduce. +@end itemize + +Time is given according to amount of work, not real time. + +@node TODO unplanned, , TODO sometime, TODO +@subsection Some things we don't have any plans to do +@itemize @bullet @item -Test that @code{perl} works by executing @code{perl -v} in a DOS shell. +Nothing; In the long run we plan to be fully ANSI 92 / ANSI 99 compliant. @end itemize -@cindex problems, installing Perl -@cindex Perl DBI/DBD, installation problems -@node Perl support problems, , Windows Perl, Perl support -@subsection Problems Using the Perl @code{DBI}/@code{DBD} Interface +@node Installing, Privilege system, Introduction, Top +@chapter Installing MySQL -If Perl reports that it can't find the @file{../mysql/mysql.so} module, -then the problem is probably that Perl can't locate the shared library -@file{libmysqlclient.so}. +@cindex installing, overview -You can fix this by any of the following methods: +@menu +* Getting MySQL:: How to get @strong{MySQL} +* Which OS:: Operating systems supported by @strong{MySQL} +* Which version:: Which @strong{MySQL} version to use +* Many versions:: How and when updates are released +* Installation layouts:: Installation layouts +* Installing binary:: Installing a @strong{MySQL} binary distribution +* Installing source:: Installing a @strong{MySQL} source distribution +* Installing source tree:: Installing @strong{MySQL} from development source tree +* Compilation problems:: Problems compiling? +* MIT-pthreads:: MIT-pthreads notes +* Perl support:: Perl installation comments +* Source install system issues:: System-specific issues +* Windows:: Windows notes +* OS/2:: OS/2 notes +* MySQL binaries:: MySQL binaries +* Post-installation:: Post-installation setup and testing +* Installing many servers:: Installing many servers on the same machine +* Upgrade:: Upgrading/Downgrading MySQL +@end menu + +This chapter describes how to obtain and install @strong{MySQL}: @itemize @bullet @item -Compile the @code{Msql-Mysql-modules} distribution with @code{perl -Makefile.PL -static -config} rather than @code{perl Makefile.PL}. +For a list of sites from which you can obtain @strong{MySQL}, see +@ref{Getting MySQL, , Getting @strong{MySQL}}. @item -Copy @code{libmysqlclient.so} to the directory where your other shared -libraries are located (probably @file{/usr/lib} or @file{/lib}). +To see which platforms are supported, see @ref{Which OS}. Please note that +not all supported system are equally good for running @strong{MySQL} on them. +On some it is much more robust and efficient than others - see @ref{Which OS} +for details. @item -On Linux you can add the pathname of the directory where -@file{libmysqlclient.so} is located to the @file{/etc/ld.so.conf} file. +Several versions of @strong{MySQL} are available in both binary and +source distributions. We also provide public access to our current +source tree for those who want to see our most recent developments and +help us test new code. To determine which version and type of +distribution you should use, see @ref{Which version}. When in doubt, +use the binary distribution. -@tindex LD_RUN_PATH environment variable -@tindex Environment variable, LD_RUN_PATH @item -Add the pathname of the directory where @file{libmysqlclient.so} is located -to the @code{LD_RUN_PATH} environment variable. +Installation instructions for binary and source distributions are described +in @ref{Installing binary} and @ref{Installing source}. Each set of +instructions includes a section on system-specific problems you may run +into. + +@item +For post-installation procedures, see @ref{Post-installation}. These +procedures apply whether you install @strong{MySQL} using a binary or +source distribution. @end itemize -If you get the following errors from @code{DBD-mysql}, -you are probably using @code{gcc} (or using an old binary compiled with -@code{gcc}): +@cindex downloading +@cindex MySQL version +@cindex version, latest +@cindex getting MySQL +@cindex mirror sites +@cindex URLS for downloading MySQL +@node Getting MySQL, Which OS, Installing, Installing +@section How to Get MySQL -@example -/usr/bin/perl: can't resolve symbol '__moddi3' -/usr/bin/perl: can't resolve symbol '__divdi3' -@end example +Check the @uref{http://www.mysql.com/, @strong{MySQL} home page} for +information about the current version and for downloading instructions. -Add @code{-L/usr/lib/gcc-lib/... -lgcc} to the link command when the -@file{mysql.so} library gets built (check the output from @code{make} for -@file{mysql.so} when you compile the Perl client). The @code{-L} option -should specify the pathname of the directory where @file{libgcc.a} is located -on your system. +Our main download mirror is located at: -Another cause of this problem may be that Perl and @strong{MySQL} aren't both -compiled with @code{gcc}. In this case, you can solve the mismatch by -compiling both with @code{gcc}. +@uref{http://download.sourceforge.net/mirrors/mysql/} -If you get the following error from @code{Msql-Mysql-modules} -when you run the tests: +If you are interested in becoming a @strong{MySQL} mirror site, you may +anonymously rsync with: @code{rsync://download.sourceforge.net/mysql/}. Please +send e-mail to @email{webmaster@@mysql.com} notifying us of your mirror to be +added to the list below. -@example -t/00base............install_driver(mysql) failed: Can't load '../blib/arch/auto/DBD/mysql/mysql.so' for module DBD::mysql: ../blib/arch/auto/DBD/mysql/mysql.so: undefined symbol: uncompress at /usr/lib/perl5/5.00503/i586-linux/DynaLoader.pm line 169. -@end example +If you have problems downloading from our main site, try using one of the +mirrors listed below. -it means that you need to include the compression library, -lz, to the -link line. This can be doing the following change in the file -@file{lib/DBD/mysql/Install.pm}: +Please report bad or out-of-date mirrors to @email{webmaster@@mysql.com}. -@example -$sysliblist .= " -lm"; +@c START_OF_MIRROR_LISTING -to +@strong{Europe:} -$sysliblist .= " -lm -lz"; -@end example +@itemize @bullet -After this, you MUST run 'make realclean' and then proceed with the -installation from the beginning. +@item +@c EMAIL: sl@iuinfo.tuwien.ac.at (Tony Sprinzl) +@image{Flags/austria} Austria [Univ. of Technology/Vienna] @ +@uref{http://gd.tuwien.ac.at/db/mysql/, WWW} +@uref{ftp://gd.tuwien.ac.at/db/mysql/, FTP} -If you want to use the Perl module on a system that doesn't support dynamic -linking (like SCO) you can generate a static version of Perl that includes -@code{DBI} and @code{DBD-mysql}. The way this works is that you generate a -version of Perl with the @code{DBI} code linked in and install it on top of -your current Perl. Then you use that to build a version of Perl that -additionally has the @code{DBD} code linked in, and install that. +@c @item +@c Not ok 20000919; Old site (Matt) +@c EMAIL: delian@naturella.com (Delian Delchev) +@c @image{Flags/bulgaria} Bulgaria [Naturella] @ +@c @uref{http://archive.nat.bg/pub/mirror/mysql/, WWW} +@c @uref{ftp://ftp.ntrl.net/pub/mirror/mysql/, FTP} -On SCO, you must have the following environment variables set: +@item +@c EMAIL: salle@online.bg (Admin) +@image{Flags/bulgaria} Bulgaria [online.bg/Sofia] @ +@uref{http://mysql.online.bg/, WWW} +@uref{ftp://mysql.online.bg/, FTP} -@example -shell> LD_LIBRARY_PATH=/lib:/usr/lib:/usr/local/lib:/usr/progressive/lib -or -shell> LD_LIBRARY_PATH=/usr/lib:/lib:/usr/local/lib:/usr/ccs/lib:/usr/progressive/lib:/usr/skunk/lib -shell> LIBPATH=/usr/lib:/lib:/usr/local/lib:/usr/ccs/lib:/usr/progressive/lib:/usr/skunk/lib -shell> MANPATH=scohelp:/usr/man:/usr/local1/man:/usr/local/man:/usr/skunk/man: -@end example +@c @item +@c Added: 990614 +@c Not ok 20000919; Out of date (Matt) +@c EMAIL: vuksan@veus.hr (Vladimir Vuksan) +@c @image{Flags/croatia} Croatia [HULK] @ +@c @uref{http://ftp.linux.hr/pub/mysql/, WWW} +@c @uref{ftp://ftp.linux.hr/pub/mysql/, FTP} -First, create a Perl that includes a statically linked @code{DBI} by running -these commands in the directory where your @code{DBI} distribution is -located: +@item +@c Added: 990614 +@c EMAIL: kas@informatics.muni.cz (Jan Kasprzak) +@image{Flags/czech-republic} Czech Republic [Masaryk University in Brno] @ +@uref{http://mysql.linux.cz/index.html, WWW} +@uref{ftp://ftp.fi.muni.cz/pub/mysql/, FTP} -@example -shell> perl Makefile.PL -static -config -shell> make -shell> make install -shell> make perl -@end example +@item +@c Added: 990920 +@c EMAIL: <radek@sopik.cz> (Radek Libovicky) +@image{Flags/czech-republic} Czech Republic [www.sopik.cz] @ +@uref{http://www.mysql.cz/, WWW} -Then you must install the new Perl. The output of @code{make perl} will -indicate the exact @code{make} command you will need to execute to perform -the installation. On SCO, this is @code{make -f Makefile.aperl inst_perl -MAP_TARGET=perl}. +@item +@c Added: 000418 +@c EMAIL: <feela@ipex.cz> (Ondrej Feela Filip) +@image{Flags/czech-republic} Czech Republic [www.gin.cz] @ +@uref{http://mysql.gin.cz/, WWW} +@uref{ftp://ftp.gin.cz/pub/MIRRORS/www.mysql.com/, FTP} -Next, use the just-created Perl to create another Perl that also includes a -statically-linked @code{DBD::mysql} by running these commands in the -directory where your @code{Msql-Mysql-modules} distribution is located: +@item +@c removed 991020 (no DNS entry). New name 991026. Added 991121 +@c Statistics at http://mirror.borsen.dk/ +@c EMAIL: mirrorman@borsen.dk (Michael Kyed) +@image{Flags/denmark} Denmark [Borsen] @ +@uref{ http://mysql.borsen.dk/, WWW} -@example -shell> perl Makefile.PL -static -config -shell> make -shell> make install -shell> make perl -@end example - -Finally, you should install this new Perl. Again, the output of @code{make -perl} indicates the command to use. - -@node Source install system issues, Windows, Perl support, Installing -@section System-specific Issues - -The following sections indicate some of the issues that have been observed to -occur on particular systems when installing @strong{MySQL} from a source -distribution. - -@menu -* Solaris:: Solaris notes -* Solaris 2.7:: Solaris 2.7 / 2.8 notes -* Solaris x86:: Solaris x86 notes -* SunOS:: SunOS 4 notes -* Linux:: Linux notes (all Linux versions) -* Alpha-DEC-UNIX:: Alpha-DEC-UNIX notes -* Alpha-DEC-OSF1:: Alpha-DEC-OSF1 notes -* SGI-Irix:: SGI-Irix notes -* FreeBSD:: FreeBSD notes -* NetBSD:: NetBSD notes -* OpenBSD:: OpenBSD 2.5 notes -* BSDI:: BSD/OS notes -* SCO:: SCO notes -* SCO Unixware:: SCO Unixware 7.0 notes -* IBM-AIX:: IBM-AIX notes -* HP-UX 10.20:: HP-UX 10.20 notes -* HP-UX 11.x:: HP-UX 11.x notes -* Mac OS X:: Mac OS X notes -* BEOS:: BeOS Notes -@end menu - - -@cindex Solaris installation problems -@cindex problems, installing on Solaris -@cindex tar, problems on Solaris -@cindex errors, directory checksum -@cindex checksum errors -@node Solaris, Solaris 2.7, Source install system issues, Source install system issues -@subsection Solaris Notes - -On Solaris, you may run into trouble even before you get the @strong{MySQL} -distribution unpacked! Solaris @code{tar} can't handle long file names, so -you may see an error like this when you unpack @strong{MySQL}: - -@example -x mysql-3.22.12-beta/bench/Results/ATIS-mysql_odbc-NT_4.0-cmp-db2,informix,ms-sql,mysql,oracle,solid,sybase, 0 bytes, 0 tape blocks -tar: directory checksum error -@end example - -In this case, you must use GNU @code{tar} (@code{gtar}) to unpack the -distribution. You can find a precompiled copy for Solaris at -@uref{http://www.mysql.com/Downloads/}. - -Sun native threads work only on Solaris 2.5 and higher. For Version 2.4 and -earlier, @strong{MySQL} will automatically use MIT-pthreads. -@xref{MIT-pthreads}. - -If you get the following error from configure: - -@example -checking for restartable system calls... configure: error can not run test -programs while cross compiling -@end example - -This means that you have something wrong with your compiler installation! -In this case you should upgrade your compiler to a newer version. You may -also be able to solve this problem by inserting the following row into the -@file{config.cache} file: - -@example -ac_cv_sys_restartable_syscalls=$@{ac_cv_sys_restartable_syscalls='no'@} -@end example - -If you are using Solaris on a SPARC, the recommended compiler is -@code{gcc} 2.95.2. You can find this at @uref{http://gcc.gnu.org/}. -Note that @code{egcs} 1.1.1 and @code{gcc} 2.8.1 don't work reliably on -SPARC! - -The recommended @code{configure} line when using @code{gcc} 2.95.2 is: +@item +@c EMAIL: mkp@socsci.auc.dk (Martin Kasper Petersen) +@image{Flags/denmark} Denmark [SunSITE] @ +@uref{http://SunSITE.auc.dk/mysql/, WWW} +@uref{ftp://SunSITE.auc.dk/pub/databases/mysql/, FTP} -@example -CC=gcc CFLAGS="-O3" \ -CXX=gcc CXXFLAGS="-O3 -felide-constructors -fno-exceptions -fno-rtti" \ -./configure --prefix=/usr/local/mysql --with-low-memory --enable-assembler -@end example +@c @item +@c Tõnu does not work there anymore :-) 990720 +@c EMAIL: tonu@tradenet.ee (Tõnu Samuel) +@c @image{Flags/estonia} Estonia [Tradenet] @ +@c @uref{http://mysql.tradenet.ee, WWW} -If you have a ultra sparc, you can get 4 % more performance by adding -"-mcpu=v8 -Wa,-xarch=v8plusa" to CFLAGS and CXXFLAGS. +@item +@c EMAIL: tonu@spam.ee (Tõnu Samuel) +@image{Flags/estonia} Estonia [OKinteractive] @ +@uref{http://mysql.mirror.ok.ee, WWW} -If you have the Sun Workshop (SunPro) 4.2 (or newer) compiler, you can -run @code{configure} like this: +@c @item +@c Changed 990531 +@c Not ok 20000919; Old site (Matt) +@c EMAIL: Steeve.Devergne@minet.net (Steeve Devergne) +@c @image{Flags/france} France [minet] @ +@c @uref{http://www.minet.net/devel/mysql/, WWW} -@example -CC=cc CFLAGS="-Xa -fast -xO4 -native -xstrconst -mt" \ -CXX=CC CXXFLAGS="-noex -xO4 -mt" \ -./configure --prefix=/usr/local/mysql --enable-assembler -@end example +@item +@c EMAIL: alex@mtesa.net (Alexandre Dupouy) +@image{Flags/france} France [mtesa.net] @ +@uref{http://mysql.mtesa.net/, WWW} -You may also have to edit the @code{configure} script to change this line: +@item +@c EMAIL: mrjack@fastorama.com (MrJacK) +@image{Flags/france} France [fastorama.com, Chatenois] @ +@uref{http://mysql.fastorama.com/, WWW} +@uref{ftp://ftp.fastorama.com/unix/mysql/, FTP} -@example -#if !defined(__STDC__) || __STDC__ != 1 -@end example +@c @item +@c Not ok 20000919; Old site (Matt) +@c EMAIL: Jaakko.Hyvatti@eunet.fi +@c @image{Flags/finland} Finland [EUnet] @ +@c @uref{http://mysql.eunet.fi/, WWW} -to this: +@c @item +@c Added 990829 +@c Not ok 20000919; Non-existent (Matt) +@c EMAIL: tomi.hakala@clinet.fi (Tomi Hakala) +@c @image{Flags/finland} Finland [clinet] @ +@c @uref{ftp://ftp.clinet.fi/mirrors/ftp.mysql.org/pub/mysql/, FTP} -@example -#if !defined(__STDC__) -@end example +@item +@c Added 20001019 +@c EMAIL: d@d.tj (Dominique L . Bouix) +@image{Flags/finland} Finland [tonnikala.net] @ +@uref{http://mysql.tonnikala.org/, WWW} -If you turn on @code{__STDC__} with the @code{-Xc} option, the Sun compiler -can't compile with the Solaris @file{pthread.h} header file. This is a Sun -bug (broken compiler or broken include file). +@item +@c Added 981208 +@c EMAIL: noel@uni-bonn.de (Noel Koethe) +@image{Flags/germany} Germany [Kernelnotes.de, Bonn] @ +@uref{http://www.kernelnotes.de/mysql/, WWW} +@uref{ftp://ftp.kernelnotes.de/pub/mirror/mysql.org/, FTP} -If @code{mysqld} issues the error message shown below when you run it, you have -tried to compile @strong{MySQL} with the Sun compiler without enabling the -multi-thread option (@code{-mt}): +@item +@c EMAIL: th@rz.fh-wolfenbuettel.de (Thorsten Ludewig) +@image{Flags/germany} Germany [Wolfenbuettel] @ +@uref{http://www.fh-wolfenbuettel.de/ftp/pub/database/mysql/, WWW} +@uref{ftp://ftp.fh-wolfenbuettel.de/pub/database/mysql/, FTP} -@example -libc internal error: _rmutex_unlock: rmutex not held -@end example +@c @item +@c Ok 980114. Removed 981208 (down > 3 days) ok 981214 +@c Not ok 20000919; Non-existent (Matt) +@c EMAIL: straub@gks.de (Hans-Peter Straub) +@c @image{Flags/germany} Germany [Staufen] @ +@c @uref{http://mysql.staufen.de/, WWW} -Add @code{-mt} to @code{CFLAGS} and @code{CXXFLAGS} and try again. +@c @item +@c Added 990614 +@c Not ok 20000919; Non-existent (Matt) +@c EMAIL: thomas.rohde@ecrc.de (Thomas Rohde) +@c @image{Flags/germany} Germany [Cable & Wireless] @ +@c @uref{ftp://ftp.ecrc.net/pub/database/mysql/, FTP} -If you get the following error when compiling @strong{MySQL} with @code{gcc}, -it means that your @code{gcc} is not configured for your version of Solaris: +@item +@c Added 981208 +@c EMAIL: christias@noc.ntua.gr (Panagiotis Christias) +@image{Flags/greece} Greece [NTUA, Athens] @ +@uref{http://www.ntua.gr/mysql/, WWW} +@uref{ftp://ftp.ntua.gr/pub/databases/mysql/, FTP} -@example -shell> gcc -O3 -g -O2 -DDBUG_OFF -o thr_alarm ... -./thr_alarm.c: In function `signal_hand': -./thr_alarm.c:556: too many arguments to function `sigwait' -@end example +@item +@c EMAIL: torlasz@xenia.sote.hu (Laszlo L. Tornoc) +@image{Flags/hungary} Hungary [Xenia] @ +@uref{http://mysql.sote.hu/, WWW} +@uref{ftp://xenia.sote.hu/pub/mirrors/www.mysql.com/, FTP} -The proper thing to do in this case is to get the newest version of -@code{gcc} and compile it with your current @code{gcc} compiler! At -least for Solaris 2.5, almost all binary versions of @code{gcc} have -old, unusable include files that will break all programs that use -threads (and possibly other programs)! +@item +@c EMAIL: hendlein@tiszanet.hu (Hendlein Peter) +@image{Flags/hungary} Hungary [TiszaneT] @ +@uref{http://mysql.tiszanet.hu/, WWW} +@uref{ftp://mysql.tiszanet.hu/pub/mirrors/mysql/, FTP} -Solaris doesn't provide static versions of all system libraries -(@code{libpthreads} and @code{libdl}), so you can't compile @strong{MySQL} -with @code{--static}. If you try to do so, you will get the error: +@item +@c EMAIL: mirrors@gm.is (Tomas Edwardsson) +@image{Flags/iceland} Iceland [GM] @ +@uref{http://mysql.gm.is/, WWW} +@uref{ftp://ftp.gm.is/pub/mysql/, FTP} -@example -ld: fatal: library -ldl: not found +@c @item +@c Out of date 990906 +@c EMAIL: bourbon@netvision.net.il (Zeev Suraski) +@c @image{Flags/israel} Israel [Netvision] @ +@c @uref{http://mysql.netvision.net.il/, WWW} -or +@c @item +@c Not working 99.03.06 +@c EMAIL: maruzz@matrice.it (Giovanni Maruzzelli) +@c @image{Flags/italy} Italy [Matrice] @ +@c @uref{http://www.matrice.it/risorse/mysql/, WWW} -undefined reference to `dlopen' -@end example +@item +@c EMAIL: gabriele@feelinglinux.com (Giansante Gabriele) +@image{Flags/italy} Italy [feelinglinux.com] @ +@uref{http://mysql.feelinglinux.com/, WWW} -If too many processes try to connect very rapidly to @code{mysqld}, you will -see this error in the @strong{MySQL} log: +@item +@c EMAIL: irena@yacc.it +@image{Flags/italy} Italy [Teta Srl] @ +@uref{http://www.teta.it/mysql/, WWW} -@example -Error in accept: Protocol error -@end example +@item +@c EMAIL: rudy@comm2000.it (Rudy) +@image{Flags/italy} Italy [tzone.it] @ +@uref{http://mysql.tzone.it/, WWW} -You might try starting the server with the @code{--set-variable back_log=50} -option as a workaround for this. @xref{Command-line options}. +@c @item +@c Added 991121 +@c Not ok 20000919; Old site (Matt) +@c EMAIL: nick@iol.ie (Nick Hilliard) +@c @image{Flags/ireland} Ireland [Ireland On-Line/Dublin] @ +@c @uref{http://mysql.iol.ie, WWW} +@c @uref{ftp://ftp.iol.ie/pub/mysql, FTP} -If you are linking your own @strong{MySQL} client, you might get the -following error when you try to execute it: +@item +@c Added 20001031 +@c EMAIL: dave@esat.net (Dave Rynne) +@image{Flags/ireland} Ireland [Esat Net] @ +@uref{http://ftp.esat.net/mirrors/download.sourceforge.net/pub/mirrors/mysql/, WWW} +@uref{ftp://ftp.esat.net/mirrors/download.sourceforge.net/pub/mirrors/mysql/, FTP} -@example -ld.so.1: ./my: fatal: libmysqlclient.so.#: open failed: No such file or directory -@end example +@item +@c Added 20010524 +@c EMAIL: arvids@parks.lv (Arvids) +@image{Flags/latvia} Latvia [linux.lv] @ +@uref{ftp://ftp.linux.lv/pub/software/mysql/, FTP} -The problem can be avoided by one of the following methods: +@item +@c Added 20001125 +@c EMAIL: mleicher@silverpoint.nl (Marcel Leicher) +@image{Flags/netherlands} Netherlands [Silverpoint] @ +@uref{http://mysql.silverpoint.nl/, WWW} -@itemize @bullet @item -Link the client with the following flag (instead of @code{-Lpath}): -@code{-Wl,r/full-path-to-libmysqlclient.so}. +@c EMAIL: mysql@widexs.nl (Wouter de Jong) +@image{Flags/netherlands} Netherlands [Widexs BV] @ +@uref{http://mysql.widexs.nl/, WWW} +@uref{ftp://mysql.widexs.nl/pub/mysql/, FTP} @item -Copy @file{libmysqclient.so} to @file{/usr/lib}. +@c EMAIL: peter@proserve.nl (ProServe - Peter Batenburg) +@image{Flags/netherlands} Netherlands [ProServe] @ +@uref{http://mysql.proserve.nl/, WWW} -@tindex LD_RUN_PATH environment variable -@tindex Environment variable, LD_RUN_PATH @item -Add the pathname of the directory where @file{libmysqlclient.so} is located -to the @code{LD_RUN_PATH} environment variable before running your client. -@end itemize +@c EMAIL: W.Sylwestrzak@icm.edu.pl (Wojtek Sylwestrzak) +@c mirroring nightly at 05:25 +@image{Flags/poland} Poland [Sunsite] @ +@uref{http://sunsite.icm.edu.pl/mysql/, WWW} +@uref{ftp://sunsite.icm.edu.pl/pub/unix/mysql/, FTP} -When using the @code{--with-libwrap} configure option, you must also -include the libraries that @file{libwrap.a} needs: +@item +@c EMAIL: admin@net.ncservice.com.pl (Marian Witkowski) +@image{Flags/poland} Poland [ncservice.com/Gdansk] @ +@uref{http://mysql.service.net.pl/, WWW} -@example ---with-libwrap="/opt/NUtcpwrapper-7.6/lib/libwrap.a -lnsl -lsocket -@end example +@c @item +@c EMAIL: melo@co.telenet.pt (Pedro Melo) +@c Temp out of service (email from Pedro) +@c @image{Flags/portugal} Portugal [IP] @ +@c @uref{http://mysql.ip.pt, WWW} -If you have problems with configure trying to link with @code{-lz} and -you don't have @code{zlib} installed, you have two options: +@c @item +@c Not ok 20000919; Non-existent (Matt) +@c EMAIL: Equipa de suporte do Leirianet <support@leirianet.pt> +@c @image{Flags/portugal} Portugal [lerianet] @ +@c @uref{http://mysql.leirianet.pt, WWW} +@c @uref{ftp://ftp.leirianet.pt/pub/mysql/,FTP} -@itemize @bullet @item -If you want to be able to use the compressed communication protocol, -you need to get and install zlib from ftp.gnu.org. +@c Added 20001031 +@c bofh@netc.pt (Bruno Rodrigues) +@image{Flags/portugal} Portugal [Netc] @ +@uref{http://ftp.netc.pt/pub/mysql/, WWW} +@uref{ftp://ftp.netc.pt/pub/mysql/, FTP} @item -Configure with @code{--with-named-z-libs=no}. -@end itemize +@c EMAIL: keeper@roedu.net (Mihai RUSU) +@image{Flags/romania} Romania [roedu.net/Bucharest] @ +@uref{ftp://ftp.roedu.net/pub/mirrors/ftp.mysql.com/, FTP} -If you are using gcc and have problems with loading @code{UDF} functions -into @strong{MySQL}, try adding @code{-lgcc} to the link line for the -@code{UDF} function. +@item +@c EMAIL: kuzmin@dn.ru (Roma Kuzmin) +@image{Flags/russia} Russia [DirectNet] @ +@uref{http://mysql.directnet.ru/, WWW} +@uref{ftp://ftp.dn.ru/pub/MySQL/, FTP} -If you would like @strong{MySQL} to start automatically, you can copy -@file{support-files/mysql.server} to @file{/etc/init.d} and create a -symbolic link to it named @file{/etc/rc3.d/S99mysql.server}. +@c @item +@c down 990113 +@c EMAIL: nikkic@cityline.ru (Nikki Chumakov) +@c @image{Flags/russia} Russia [Cityline] @ +@c @uref{ftp://mysql.cityline.ru/pub/mysql, FTP} +@c @uref{http://mysql.cityline.ru, WWW} +@c EMAIL: bar@izhcom.ru (Alexander I Barkov) -@node Solaris 2.7, Solaris x86, Solaris, Source install system issues -@subsection Solaris 2.7/2.8 Notes +@c @item +@c Not ok 20000919; Non-existent (Matt) +@c @image{Flags/russia} Russia [IZHCOM] @ +@c @uref{http://mysql.udm.net/, WWW} +@c @uref{ftp://ftp.izhcom.ru/pub/mysql/,FTP} -You can normally use a Solaris 2.6 binary on Solaris 2.7 and 2.8. Most -of the Solaris 2.6 issues also apply for Solaris 2.7 and 2.8. +@item +@c Added 990507 +@c EMAIL: demon@gpad.ac.ru (Dima Sivachenko) +@image{Flags/russia} Russia [Scientific Center/Chernogolovka] @ +@uref{ftp://ftp.chg.ru/pub/databases/mysql/, FTP} -Note that @strong{MySQL} Version 3.23.4 and above should be able to autodetect -new versions of Solaris and enable workarounds for the following problems! +@c @item +@c Not ok 20000919; Old site (Matt) +@c EMAIL: sebi@dnttm.ro (Sebastian DEAC) +@c @image{Flags/romania} Romania [Timisoara] @ +@c @uref{http://download.dnttm.ro/mysql, WWW} +@c @uref{ftp://ftp.dnttm.ro/pub/mysql, FTP} -Solaris 2.7 / 2.8 has some bugs in the include files. You may see the -following error when you use @code{gcc}: +@c @item +@c Not ok 20001123; Non-existent (Matt) +@c EMAIL: tim@lbi.ro (Bogdan Surdu) +@c @image{Flags/romania} +@c Romania [Bucharest] @ +@c @uref{http://www.lbi.ro/MySQL/, WWW} +@c @uref{ftp://ftp.lbi.ro/mirrors/ftp.tcx.se, FTP} -@example -/usr/include/widec.h:42: warning: `getwc' redefined -/usr/include/wchar.h:326: warning: this is the location of the previous -definition -@end example +@c @item +@c Removed 20000521 because there is no mirror here. +@c EMAIL: jips@masterd.es (Juan Ignacio Pérez Sacristán) +@c @image{Flags/spain} Spain [MasterD] +@c @uref{http://mysql.masterd.es, WWW} -If this occurs, you can do the following to fix the problem: +@c @item +@c Not ok 20000919; Old site (Matt) +@c EMAIL: Patrik.Karen@sdi.slu.se (Patrik Karen) +@c ftp -> remove old files +@c @image{Flags/sweden} Sweden [Sunet] @ +@c @uref{http://ftp.sunet.se/pub/unix/databases/relational/mysql/, WWW} +@c @uref{ftp://ftp.sunet.se/pub/unix/databases/relational/mysql/, FTP} -Copy @code{/usr/include/widec.h} to -@code{.../lib/gcc-lib/os/gcc-version/include} and change line 41 from: +@item +@c EMAIL: archive@sunsite.cnlab-switch.ch (Thomas Lenggenhager) +@image{Flags/switzerland} Switzerland [Sunsite] @ +@uref{http://sunsite.cnlab-switch.ch/ftp/mirror/mysql/, WWW} +@uref{ftp://sunsite.cnlab-switch.ch/mirror/mysql/, FTP} -@example -#if !defined(lint) && !defined(__lint) +@c @item +@c Not ok 20000919; Non-existent (Matt) +@c simon@oyster.co.uk (Simon Gornall) +@c @image{Flags/great-britain} UK [Oyster/UK] @ +@c @uref{ftp://ftp.oyster.co.uk/pub/mysql, FTP} -to +@item +@c EMAIL: gareth@omnipotent.net (Gareth Watts) +@image{Flags/great-britain} UK [Omnipotent/UK] @ +@uref{http://mysql.omnipotent.net/, WWW} +@uref{ftp://mysql.omnipotent.net/, FTP} -#if !defined(lint) && !defined(__lint) && !defined(getwc) -@end example +@item +@c EMAIL: keet@mordor.plig.net (Christiaan Keet) +@image{Flags/great-britain} UK [PLiG/UK] @ +@uref{http://ftp.plig.org/pub/mysql/, WWW} +@uref{ftp://ftp.plig.org/pub/mysql/, FTP} -Alternatively, you can edit @file{/usr/include/widec.h} directly. Either -way, after you make the fix, you should remove @file{config.cache} and run -@code{configure} again! +@item +@c EMAIL: sean@telekon.co.uk (Sean Gibson) +@image{Flags/great-britain} UK [Telekon Internet/UK] @ +@uref{ftp://ftp.telekon.co.uk/pub/mysql/, FTP} -If you get errors like this when you run @code{make}, it's because -@code{configure} didn't detect the @file{curses.h} file (probably -because of the error in @file{/usr/include/widec.h}): +@c @item +@c lance@uklinux.net (Lance) +@c @image{Flags/great-britain} UK [uklinux.net] @ +@c @uref{http://mirror.uklinux.net/mysql/, WWW} +@c @uref{ftp://mirror.uklinux.net/pub/mysql/, FTP} -@example -In file included from mysql.cc:50: -/usr/include/term.h:1060: syntax error before `,' -/usr/include/term.h:1081: syntax error before `;' -@end example +@c @item +@c unknown +@c @image{Flags/great-britain} UK [MicroMuse] @ +@c @uref{ftp://ftp.micromuse.co.uk/pub/packages/unix/databases/mysql/, FTP} -The solution to this is to do one of the following: +@c @item +@c Not ok 20000919; Old site (Matt) +@c lmjm@icparc.ic.ac.uk (Lee McLoughlin) +@c @image{Flags/great-britain} UK [SunSITE] @ +@c @uref{http://sunsite.org.uk/packages/mysql/, WWW} +@c @uref{ftp://sunsite.org.uk/packages/mysql/, FTP} -@itemize @bullet @item -Configure with @code{CFLAGS=-DHAVE_CURSES_H CXXFLAGS=-DHAVE_CURSES_H ./configure}. +@c sander@paco.net (Alexander Ivanov) +@image{Flags/ukraine} Ukraine [PACO] @ +@uref{http://mysql.paco.net.ua, WWW} +@uref{ftp://mysql.paco.net.ua/, FTP} @item -Edit @file{/usr/include/widec.h} as indicted above and rerun configure. +@c mizi@alkar.net (Alexander Ryumshin) +@image{Flags/ukraine} Ukraine [ISP Alkar Teleport/Dnepropetrovsk] @ +@uref{http://mysql.dp.ua/, WWW} -@item -Remove the @code{#define HAVE_TERM} line from @file{config.h} file and -run @code{make} again. @end itemize -If you get a problem that your linker can't find @code{-lz} when linking -your client program, the problem is probably that your @file{libz.so} file is -installed in @file{/usr/local/lib}. You can fix this by one of the -following methods: +@strong{North America:} @itemize @bullet -@item -Add @file{/usr/local/lib} to @code{LD_LIBRARY_PATH}. +@c @item +@c Not ok 990101 (only to 981007) +@c EMAIL: sysop@polarcom.com (Seamus Venasse) +@c @image{Flags/canada} Canada [Polaris Computing] @ +@c @uref{http://mysql.polaris.ca/, WWW} @item -Add a link to @file{libz.so} from @file{/lib}. +@c Ok 980109 +@c EMAIL: wojtek@tryc.on.ca (Wojtjeck Tryc) +@image{Flags/canada} Canada [Tryc] @ +@uref{http://web.tryc.on.ca/mysql/, WWW} -@item -If you are using Solaris 8, you can install the optional zlib from your -Solaris 8 CD distribution. +@c @item +@c not updated 990218. Added again 990918 +@c Not ok 20000919; Non-existent (Matt) +@c EMAIL: rhooper@cyberus.ca (Roy Hooper) +@c @image{Flags/canada} Canada [Cyberus] @ +@c @uref{http://mysql.cyberus.ca/, WWW} +@c @uref{ftp://mysql.cyberus.ca/, FTP} @item -Configure @strong{MySQL} with the @code{--with-named-z-libs=no} option. -@end itemize +@c EMAIL: mleber@he.net (Mike Leber) +@c Added 980312 +@image{Flags/usa} USA [Hurricane Electric/San Jose] @ +@uref{http://mysql.he.net/, WWW} -@node Solaris x86, SunOS, Solaris 2.7, Source install system issues -@subsection Solaris x86 Notes +@item +@c EMAIL: ask@valueclick.com (Ask Bjoern Hansen) +@c Added 20000925 +@image{Flags/usa} USA [ValueClick, Los Angeles CA] @ +@uref{http://mysql.valueclick.com/, WWW} +@uref{ftp://mysql.valueclick.com/pub/mysql/Downloads/, FTP} -On Solaris 2.8 on x86, @strong{mysqld} will core dump if you run -'strip' in. +@c @item +@c Not ok 20000919; Non-existent (Matt) +@c EMAIL: meltzer@icsnet.com (Jeffrey Meltzer) +@c Added 000108 +@c @image{Flags/usa} USA [Meltzer/New York State] @ +@c @uref{ftp://ftp.meltzer.org/pub/mysql/, FTP} -If you are using @code{gcc} or @code{egcs} on Solaris x86 and you -experience problems with core dumps under load, you should use the -following @code{configure} command: +@c @item +@c No such directory 990830 +@c EMAIL: tps@users.buoy.com (Tim Sailer) +@c @image{Flags/usa} USA [Buoy/New York] @ +@c @uref{http://www.buoy.com/mysql/, WWW} -@example -CC=gcc CFLAGS="-O3 -fomit-frame-pointer -DHAVE_CURSES_H" \ -CXX=gcc \ -CXXFLAGS="-O3 -fomit-frame-pointer -felide-constructors -fno-exceptions -fno-rtti -DHAVE_CURSES_H" \ -./configure --prefix=/usr/local/mysql -@end example +@c @item +@c EMAIL: db@hpnc.com (Douglas Bowyer) +@c Added 980107, removed 981124 because of 'file not found' +@c @image{Flags/usa} USA [Hypernet Communications/Dallas] @ +@c @uref{http://epsilon.hpnc.com/mysql, WWW} -This will avoid problems with the @code{libstdc++} library and with C++ -exceptions. +@c @item @c ********************************** +@c Not updated 980106 +@c EX: twh@iquest.net (Thomas Holt) who no longer works there +@c @image{Flags/usa} USA [IQuest/Indiana] @ +@c @uref{http://mirrors.iquest.net/mysql/, WWW} -If this doesn't help, you should compile a debug version and run -it with a trace file or under @code{gdb}. @xref{Using gdb on mysqld}. +@c @item @c ********************************** +@c Only a partial mirror so we exclude it from the list +@c EX: lindberg@id.wustl.edu (Fred Lindberg) +@c @image{Flags/usa} USA [Washington University/St. Louis] @ +@c @uref{ftp://ftp.id.wustl.edu/pub/database/mysql/, FTP} -@node SunOS, Linux, Solaris x86, Source install system issues -@subsection SunOS 4 Notes +@c removed 991111 -> no answer +@c @item +@c EMAIL: andrew@netcasting.net (Andrew Sawyers) +@c @image{Flags/usa} USA [Netcasting/West Coast] @ +@c @uref{ftp://ftp.netcasting.net/pub/mysql/, FTP} -On SunOS 4, MIT-pthreads is needed to compile @strong{MySQL}, which in turn -means you will need GNU @code{make}. +@c @item +@c No mirror! 980809 David +@c EMAIL: savages@savages.com (Shaun Savage) +@c @image{Flags/usa} USA [Savages/Oregon] @ +@c @uref{http://mysql.savages.com, WWW} -Some SunOS 4 systems have problems with dynamic libraries and @code{libtool}. -You can use the following @code{configure} line to avoid this problem: +@c @item +@c Not ok 20000919; Non-existent (Matt) +@c EMAIL: paul@gina.net (Paul Vining) +@c mirrors ftp.sunet.se +@c @image{Flags/usa} USA [Gina net/Florida] @ +@c @uref{http://www.gina.net/mysql/, WWW} -@example -shell> ./configure --disable-shared --with-mysqld-ldflags=-all-static -@end example +@c Out of date 2000-01-08 (Not updated since 1999-10) +@c @item +@c EMAIL: wswanson@pingzero.net (Wylie Swanson) +@c mirrors mysql.org +@c @image{Flags/usa} USA [pingzero/Los Angeles] @ +@c @uref{http://mysql.pingzero.net/, WWW} -When compiling @code{readline}, you may get warnings about duplicate defines. -These may be ignored. +@item +@c EMAIL: ftpkeeper@mirror.sit.wisc.edu +@image{Flags/usa} USA [Wisconsin University/Wisconsin] @ +@uref{http://mirror.sit.wisc.edu/mysql/, WWW} +@uref{ftp://mirror.sit.wisc.edu/mirrors/mysql/, FTP} -When compiling @code{mysqld}, there will be some @code{implicit declaration -of function} warnings. These may be ignored. +@c @item +@c Not ok 20000919; Out of date (Matt) +@c EMAIL: ftp-admin@digex.net +@c @image{Flags/usa} USA [DIGEX] @ +@c @uref{ftp://ftp.digex.net/pub/packages/database/mysql/, FTP} -@node Linux, Alpha-DEC-UNIX, SunOS, Source install system issues -@subsection Linux Notes (All Linux Versions) +@item +@c EMAIL: andrew.sawyers@thelinuxstore.com +@image{Flags/usa} USA [LinuxWired/Scottsdale, AZ] @ +@uref{http://mysql.linuxwired.net/, WWW} +@uref{ftp://ftp.linuxwired.net/pub/mirrors/mysql/, FTP} -The notes below regarding @strong{glibc} apply only to the situation -when you build @strong{MySQL} -yourself. If you are running Linux on an x86 machine, in most cases it is -much better for you to just use our binary. We link our binaries against -the best patched version of @strong{glibc} we can come up with and with the -best compiler options, in an attempt to make it suitable for a high-load -server. So if you read the text below, and are in doubt about -what you should do, try our binary first to see if it meets your needs, and -worry about your own build only after you have discovered that our binary is -not good enough. In that case, we would appreciate a note about it, so we -can build a better binary next time. For a typical user, even for setups with -a lot of concurrent connections and/or tables exceeding 2GB limit, our -binary in most cases is the best choice. +@c @item +@c EMAIL: dan@surfsouth.com (Dan Muntz) +@c @image{Flags/usa} USA [Venoma.Org/Valdosta, GA] @ +@c @uref{http://mysql.venoma.org/, WWW} -@strong{MySQL} uses LinuxThreads on Linux. If you are using an old -Linux version that doesn't have @code{glibc2}, you must install -LinuxThreads before trying to compile @strong{MySQL}. You can get -LinuxThreads at @uref{http://www.mysql.com/Downloads/Linux}. +@item +@c EMAIL: hkind@adgrafix.com (Hans Kind) +@image{Flags/usa} USA [adgrafix.com/Boston, MA] @ +@uref{http://mysql.adgrafix.com/, WWW} -@strong{NOTE:} We have seen some strange problems with Linux 2.2.14 and -@strong{MySQL} on SMP systems; If you have a SMP system, we recommend -you to upgrade to Linux 2.4 ASAP! Your system will be faster and more -stable by doing this! +@end itemize -Note that @code{glibc} versions before and including Version 2.1.1 have -a fatal bug in @code{pthread_mutex_timedwait} handling, which is used -when you do @code{INSERT DELAYED}. We recommend you to not use -@code{INSERT DELAYED} before upgrading glibc. +@strong{South America:} -If you plan to have 1000+ concurrent connections, you will need to make -some changes to LinuxThreads, recompile it, and relink @strong{MySQL} against -the new @file{libpthread.a}. Increase @code{PTHREAD_THREADS_MAX} in -@file{sysdeps/unix/sysv/linux/bits/local_lim.h} to 4096 and decrease -@code{STACK_SIZE} in @file{linuxthreads/internals.h} to 256 KB. The paths are -relative to the root of @code{glibc} Note that @strong{MySQL} will not be -stable with around 600-1000 connections if @code{STACK_SIZE} is the default -of 2 MB. +@itemize @bullet -If you have a problem with that @strong{MySQL} can't open enough files, -or connections, it may be that you haven't configured Linux to handle -enough files. +@item +@c Added 20001102 +@c EMAIL: nico@bannerlandia.com (Nicolas Moldavsky) +@image{Flags/argentina} Argentina [bannerlandia.com] @ +@uref{http://mysql.bannerlandia.com.ar/, WWW} +@uref{ftp://mysql.bannerlandia.com.ar/mirrors/mysql/, FTP} -In Linux 2.2 and forwards, you can check the number of allocated -file handlers by doing: +@c @item +@c Not ok 20000919; Non-existent (Matt) +@c EMAIL: gaiser@matrix.com.br (Roberto Gaiser) +@c @image{Flags/brazil} Brazil [Matrix] @ +@c @uref{http://mysql.matrix.com.br, WWW} -@example -cat /proc/sys/fs/file-max -cat /proc/sys/fs/dquot-max -cat /proc/sys/fs/super-max -@end example +@item +@c jpabuyer@vision.cl +@image{Flags/chile} Chile [Vision] @ +@uref{http://mysql.vision.cl/, WWW} -If you have more than 16M of memory, you should add something like the -following in your boot script (@file{/etc/rc/boot.local} on SuSE): +@item +@c EMAIL: bartschm@psi.com (Marcelo Bartsch) +@image{Flags/chile} Chile [PSINet] @ +@uref{http://mysql.psinet.cl/, WWW} +@uref{ftp://ftp.psinet.cl/pub/database/mysql/, FTP} -@example -echo 65536 > /proc/sys/fs/file-max -echo 8192 > /proc/sys/fs/dquot-max -echo 1024 > /proc/sys/fs/super-max -@end example +@item +@c EMAIL: jpabuyer@tecnoera.com (Juan Pablo Abuyeres) +@image{Flags/chile} Chile [Tecnoera] @ +@uref{http://mysql.tecnoera.com/, WWW} -You can also run the above from the command line as root, but in this case -your old limits will be used next time your computer reboots. +@c @item +@c Removed 990730 +@c @c EMAIL: dan@amerikanclaris.com (Danilo Lotina F.) +@c @image{Flags/chile} Chile [Amerikanclaris] @ +@c @uref{http://www.labs.amerikanclaris.cl/mysql, WWW} +@c @uref{ftp://ftp.amerikanclaris.cl/pub/mysql, FTP} +@end itemize -You should also add /etc/my.cnf: +@strong{Asia:} -@example -[safe_mysqld] -open_files_limit=8192 -@end example +@itemize @bullet +@item +@c EMAIL: mirnshi@263.net (Meng Lingbo) +@image{Flags/china} China [Freecode] @ +@uref{http://www.freecode.net.cn/mirror/mysql/, WWW} -The above should allow @strong{MySQL} to create up to 8192 connections/files. +@item +@c EMAIL: yusun@atwell.co.jp (yu sun) +@image{Flags/china} China [linuxforum.net] @ +@uref{http://www2.linuxforum.net/mirror/mysql/, WWW} -The @code{STACK_SIZE} constant in LinuxThreads controls the spacing of thread -stacks in the address space. It needs to be large enough so that there will -be plenty of room for the stack of each individual thread, but small enough -to keep the stack of some thread from running into the global @code{mysqld} -data. Unfortunately, the Linux implementation of @code{mmap()}, as we have -experimentally discovered, will successfully unmap an already mapped region -if you ask it to map out an address already in use, zeroing out the data -on the entire page, instead of returning an error. So, the safety of -@code{mysqld} or any other threaded application depends on the "gentleman" -behavior of the code that creates threads. The user must take measures to -make sure the number of running threads at any time is sufficiently low for -thread stacks to stay away from the global heap. With @code{mysqld}, you -should enforce this "gentleman" behavior by setting a reasonable value for -the @code{max_connections} variable. +@item +@c EMAIL: Vincent_Fong@innovator.com.hk (Vincent Fong) +@image{Flags/china} China [ISL/Hong Kong] @ +@uref{http://mysql.islnet.net, WWW} -If you build @strong{MySQL} yourself and do not what to mess with patching -LinuxThreads, you should set @code{max_connections} to a value no higher -than 500. It should be even less if you have a large key buffer, large -heap tables, or some other things that make @strong{mysqld} allocate a lot -of memory or if you are running a 2.2 kernel with a 2GB patch. If you are -using our binary or RPM version 3.23.25 or later, you can safely set -@code{max_connections} at 1500, assuming no large key buffer or heap tables -with lots of data. The more you reduce @code{STACK_SIZE} in LinuxThreads -the more threads you can safely create. We recommend the values between -128K and 256K. +@item +@c EMAIL: xcyber@yahoo.com (xcyber) +@image{Flags/china} China [xcyber.org/Hong Kong] @ +@uref{http://mysql.xcyber.org/, WWW} -If you use a lot of concurrent connections, you may suffer from a "feature" -in the 2.2 kernel that penalizes a process for forking or cloning a child -in an attempt to prevent a fork bomb attack. This will cause @strong{MySQL} -not to scale well as you increase the number of concurrent clients. On -single CPU systems, we have seen this manifested in a very slow thread -creation, which means it may take a long time to connect to @strong{MySQL} -(as long as 1 minute), and it may take just as long to shut it down. On -multiple CPU systems, we have observed a gradual drop in query speed as -the number of clients increases. In the process of trying to find a -solution, we have received a kernel patch from one of our users, who -claimed it made a lot of difference for his site. The patch is available here -(@uref{http://www.mysql.com/Downloads/Patches/linux-fork.patch}). We have -now done rather extensive testing of this patch on both development and -production systems. It has significantly -improved @code{MySQL} performance without causing any problems and we now -recommend it to our users who are still running high-load servers on -2.2 kernels. This issue has been fixed in the 2.4 kernel, so if you are not -satisfied with -the current performance of your system, rather than patching your 2.2 kernel, -it might be easier to just upgrade to 2.4, which will also give you a nice -SMP boost in addition to fixing this fairness bug. +@c @item +@c Not ok 20010330; Non-existent! (Matt) +@c EMAIL: marquischan@hotmail.com (Marquis Chan) +@c @image{Flags/china} China [TraLand.com/Hong Kong] @ +@c @uref{http://www.traland.com/mysql/, WWW} -We have tested @strong{MySQL} on the 2.4 kernel on a 2 CPU machine and -found @strong{MySQL} scales MUCH better - there was virtually no slowdown -on query throughput all the way up -to 1000 clients, and @strong{MySQL} scaling factor ( computed as the ratio of -maximum throughput to the throughput with one client) was 180%. -We have observed similar results on a 4-CPU system - virtually no -slowdown as the number of -clients was increased up to 1000, and 300% scaling factor. So for a high-load -SMP server we would definitely recommend the 2.4 kernel at this point. We -have discovered that it is essential to run @code{mysqld} process with the -highest possible priority on the 2.4 kernel to achieve maximum performance. -This can be done by adding -@code{renice -20 $$} command to @code{safe_mysqld}. In our testing on a -4-CPU machine, increasing the priority gave 60% increase in throughput with -400 clients. +@c @item +@c Not ok 20000919; Non-existent (Matt) +@c EMAIL: george@netfirm.net (Hongsheng Zhu) +@c @image{Flags/china} China [Netfirm] @ +@c @uref{http://mysql.netfirm.net, WWW} -We are currently also trying to collect -more info on how well @code{MySQL} performs on 2.4 kernel on 4-way and 8-way -systems. If you have access such a system and have done some benchmarks, -please send a mail to @email{docs@@mysql.com} with the results - we will -include them in the manual. +@c @item +@c Not ok 20000919; Old site (Matt) +@c EMAIL: ahmlhs@nmsvr.chosun.com (Ho-sun Lee) +@c @image{Flags/south-korea} South Korea [KREONet] @ +@c @uref{http://linux.kreonet.re.kr/mysql/, WWW} -There is another issue that greatly hurts @strong{MySQL} performance, -especially on SMP systems. The implementation of mutex in -LinuxThreads in @strong{glibc-2.1} is very bad for programs with many -threads that only -hold the mutex for a short time. On an SMP system, ironic as it is, if -you link @strong{MySQL} against unmodified @strong{LinuxThreads}, -removing processors from the machine improves @strong{MySQL} performance -in many cases. We have made a patch available for @strong{glibc 2.1.3}, -@uref{http://www.mysql.com/Downloads/Linux/linuxthreads-2.1-patch,linuxthreads-2.1-patch} -to correct this behavior. +@item +@c EMAIL: jasper@webiiz.com (Kang, Tae-jin) +@image{Flags/south-korea} South Korea [Webiiz] @ +@uref{http://mysql.webiiz.com/, WWW} -With @strong{glibc-2.2.2} -@strong{MySQL} version 3.23.36 will use the adaptive mutex, which is much -better than even the patched one in @strong{glibc-2.1.3}. Be warned, however, -that under some conditions, the current mutex code in @strong{glibc-2.2.2} -overspins, which hurts @strong{MySQL} performance. The chance of this -condition can be reduced by renicing @code{mysqld} process to the highest -priority. We have also been able to correct the overspin behavior with -a patch, available @uref{http://www.mysql.com/Downloads/Linux/linuxthreads-2.2.2.patch,here}. It combines the correction of overspin, maximum number of -threads, and stack spacing all in one. You will need to apply it in the -@code{linuxthreads} directory with -@code{patch -p0 </tmp/linuxthreads-2.2.2.patch}. -We hope it will be included in -some form in to the future releases of @code{glibc-2.2}. In any case, if -you link against @code{glibc-2.2.2} you still need to correct -@code{STACK_SIZE} and @code{PTHREAD_THREADS_MAX}. We hope that the defaults -will be corrected to some more acceptable values for high-load -@strong{MySQL} setup in the future, so that your own build can be reduced -to @code{./configure; make; make install}. +@item +@c EMAIL: hollywar@holywar.net (Oh Junseon) +@image{Flags/south-korea} South Korea [PanworldNet] @ +@uref{http://mysql.holywar.net/, WWW} -We recommend that you use the above patches to build a special static -version of @code{libpthread.a} and use it only for statically linking -against @code{MySQL}. We know that the patches are safe for @code{MySQL} -and significantly improve its performance, but we cannot say anything -about other applications. If you link other applications against the -patched version of the library, or build a patched shared version and -install it on your system, you are doing it at your own risk with regard -to other applications that depend on @code{LinuxThreads}. +@c @item +@c ftp -> remove old files +@c EX: ahmlhs@nmsvr.chosun.com (Ho-sun Lee) +@c @image{Flags/south-korea} South Korea [KREONet] @ +@c @uref{ftp://linux.kreonet.re.kr/pub/tools/db/mysql/, FTP} -If you experience any strange problems during the installation of -@strong{MySQL}, or with some common utilties hanging, it is very likely that -they are either library or compiler related. If this is the case, using our -binary will resolve them. +@item +@c Ok 980805 +@c EMAIL: takeshi@SoftAgency.co.jp +@image{Flags/japan} Japan [Soft Agency] @ +@uref{http://www.softagency.co.jp/MySQL, WWW} -One known problem with the binary distribution is that with older Linux -systems that use @code{libc} (like RedHat 4.x or Slackware), you will get -some non-fatal problems with hostname resolution. -@xref{Binary notes-Linux}. +@item +@c EMAIL: takafumi@u-aizu.ac.jp (Takafumi Hayashi) +@image{Flags/japan} Japan [u-aizu.ac.jp/Aizu] @ +@uref{ftp://ftp.u-aizu.ac.jp/ftp/pub/dbms/mysql/mysql.com, FTP} -When using LinuxThreads you will see a minimum of three processes -running. These are in fact threads. There will be one thread for the -LinuxThreads manager, one thread to handle connections, and one thread -to handle alarms and signals. +@c @item +@c Ok 980109 Removed 990730 +@c EMAIL: satoshi@HappySize.co.jp (Satoshi TATSUOKA) +@c @image{Flags/japan} Japan [HappySize] @ +@c @uref{http://www.happysize.co.jp/mysql/, WWW} +@c @uref{ftp://ftp.happysize.co.jp/pub/mysql/, FTP} -Note that the Linux kernel and the LinuxThread library can by default -only have 1024 threads. This means that you can only have up to 1021 -connections to @strong{MySQL} on an unpatched system. The page -@uref{http://www.volano.com/linuxnotes.html} contains information how to -go around this limit. +@c @item +@c Ok 981204 +@c Not ok 20000919; Old site (Matt) +@c EMAIL: hiroyuki@nucba.ac.jp (hiroyuki kurimoto) +@c @image{Flags/japan} Japan [Nagoya Syouka University] @ +@c @uref{http://mirror.nucba.ac.jp/mirror/mysql, WWW} +@c @uref{ftp://mirror.nucba.ac.jp/mirror/mysql, FTP} -If you see a dead @code{mysqld} daemon process with @code{ps}, this usually -means that you have found a bug in @strong{MySQL} or you have a corrupted -table. @xref{Crashing}. +@c @item +@c Removed 990308 +@c EMAIL: terence@com5.net (Terence Chan) +@c @image{Flags/singapore} Singapore [Com5 Productions] @ +@c @uref{http://mysql.com5.net, WWW} +@c @uref{ftp://ftp.com5.net/pub/mysql, FTP} -To get a core dump on Linux if @code{mysqld} dies with a SIGSEGV -signal, you can start @code{mysqld} with the @code{--core-file} option. Note -that you also probably need to raise the @code{core file size} by adding -@code{ulimit -c 1000000} to @code{safe_mysqld} or starting @code{safe_mysqld} -with @code{--core-file-sizes=1000000}. @xref{safe_mysqld, , @code{safe_mysqld}}. +@item +@c EMAIL: csy@hjc.edu.sg +@image{Flags/singapore} Singapore [HJC] @ +@uref{http://mysql.hjc.edu.sg, WWW} +@uref{ftp://ftp.hjc.edu.sg/mysql, FTP} -To get a core dump on Linux if @code{mysqld} dies with a SIGSEGV signal, you can -start @code{mysqld} with the @code{--core-file} option. Note that you also probably -need to raise the @code{core file size} by adding @code{ulimit -c 1000000} to -@code{safe_mysqld} or starting @code{safe_mysqld} with -@code{--core-file-sizes=1000000}. @xref{safe_mysqld, , @code{safe_mysqld}}. +@c @item +@c Not ok 20000919; Old site (Matt) +@c 991118: Removed because a user complained about that the page contains +@c nothing about MySQL. 991119: Added again because it is a mirror again +@c EMAIL: dean@ht.net.tw (Dean Lin) +@c @image{Flags/taiwan} Taiwan [HT] @ +@c @uref{http://mysql.ht.net.tw, WWW} -If you are linking your own @strong{MySQL} client and get the error: +@item +@c EMAIL: linda@ttn.com.tw (Linda Hu) +@image{Flags/taiwan} Taiwan [TTN] @ +@uref{http://mysql.ttn.net, WWW} -@example -ld.so.1: ./my: fatal: libmysqlclient.so.4: open failed: No such file or directory -@end example +@c @item +@c Ok 980321 No connect -> removed 990730 +@c EMAIL: tby@ccca.nctu.edu.tw (Bao-Yi Tuang) +@c @image{Flags/taiwan} Taiwan [NCTU] @ +@c @uref{http://mysql.taconet.com.tw, WWW} -When executing them, the problem can be avoided by one of the following -methods: +@c @item +@c Out of date 990905 (~2 months) +@c @item @c ********************************** +@c Error 980106 +@c EX: WolfySu@acer.net (Wolfy Su) +@c @image{Flags/taiwan} Taiwan [Acer] @ +@c @uref{http://mysql.acer.net/, WWW} -@itemize @bullet -@item -Link the client with the following flag (instead of @code{-Lpath}): -@code{-Wl,r/path-libmysqlclient.so}. +@c @item @c ********************************** +@c files to delete +@c EX: service@wownet.net +@c @image{Flags/taiwan} Taiwan [Wownet] @ +@c @uref{ftp://ftp.wownet.net/mysql/, FTP} -@item -Copy @code{libmysqclient.so} to @file{/usr/lib}. +@c @item @c ********************************** +@c No contact 980106 +@c EX: serge@oneway.net +@c @image{Flags/taiwan} Taiwan [Oneway] @ +@c @uref{ftp://ftp.oneway.com.tw/pub/mysql/, FTP} -@tindex LD_RUN_PATH environment variable -@tindex environment variable, LD_RUN_PATH @item -Add the pathname of the directory where @code{libmysqlclient.so} is located -to the @code{LD_RUN_PATH} environment variable before running your client. +@c Added 20001031 +@c EMAIL: ijliao@php.nctu.edu.tw (Ying-Chieh Liao) +@image{Flags/taiwan} Taiwan [nctu.edu/HsinChu] @ +@uref{http://mysql.nctu.edu.tw/, WWW} + @end itemize -If you are using the Fujitsu compiler @code{(fcc / FCC)} you will have -some problems compiling @strong{MySQL} because the Linux header files are very -@code{gcc} oriented. +@strong{Australia:} -The following @code{configure} line should work with @code{fcc/FCC}: +@itemize @bullet +@item +@c Added 980610 +@c EMAIL: jason@dstc.edu.au (Jason Andrade) +@image{Flags/australia} Australia [AARNet/Queensland] @ +@uref{http://mysql.mirror.aarnet.edu.au/, WWW} +@uref{ftp://mysql.mirror.aarnet.edu.au/, FTP} -@example -CC=fcc CFLAGS="-O -K fast -K lib -K omitfp -Kpreex -D_GNU_SOURCE -DCONST=const -DNO_STRTOLL_PROTO" CXX=FCC CXXFLAGS="-O -K fast -K lib -K omitfp -K preex --no_exceptions --no_rtti -D_GNU_SOURCE -DCONST=const -Dalloca=__builtin_alloca -DNO_STRTOLL_PROTO '-D_EXTERN_INLINE=static __inline'" ./configure --prefix=/usr/local/mysql --enable-assembler --with-mysqld-ldflags=-all-static --disable-shared --with-low-memory -@end example +@c @item +@c Added 980805. Removed 000102 'no such directory' +@c EMAIL: sdd@ntccc.tas.gov.au (Scott Donovan) +@c @image{Flags/australia} Australia [Tas] @ +@c @uref{http://ftp.tas.gov.au/mysql, WWW} +@c @uref{ftp://ftp.tas.gov.au/pub/mysql, FTP} -@menu -* Linux-x86:: Linux-x86 notes -* Linux-RedHat50:: RedHat 5.0 notes -* Linux-RedHat51:: RedHat 5.1 notes -* Linux-SPARC:: Linux-SPARC notes -* Linux-Alpha:: Linux-Alpha notes -* MKLinux:: MkLinux notes -* Qube2:: Qube2 Linux notes -* Linux-Ia64:: Linux-Ia64 notes -@end menu +@c @item +@c Ok 980623 +@c Not ok 20000919; Old site (Matt) +@c EMAIL: samh@bluep.com (Sam Hadzajlic) +@c @image{Flags/australia} Australia [Blue Planet/Melbourne] @ +@c @uref{http://mysql.bluep.com/, WWW} +@c removed because ftp was not working 990729 & 30 +@c @uref{ftp://mysql.bluep.com/pub/mirror1/mysql/, FTP} -@node Linux-x86, Linux-RedHat50, Linux, Linux -@subsubsection Linux-x86 Notes +@c @item +@c Added 990531 +@c EMAIL: gavin@itworks.com.au (Gavin Cameron) +@c @image{Flags/australia} Australia [ITworks Consulting/Victoria] @ +@c @uref{http://mysql.itworks.com.au, WWW} -@strong{MySQL} requires @code{libc} Version 5.4.12 or newer. It's known to -work with @code{libc} 5.4.46. @code{glibc} Version 2.0.6 and later should -also work. There have been some problems with the @code{glibc} RPMs from -RedHat, so if you have problems, check whether or not there are any updates! -The @code{glibc} 2.0.7-19 and 2.0.7-29 RPMs are known to work. - -On some older Linux distributions, @code{configure} may produce an error -like this: - -@example -Syntax error in sched.h. Change _P to __P in the /usr/include/sched.h file. -See the Installation chapter in the Reference Manual. -@end example - -Just do what the error message says and add an extra underscore to the -@code{_P} macro that has only one underscore, then try again. - -You may get some warnings when compiling; those shown below can be ignored: - -@example -mysqld.cc -o objs-thread/mysqld.o -mysqld.cc: In function `void init_signals()': -mysqld.cc:315: warning: assignment of negative value `-1' to `long unsigned int' -mysqld.cc: In function `void * signal_hand(void *)': -mysqld.cc:346: warning: assignment of negative value `-1' to `long unsigned int' -@end example +@c @item +@c 980610 Only the toplevel dir! +@c EMAIL: lucifer@maths.uq.edu.au (David Conran) +@c @image{Flags/australia} Australia FTP @ +@c @uref{ftp://ftp.sage-au.org.au/pub/database/mysql, [Sage]} +@end itemize -In Debian GNU/Linux, if you want @strong{MySQL} to start automatically when -the system boots, do the following: +@strong{Africa:} -@example -shell> cp support-files/mysql.server /etc/init.d/mysql.server -shell> /usr/sbin/update-rc.d mysql.server defaults 99 -@end example +@itemize @bullet -@code{mysql.server} can be found in the @file{share/mysql} directory -under the @strong{MySQL} installation directory or in the -@file{support-files} directory of the @strong{MySQL} source tree. +@item +@c EMAIL: ftp-admin@mweb.com (Warren Baker) +@image{Flags/south-africa1} South-Africa [Mweb] @ +@uref{http://www.mysql.mweb.co.za/, WWW} -If @code{mysqld} always core dumps when it starts up, the problem may be that -you have an old @file{/lib/libc.a}. Try renaming it, then remove -@file{sql/mysqld} and do a new @code{make install} and try again. This -problem has been reported on some Slackware installations. RedHat Version 5.0 -also has a similar problem with some new @code{glibc} versions. -@xref{Linux-RedHat50}. +@item +@c Ok 981010 +@c EMAIL: oskar@is.co.za (Oskar Pearson) +@image{Flags/south-africa1} South Africa [The Internet Solution/Johannesburg] @ +@uref{ftp://ftp.is.co.za/linux/mysql/, FTP} -If you get the following error when linking @code{mysqld}, -it means that your @file{libg++.a} is not installed correctly: +@end itemize -@example -/usr/lib/libc.a(putc.o): In function `_IO_putc': -putc.o(.text+0x0): multiple definition of `_IO_putc' -@end example +@c END_OF_MIRROR_LISTING -You can avoid using @file{libg++.a} by running @code{configure} like this: -@example -shell> CXX=gcc ./configure -@end example +@cindex operating systems, supported +@cindex native thread support +@cindex thread support +@cindex process support +@cindex support, for operating systems +@node Which OS, Which version, Getting MySQL, Installing +@section Operating Systems Supported by MySQL -@node Linux-RedHat50, Linux-RedHat51, Linux-x86, Linux -@subsubsection RedHat Version 5.0 Notes +We use GNU Autoconf, so it is possible to port @strong{MySQL} to all modern +systems with working Posix threads and a C++ compiler. (To compile only the +client code, a C++ compiler is required but not threads.) We use and develop +the software ourselves primarily on Sun Solaris (Versions 2.5 - 2.7) and +SuSE Linux Version 7.x. -If you have any problems with @strong{MySQL} on RedHat, you should start by -upgrading @code{glibc} to the newest possible version! +Note that for many operating systems, the native thread support works only +in the latest versions. @strong{MySQL} has been reported to compile +successfully on the following operating system/thread package combinations: -If you install all the official RedHat patches (including -@code{glibc-2.0.7-19} and @code{glibc-devel-2.0.7-19}), both the -binary and source distributions of @strong{MySQL} should work without -any trouble! +@itemize @bullet +@item +AIX 4.x with native threads. @xref{IBM-AIX}. +@item +Amiga. +@item +BSDI 2.x with the included MIT-pthreads package. @xref{BSDI}. +@item +BSDI 3.0, 3.1 and 4.x with native threads. @xref{BSDI}. +@item +DEC Unix 4.x with native threads. @xref{Alpha-DEC-UNIX}. +@item +FreeBSD 2.x with the included MIT-pthreads package. @xref{FreeBSD}. +@item +FreeBSD 3.x and 4.x with native threads. @xref{FreeBSD}. +@item +HP-UX 10.20 with the included MIT-pthreads package. @xref{HP-UX 10.20}. +@item +HP-UX 11.x with the native threads. @xref{HP-UX 11.x}. +@item +Linux 2.0+ with LinuxThreads 0.7.1+ or @code{glibc} 2.0.7+. @xref{Linux}. +@item +Mac OS X Server. @xref{Mac OS X}. +@item +NetBSD 1.3/1.4 Intel and NetBSD 1.3 Alpha (Requires GNU make). @xref{NetBSD}. +@item +OpenBSD > 2.5 with native therads. OpenBSD < 2.5 with the included +MIT-pthreads package. @xref{OpenBSD}. +@item +OS/2 Warp 3, FixPack 29 and OS/2 Warp 4, FixPack 4. @xref{OS/2}. +@item +SGI Irix 6.x with native threads. @xref{SGI-Irix}. +@item +Solaris 2.5 and above with native threads on SPARC and x86. @xref{Solaris}. +@item +SunOS 4.x with the included MIT-pthreads package. @xref{Solaris}. +@item +SCO OpenServer with a recent port of the FSU Pthreads package. @xref{SCO}. +@item +SCO UnixWare 7.0.1. @xref{SCO Unixware}. +@item +Tru64 Unix +@item +Win95, Win98, NT, and Win2000. @xref{Windows}. +@end itemize -The updates are needed because there is a bug in @code{glibc} 2.0.5 in how -@code{pthread_key_create} variables are freed. With @code{glibc} 2.0.5, you -must use a statically linked @strong{MySQL} binary distribution. If you -want to compile from source, you must install the corrected version of -LinuxThreads from @uref{http://www.mysql.com/Downloads/Linux} or upgrade your -@code{glibc}. +Note that not all platforms are suited equally well for running +@strong{MySQL}. How well a certain platform is suited for a high-load +mission critical @strong{MySQL} server is determined by the following +factors: -If you have an incorrect version of @code{glibc} or LinuxThreads, the symptom -is that @code{mysqld} crashes after each connection. For example, -@code{mysqladmin version} will crash @code{mysqld} when it finishes! +@itemize +@item +General stability of the thread library. A platform may have excellent +reputation otherwise, but if the thread library is unstable in the code +that is called by @strong{MySQL}, even if +everything else is perfect, @strong{MySQL} will be only as stable as the +thread library. +@item +The ability of the kernel and/or thread library to take advantage of +@strong{SMP} on +multi-processor systems. In other words, when a process creates a thread, it +should be possible for that thread to run on a different CPU than the original +process. +@item +The ability of the kernel and/or the thread library to run many threads which +acquire/release a mutex over a short critical region frequently without +excessive context switches. In other words, if the implementation of +@code{pthread_mutex_lock()} is too anxious to yield CPU, this will hurt +@strong{MySQL} tremendously. If this issue +is not taken care of, adding extra CPUs will actually make @strong{MySQL} +slower. +@item +General file system stability/performance. +@item +Ability of the file system to deal with large files at all and deal with them +efficiently, if your tables are big. +@item +Our level of expertise here at @strong{MySQL AB} with the platform. If we know +a platform well, we introduce platform-specific optimizations/fixes enabled at +compile time. We can also provide advice on configuring your system optimally +for @strong{MySQL}. +@item +The amount of testing of similar configurations we have done internally. +@item +The number of users that have successfully run @strong{MySQL} on that +platform in similar configurations. If this number is high, the chances of +hitting some platform-specific surprise are much smaller. +@end itemize -Another symptom of incorrect libraries is that @code{mysqld} crashes at -once when it starts. On some Linux systems, this can be fixed by configuring -like this: +Based on the above criteria, the best platforms for running +@strong{MySQL} at this point are x86 with SuSE Linux 7.1, 2.4 kernel and +ReiserFS (or any similar Linux distribution) and Sparc with Solaris 2.7 +or 2.8. FreeBSD comes third, but we really hope it will join the top +club once the thread library is improved. We also hope that at some +point we will be able to include all other platforms on which +@strong{MySQL} compiles, runs ok, but not quite with the same level of +stability and performance, into the top category. This will require some +effort on our part in cooperation with the developers of the OS/library +components @strong{MySQL} depends upon. If you are interested in making +one of those components better, are in a position to influence their +development, and need more detailed instructions on what @strong{MySQL} +needs to run better, send an e-mail to +@email{internals@@lists.mysql.com}. -@example -shell> ./configure --with-mysqld-ldflags=-all-static -@end example +Please note that the comparison above is not to say that one OS is better or +worse than the other in general. We are talking about choosing a particular OS +for a dedicated purpose - running @strong{MySQL}, and compare platforms in that +regard only. With this in mind, the result of this comparison +would be different if we included more issues into it. And in some cases, +the reason one OS is better than the other could simply be that we have put +forth more effort into testing on and optimizing for that particular platform. +We are just stating our observations to help you make a +decision on which platform to use @strong{MySQL} on in your setup. -On Redhat Version 5.0, the easy way out is to install the @code{glibc} -2.0.7-19 RPM and run @code{configure} @strong{without} the -@code{--with-mysqld-ldflags=-all-static} option. +@cindex MySQL binary distribution +@cindex MySQL source distribution +@cindex release numbers +@cindex version, choosing +@cindex choosing, a MySQL version +@node Which version, Many versions, Which OS, Installing +@section Which MySQL Version to Use -For the source distribution of @code{glibc} 2.0.7, a patch that is easy to -apply and is tested with @strong{MySQL} may be found at: +The first decision to make is whether you want to use the latest development +release or the last stable release: -@example -@uref{http://www.mysql.com/Downloads/Linux/glibc-2.0.7-total-patch.tar.gz} -@end example +@itemize @bullet +@item +Normally, if you are beginning to use @strong{MySQL} for the first time +or trying to port it to some system for which there is no binary +distribution, we recommend going with the stable release (currently +Version @value{mysql_version}. Note that all @strong{MySQL} releases are +checked with the @strong{MySQL} benchmarks and an extensive test suite +before each release. -If you experience crashes like these when you build @strong{MySQL}, you can -always download the newest binary version of @strong{MySQL}. This is -statically-linked to avoid library conflicts and should work on all Linux -systems! +@item +Otherwise, if you are running an old system and want to upgrade, but +don't want to take chances with a non-seamless upgrade, you should +upgrade to the latest in the same branch you are using (where only the +last version number is newer than yours). We have tried to fix only +fatal bugs and make small, relatively safe changes to that version. +@end itemize -@strong{MySQL} comes with an internal debugger that can generate -trace files with a lot of information that can be used to find and solve a -wide range of different problems. -@xref{Debugging server}. +The second decision to make is whether you want to use a source +distribution or a binary distribution. In most cases you should probably +use a binary distribution, if one exists for your platform, as this +generally will be easier to install than a source distribution. -@node Linux-RedHat51, Linux-SPARC, Linux-RedHat50, Linux -@subsubsection RedHat Version 5.1 notes +In the following cases you probably will be better off with a source +installation: -The @code{glibc} of RedHat Version 5.1 (@code{glibc} 2.0.7-13) has a memory -leak, so to get a stable @strong{MySQL} version, you must upgrade @code{glibc}, -to 2.0.7-19, downgrade @code{glibc} or use a binary version of @code{mysqld}. -If you don't do this, you will encounter memory problems (out of memory, etc.). -The most common error in this case is: +@itemize @bullet +@item +If you want to install @strong{MySQL} at some explicit location. (The standard +binary distributions are ``ready to run'' at any place, but you may want +to get even more flexibility). -@example -Can't create a new thread (errno 11). If you are not out of available -memory, you can consult the manual for any possible OS dependent bug -@end example +@item +To be able to satisfy different user requirements, we are providing two +different binary versions; One compiled with the non-transactional table +handlers, (a small, fast binary), and one configured with the most +important extended options like transaction-safe tables. Both versions +are compiled from the same source distribution. All native @code{MySQL} +clients can connect to both @strong{MySQL} versions. -After you have upgraded to @code{glibc} 2.0.7-19, you can configure -@strong{MySQL} with dynamic linking (the default), but you @strong{cannot} -run @code{configure} with the @code{--with-mysqld-ldflags=-all-static} option -until you have installed @code{glibc} 2.0.7-19 from source! +The extended @strong{MySQL} binary distribution is marked with the +@code{-max} suffix and is configured with the same options as +@code{mysqld-max}. @xref{mysqld-max, , @code{mysqld-max}}. -You can check which version of @code{glibc} you have with @code{rpm -q glibc}. +If you want to use the @code{MySQL-Max} RPM, you must first +install the standard @code{MySQL} RPM. -Another reason for the above error is if you try to use more threads -than your Linux kernel is configured for. In this case you should raise -the limits in @file{include/linux/tasks.h} and recompile your kernel! +@item +If you want to configure @code{mysqld} with some extra feature that are NOT in +the standard binary distributions. Here is a list of the most common +extra options that you may want to use: -@node Linux-SPARC, Linux-Alpha, Linux-RedHat51, Linux -@subsubsection Linux-SPARC Notes +@itemize @bullet +@item --with-berkeley-db +@item --with-innodb +@item --with-raid +@item --with-libwrap +@item --with-named-z-lib (This is done for some of the binaries) +@item --with-debug[=full] +@end itemize -In some implementations, @code{readdir_r()} is broken. The symptom is that -@code{SHOW DATABASES} always returns an empty set. This can be fixed by -removing @code{HAVE_READDIR_R} from @file{config.h} after configuring and -before compiling. +@item +The default binary distribution is normally compiled with support +for all characters sets and should work on a variety of processors from +the same processor family. -Some problems will require patching your Linux installation. The patch can -be found at -@uref{http://www.mysql.com/Downloads/patches/Linux-sparc-2.0.30.diff}. -This patch is against the Linux distribution @file{sparclinux-2.0.30.tar.gz} -that is available at @code{vger.rutgers.edu} (a version of Linux that was -never merged with the official 2.0.30). You must also install LinuxThreads -Version 0.6 or newer. +If you want a faster @strong{MySQL} server you may want to recompile it +with support for only the character sets you need, use a better compiler +(like @code{pgcc}) or use compiler options that are better optimized for your +processor. -@node Linux-Alpha, MKLinux, Linux-SPARC, Linux -@subsubsection Linux-Alpha Notes +@item +If you have found a bug and reported it to the @strong{MySQL} +development team you will probably receive a patch that you need to apply to +the source distribution to get the bug fixed. -@strong{MySQL} Version 3.23.12 is the first @strong{MySQL} version that is -tested on Linux-Alpha. If you plan to use @strong{MySQL} on Linux-Alpha, -you should ensure that you have this version or newer. +@item +If you want to read (and/or modify) the C and C++ code that makes up +@strong{MySQL}, you should get a source distribution. The source code is +always the ultimate manual. Source distributions also contain more +tests and examples than binary distributions. +@end itemize -We have tested @strong{MySQL} on Alpha with our benchmarks and test suite, -and it appears to work nicely. The main thing we haven't yet had time to -test is how things works with many concurrent users. +@cindex naming, releases of MySQL +@cindex releases, naming scheme -When we compiled the standard @strong{MySQL} binary we are using SuSE 6.4, -kernel 2.2.13-SMP, Compaq C compiler (V6.2-504) and Compaq C++ compiler -(V6.3-005) on a Comaq DS20 machine with an Alpha EV6 processor. +The @strong{MySQL} naming scheme uses release numbers that consist of three +numbers and a suffix. For example, a release name like +@code{mysql-3.21.17-beta} is interpreted like this: -You can find the above compilers at -@uref{http://www.support.compaq.com/alpha-tools}). By using these compilers, -instead of gcc, we get about 9-14 % better performance with @strong{MySQL}. +@itemize @bullet +@item +The first number (@code{3}) describes the file format. All Version 3 +releases have the same file format. -Note that the configure line optimized the binary for the current CPU; This -means you can only use our binary if you have an Alpha EV6 processor. We also -compile statically to avoid library problems. +@item +The second number (@code{21}) is the release level. Normally there are two to +choose from. One is the release/stable branch (currently @code{23}) and the +other is the development branch (currently @code{4.0}). Normally both are +stable, but the development version may have quirks, missing documentation on +new features, or may fail to compile on some systems. -@example -CC=ccc CFLAGS="-fast" CXX=cxx CXXFLAGS="-fast -noexceptions -nortti" ./configure --prefix=/usr/local/mysql --disable-shared --with-extra-charsets=complex --enable-thread-safe-client --with-mysqld-ldflags=-non_shared --with-client-ldflags=-non_shared -@end example +@item +The third number (@code{17}) is the version number within the +release level. This is incremented for each new distribution. Usually you +want the latest version for the release level you have chosen. -If you want to use egcs the following configure line worked for us: +@item +The suffix (@code{beta}) indicates the stability level of the release. +The possible suffixes are: -@example -CFLAGS="-O3 -fomit-frame-pointer" CXX=gcc CXXFLAGS="-O3 -fomit-frame-pointer -felide-constructors -fno-exceptions -fno-rtti" ./configure --prefix=/usr/local/mysql --disable-shared -@end example +@itemize @minus +@item +@code{alpha} indicates that the release contains some large section of +new code that hasn't been 100% tested. Known bugs (usually there are none) +should be documented in the News section. @xref{News}. There are also new +commands and extensions in most alpha releases. Active development that +may involve major code changes can occur on an alpha release, but everything +will be tested before doing a release. There should be no known bugs in any +@strong{MySQL} release. -Some known problems when running @strong{MySQL} on Linux-Alpha: +@item +@code{beta} means that all new code has been tested. No major new +features that could cause corruption on old code are added. There should +be no known bugs. A version changes from alpha to beta when there +haven't been any reported fatal bugs within an alpha version for at least +a month and we don't plan to add any features that could make any old command +more unreliable. -@itemize @bullet @item -Debugging threaded applications like @strong{MySQL} will not work with -@code{gdb 4.18}. You should download and use gdb 5.0 instead! +@code{gamma} is a beta that has been around a while and seems to work fine. +Only minor fixes are added. This is what many other companies call a release. @item -If you try linking @code{mysqld} statically when using @code{gcc}, the -resulting image will core dump at start. In other words, @strong{DON'T} -use @code{--with-mysqld-ldflags=-all-static} with @code{gcc}. +If there is no suffix, it means that the version has been run for a +while at many different sites with no reports of bugs other than +platform-specific bugs. Only critical bug fixes are applied to the +release. This is what we call a stable release. +@end itemize @end itemize -@node MKLinux, Qube2, Linux-Alpha, Linux -@subsubsection MkLinux Notes - -@strong{MySQL} should work on MkLinux with the newest @code{glibc} package -(tested with @code{glibc} 2.0.7). +All versions of @strong{MySQL} are run through our standard tests and +benchmarks to ensure that they are relatively safe to use. Because the +standard tests are extended over time to check for all previously found bugs, +the test suite keeps getting better. -@node Qube2, Linux-Ia64, MKLinux, Linux -@subsubsection Qube2 Linux Notes +@cindex releases, testing +@cindex testing, of MySQL releases -To get @strong{MySQL} to work on Qube2, (Linux Mips), you need the -newest @code{glibc} libraries (@code{glibc-2.0.7-29C2} is known to -work). You must also use the @code{egcs} C++ compiler -(@code{egcs-1.0.2-9}, @code{gcc 2.95.2} or newer). +Note that all releases have been tested at least with: -@node Linux-Ia64, , Qube2, Linux -@subsubsection Linux IA64 Notes +@table @asis +@item An internal test suite +This is part of a production system for a customer. It has many tables with +hundreds of megabytes of data. -To get @strong{MySQL} to compile on Linux Ia64, we had to do the following -(we assume that this will be easier when next gcc version for ia64 is -released). +@item The @strong{MySQL} benchmark suite +This runs a range of common queries. It is also a test to see whether the +latest batch of optimizations actually made the code faster. +@xref{Benchmarks}. -Using @code{gcc-2.9-final}: +@item The @code{crash-me} test +This tries to determine what features the database supports and what its +capabilities and limitations are. @xref{Benchmarks}. +@end table -@example -CFLAGS="-O2" CXX=gcc CXXFLAGS="-O2 -felide-constructors -fno-exceptions -fno-rtti" ./configure --prefix=/usr/local/mysql --enable-assembler --with-mysqld-ldflags=-all-static --disable-shared --with-extra-charsets=complex -@end example +Another test is that we use the newest @strong{MySQL} version in our internal +production environment, on at least one machine. We have more than 100 +gigabytes of data to work with. -After @code{make} you will get an error that @code{sql/opt_range.cc} -will not compile (internal compiler error). To fix this, go to the sql -directory and type @code{make} again. Copy the compile line, but change --O2 to -O0. The file should now compile. +@cindex releases, updating +@cindex updating, releases of MySQL +@node Many versions, Installation layouts, Which version, Installing +@section How and When Updates Are Released -Now you can do: +@strong{MySQL} is evolving quite rapidly here at @strong{MySQL AB} and we +want to share this with other @strong{MySQL} users. We try to make a release +when we have very useful features that others seem to have a need for. -@example -cd .. -make -make_install -@end example +We also try to help out users who request features that are easy to +implement. We take note of what our licensed users want to have, and +we especially take note of what our extended e-mail supported customers +want and try to help them out. -and @strong{mysqld} should be ready to run. +No one has to download a new release. The News section will tell you if +the new release has something you really want. @xref{News}. -@node Alpha-DEC-UNIX, Alpha-DEC-OSF1, Linux, Source install system issues -@subsection Alpha-DEC-UNIX Notes (Tru64) +We use the following policy when updating @strong{MySQL}: -If you are using egcs 1.1.2 on Digital Unix, you should upgrade to gcc -2.95.2, as egcs on DEC has some serious bugs! +@itemize @bullet +@item +For each minor update, the last number in the version string is incremented. +When there are major new features or minor incompatibilities with previous +versions, the second number in the version string is incremented. When the +file format changes, the first number is increased. -When compiling threaded programs under Digital Unix, the documentation -recommends using the @code{-pthread} option for @code{cc} and @code{cxx} and -the libraries @code{-lmach -lexc} (in addition to @code{-lpthread}). You -should run @code{configure} something like this: +@item +Stable tested releases are meant to appear about 1-2 times a year, but +if small bugs are found, a release with only bug fixes will be released. -@example -CC="cc -pthread" CXX="cxx -pthread -O" \ -./configure --with-named-thread-libs="-lpthread -lmach -lexc -lc" -@end example +@item +Working releases are meant to appear about every 1-8 weeks. -When compiling @code{mysqld}, you may see a couple of warnings like this: +@item +Binary distributions for some platforms will be made by us for major releases. +Other people may make binary distributions for other systems but probably +less frequently. -@example -mysqld.cc: In function void handle_connections()': -mysqld.cc:626: passing long unsigned int *' as argument 3 of -accept(int,sockadddr *, int *)' -@end example +@item +We usually make patches available as soon as we have located and fixed +small bugs. -You can safely ignore these warnings. They occur because @code{configure} -can detect only errors, not warnings. +@item +For non-critical but annoying bugs, we will make patches available if they +are sent to us. Otherwise we will combine many of them into a larger +patch. -If you start the server directly from the command line, you may have problems -with it dying when you log out. (When you log out, your outstanding processes -receive a @code{SIGHUP} signal.) If so, try starting the server like this: +@item +If there is, by any chance, a fatal bug in a release we will make a new +release as soon as possible. We would like other companies to do this, +too. +@end itemize -@example -shell> nohup mysqld [options] & -@end example +The current stable release is Version 3.23; We have already moved active +development to Version 4.0. Bugs will still be fixed in the stable version. +We don't believe in a complete freeze, as this also leaves out bug fixes +and things that ``must be done.'' ``Somewhat frozen'' means that we may +add small things that ``almost surely will not affect anything that's +already working.'' -@code{nohup} causes the command following it to ignore any @code{SIGHUP} -signal sent from the terminal. Alternatively, start the server by running -@code{safe_mysqld}, which invokes @code{mysqld} using @code{nohup} for you. -@xref{safe_mysqld, , @code{safe_mysqld}}. +@node Installation layouts, Installing binary, Many versions, Installing +@section Installation Layouts -If you get a problem when compiling mysys/get_opt.c, just remove the -line #define _NO_PROTO from the start of that file! +@cindex installation layouts +@cindex layout of installation +@cindex directory structure, default +@cindex default installation location -If you are using Compac's CC compiler, the following configure line should -work: +This section describes the default layout of the directories created by +installing binary and source distributions. -@example -CC="cc -pthread" -CFLAGS="-O4 -ansi_alias -ansi_args -fast -inline speed -speculate all -arch host" -CXX="cxx -pthread" -CXXFLAGS="-O4 -ansi_alias -ansi_args -fast -inline speed -speculate all -arch host" -export CC CFLAGS CXX CXXFLAGS -./configure \ ---prefix=/usr/local/mysql \ ---with-low-memory \ ---enable-large-files \ ---enable-shared=yes \ ---with-named-thread-libs="-lpthread -lmach -lexc -lc" -gnumake -@end example +A binary distribution is installed by unpacking it at the installation +location you choose (typically @file{/usr/local/mysql}) and creates the +following directories in that location: -If you get a problem with libtool, when compiling with shared libraries -as above, when linking @code{mysql}, you should be able to get around -this by issuing: +@multitable @columnfractions .3 .7 +@item @strong{Directory} @tab @strong{Contents of directory} +@item @file{bin} @tab Client programs and the @code{mysqld} server +@item @file{data} @tab Log files, databases +@item @file{include} @tab Include (header) files +@item @file{lib} @tab Libraries +@item @file{scripts} @tab @code{mysql_install_db} +@item @file{share/mysql} @tab Error message files +@item @file{sql-bench} @tab Benchmarks +@end multitable -@example -cd mysql -/bin/sh ../libtool --mode=link cxx -pthread -O3 -DDBUG_OFF \ --O4 -ansi_alias -ansi_args -fast -inline speed \ --speculate all \ -arch host -DUNDEF_HAVE_GETHOSTBYNAME_R \ --o mysql mysql.o readline.o sql_string.o completion_hash.o \ -../readline/libreadline.a -lcurses \ -../libmysql/.libs/libmysqlclient.so -lm -cd .. -gnumake -gnumake install -scripts/mysql_install_db -@end example +A source distribution is installed after you configure and compile it. By +default, the installation step installs files under @file{/usr/local}, in the +following subdirectories: -@node Alpha-DEC-OSF1, SGI-Irix, Alpha-DEC-UNIX, Source install system issues -@subsection Alpha-DEC-OSF1 Notes +@multitable @columnfractions .3 .7 +@item @strong{Directory} @tab @strong{Contents of directory} +@item @file{bin} @tab Client programs and scripts +@item @file{include/mysql} @tab Include (header) files +@item @file{info} @tab Documentation in Info format +@item @file{lib/mysql} @tab Libraries +@item @file{libexec} @tab The @code{mysqld} server +@item @file{share/mysql} @tab Error message files +@item @file{sql-bench} @tab Benchmarks and @code{crash-me} test +@item @file{var} @tab Databases and log files +@end multitable -If you have problems compiling and have DEC @code{CC} and @code{gcc} -installed, try running @code{configure} like this: +Within an installation directory, the layout of a source installation differs +from that of a binary installation in the following ways: -@example -CC=cc CFLAGS=-O CXX=gcc CXXFLAGS=-O3 \ -./configure --prefix=/usr/local/mysql -@end example +@itemize @bullet +@item +The @code{mysqld} server is installed in the @file{libexec} +directory rather than in the @file{bin} directory. -If you get problems with the @file{c_asm.h} file, you can create and use -a 'dummy' @file{c_asm.h} file with: +@item +The data directory is @file{var} rather than @file{data}. -@example -touch include/c_asm.h -CC=gcc CFLAGS=-I./include \ -CXX=gcc CXXFLAGS=-O3 \ -./configure --prefix=/usr/local/mysql -@end example +@item +@code{mysql_install_db} is installed in the @file{/usr/local/bin} directory +rather than in @file{/usr/local/mysql/scripts}. -Note that the following problems with the @code{ld} program can be fixed -by downloading the latest DEC (Compaq) patch kit from: -@uref{http://ftp.support.compaq.com/public/unix/}. +@item +The header file and library directories are @file{include/mysql} and +@file{lib/mysql} rather than @file{include} and @file{lib}. +@end itemize -On OSF1 V4.0D and compiler "DEC C V5.6-071 on Digital Unix V4.0 (Rev. 878)" -the compiler had some strange behavior (undefined @code{asm} symbols). -@code{/bin/ld} also appears to be broken (problems with @code{_exit -undefined} errors occuring while linking @code{mysqld}). On this system, we -have managed to compile @strong{MySQL} with the following @code{configure} -line, after replacing @code{/bin/ld} with the version from OSF 4.0C: +You can create your own binary installation from a compiled source +distribution by executing the script @file{scripts/make_binary_distribution}. -@example -CC=gcc CXX=gcc CXXFLAGS=-O3 ./configure --prefix=/usr/local/mysql -@end example +@cindex installing, binary distribution +@cindex binary distributions, installing +@node Installing binary, Installing source, Installation layouts, Installing +@section Installing a MySQL Binary Distribution -With the Digital compiler "C++ V6.1-029", the following should work: +@menu +* Linux-RPM:: Linux RPM files +* Building clients:: Building client programs +* Binary install system issues:: System-specific issues +@end menu -@example -CC=cc -pthread -CFLAGS=-O4 -ansi_alias -ansi_args -fast -inline speed -speculate all -arch host -CXX=cxx -pthread -CXXFLAGS=-O4 -ansi_alias -ansi_args -fast -inline speed -speculate all -arch host -noexceptions -nortti -export CC CFLAGS CXX CXXFLAGS -./configure --prefix=/usr/mysql/mysql --with-mysqld-ldflags=-all-static --disable-shared --with-named-thread-libs="-lmach -lexc -lc" -@end example +You need the following tools to install a @strong{MySQL} binary distribution: -In some versions of OSF1, the @code{alloca()} function is broken. Fix -this by removing the line in @file{config.h} that defines @code{'HAVE_ALLOCA'}. +@itemize @bullet +@item +GNU @code{gunzip} to uncompress the distribution. -The @code{alloca()} function also may have an incorrect prototype in -@code{/usr/include/alloca.h}. This warning resulting from this can be ignored. +@item +A reasonable @code{tar} to unpack the distribution. GNU @code{tar} is +known to work. Sun @code{tar} is known to have problems. +@end itemize -@code{configure} will use the following thread libraries automatically: -@code{--with-named-thread-libs="-lpthread -lmach -lexc -lc"}. +@cindex RPM, defined +@cindex RedHat Package Manager +An alternative installation method under Linux is to use RPM (RedHat Package +Manager) distributions. @xref{Linux-RPM}. -When using @code{gcc}, you can also try running @code{configure} like this: +@c texi2html fails to split chapters if I use strong for all of this. +If you run into problems, @strong{PLEASE ALWAYS USE} @code{mysqlbug} when +posting questions to @email{mysql@@lists.mysql.com}. Even if the problem +isn't a bug, @code{mysqlbug} gathers system information that will help others +solve your problem. By not using @code{mysqlbug}, you lessen the likelihood +of getting a solution to your problem! You will find @code{mysqlbug} in the +@file{bin} directory after you unpack the distribution. @xref{Bug reports}. + +@cindex commands, for binary distribution +The basic commands you must execute to install and use a @strong{MySQL} +binary distribution are: @example -shell> CFLAGS=-D_PTHREAD_USE_D4 CXX=gcc CXXFLAGS=-O3 ./configure .... +shell> groupadd mysql +shell> useradd -g mysql mysql +shell> cd /usr/local +shell> gunzip < /path/to/mysql-VERSION-OS.tar.gz | tar xvf - +shell> ln -s mysql-VERSION-OS mysql +shell> cd mysql +shell> scripts/mysql_install_db +shell> chown -R root /usr/local/mysql +shell> chown -R mysql /usr/local/mysql/data +shell> chgrp -R mysql /usr/local/mysql +shell> chown -R root /usr/local/mysql/bin +shell> bin/safe_mysqld --user=mysql & @end example -If you have problems with signals (@strong{MySQL} dies unexpectedly -under high load), you may have found an OS bug with threads and -signals. In this case you can tell @strong{MySQL} not to use signals by -configuring with: +@cindex adding, new users +@cindex new users, adding +@cindex users, adding -@example -shell> CFLAGS=-DDONT_USE_THR_ALARM \ - CXXFLAGS=-DDONT_USE_THR_ALARM \ - ./configure ... -@end example +You can add new users using the @code{bin/mysql_setpermission} script if +you install the @code{DBI} and @code{Msql-Mysql-modules} Perl modules. -This doesn't affect the performance of @strong{MySQL}, but has the side -effect that you can't kill clients that are ``sleeping'' on a connection with -@code{mysqladmin kill} or @code{mysqladmin shutdown}. Instead, the client -will die when it issues its next command. +A more detailed description follows. -With @code{gcc} 2.95.2, you will probably run into the following compile error: +To install a binary distribution, follow the steps below, then proceed +to @ref{Post-installation}, for post-installation setup and testing: -@example -sql_acl.cc:1456: Internal compiler error in `scan_region', at except.c:2566 -Please submit a full bug report. -@end example +@enumerate +@item +Pick the directory under which you want to unpack the distribution, and move +into it. In the example below, we unpack the distribution under +@file{/usr/local} and create a directory @file{/usr/local/mysql} into which +@strong{MySQL} is installed. (The following instructions therefore assume +you have permission to create files in @file{/usr/local}. If that directory +is protected, you will need to perform the installation as @code{root}.) -To fix this you should change to the @code{sql} directory and do a ``cut -and paste'' of the last @code{gcc} line, but change @code{-O3} to -@code{-O0} (or add @code{-O0} immediately after @code{gcc} if you don't -have any @code{-O} option on your compile line.) After this is done you -can just change back to the top-level directly and run @code{make} -again. +@item +Obtain a distribution file from one of the sites listed in +@ref{Getting MySQL, , Getting @strong{MySQL}}. -@node SGI-Irix, FreeBSD, Alpha-DEC-OSF1, Source install system issues -@subsection SGI-Irix Notes +@strong{MySQL} binary distributions are provided as compressed @code{tar} +archives and have names like @file{mysql-VERSION-OS.tar.gz}, where +@code{VERSION} is a number (for example, @code{3.21.15}), and @code{OS} +indicates the type of operating system for which the distribution is intended +(for example, @code{pc-linux-gnu-i586}). -If you are using Irix Version 6.5.3 or newer @code{mysqld} will only be able to -create threads if you run it as a user with @code{CAP_SCHED_MGT} -privileges (like @code{root}) or give the @code{mysqld} server this privilege -with the following shell command: +@item +If you see a binary distribution marked with the @code{-max} prefix, this +means that the binary has support for transaction-safe tables and other +features. @xref{mysqld-max, , @code{mysqld-max}}. Note that all binaries +are built from the same @strong{MySQL} source distribution. + +@item +Add a user and group for @code{mysqld} to run as: @example -shell> chcap "CAP_SCHED_MGT+epi" /opt/mysql/libexec/mysqld +shell> groupadd mysql +shell> useradd -g mysql mysql @end example -You may have to undefine some things in @file{config.h} after running -@code{configure} and before compiling. +These commands add the @code{mysql} group and the @code{mysql} user. The +syntax for @code{useradd} and @code{groupadd} may differ slightly on different +versions of Unix. They may also be called @code{adduser} and @code{addgroup}. +You may wish to call the user and group something else instead of @code{mysql}. -In some Irix implementations, the @code{alloca()} function is broken. If the -@code{mysqld} server dies on some @code{SELECT} statements, remove the lines -from @file{config.h} that define @code{HAVE_ALLOC} and @code{HAVE_ALLOCA_H}. -If @code{mysqladmin create} doesn't work, remove the line from @file{config.h} -that defines @code{HAVE_READDIR_R}. You may have to remove the -@code{HAVE_TERM_H} line as well. +@item +Change into the intended installation directory: -SGI recommends that you install all of the patches on this page as a set: -http://support.sgi.com/surfzone/patches/patchset/6.2_indigo.rps.html +@example +shell> cd /usr/local +@end example -At the very minimum, you should install the latest kernel rollup, the -latest @code{rld} rollup, and the latest @code{libc} rollup. +@item +Unpack the distribution and create the installation directory: -You definitely need all the POSIX patches on this page, for pthreads support: +@example +shell> gunzip < /path/to/mysql-VERSION-OS.tar.gz | tar xvf - +shell> ln -s mysql-VERSION-OS mysql +@end example -@uref{http://support.sgi.com/surfzone/patches/patchset/6.2_posix.rps.html} +The first command creates a directory named @file{mysql-VERSION-OS}. The +second command makes a symbolic link to that directory. This lets you refer +more easily to the installation directory as @file{/usr/local/mysql}. -If you get the something like the following error when compiling -@file{mysql.cc}: +@item +Change into the installation directory: @example -"/usr/include/curses.h", line 82: error(1084): invalid combination of type +shell> cd mysql @end example -Type the following in the top-level directory of your @strong{MySQL} source -tree: +You will find several files and subdirectories in the @code{mysql} directory. +The most important for installation purposes are the @file{bin} and +@file{scripts} subdirectories. -@example -shell> extra/replace bool curses_bool < /usr/include/curses.h > include/curses.h -shell> make -@end example +@table @file +@item bin +@tindex PATH environment variable +@tindex environment variable, PATH +This directory contains client programs and the server +You should add the full pathname of this directory to your +@code{PATH} environment variable so that your shell finds the @strong{MySQL} +programs properly. @xref{Environment variables}. -There have also been reports of scheduling problems. If only one thread is -running, things go slow. Avoid this by starting another client. This may -lead to a 2-to-10-fold increase in execution speed thereafter for the other -thread. This is a poorly understood problem with Irix threads; you may have -to improvise to find solutions until this can be fixed. +@item scripts +This directory contains the @code{mysql_install_db} script used to initialize +the @code{mysql} database containing the grant tables that store the server +access permissions. +@end table -If you are compiling with @code{gcc}, you can use the following -@code{configure} command: +@item +If you would like to use @code{mysqlaccess} and have the @strong{MySQL} +distribution in some non-standard place, you must change the location where +@code{mysqlaccess} expects to find the @code{mysql} client. Edit the +@file{bin/mysqlaccess} script at approximately line 18. Search for a line +that looks like this: @example -CC=gcc CXX=gcc CXXFLAGS=-O3 \ -./configure --prefix=/usr/local/mysql --with-thread-safe-client --with-named-thread-libs=-lpthread +$MYSQL = '/usr/local/bin/mysql'; # path to mysql executable @end example -On Irix 6.5.11 with native Irix C and C++ compilers ver. 7.3.1.2, the -following is reported to work +Change the path to reflect the location where @code{mysql} actually is +stored on your system. If you do not do this, you will get a @code{Broken +pipe} error when you run @code{mysqlaccess}. +@item +Create the @strong{MySQL} grant tables (necessary only if you haven't +installed @strong{MySQL} before): @example -CC=cc CXX=CC CFLAGS='-O3 -n32 -TARG:platform=IP22 -I/usr/local/include \ --L/usr/local/lib' CXXFLAGS='-O3 -n32 -TARG:platform=IP22 \ --I/usr/local/include -L/usr/local/lib' ./configure --prefix=/usr/local/mysql \ ---with-berkeley-db --with-innodb \ ---with-libwrap=/usr/local --with-named-curses-libs=/usr/local/lib/libncurses.a +shell> scripts/mysql_install_db @end example -@node FreeBSD, NetBSD, SGI-Irix, Source install system issues -@subsection FreeBSD Notes - -FreeBSD 3.x is recommended for running @strong{MySQL} since the thread package -is much more integrated. - -The easiest and therefor the preferred way to install is to use the -mysql-server and mysql-client ports available on @uref{http://www.freebsd.org}. +Note that @strong{MySQL} versions older than Version 3.22.10 started the +@strong{MySQL} server when you run @code{mysql_install_db}. This is no +longer true! -Using these gives you: -@itemize @bullet @item -A working @strong{MySQL} with all optimizations known to work on your version -of FreeBSD enabled. +Change ownership of binaries to @code{root} and ownership of the data +directory to the user that you will run @code{mysqld} as: -@item -Automatic configuration and build. +@example +shell> chown -R root /usr/local/mysql +shell> chown -R mysql /usr/local/mysql/data +shell> chgrp -R mysql /usr/local/mysql +@end example -@item -Startup scripts installed in /usr/local/etc/rc.d. +The first command changes the @code{owner} attribute of the files to the +@code{root} user, the second one changes the @code{owner} attribute of the +data directory to the @code{mysql} user, and the third one changes the +@code{group} attribute to the @code{mysql} group. @item -Ability to see which files that are installed with pkg_info -L. And to -remove them all with pkg_delete if you no longer want @strong{MySQL} on that -machine. -@end itemize +If you want to install support for the Perl @code{DBI}/@code{DBD} interface, +see @ref{Perl support}. -It is recommended you use MIT-pthreads on FreeBSD 2.x and native threads on -Versions 3 and up. It is possible to run with native threads on some late -2.2.x versions but you may encounter problems shutting down @code{mysqld}. +@item +If you would like @strong{MySQL} to start automatically when you boot your +machine, you can copy @code{support-files/mysql.server} to the location where +your system has its startup files. More information can be found in the +@code{support-files/mysql.server} script itself and in +@ref{Automatic start}. -The @strong{MYSQL} Makefiles require GNU make (@code{gmake}) to work. -If you want to compile @strong{MYSQL} you need to install GNU make -first. +@end enumerate -Be sure to have your name resolver setup correct. Otherwise you may -experience resolver delays or failures when connecting to @code{mysqld}. +After everything has been unpacked and installed, you should initialize +and test your distribution. -Make sure that the @code{localhost} entry in the @file{/etc/hosts} file is -correct (otherwise you will have problems connecting to the database). The -@file{/etc/hosts} file should start with a line: +You can start the @strong{MySQL} server with the following command: @example -127.0.0.1 localhost localhost.your.domain +shell> bin/safe_mysqld --user=mysql & @end example -If you notice that @code{configure} will use MIT-pthreads, you should read -the MIT-pthreads notes. @xref{MIT-pthreads}. +@xref{safe_mysqld, , @code{safe_mysqld}}. -If you get an error from @code{make install} that it can't find -@file{/usr/include/pthreads}, @code{configure} didn't detect that you need -MIT-pthreads. This is fixed by executing these commands: +@xref{Post-installation}. -@example -shell> rm config.cache -shell> ./configure --with-mit-threads -@end example +@cindex RPM file +@cindex RedHat Package Manager +@c This node name is special +@node Linux-RPM, Building clients, Installing binary, Installing binary +@subsection Linux RPM Notes -FreeBSD is also known to have a very low default file handle limit. -@xref{Not enough file handles}. Uncomment the ulimit -n section in -safe_mysqld or raise the limits for the @code{mysqld} user in /etc/login.conf -(and rebuild it with cap_mkdb /etc/login.conf). Also be sure you set the -appropriate class for this user in the password file if you are not -using the default (use: chpass mysqld-user-name). @xref{safe_mysqld, , -@code{safe_mysqld}}. +The recommended way to install @strong{MySQL} on Linux is by using an RPM +file. The @strong{MySQL} RPMs are currently being built on a RedHat Version +6.2 system but should work on other versions of Linux that support @code{rpm} +and use @code{glibc}. -If you get problems with the current date in @strong{MySQL}, setting the -@code{TZ} variable will probably help. @xref{Environment variables}. +If you have problems with an RPM file, for example, if you receive the error +``@code{Sorry, the host 'xxxx' could not be looked up}'', see +@ref{Binary notes-Linux}. -To get a secure and stable system you should only use FreeBSD kernels -that are marked @code{-STABLE}. +The RPM files you may want to use are: -@node NetBSD, OpenBSD, FreeBSD, Source install system issues -@subsection NetBSD notes +@itemize @bullet +@item @code{MySQL-VERSION.i386.rpm} -To compile on NetBSD you need GNU @code{make}. Otherwise the compile will -crash when @code{make} tries to run @code{lint} on C++ files. +The @strong{MySQL} server. You will need this unless you only want to +connect to a @strong{MySQL} server running on another machine. -@node OpenBSD, BSDI, NetBSD, Source install system issues -@subsection OpenBSD Notes +@item @code{MySQL-client-VERSION.i386.rpm} -@menu -* OpenBSD 2.5:: OpenBSD 2.5 Notes -* OpenBSD 2.8:: OpenBSD 2.8 Notes -@end menu +The standard @strong{MySQL} client programs. You probably always want to +install this package. -@node OpenBSD 2.5, OpenBSD 2.8, OpenBSD, OpenBSD -@subsubsection OpenBSD 2.5 Notes +@item @code{MySQL-bench-VERSION.i386.rpm} -On OpenBSD Version 2.5, you can compile @strong{MySQL} with native threads -with the following options: +Tests and benchmarks. Requires Perl and msql-mysql-modules RPMs. + +@item @code{MySQL-devel-VERSION.i386.rpm} + +Libraries and include files needed if you want to compile other +@strong{MySQL} clients, such as the Perl modules. + +@item @code{MySQL-VERSION.src.rpm} +This contains the source code for all of the above packages. It can also +be used to try to build RPMs for other architectures (for example, Alpha +or SPARC). +@end itemize + +To see all files in an RPM package, run: @example -CFLAGS=-pthread CXXFLAGS=-pthread ./configure --with-mit-threads=no +shell> rpm -qpl MySQL-VERSION.i386.rpm @end example -@node OpenBSD 2.8, , OpenBSD 2.5, OpenBSD -@subsubsection OpenBSD 2.8 Notes +To perform a standard minimal installation, run: -Our users have reported that OpenBSD 2.8 has a threading bug which causes -problems with @strong{MySQL}. The OpenBSD Developers have fixed the problem, -but as of January 25th, 2001, it's only available in the ``-current'' branch. -The symptoms of this threading bug are: slow response, high load, high CPU -usage, and crashes. +@example +shell> rpm -i MySQL-VERSION.i386.rpm MySQL-client-VERSION.i386.rpm +@end example -@node BSDI, SCO, OpenBSD, Source install system issues -@subsection BSD/OS Notes +To install just the client package, run: -@menu -* BSDI2:: BSD/OS 2.x notes -* BSDI3:: BSD/OS 3.x notes -* BSDI4:: BSD/OS 4.x notes -@end menu +@example +shell> rpm -i MySQL-client-VERSION.i386.rpm +@end example -@node BSDI2, BSDI3, BSDI, BSDI -@subsubsection BSD/OS Version 2.x Notes +The RPM places data in @file{/var/lib/mysql}. The RPM also creates the +appropriate entries in @file{/etc/rc.d/} to start the server automatically +at boot time. (This means that if you have performed a previous +installation, you may want to make a copy of your previously installed +@strong{MySQL} startup file if you made any changes to it, so you don't lose +your changes.) -If you get the following error when compiling @strong{MySQL}, your -@code{ulimit} value for virtual memory is too low: +After installing the RPM file(s), the @code{mysqld} daemon should be running +and you should now be able to start using @strong{MySQL}. +@xref{Post-installation}. -@example -item_func.h: In method `Item_func_ge::Item_func_ge(const Item_func_ge &)': -item_func.h:28: virtual memory exhausted -make[2]: *** [item_func.o] Error 1 -@end example +If something goes wrong, you can find more information in the binary +installation chapter. @xref{Installing binary}. -Try using @code{ulimit -v 80000} and run @code{make} again. If this -doesn't work and you are using @code{bash}, try switching to @code{csh} -or @code{sh}; some BSDI users have reported problems with @code{bash} -and @code{ulimit}. +@cindex client programs, building +@cindex linking +@cindex building, client programs +@cindex programs, client +@node Building clients, Binary install system issues, Linux-RPM, Installing binary +@subsection Building Client Programs -If you are using @code{gcc}, you may also use have to use the -@code{--with-low-memory} flag for @code{configure} to be able to compile -@file{sql_yacc.cc}. +If you compile @strong{MySQL} clients that you've written yourself or that +you obtain from a third party, they must be linked using the +@code{-lmysqlclient -lz} option on the link command. You may also need to +specify a @code{-L} option to tell the linker where to find the library. For +example, if the library is installed in @file{/usr/local/mysql/lib}, use +@code{-L/usr/local/mysql/lib -lmysqlclient -lz} on the link command. -If you get problems with the current date in @strong{MySQL}, setting the -@code{TZ} variable will probably help. @xref{Environment variables}. +For clients that use @strong{MySQL} header files, you may need to specify a +@code{-I} option when you compile them (for example, +@code{-I/usr/local/mysql/include}), so the compiler can find the header +files. -@node BSDI3, BSDI4, BSDI2, BSDI -@subsubsection BSD/OS Version 3.x Notes +@node Binary install system issues, , Building clients, Installing binary +@subsection System-specific Issues -Upgrade to BSD/OS Version 3.1. If that is not possible, install -BSDIpatch M300-038. +@menu +* Binary notes-Linux:: Linux notes for binary distribution +* Binary notes-HP-UX:: HP-UX notes for binary distribution +@end menu -Use the following command when configuring @strong{MySQL}: +The following sections indicate some of the issues that have been observed +on particular systems when installing @strong{MySQL} from a binary +distribution or from RPM files. + +@cindex binary distributions, on Linux +@cindex Linux, binary distribution +@node Binary notes-Linux, Binary notes-HP-UX, Binary install system issues, Binary install system issues +@subsubsection Linux Notes for Binary Distributions + +@strong{MySQL} needs at least Linux Version 2.0. + +The binary release is linked with @code{-static}, which means you do not +normally need to worry about which version of the system libraries you +have. You need not install LinuxThreads, either. A program linked with +@code{-static} is slightly bigger than a dynamically linked program but +also slightly faster (3-5%). One problem, however, is that you can't use +user-definable functions (UDFs) with a statically linked program. If +you are going to write or use UDF functions (this is something only for +C or C++ programmers), you must compile @strong{MySQL} yourself, using +dynamic linking. + +If you are using a @code{libc}-based system (instead of a @code{glibc2} +system), you will probably get some problems with hostname resolving and +@code{getpwnam()} with the binary release. (This is because @code{glibc} +unfortunately depends on some external libraries to resolve hostnames +and @code{getpwent()}, even when compiled with @code{-static}). In this +case you probably get the following error message when you run +@code{mysql_install_db}: @example -shell> env CXX=shlicc++ CC=shlicc2 \ - ./configure \ - --prefix=/usr/local/mysql \ - --localstatedir=/var/mysql \ - --without-perl \ - --with-unix-socket-path=/var/mysql/mysql.sock +Sorry, the host 'xxxx' could not be looked up @end example -The following is also known to work: +or the following error when you try to run @code{mysqld} with the @code{--user} +option: @example -shell> env CC=gcc CXX=gcc CXXFLAGS=-O3 \ - ./configure \ - --prefix=/usr/local/mysql \ - --with-unix-socket-path=/var/mysql/mysql.sock +getpwnam: No such file or directory @end example -You can change the directory locations if you wish, or just use the -defaults by not specifying any locations. - -If you have problems with performance under heavy load, try using the -@code{--skip-thread-priority} option to @code{mysqld}! This will run -all threads with the same priority; on BSDI Version 3.1, this gives better -performance (at least until BSDI fixes their thread scheduler). +You can solve this problem in one of the following ways: -If you get the error @code{virtual memory exhausted} while compiling, -you should try using @code{ulimit -v 80000} and run @code{make} again. -If this doesn't work and you are using @code{bash}, try switching to -@code{csh} or @code{sh}; some BSDI users have reported problems with -@code{bash} and @code{ulimit}. +@itemize @bullet +@item +Get a @strong{MySQL} source distribution (an RPM or the @code{tar.gz} +distribution) and install this instead. +@item +Execute @code{mysql_install_db --force}; This will not execute the +@code{resolveip} test in @code{mysql_install_db}. The downside is that +you can't use host names in the grant tables; you must use IP numbers +instead (except for @code{localhost}). If you are using an old @strong{MySQL} +release that doesn't support @code{--force}, you have to remove the +@code{resolveip} test in @code{mysql_install} with an editor. +@item +Start @code{mysqld} with @code{su} instead of using @code{--user}. +@end itemize -@node BSDI4, , BSDI3, BSDI -@subsubsection BSD/OS Version 4.x Notes +The Linux-Intel binary and RPM releases of @strong{MySQL} are configured +for the highest possible speed. We are always trying to use the fastest +stable compiler available. -BSDI Version 4.x has some thread-related bugs. If you want to use -@strong{MySQL} on this, you should install all thread-related patches. At -least M400-023 should be installed. +@strong{MySQL} Perl support requires Version Perl 5.004_03 or newer. -On some BSDI Version 4.x systems, you may get problems with shared libraries. -The symptom is that you can't execute any client programs, for example, -@code{mysqladmin}. In this case you need to reconfigure not to use -shared libraries with the @code{--disable-shared} option to configure. +On some Linux 2.2 versions, you may get the error @code{Resource +temporarily unavailable} when you do a lot of new connections to a +@code{mysqld} server over TCP/IP. -Some customers have had problems on BSDI 4.0.1 that the @code{mysqld} -binary after a while can't open tables. This is because some -library/system related bug causes @code{mysqld} to change current -directory without asking for this! +The problem is that Linux has a delay between when you close a TCP/IP +socket and until this is actually freed by the system. As there is only +room for a finite number of TCP/IP slots, you will get the above error if +you try to do too many new TCP/IP connections during a small time, like +when you run the @strong{MySQL} @file{test-connect} benchmark over +TCP/IP. -The fix is to either upgrade to 3.23.34 or after running @code{configure} -remove the line @code{#define HAVE_REALPATH} from @code{config.h} -before running make. +We have mailed about this problem a couple of times to different Linux +mailing lists but have never been able to resolve this properly. -Note that the above means that you can't symbolic link a database directories -to another database directory or symbolic link a table to another database -on BSDI! (Making a symbolic link to another disk is ok). +The only known 'fix' to this problem is to use persistent connections in +your clients or use sockets, if you are running the database server +and clients on the same machine. We hope that the @code{Linux 2.4} +kernel will fix this problem in the future. -@node SCO, SCO Unixware, BSDI, Source install system issues -@subsection SCO Notes +@cindex HP-UX, binary distribution +@cindex binary distributions, on HP-UX +@node Binary notes-HP-UX, , Binary notes-Linux, Binary install system issues +@subsubsection HP-UX Notes for Binary Distributions -The current port is tested only on a ``sco3.2v5.0.4'' and -``sco3.2v5.0.5'' system. There has also been a lot of progress on a -port to ``sco 3.2v4.2''. +Some of the binary distributions of @strong{MySQL} for HP-UX is +distributed as an HP depot file and as a tar file. To use the depot +file you must be running at least HP-UX 10.x to have access to HP's +software depot tools. -For the moment the recommended compiler on OpenServer is gcc 2.95.2. With this -you should be able to compile @strong{MySQL} with just: +The HP version of @strong{MySQL} was compiled on an HP 9000/8xx server +under HP-UX 10.20, and uses MIT-pthreads. It is known to work well under +this configuration. @strong{MySQL} Version 3.22.26 and newer can also be +built with HP's native thread package. -@example -CC=gcc CXX=gcc ./configure ... (options) -@end example +Other configurations that may work: -@enumerate +@itemize @bullet @item -For OpenServer 5.0.X you need to use GDS in Skunkware 95 (95q4c). This -is necessary because GNU @code{gcc} 2.7.2 in Skunkware 97 does not have -GNU @code{as}. You can also use @code{egcs} 1.1.2 or newer -@uref{http://www.egcs.com/}. If you are using @code{egcs} 1.1.2 you have -to execute the following command: - -@example -shell> cp -p /usr/include/pthread/stdtypes.h /usr/local/lib/gcc-lib/i386-pc-sco3.2v5.0.5/egcs-2.91.66/include/pthread/ -@end example - +HP 9000/7xx running HP-UX 10.20+ @item -You need the port of GCC 2.5.x for this product and the Development -system. They are required on this version of SCO Unix. You cannot -just use the GCC Dev system. +HP 9000/8xx running HP-UX 10.30 +@end itemize -@item -You should get the FSU Pthreads package and install it first. This can be -found at -@uref{http://www.cs.wustl.edu/~schmidt/ACE_wrappers/FSU-threads.tar.gz}. -You can also get a precompiled package from -@uref{http://www.mysql.com/Downloads/SCO/FSU-threads-3.5c.tar.gz}. +The following configurations almost definitely won't work: +@itemize @bullet @item -FSU Pthreads can be compiled with SCO Unix 4.2 with tcpip. Or -OpenServer 3.0 or Open Desktop 3.0 (OS 3.0 ODT 3.0), with the SCO -Development System installed using a good port of GCC 2.5.x ODT or OS -3.0 you will need a good port of GCC 2.5.x There are a lot of problems -without a good port. The port for this product requires the SCO Unix -Development system. Without it, you are missing the libraries and the -linker that is needed. - +HP 9000/7xx or 8xx running HP-UX 10.x where x < 2 @item -To build FSU Pthreads on your system, do the following: +HP 9000/7xx or 8xx running HP-UX 9.x +@end itemize -@enumerate a -@item -Run @code{./configure} in the @file{threads/src} directory and select -the SCO OpenServer option. This command copies @file{Makefile.SCO5} to -@file{Makefile}. +To install the distribution, use one of the commands below, where +@code{/path/to/depot} is the full pathname of the depot file: +@itemize @bullet @item -Run @code{make}. +To install everything, including the server, client and development tools: -@item -To install in the default @file{/usr/include} directory, login as root, -then @code{cd} to the @file{thread/src} directory, and run @code{make -install}. -@end enumerate +@example +shell> /usr/sbin/swinstall -s /path/to/depot mysql.full +@end example @item -Remember to use GNU @code{make} when making @strong{MySQL}. +To install only the server: -@item -If you don't start @code{safe_mysqld} as root, you probably will get only the -default 110 open files per process. @code{mysqld} will write a note about this -in the log file. +@example +shell> /usr/sbin/swinstall -s /path/to/depot mysql.server +@end example @item -With SCO 3.2V5.0.5, you should use FSU Pthreads version 3.5c or newer. -You should also use gcc 2.95.2 or newer! - -The following @code{configure} command should work: +To install only the client package: @example -shell> ./configure --prefix=/usr/local/mysql --disable-shared +shell> /usr/sbin/swinstall -s /path/to/depot mysql.client @end example @item -With SCO 3.2V4.2, you should use FSU Pthreads version 3.5c or newer. -The following @code{configure} command should work: +To install only the development tools: @example -shell> CFLAGS="-D_XOPEN_XPG4" CXX=gcc CXXFLAGS="-D_XOPEN_XPG4" \ - ./configure \ - --prefix=/usr/local/mysql \ - --with-named-thread-libs="-lgthreads -lsocket -lgen -lgthreads" \ - --with-named-curses-libs="-lcurses" +shell> /usr/sbin/swinstall -s /path/to/depot mysql.developer @end example +@end itemize -You may get some problems with some include files. In this case, you can -find new SCO-specific include files at -@uref{http://www.mysql.com/Downloads/SCO/SCO-3.2v4.2-includes.tar.gz}. -You should unpack this file in the @file{include} directory of your -@strong{MySQL} source tree. -@end enumerate +The depot places binaries and libraries in @file{/opt/mysql} and data in +@file{/var/opt/mysql}. The depot also creates the appropriate entries in +@file{/etc/init.d} and @file{/etc/rc2.d} to start the server automatically +at boot time. Obviously, this entails being @code{root} to install. -SCO development notes: +To install the HP-UX tar.gz distribution, you must have a copy of GNU +@code{tar}. + +@cindex installing, source distribution +@cindex source distribution, installing +@cindex installation overview +@node Installing source, Installing source tree, Installing binary, Installing +@section Installing a MySQL Source Distribution + +Before you proceed with the source installation, check first to see if our +binary is available for your platform and if it will work for you. We +put in a lot of effort into making sure that our binaries are built with the +best possible options. + +You need the following tools to build and install @strong{MySQL} from source: @itemize @bullet @item -@strong{MySQL} should automatically detect FSU Pthreads and link @code{mysqld} -with @code{-lgthreads -lsocket -lgthreads}. +GNU @code{gunzip} to uncompress the distribution. @item -The SCO development libraries are re-entrant in FSU Pthreads. SCO claims -that its libraries' functions are re-entrant, so they must be reentrant with -FSU Pthreads. FSU Pthreads on OpenServer tries to use the SCO scheme to -make re-entrant libraries. +A reasonable @code{tar} to unpack the distribution. GNU @code{tar} is +known to work. Sun @code{tar} is known to have problems. @item -FSU Pthreads (at least the version at @uref{http://www.mysql.com/}) comes -linked with GNU @code{malloc}. If you encounter problems with memory usage, -make sure that @file{gmalloc.o} is included in @file{libgthreads.a} and -@file{libgthreads.so}. +A working ANSI C++ compiler. @code{gcc} >= 2.95.2, @code{egcs} >= 1.0.2 +or @code{egcs 2.91.66}, SGI C++, and SunPro C++ are some of the +compilers that are known to work. @code{libg++} is not needed when +using @code{gcc}. @code{gcc} 2.7.x has a bug that makes it impossible +to compile some perfectly legal C++ files, such as +@file{sql/sql_base.cc}. If you only have @code{gcc} 2.7.x, you must +upgrade your @code{gcc} to be able to compile @strong{MySQL}. @code{gcc} +2.8.1 is also known to have problems on some platforms so it should be +avoided if there exists a new compiler for the platform.. + +@code{gcc} >= 2.95.2 is recommended when compiling @strong{MySQL} +Version 3.23.x. @item -In FSU Pthreads, the following system calls are pthreads-aware: @code{read()}, -@code{write()}, @code{getmsg()}, @code{connect()}, @code{accept()}, -@code{select()}, and @code{wait()}. +A good @code{make} program. GNU @code{make} is always recommended and is +sometimes required. If you have problems, we recommend trying GNU +@code{make} 3.75 or newer. @end itemize -If you want to install DBI on SCO, you have to edit the @file{Makefile} in -DBI-xxx and each subdirectory. - -Note that the following assumes gcc 2.95.2 or newer: +If you are using a recent version of @strong{gcc}, recent enough to understand +@code{-fno-exceptions} option, it is @strong{VERY IMPORTANT} that you use +it. Otherwise, you may compile a binary that crashes randomly. We also +recommend that you use @code{-felide-contructors} and @code{-fno-rtti} along +with @code{-fno-exceptions}. When in doubt, do the following: @example -OLD: NEW: -CC = cc CC = gcc -CCCDLFLAGS = -KPIC -W1,-Bexport CCCDLFLAGS = -fpic -CCDLFLAGS = -wl,-Bexport CCDLFLAGS = -LD = ld LD = gcc -G -fpic -LDDLFLAGS = -G -L/usr/local/lib LDDLFLAGS = -L/usr/local/lib -LDFLAGS = -belf -L/usr/local/lib LDFLAGS = -L/usr/local/lib +CFLAGS="-O3" CXX=gcc CXXFLAGS="-O3 -felide-constructors -fno-exceptions -fno-rtti" ./configure --prefix=/usr/local/mysql --enable-assembler --with-mysqld-ldflags=-all-static -LD = ld LD = gcc -G -fpic -OPTIMISE = -Od OPTIMISE = -O1 +@end example -OLD: -CCCFLAGS = -belf -dy -w0 -U M_XENIX -DPERL_SCO5 -I/usr/local/include +On most systems this will give you a fast and stable binary. -NEW: -CCFLAGS = -U M_XENIX -DPERL_SCO5 -I/usr/local/include -@end example +@c texi2html fails to split chapters if I use strong for all of this. +If you run into problems, @strong{PLEASE ALWAYS USE @code{mysqlbug}} when +posting questions to @email{mysql@@lists.mysql.com}. Even if the problem +isn't a bug, @code{mysqlbug} gathers system information that will help others +solve your problem. By not using @code{mysqlbug}, you lessen the likelihood +of getting a solution to your problem! You will find @code{mysqlbug} in the +@file{scripts} directory after you unpack the distribution. +@xref{Bug reports}. -This is because the Perl dynaloader will not load the @code{DBI} modules -if they were compiled with @code{icc} or @code{cc}. +@menu +* Quick install:: Quick installation overview +* Applying patches:: Applying patches +* configure options:: Typical @code{configure} options +@end menu -Perl works best when compiled with @code{cc}. +@node Quick install, Applying patches, Installing source, Installing source +@subsection Quick Installation Overview -@node SCO Unixware, IBM-AIX, SCO, Source install system issues -@subsection SCO Unixware Version 7.0 Notes +The basic commands you must execute to install a @strong{MySQL} source +distribution are: -You must use a version of @strong{MySQL} at least as recent as Version 3.22.13 -because that version fixes some portability problems under Unixware. +@example +shell> groupadd mysql +shell> useradd -g mysql mysql +shell> gunzip < mysql-VERSION.tar.gz | tar -xvf - +shell> cd mysql-VERSION +shell> ./configure --prefix=/usr/local/mysql +shell> make +shell> make install +shell> scripts/mysql_install_db +shell> chown -R root /usr/local/mysql +shell> chown -R mysql /usr/local/mysql/var +shell> chgrp -R mysql /usr/local/mysql +shell> cp support-files/my-medium.cnf /etc/my.cnf +shell> /usr/local/mysql/bin/safe_mysqld --user=mysql & +@end example -We have been able to compile @strong{MySQL} with the following @code{configure} -command on Unixware Version 7.0.1: +If you want have support for InnoDB tables, you should edit the +@code{/etc/my.cnf} file and remove the @code{#} character before the +parameters that starts with @code{innodb_...}. @xref{Option +files}. @xref{InnoDB start}. + +If you start from a source RPM, then do the following: @example -CC=cc CXX=CC ./configure --prefix=/usr/local/mysql +shell> rpm --rebuild MySQL-VERSION.src.rpm @end example -If you want to use @code{gcc}, you must use @code{gcc} 2.95.2 or newer. +This will make a binary RPM that you can install. -@node IBM-AIX, HP-UX 10.20, SCO Unixware, Source install system issues -@subsection IBM-AIX notes -@cindex problems, installing on IBM-AIX +You can add new users using the @code{bin/mysql_setpermission} script if +you install the @code{DBI} and @code{Msql-Mysql-modules} Perl modules. -Automatic detection of @code{xlC} is missing from Autoconf, so a -@code{configure} command something like this is needed when compiling -@strong{MySQL} (This example uses the IBM compiler): +A more detailed description follows. -@example -export CC="xlc_r -ma -O3 -qstrict -qoptimize=3 -qmaxmem=8192 " -export CXX="xlC_r -ma -O3 -qstrict -qoptimize=3 -qmaxmem=8192" -export CFLAGS="-I /usr/local/include" -export LDLFAGS="-L /usr/local/lib" -export CPPFLAGS=$CFLAGS -export CXXFLAGS=$CFLAGS +To install a source distribution, follow the steps below, then proceed +to @ref{Post-installation}, for post-installation initialization and testing: -./configure --prefix=/usr/local \ - --localstatedir=/var/mysql \ - --sysconfdir=/etc/mysql \ - --sbindir='/usr/local/bin' \ - --libexecdir='/usr/local/bin' \ - --enable-thread-safe-client \ - --enable-large-files -@end example +@enumerate +@item +Pick the directory under which you want to unpack the distribution, and move +into it. -Above are the options used to compile the @strong{MySQL} distribution that -can be found at @uref{http://www-frec.bull.com/}. +@item +Obtain a distribution file from one of the sites listed in +@ref{Getting MySQL, , Getting @strong{MySQL}}. -If you change the @code{-O3} to @code{-O2} in the above configure line, -you must also remove the @code{-qstrict} option (this is a limitation in -the IBM C compiler). +@item +If you are interested in using Berkeley DB tables with MySQL, you +will need to obtain a patched version of the Berkeley DB source +code. Please read the chapter on Berkeley DB tables before +proceeding. @xref{BDB}. -If you are using @code{gcc} or @code{egcs} to compile @strong{MySQL}, you -@strong{MUST} use the @code{-fno-exceptions} flag, as the exception -handling in @code{gcc}/@code{egcs} is not thread safe! (This is tested with -@code{egcs} 1.1.). There are also some known problems with IBM's assembler, -which may cause it to generate bad code when used with gcc. +@strong{MySQL} source distributions are provided as compressed @code{tar} +archives and have names like @file{mysql-VERSION.tar.gz}, where +@code{VERSION} is a number like @value{mysql_version}. -We recommend the following @code{configure} line with @code{egcs} and -@code{gcc 2.95} on AIX: +@item +Add a user and group for @code{mysqld} to run as: @example -CC="gcc -pipe -mcpu=power -Wa,-many" \ -CXX="gcc -pipe -mcpu=power -Wa,-many" \ -CXXFLAGS="-felide-constructors -fno-exceptions -fno-rtti" \ -./configure --prefix=/usr/local/mysql --with-low-memory +shell> groupadd mysql +shell> useradd -g mysql mysql @end example -The @code{-Wa,-many} is necessary for the compile to be successful. IBM is -aware of this problem but is in to hurry to fix it because of the workaround -available. We don't know if the @code{-fno-exceptions} is required with -@code{gcc 2.95}, but as @strong{MySQL} doesn't use exceptions and the above -option generates faster code, we recommend that you should always use this -option with @code{egcs / gcc}. +These commands add the @code{mysql} group, and the @code{mysql} user. The +syntax for @code{useradd} and @code{groupadd} may differ slightly on different +versions of Unix. They may also be called @code{adduser} and @code{addgroup}. +You may wish to call the user and group something else instead of @code{mysql}. -If you get a problem with assembler code try changing the -mcpu=xxx to -match your cpu. Typically power2, power, or powerpc may need to be used, -alternatively you might need to use 604 or 604e. I'm not positive but I -would think using "power" would likely be safe most of the time, even on -a power2 machine. +@item +Unpack the distribution into the current directory: +@example +shell> gunzip < /path/to/mysql-VERSION.tar.gz | tar xvf - +@end example -If you don't know what your cpu is then do a "uname -m", this will give -you back a string that looks like "000514676700", with a format of -xxyyyyyymmss where xx and ss are always 0's, yyyyyy is a unique system -id and mm is the id of the CPU Planar. A chart of these values can be -found at -@uref{http://www.rs6000.ibm.com/doc_link/en_US/a_doc_lib/cmds/aixcmds5/uname.htm}. -This will give you a machine type and a machine model you can use to -determine what type of cpu you have. +This command creates a directory named @file{mysql-VERSION}. -If you have problems with signals (@strong{MySQL} dies unexpectedly -under high load) you may have found an OS bug with threads and -signals. In this case you can tell @strong{MySQL} not to use signals by -configuring with: +@item +Change into the top-level directory of the unpacked distribution: @example -shell> CFLAGS=-DDONT_USE_THR_ALARM CXX=gcc \ - CXXFLAGS="-felide-constructors -fno-exceptions -fno-rtti -DDONT_USE_THR_ALARM" \ - ./configure --prefix=/usr/local/mysql --with-debug --with-low-memory +shell> cd mysql-VERSION @end example -This doesn't affect the performance of @strong{MySQL}, but has the side -effect that you can't kill clients that are ``sleeping'' on a connection with -@code{mysqladmin kill} or @code{mysqladmin shutdown}. Instead, the client -will die when it issues its next command. - -On some versions of AIX, linking with @code{libbind.a} makes -@code{getservbyname} core dump. This is an AIX bug and should be reported -to IBM. - -For AIX 4.2.1 and gcc you have to do the following changes. +Note that currently you must configure and build @strong{MySQL} from +this top-level directory. You can not build it in a different +directory. -After configuring, edit @file{config.h} and @file{include/my_config.h} -and change the line that says +@item +Configure the release and compile everything: @example -#define HAVE_SNPRINTF 1 +shell> ./configure --prefix=/usr/local/mysql +shell> make @end example -to +When you run @code{configure}, you might want to specify some options. +Run @code{./configure --help} for a list of options. +@ref{configure options, , @code{configure} options}, discusses some of the +more useful options. -@example -#undef HAVE_SNPRINTF -@end example +If @code{configure} fails, and you are going to send mail to +@email{mysql@@lists.mysql.com} to ask for assistance, please include any +lines from @file{config.log} that you think can help solve the problem. Also +include the last couple of lines of output from @code{configure} if +@code{configure} aborts. Post the bug report using the @code{mysqlbug} +script. @xref{Bug reports}. -And finally, in @file{mysqld.cc} you need to add a prototype for initgoups. +If the compile fails, see @ref{Compilation problems}, for help with +a number of common problems. + +@item +Install everything: @example -#ifdef _AIX41 -extern "C" int initgroups(const char *,int); -#endif +shell> make install @end example -@node HP-UX 10.20, HP-UX 11.x, IBM-AIX, Source install system issues -@subsection HP-UX Version 10.20 Notes +You might need to run this command as @code{root}. -There are a couple of small problems when compiling @strong{MySQL} on -HP-UX. We recommend that you use @code{gcc} instead of the HP-UX native -compiler, because @code{gcc} produces better code! +@item +Create the @strong{MySQL} grant tables (necessary only if you haven't +installed @strong{MySQL} before): -We recommend using gcc 2.95 on HP-UX. Don't use high optimization -flags (like -O6) as this may not be safe on HP-UX. +@example +shell> scripts/mysql_install_db +@end example -Note that MIT-pthreads can't be compiled with the HP-UX compiler -because it can't compile @code{.S} (assembler) files. +Note that @strong{MySQL} versions older than Version 3.22.10 started the +@strong{MySQL} server when you run @code{mysql_install_db}. This is no +longer true! -The following configure line should work: +@item +Change ownership of binaries to @code{root} and ownership of the data +directory to the user that you will run @code{mysqld} as: @example -CFLAGS="-DHPUX -I/opt/dce/include" CXXFLAGS="-DHPUX -I/opt/dce/include -felide-constructors -fno-exceptions -fno-rtti" CXX=gcc ./configure --with-pthread --with-named-thread-libs='-ldce' --prefix=/usr/local/mysql --disable-shared +shell> chown -R root /usr/local/mysql +shell> chown -R mysql /usr/local/mysql/var +shell> chgrp -R mysql /usr/local/mysql @end example -If you are compiling @code{gcc} 2.95 yourself, you should NOT link it with -the DCE libraries (@code{libdce.a} or @code{libcma.a}) if you want to compile -@strong{MySQL} with MIT-pthreads. If you mix the DCE and MIT-pthreads -packages you will get a @code{mysqld} to which you cannot connect. Remove -the DCE libraries while you compile @code{gcc} 2.95! +The first command changes the @code{owner} attribute of the files to the +@code{root} user, the second one changes the @code{owner} attribute of the +data directory to the @code{mysql} user, and the third one changes the +@code{group} attribute to the @code{mysql} group. -@node HP-UX 11.x, Mac OS X, HP-UX 10.20, Source install system issues -@subsection HP-UX Version 11.x Notes +@item +If you want to install support for the Perl @code{DBI}/@code{DBD} interface, +see @ref{Perl support}. -For HP-UX Version 11.x we recommend @strong{MySQL} Version 3.23.15 or later. +@item +If you would like @strong{MySQL} to start automatically when you boot your +machine, you can copy @code{support-files/mysql.server} to the location where +your system has its startup files. More information can be found in the +@code{support-files/mysql.server} script itself and in +@ref{Automatic start}. +@end enumerate -Because of some critical bugs in the standard HP-UX libraries, you should -install the following patches before trying to run @strong{MySQL} on HP-UX 11.0: +After everything has been installed, you should initialize and test your +distribution: @example -PHKL_22840 Streams cumulative -PHNE_22397 ARPA cumulative +shell> /usr/local/mysql/bin/safe_mysqld --user=mysql & @end example -This will solve a problem that one gets @code{EWOULDBLOCK} from @code{recv()} -and @code{EBADF} from @code{accept()} in threaded applications. +If that command fails immediately with @code{mysqld daemon ended} then you can +find some information in the file @file{mysql-data-directory/'hostname'.err}. +The likely reason is that you already have another @code{mysqld} server +running. @xref{Multiple servers}. -If you are using @code{gcc} 2.95.1 on an unpatched HP-UX 11.x system, -you will get the error: +@xref{Post-installation}. -@example -In file included from /usr/include/unistd.h:11, - from ../include/global.h:125, - from mysql_priv.h:15, - from item.cc:19: -/usr/include/sys/unistd.h:184: declaration of C function ... -/usr/include/sys/pthread.h:440: previous declaration ... -In file included from item.h:306, - from mysql_priv.h:158, - from item.cc:19: -@end example +@cindex patches, applying +@cindex applying, patches +@node Applying patches, configure options, Quick install, Installing source +@subsection Applying Patches -The problem is that HP-UX doesn't define @code{pthreads_atfork()} consistently. -It has conflicting prototypes in -@file{/usr/include/sys/unistd.h}:184 and -@file{/usr/include/sys/pthread.h}:440 (details below). +Sometimes patches appear on the mailing list or are placed in the +@uref{http://www.mysql.com/Downloads/Patches, patches area} of the +@strong{MySQL} Web site. -One solution is to copy @file{/usr/include/sys/unistd.h} into -@file{mysql/include} and edit @file{unistd.h} and change it to match -the definition in @file{pthread.h}. Here's the diff: +To apply a patch from the mailing list, save the message in which the patch +appears in a file, change into the top-level directory of your @strong{MySQL} +source tree, and run these commands: @example -183,184c183,184 -< extern int pthread_atfork(void (*prepare)(), void (*parent)(), -< void (*child)()); ---- -> extern int pthread_atfork(void (*prepare)(void), void (*parent)(void), -> void (*child)(void)); +shell> patch -p1 < patch-file-name +shell> rm config.cache +shell> make clean @end example -After this, the following configure line should work: +Patches from the FTP site are distributed as plain text files or as files +compressed with @code{gzip}. Apply a plain patch as shown above for +mailing list patches. To apply a compressed patch, change into the +top-level directory of your @strong{MySQL} source tree and run these +commands: @example -CFLAGS="-fomit-frame-pointer -O3 -fpic" CXX=gcc CXXFLAGS="-felide-constructors -fno-exceptions -fno-rtti -O3" ./configure --prefix=/usr/local/mysql --disable-shared +shell> gunzip < patch-file-name.gz | patch -p1 +shell> rm config.cache +shell> make clean @end example -Here is some information that a HP-UX Version 11.x user sent us about compiling -@strong{MySQL} with HP-UX:x compiler: +After applying a patch, follow the instructions for a normal source install, +beginning with the @code{./configure} step. After running the @code{make +install} step, restart your @strong{MySQL} server. -@example - Environment: - proper compilers. - setenv CC cc - setenv CXX aCC - flags - setenv CFLAGS -D_REENTRANT - setenv CXXFLAGS -D_REENTRANT - setenv CPPFLAGS -D_REENTRANT - % aCC -V - aCC: HP ANSI C++ B3910B X.03.14.06 - % cc -V /tmp/empty.c - cpp.ansi: HP92453-01 A.11.02.00 HP C Preprocessor (ANSI) - ccom: HP92453-01 A.11.01.00 HP C Compiler - cc: "/tmp/empty.c", line 1: warning 501: Empty source file. +You may need to bring down any currently running server before you run +@code{make install}. (Use @code{mysqladmin shutdown} to do this.) Some +systems do not allow you to install a new version of a program if it replaces +the version that is currently executing. - configuration: - ./configure --with-pthread \ - --prefix=/source-control/mysql \ - --with-named-thread-libs=-lpthread \ - --with-low-memory +@cindex @code{configure} script +@cindex options, configure +@cindex configuration options +@findex without-server option +@cindex log files +@cindex files, log +@node configure options, , Applying patches, Installing source +@subsection Typical @code{configure} Options - added '#define _CTYPE_INCLUDED' to include/m_ctype.h. This - symbol is the one defined in HP's /usr/include/ctype.h: +The @code{configure} script gives you a great deal of control over how +you configure your @strong{MySQL} distribution. Typically you do this +using options on the @code{configure} command line. You can also affect +@code{configure} using certain environment variables. @xref{Environment +variables}. For a list of options supported by @code{configure}, run +this command: - /* Don't include std ctype.h when this is included */ - #define _CTYPE_H - #define __CTYPE_INCLUDED - #define _CTYPE_INCLUDED - #define _CTYPE_USING /* Don't put names in global namespace. */ +@example +shell> ./configure --help @end example +Some of the more commonly-used @code{configure} options are described below: + @itemize @bullet @item -I had to use the compile-time flag @code{-D_REENTRANT} to get the compiler -to recognize the prototype for @code{localtime_r}. Alternatively I could have -supplied the prototype for @code{localtime_r}. But I wanted to catch other -bugs without needing to run into them. I wasn't sure where I needed it, so I -added it to all flags. -@item -The optimization flags used by @strong{MySQL} (-O3) are not recognized by HP's -compilers. I did not change the flags. -@end itemize - -If you get the following error from @code{configure} +To compile just the @strong{MySQL} client libraries and client programs and +not the server, use the @code{--without-server} option: @example -checking for cc option to accept ANSI C... no -configure: error: MySQL requires a ANSI C compiler (and a C++ compiler). Try gcc. See the Installation chapter in the Reference Manual. +shell> ./configure --without-server @end example -Check that you don't have the path to the K&R compiler before the path -to the HP-UX C and C++ compiler. - -@node Mac OS X, BEOS, HP-UX 11.x, Source install system issues -@subsection Mac OS X Notes - -@menu -* Mac OS X Public Data:: -* Mac OS X Server:: -@end menu +@c FIX incorrect.. -@node Mac OS X Public Data, Mac OS X Server, Mac OS X, Mac OS X -@subsubsection Mac OS X Public beta +If you don't have a C++ compiler, @code{mysql} will not compile (it is the +one client program that requires C++). In this case, +you can remove the code in @code{configure} that tests for the C++ compiler +and then run @code{./configure} with the @code{--without-server} option. The +compile step will still try to build @code{mysql}, but you can ignore any +warnings about @file{mysql.cc}. (If @code{make} stops, try @code{make -k} +to tell it to continue with the rest of the build even if errors occur.) -@strong{MySQL} should work without any problems on Mac OS X Public Beta -(Darwin). You don't need the pthread patches for this OS! +@item +If you don't want your log files and database directories located under +@file{/usr/local/var}, use a @code{configure} command, something like one +of these: -@node Mac OS X Server, , Mac OS X Public Data, Mac OS X -@subsubsection Mac OS X Server +@example +shell> ./configure --prefix=/usr/local/mysql +shell> ./configure --prefix=/usr/local \ + --localstatedir=/usr/local/mysql/data +@end example -Before trying to configure @strong{MySQL} on Mac OS X server you must -first install the pthread package from -@uref{http://www.prnet.de/RegEx/mysql.html}. +The first command changes the installation prefix so that everything is +installed under @file{/usr/local/mysql} rather than the default of +@file{/usr/local}. The second command preserves the default installation +prefix, but overrides the default location for database directories +(normally @file{/usr/local/var}) and changes it to +@code{/usr/local/mysql/data}. -Our binary for Mac OS X is compiled on Rhapsody 5.5 with the following -configure line: +@cindex changing socket location +@cindex socket location, changing +@cindex mysql.sock, changing location of +@item +If you are using Unix and you want the @strong{MySQL} socket located somewhere +other than the default location (normally in the directory @file{/tmp} or +@file{/var/run}) use a @code{configure} command like this: @example -CC=gcc CFLAGS="-O2 -fomit-frame-pointer" CXX=gcc CXXFLAGS="-O2 -fomit-frame-pointer" ./configure --prefix=/usr/local/mysql "--with-comment=Official MySQL binary" --with-extra-charsets=complex --disable-shared +shell> ./configure --with-unix-socket-path=/usr/local/mysql/tmp/mysql.sock @end example -You might want to also add aliases to your shell's resource file to -access @code{mysql} and @code{mysqladmin} from the command line: +Note that the given file must be an absolute pathname! + +@cindex compiling statically +@cindex statically, compiling +@item +If you want to compile statically linked programs (for example, to make a +binary distribution, to get more speed, or to work around problems with some +RedHat Linux distributions), run @code{configure} like this: @example -alias mysql '/usr/local/mysql/bin/mysql' -alias mysqladmin '/usr/local/mysql/bin/mysqladmin' +shell> ./configure --with-client-ldflags=-all-static \ + --with-mysqld-ldflags=-all-static @end example -@node BEOS, , Mac OS X, Source install system issues -@subsection BeOS Notes - -We are really interested in getting @strong{MySQL} to work on BeOS, but -unfortunately we don't have any person who knows BeOS or has time to do -a port. +@tindex CC environment variable +@tindex environment variable, CC +@tindex CXX environment variable +@tindex environment variable, CXX +@cindex gcc +@cindex C++ complier, gcc +@cindex compiler, C++ gcc +@item +If you are using @code{gcc} and don't have @code{libg++} or @code{libstdc++} +installed, you can tell @code{configure} to use @code{gcc} as your C++ +compiler: -We are interested in finding someone to do a port, and we will help them -with any technical questions they may have while doing the port. +@example +shell> CC=gcc CXX=gcc ./configure +@end example -We have previously talked with some BeOS developers that have said that -@strong{MySQL} is 80% ported to BeOS, but we haven't heard from them -in a while. +When you use @code{gcc} as your C++ compiler, it will not attempt to link in +@code{libg++} or @code{libstdc++}. -@node Windows, OS/2, Source install system issues, Installing -@section Windows Notes +Here is some common environment variables to set depending on +the compiler you are using: -This section describes installation and use of @strong{MySQL} on Windows. -This information is also provided in the @file{README} file that comes -with the @strong{MySQL} Windows distribution. +@tindex CXXFLAGS environment variable +@tindex environment variable, CXXFLAGS +@multitable @columnfractions .20 .80 +@item gcc 2.7.2.1 @tab +CC=gcc CXX=gcc CXXFLAGS="-O3 -felide-constructors" +@item egcs 1.0.3a @tab +CC=gcc CXX=gcc CXXFLAGS="-O3 -felide-constructors -fno-exceptions -fno-rtti" +@item gcc 2.95.2 @tab +CFLAGS="-O3 -mpentiumpro" CXX=gcc CXXFLAGS="-O3 -mpentiumpro -felide-constructors -fno-exceptions -fno-rtti" +@item pgcc 2.90.29 or newer @tab +CFLAGS="-O3 -mpentiumpro -mstack-align-double" CXX=gcc CXXFLAGS="-O3 -mpentiumpro -mstack-align-double -felide-constructors -fno-exceptions -fno-rtti" +@end multitable -@menu -* Windows installation:: Installing @strong{MySQL} on Windows -* Win95 start:: Starting @strong{MySQL} on Win95 / Win98 -* NT start:: Starting @strong{MySQL} on NT / Win2000 -* Windows running:: Running @strong{MySQL} on Windows -* Windows and SSH:: Connecting to a remote @strong{MySQL} from Windows with SSH -* Windows symbolic links:: Splitting data across different disks under Win32 -* Windows compiling:: Compiling MySQL clients on Windows. -* Windows vs Unix:: @strong{MySQL}-Windows compared to Unix @strong{MySQL} -@end menu +In most cases you can get a reasonably optimal @strong{MySQL} binary by +using the options from the above and adding the following options to the +configure line: -@node Windows installation, Win95 start, Windows, Windows -@subsection Installing MySQL on Windows +@example +--prefix=/usr/local/mysql --enable-assembler --with-mysqld-ldflags=-all-static +@end example -The following instructions apply to precompiled binary distributions. -If you download a source distribution, you will have to compile and install -it yourself. +The full configure line would in other words be something like the +following for all recent gcc versions: -If you don't have a copy of the @strong{MySQL} distribution, you should -first download one from @uref{http://www.mysql.com/downloads/mysql-3.23.html}. +@example +CFLAGS="-O3 -mpentiumpro" CXX=gcc CXXFLAGS="-O3 -mpentiumpro -felide-constructors -fno-exceptions -fno-rtti" ./configure --prefix=/usr/local/mysql --enable-assembler --with-mysqld-ldflags=-all-static +@end example -If you plan to connect to @strong{MySQL} from some other program, you will -probably also need the @strong{MyODBC} driver. You can find this at the -@strong{MyODBC} download page -(@uref{http://www.mysql.com/downloads/api-myodbc.html}). +The binaries we provide on the @strong{MySQL} Web site at +@uref{http://www.mysql.com} are all compiled with full optimization and +should be perfect for most users. @xref{MySQL binaries}. There are some +things you can tweak to make an even faster binary, but this is only for +advanced users. @xref{Compile and link options}. -To install either distribution, unzip it in some empty directory and run the -@code{Setup.exe} program. +If the build fails and produces errors about your compiler or linker not +being able to create the shared library @file{libmysqlclient.so.#} (@samp{#} +is a version number), you can work around this problem by giving the +@code{--disable-shared} option to @code{configure}. In this case, +@code{configure} will not build a shared @code{libmysqlclient.so.#} library. -By default, @strong{MySQL}-Windows is configured to be installed in -@file{C:\mysql}. If you want to install @strong{MySQL} elsewhere, -install it in @file{C:\mysql} first, then move the installation to -where you want it. If you do move @strong{MySQL}, you must indicate -where everything is located by supplying a @code{--basedir} option when -you start the server. For example, if you have moved the @strong{MySQL} -distribution to @file{D:\programs\mysql}, you must start @code{mysqld} -like this: +@cindex default values, suppression +@cindex suppression, default values +@item +You can configure @strong{MySQL} not to use @code{DEFAULT} column values for +non-@code{NULL} columns (that is, columns that are not allowed to be +@code{NULL}). This causes @code{INSERT} statements to generate an error +unless you explicitly specify values for all columns that require a +non-@code{NULL} value. To suppress use of default values, run +@code{configure} like this: +@tindex CXXFLAGS environment variable +@tindex environment variable, CXXFLAGS @example -C:\> D:\programs\mysql\bin\mysqld --basedir D:\programs\mysql +shell> CXXFLAGS=-DDONT_USE_DEFAULT_FIELDS ./configure @end example -Use @code{mysqld --help} to display all the options that @code{mysqld} -understands! - -With all newer @strong{MySQL} versions, you can also create a -@file{C:\my.cnf} file that holds any default options for the -@strong{MySQL} server. Copy the file @file{\mysql\my-xxxxx.cnf} to -@file{C:\my.cnf} and edit it to suit your setup. Note that you should -specify all paths with @samp{/} instead of @samp{\}. If you use -@samp{\}, you need to specify it twice, because @samp{\} is the escape -character in @strong{MySQL}. @xref{Option files}. +@cindex character sets +@findex configure option, --with-charset +@findex configure option, --with-extra-charset +@item +By default, @strong{MySQL} uses the ISO-8859-1 (Latin1) character set. To +change the default set, use the @code{--with-charset} option: +@example +shell> ./configure --with-charset=CHARSET +@end example +@code{CHARSET} may be one of @code{big5}, @code{cp1251}, @code{cp1257}, +@code{czech}, @code{danish}, @code{dec8}, @code{dos}, @code{euc_kr}, +@code{gb2312}, @code{gbk}, @code{german1}, @code{hebrew}, @code{hp8}, +@code{hungarian}, @code{koi8_ru}, @code{koi8_ukr}, @code{latin1}, +@code{latin2}, @code{sjis}, @code{swe7}, @code{tis620}, @code{ujis}, +@code{usa7}, or @code{win1251ukr}. +@xref{Character sets}. -Starting with @strong{MySQL} 3.23.38, the Windows distribution includes -both the normal and the @strong{MySQL-Max} binaries. The main benefit -of using the normal @code{mysqld.exe} binary is that it's a little -faster and uses less resources. +If you want to convert characters between the server and the client, +you should take a look at the @code{SET OPTION CHARACTER SET} command. +@xref{SET OPTION, , @code{SET OPTION}}. -Here is a list of the different @strong{MySQL} servers you can use: +@cindex @code{myisamchk} +@strong{Warning:} If you change character sets after having created any +tables, you will have to run @code{myisamchk -r -q} on every table. Your +indexes may be sorted incorrectly otherwise. (This can happen if you +install @strong{MySQL}, create some tables, then reconfigure +@strong{MySQL} to use a different character set and reinstall it.) -@multitable @columnfractions .25 .75 -@item @code{mysqld} @tab -Compiled with full debugging and automatic memory allocation checking, -symbolic links, BDB and InnoDB tables. -@item @code{mysqld-opt} @tab -Optimized binary with no support for transactional tables. -@item @code{mysqld-nt} @tab -Optimized binary for NT with support for named pipes. You can run this -version on Win98, but in this case no named pipes are created and you must -have TCP/IP installed. -@item @code{mysqld-max} @tab -Optimized binary with support for symbolic links, BDB and InnoDB tables. -@item @code{mysqld-max-nt} @tab -Like @code{mysqld-max}, but compiled with support for named pipes. -@end multitable +With the option @code{--with-extra-charset=LIST} you can define +which additional character sets should be incompiled in the server. -All of the above binaries are optimized for the Pentium Pro processor but -should work on any Intel processor >= i386. +Here @code{LIST} is either a list of character set separated with space, +@code{complex} to include all characters that can't be dynamically loaded +or @code{all} to include all character sets into the binaries. -NOTE: If you want to use InnoDB tables, there are certain startup -options that must be specified in your @file{my.ini} file! @xref{InnoDB start}. +@item +To configure @strong{MySQL} with debugging code, use the @code{--with-debug} +option: +@example +shell> ./configure --with-debug +@end example +This causes a safe memory allocator to be included that can find some errors +and that provides output about what is happening. +@xref{Debugging server}. -@node Win95 start, NT start, Windows installation, Windows -@subsection Starting MySQL on Windows 95 or Windows 98 +@item +If your client programs are using threads, you need to also compile a +thread-safe version of the @strong{MySQL} client library with the +@code{--with-thread-safe-client} configure options. This will create a +@code{libmysqlclient_r} library with which you should link your threaded +applications. @xref{Thread-safe clients}. -@strong{MySQL} uses TCP/IP to connect a client to a server. (This will -allow any machine on your network to connect to your @strong{MySQL} -server.) Because of this, you must install TCP/IP on your machine before -starting @strong{MySQL}. You can find TCP/IP on your Windows CD-ROM. +@item +Options that pertain to particular systems can be found in the +system-specific sections later in this chapter. @xref{Source install +system issues}. +@end itemize -Note that if you are using an old Win95 release (for example OSR2), it's -likely that you have an old Winsock package! @strong{MySQL} requires -Winsock 2! You can get the newest Winsock from -@uref{http://www.microsoft.com/}. Win98 has the new Winsock 2 library, so -the above doesn't apply for Win98. +@node Installing source tree, Compilation problems, Installing source, Installing +@section Installing from the Development Source Tree -To start the @code{mysqld} server, you should start an MS-DOS window and type: +@cindex development source tree +@cindex BitKeeper tree +@cindex cvs tree -@example -C:\> C:\mysql\bin\mysqld -@end example +@strong{CAUTION:} You should read this section only if you are interested +in helping us test our new code. If you just want to get @strong{MySQL} up +and running on your system, you should use a standard release distribution +(either a source or binary distribution will do). -This will start @code{mysqld} in the background without a window. +To obtain our most recent development source tree, use these instructions: -You can kill the @strong{MySQL} server by executing: +@enumerate +@item +Download @strong{BitKeeper} from +@uref{http://www.bitmover.com/cgi-bin/download.cgi}. You will need +@strong{Bitkeeper} 2.0 or newer to access our repository. + +@item +Follow the instructions to install it. + +@item +After @strong{BitKeeper} is installed, use this command if you want to clone +the @strong{MySQL} 3.23 branch: @example -C:\> C:\mysql\bin\mysqladmin -u root shutdown +shell> bk clone bk://work.mysql.com:7000 mysql @end example -Note that Win95 and Win98 don't support creation of named pipes. -On Win95 and Win98, you can only use named pipes to connect to a -remote @strong{MySQL} server running on a Windows NT server host. -(The @strong{MySQL} server must also support named pipes, of -course. For example, using @code{mysqld-opt} under NT will not allow -named pipe connections. You should use either @code{mysqld-nt} or -@code{mysqld-max-nt}.) - -If @code{mysqld} doesn't start, please check the -@file{\mysql\data\mysql.err} file to see if the server wrote any message -there to indicate the cause of the problem. You can also try to start -the server with @code{mysqld --standalone}; In this case, you may get -some useful information on the screen that may help solve the problem. +To clone the 4.0 branch, use this command instead: -The last option is to start @code{mysqld} with @code{--standalone ---debug}. In this case @code{mysqld} will write a log file -@file{C:\mysqld.trace} that should contain the reason why @code{mysqld} -doesn't start. @xref{Making trace files}. +@example +shell> bk clone bk://work.mysql.com:7001 mysql-4.0 +@end example -@node NT start, Windows running, Win95 start, Windows -@subsection Starting MySQL on Windows NT or Windows 2000 +The initial download of the source tree may take a while, depending on the +speed of your connection; be patient. -The Win95/Win98 section also applies to @strong{MySQL} on NT/Win2000, with -the following differences: +@item +You will need GNU @code{autoconf}, @code{automake}, @code{libtool}, and +@code{m4} to run the next set of commands. +If you get some strange error during this stage, check that you really +have @code{libtool} installed! -To get @strong{MySQL} to work with TCP/IP on NT, you must install -service pack 3 (or newer)! +@example +shell> cd mysql +shell> bk -r edit +shell> aclocal; autoheader; autoconf; automake; +shell> ./configure # Add your favorite options here +shell> make +@end example -Note that everything in the following that applies for NT also applies -for Win2000! +A collection of our standard configure scripts is located in the +@file{BUILD/} subdirectory. If you are lazy, you can use +@file{BUILD/compile-pentium-debug}. To compile on a different architecture, +modify the script removing flags that are Pentium-specific. -For NT/Win2000, the server name is @code{mysqld-nt}. Normally you -should install @strong{MySQL} as a service on NT/Win2000: +@item +When the build is done, run @code{make install}. Be careful with this +on a production machine; the command may overwrite your live release +installation. If you have another installation of @strong{MySQL}, we +recommand that you run @code{./configure} with different values for the +@code{prefix}, @code{tcp-port}, and @code{unix-socket-path} options than +those used for your production server. -@example -C:\> C:\mysql\bin\mysqld-nt --install -@end example +@item +Play hard with your new installation and try to make the new features +crash. Start by running @code{make test}. @xref{MySQL test suite}. -or +@item +If you have gotten to the @code{make} stage and the distribution does +not compile, please report it to @email{bugs@@lists.mysql.com}. If you +have installed the latest versions of the required GNU tools, and they +crash trying to process our configuration files, please report that also. +However, if you execute @code{aclocal} and get a @code{command not found} +error or a similar problem, do not report it. Instead, make sure all +the necessary tools are installed and that your @code{PATH} variable is +set correctly so your shell can find them. -@example -C:\> C:\mysql\bin\mysqld-max-nt --install -@end example +@item +After the initial @code{bk clone} operation to get the source tree, you +should run @code{bk pull} periodically to get the updates. -(Under Windows NT, you can actually install any of the server binaries -as a service, but only those having names that end with @code{-nt.exe} -provide support for named pipes.) +@item +You can examine the change history for the tree with all the diffs by using +@code{bk sccstool}. If you see some funny diffs or code that you have a +question about, do not hesitate to send e-mail to +@email{internals@@lists.mysql.com}. Also, if you think you have a better idea +on how to do something, send an email to the same address with a patch. +@code{bk diffs} will produce a patch for you after you have made changes +to the source. If you do not have the time to code your idea, just send +a description. -You can start and stop the @strong{MySQL} service with these commands: +@item +@strong{BitKeeper} has a nice help utility that you can access via +@code{bk helptool}. -@example -C:\> NET START mysql -C:\> NET STOP mysql -@end example +@end enumerate -Note that in this case you can't use any other options for @code{mysqld-nt}! +@cindex compiling, problems +@cindex problems, compiling +@cindex reconfiguring +@cindex @code{config.cache} file +@cindex files, @code{config.cache} +@node Compilation problems, MIT-pthreads, Installing source tree, Installing +@section Problems Compiling? -You can also run @code{mysqld-nt} as a stand-alone program on NT if you need -to start @code{mysqld-nt} with any options! If you start @code{mysqld-nt} -without options on NT, @code{mysqld-nt} tries to start itself as a service -with the default service options. If you have stopped @code{mysqld-nt}, you -have to start it with @code{NET START mysql}. +All @strong{MySQL} programs compile cleanly for us with no warnings on +Solaris using @code{gcc}. On other systems, warnings may occur due to +differences in system include files. See @ref{MIT-pthreads} for warnings +that may occur when using MIT-pthreads. For other problems, check the list +below. -The service is installed with the name @code{MySQL}. Once installed, it must -be started using the Services Control Manager (SCM) Utility found in the -Control Panel, or by using the @code{NET START MySQL} command. If any options -are desired, they must be specified as ``Startup parameters'' in the SCM utility -before you start the @strong{MySQL} service. Once running, @code{mysqld-nt} -can be stopped using @code{mysqladmin}, or from the SCM utility or by using -the command @code{NET STOP MySQL}. If you use SCM to stop @code{mysqld-nt}, -there is a strange message from SCM about @code{mysqld shutdown normally}. -When run as a service, @code{mysqld-nt} has no access to a console and so no -messages can be seen. +The solution to many problems involves reconfiguring. If you do need to +reconfigure, take note of the following: -On NT you can get the following service error messages: +@cindex running @code{configure} after prior invocation +@cindex @code{configure}, running after prior invocation +@cindex reconfiguring +@cindex config.cache +@itemize @bullet +@item +If @code{configure} is run after it already has been run, it may use +information that was gathered during its previous invocation. This +information is stored in @file{config.cache}. When @code{configure} starts +up, it looks for that file and reads its contents if it exists, on the +assumption that the information is still correct. That assumption is invalid +when you reconfigure. -@multitable @columnfractions .3 .7 -@item Permission Denied @tab Means that it cannot find @code{mysqld-nt.exe}. -@item Cannot Register @tab Means that the path is incorrect. -@item Failed to install service. @tab Means that the service is already installed or that the Service Control Manager is in bad state. -@end multitable +@item +Each time you run @code{configure}, you must run @code{make} again +to recompile. However, you may want to remove old object files from previous +builds first, because they were compiled using different configuration options. +@end itemize -If you have problems installing @code{mysqld-nt} as a service, try starting -it with the full path: +To prevent old configuration information or object files from being used, +run these commands before rerunning @code{configure}: @example -C:\> C:\mysql\bin\mysqld-nt --install +shell> rm config.cache +shell> make clean @end example -If this doesn't work, you can get @code{mysqld-nt} to start properly by fixing -the path in the registry! +Alternatively, you can run @code{make distclean}. -If you don't want to start @code{mysqld-nt} as a service, you can start it as -follows: +The list below describes some of the problems compiling @strong{MySQL} +that have been found to occur most often: + +@itemize @bullet +@item +@cindex @code{cc1plus} problems +@cindex @code{fatal signal 11} +@cindex @code{sql_yacc.cc} problems +@cindex internal compiler errors +@cindex virtual memory, problems while compiling +@cindex configure option, --with-low-memory +If you get errors when compiling @file{sql_yacc.cc}, such as the ones shown +below, you have probably run out of memory or swap space: @example -C:\> C:\mysql\bin\mysqld-nt --standalone +Internal compiler error: program cc1plus got fatal signal 11 + or +Out of virtual memory + or +Virtual memory exhausted @end example -or +The problem is that @code{gcc} requires huge amounts of memory to compile +@file{sql_yacc.cc} with inline functions. Try running @code{configure} with +the @code{--with-low-memory} option: @example -C:\> C:\mysql\bin\mysqld --standalone --debug +shell> ./configure --with-low-memory @end example -The last version gives you a debug trace in @file{C:\mysqld.trace}. -@xref{Making trace files}. - -@node Windows running, Windows and SSH, NT start, Windows -@subsection Running MySQL on Windows - -@cindex TCP/IP -@cindex named pipes +This option causes @code{-fno-inline} to be added to the compile line if you +are using @code{gcc} and @code{-O0} if you are using something else. You +should try the @code{--with-low-memory} option even if you have so much +memory and swap space that you think you can't possibly have run out. This +problem has been observed to occur even on systems with generous hardware +configurations, and the @code{--with-low-memory} option usually fixes it. -@strong{MySQL} supports TCP/IP on all Windows platforms and named pipes on NT. -The default is to use named pipes for local connections on NT and TCP/IP for -all other cases if the client has TCP/IP installed. The host name specifies -which protocol is used: +@item +By default, @code{configure} picks @code{c++} as the compiler name and +GNU @code{c++} links with @code{-lg++}. If you are using @code{gcc}, +that behavior can cause problems during configuration such as this: -@multitable @columnfractions .3 .7 -@strong{Host name} @tab @strong{Protocol} -@item NULL (none) @tab On NT, try named pipes first; if that doesn't work, use TCP/IP. On Win95/Win98, TCP/IP is used. -@item . @tab Named pipes -@item localhost @tab TCP/IP to current host -@item hostname @tab TCP/IP -@end multitable +@cindex C++ compiler cannot create executables +@example +configure: error: installation or configuration problem: +C++ compiler cannot create executables. +@end example -You can force a @strong{MySQL} client to use named pipes by specifying the -@code{--pipe} option or by specifying @code{.} as the host name. Use the -@code{--socket} option to specify the name of the pipe. +@tindex CXX environment variable +@tindex Environment variable, CXX +You might also observe problems during compilation related to +@code{g++}, @code{libg++}, or @code{libstdc++}. -You can test whether or not @strong{MySQL} is working by executing the -following commands: +One cause of these problems is that you may not have @code{g++}, or you may +have @code{g++} but not @code{libg++}, or @code{libstdc++}. Take a look at +the @file{config.log} file. It should contain the exact reason why your c++ +compiler didn't work! To work around these problems, you can use @code{gcc} +as your C++ compiler. Try setting the environment variable @code{CXX} to +@code{"gcc -O3"}. For example: +@tindex CXX environment variable +@tindex Environment variables, CXX @example -C:\> C:\mysql\bin\mysqlshow -C:\> C:\mysql\bin\mysqlshow -u root mysql -C:\> C:\mysql\bin\mysqladmin version status proc -C:\> C:\mysql\bin\mysql test +shell> CXX="gcc -O3" ./configure @end example -If @code{mysqld} is slow to answer to connections on Win95/Win98, there is -probably a problem with your DNS. In this case, start @code{mysqld} with -@code{--skip-name-resolve} and use only @code{localhost} and IP numbers in -the @strong{MySQL} grant tables. You can also avoid DNS when connecting to a -@code{mysqld-nt} @strong{MySQL} server running on NT by using the -@code{--pipe} argument to specify use of named pipes. This works for most -@strong{MySQL} clients. - -There are two versions of the @strong{MySQL} command-line tool: -@multitable @columnfractions .25 .75 -@item @code{mysql} @tab Compiled on native Windows, which offers very limited text editing capabilities. -@item @code{mysqlc} @tab Compiled with the Cygnus GNU compiler and libraries, which offers @code{readline} editing. -@end multitable - -If you want to use @code{mysqlc.exe}, you must copy -@file{C:\mysql\lib\cygwinb19.dll} to your Windows system directory -(@file{\windows\system} or similar place). +This works because @code{gcc} compiles C++ sources as well as @code{g++} +does, but does not link in @code{libg++} or @code{libstdc++} by default. -The default privileges on Windows give all local users full privileges -to all databases without specifying a password. To make @strong{MySQL} -more secure, you should set a password for all users and remove the row in -the @code{mysql.user} table that has @code{Host='localhost'} and -@code{User=''}. +Another way to fix these problems, of course, is to install @code{g++}, +@code{libg++} and @code{libstdc++}. -You should also add a password for the @code{root} user. The following -example starts by removing the anonymous user that can be used by anyone -to access the @code{test} database, then sets a @code{root} user password: +@item +If your compile fails with errors, such as any of the following, +you must upgrade your version of @code{make} to GNU @code{make}: @example -C:\> C:\mysql\bin\mysql mysql -mysql> DELETE FROM user WHERE Host='localhost' AND User=''; -mysql> QUIT -C:\> C:\mysql\bin\mysqladmin reload -C:\> C:\mysql\bin\mysqladmin -u root password your_password +making all in mit-pthreads +make: Fatal error in reader: Makefile, line 18: +Badly formed macro assignment + or +make: file `Makefile' line 18: Must be a separator (: + or +pthread.h: No such file or directory @end example -After you've set the password, if you want to take down the @code{mysqld} -server, you can do so using this command: +@cindex Solaris troubleshooting +@cindex FreeBSD troubleshooting +@cindex troubleshooting, Solaris +@cindex troubleshooting, FreeBSD +Solaris and FreeBSD are known to have troublesome @code{make} programs. + +GNU @code{make} Version 3.75 is known to work. + +@tindex CC environment variable +@tindex environment variable, CC +@tindex CXX environment variable +@tindex environment variable, CXX +@tindex CFLAGS environment variable +@tindex environment variable, CFLAGS +@tindex CXXFLAGS environment variable +@tindex environment variable, CXXFLAGS +@item +If you want to define flags to be used by your C or C++ compilers, do so by +adding the flags to the @code{CFLAGS} and @code{CXXFLAGS} environment +variables. You can also specify the compiler names this way using @code{CC} +and @code{CXX}. For example: @example -C:\> mysqladmin --user=root --password=your_password shutdown +shell> CC=gcc +shell> CFLAGS=-O3 +shell> CXX=gcc +shell> CXXFLAGS=-O3 +shell> export CC CFLAGS CXX CXXFLAGS @end example -If you are using the old shareware version of @strong{MySQL} Version -3.21 under Windows, the above command will fail with an error: -@code{parse error near 'SET OPTION password'}. The fix is in to upgrade -to the current @strong{MySQL} version, which is freely available. +See @ref{MySQL binaries}, for a list of flag definitions that have been found +to be useful on various systems. -With the current @strong{MySQL} versions you can easily add new users -and change privileges with @code{GRANT} and @code{REVOKE} commands. -@xref{GRANT}. +@item +If you get an error message like this, +you need to upgrade your @code{gcc} compiler: -@c FIX this is ugly, real ugly. -@cindex SSH -@cindex connecting, remotely with SSH -@node Windows and SSH, Windows symbolic links, Windows running, Windows -@subsection Connecting to a Remote MySQL from Windows with SSH +@example +client/libmysql.c:273: parse error before `__attribute__' +@end example -Here is a note about how to connect to get a secure connection to remote -@strong{MySQL} server with SSH (by David Carlson @email{dcarlson@@mplcomm.com}): +@code{gcc} 2.8.1 is known to work, but we recommend using @code{gcc} 2.95.2 or +@code{egcs} 1.0.3a instead. -@itemize @bullet @item -Install an SSH client on your Windows machine. As a user, the best non-free -one I've found is from @code{SecureCRT} from @uref{http://www.vandyke.com/}. -Another option is @code{f-secure} from @uref{http://www.f-secure.com/}. You -can also find some free ones on @strong{Google} at -@uref{http://directory.google.com/Top/Computers/Security/Products_and_Tools/Cryptography/SSH/Clients/Windows/}. +If you get errors such as those shown below when compiling @code{mysqld}, +@code{configure} didn't correctly detect the type of the last argument to +@code{accept()}, @code{getsockname()}, or @code{getpeername()}: -@item -Start your Windows SSH client. -Set @code{Host_Name = yourmysqlserver_URL_or_IP}. -Set @code{userid=your_userid} to log in to your server (probably not the same -as your @strong{MySQL} login/password. +@example +cxx: Error: mysqld.cc, line 645: In this statement, the referenced + type of the pointer value "&length" is "unsigned long", which + is not compatible with "int". +new_sock = accept(sock, (struct sockaddr *)&cAddr, &length); +@end example -@item -Set up port forwarding. Either do a remote forward (Set @code{local_port: 3306}, @code{remote_host: yourmysqlservername_or_ip}, @code{remote_port: 3306} ) -or a local forward (Set @code{port: 3306}, @code{host: localhost}, @code{remote port: 3306}). +To fix this, edit the @file{config.h} file (which is generated by +@code{configure}). Look for these lines: -@item -Save everything, otherwise you'll have to redo it the next time. +@example +/* Define as the base type of the last arg to accept */ +#define SOCKET_SIZE_TYPE XXX +@end example -@item -Log in to your server with SSH session you just created. +Change @code{XXX} to @code{size_t} or @code{int}, depending on your +operating system. (Note that you will have to do this each time you run +@code{configure}, because @code{configure} regenerates @file{config.h}.) @item -On your Windows machine, start some ODBC application (such as Access). +The @file{sql_yacc.cc} file is generated from @file{sql_yacc.yy}. Normally +the build process doesn't need to create @file{sql_yacc.cc}, because +@strong{MySQL} comes with an already-generated copy. However, if you do need +to re-create it, you might encounter this error: + +@example +"sql_yacc.yy", line xxx fatal: default action causes potential... +@end example + +This is a sign that your version of @code{yacc} is deficient. +You probably need to install @code{bison} (the GNU version of @code{yacc}) +and use that instead. @item -Create a new file in Windows and link to @strong{MySQL} using the ODBC -driver the same way you normally do, EXCEPT type in @code{localhost} -for the @strong{MySQL} host server --- not @code{yourmysqlservername}. +If you need to debug @code{mysqld} or a @strong{MySQL} client, run +@code{configure} with the @code{--with-debug} option, then recompile and +link your clients with the new client library. @xref{Debugging client}. @end itemize -You should now have an ODBC connection to @strong{MySQL}, encrypted using SSH. - -@cindex symbolic links -@cindex using multiple disks to start data -@cindex disks, splitting data across -@node Windows symbolic links, Windows compiling, Windows and SSH, Windows -@subsection Splitting Data Across Different Disks on Windows +@cindex MIT-pthreads +@cindex thread support, non-native +@node MIT-pthreads, Perl support, Compilation problems, Installing +@section MIT-pthreads Notes -Beginning with @strong{MySQL} Version 3.23.16, the @code{mysqld-max} -and @code{mysql-max-nt} servers in the @strong{MySQL} distribution are -compiled with the @code{-DUSE_SYMDIR} option. This allows you to put a -database on different disk by adding a symbolic link to it -(in a manner similar to the way that symbolic links work on Unix). +This section describes some of the issues involved in using MIT-pthreads. -On Windows, you make a symbolic link to a database by creating a file -that contains the path to the destination directory and saving this in -the @file{mysql_data} directory under the filename @file{database.sym}. -Note that the symbolic link will be used only if the directory -@file{mysql_data_dir\database} doesn't exist. +Note that on Linux you should NOT use MIT-pthreads but install LinuxThreads! +@xref{Linux}. -For example, if the @strong{MySQL} data directory is @file{C:\mysql\data} -and you want to have database @code{foo} located at @file{D:\data\foo}, you -should create the file @file{C:\mysql\data\foo.sym} that contains the -text @code{D:\data\foo\}. After that, all tables created in the database -@code{foo} will be created in @file{D:\data\foo}. +If your system does not provide native thread support, you will need to +build @strong{MySQL} using the MIT-pthreads package. This includes +older FreeBSD systems, SunOS 4.x, Solaris 2.4 and earlier, and some others. +@xref{Which OS}. -Note that because of the speed penalty you get when opening every table, -we have not enabled this by default even if you have compiled -@strong{MySQL} with support for this. To enable symlinks you should put -in your @code{my.cnf} or @code{my.ini} file the following entry: +@itemize @bullet +@item +On most systems, you can force MIT-pthreads to be used by running +@code{configure} with the @code{--with-mit-threads} option: @example -[mysqld] -use-symbolic-links +shell> ./configure --with-mit-threads @end example -In @strong{MySQL} 4.0 we will enable symlinks by default. Then you -should instead use the @code{skip-symlink} option if you want to -disable this. +Building in a non-source directory is not supported when using +MIT-pthreads, because we want to minimize our changes to this code. -@cindex compiling, on Windows -@cindex Windows, compiling on -@node Windows compiling, Windows vs Unix, Windows symbolic links, Windows -@subsection Compiling MySQL Clients on Windows +@item +The checks that determine whether or not to use MIT-pthreads occur only +during the part of the configuration process that deals with the server +code. If you have configured the distribution using @code{--without-server} +to build only the client code, clients will not know whether or not +MIT-pthreads is being used and will use Unix socket connections by default. +Because Unix sockets do not work under MIT-pthreads, this means you will need +to use @code{-h} or @code{--host} when you run client programs. -In your source files, you should include @file{windows.h} before you include -@file{mysql.h}: +@item +When @strong{MySQL} is compiled using MIT-pthreads, system locking is +disabled by default for performance reasons. You can tell the server to use +system locking with the @code{--use-locking} option. + +@item +Sometimes the pthread @code{bind()} command fails to bind to a socket without +any error message (at least on Solaris). The result is that all connections +to the server fail. For example: @example -#if defined(_WIN32) || defined(_WIN64) -#include <windows.h> -#endif -#include <mysql.h> +shell> mysqladmin version +mysqladmin: connect to server at '' failed; +error: 'Can't connect to mysql server on localhost (146)' @end example -You can either link your code with the dynamic @file{libmysql.lib} library, -which is just a wrapper to load in @file{libmysql.dll} on demand, or link -with the static @file{mysqlclient.lib} library. +The solution to this is to kill the @code{mysqld} server and restart it. +This has only happened to us when we have forced the server down and done +a restart immediately. -Note that as the mysqlclient libraries are compiled as threaded libraries, -you should also compile your code to be multi-threaded! +@item +With MIT-pthreads, the @code{sleep()} system call isn't interruptible with +@code{SIGINT} (break). This is only noticeable when you run +@code{mysqladmin --sleep}. You must wait for the @code{sleep()} call to +terminate before the interrupt is served and the process stops. -@cindex Windows, versus Unix -@cindex operating systems, Windows versus Unix -@node Windows vs Unix, , Windows compiling, Windows -@subsection MySQL-Windows Compared to Unix MySQL +@item +When linking, you may receive warning messages like these (at least on +Solaris); they can be ignored: -@strong{MySQL}-Windows has by now proven itself to be very stable. This version -of @strong{MySQL} has the same features as the corresponding Unix version -with the following exceptions: - -@table @strong -@item Win95 and threads -Win95 leaks about 200 bytes of main memory for each thread creation. -Each connection in @strong{MySQL} creates a new thread, so you shouldn't -run @code{mysqld} for an extended time on Win95 if your server handles -many connections! WinNT and Win98 don't suffer from this bug. +@example +ld: warning: symbol `_iob' has differing sizes: + (file /my/local/pthreads/lib/libpthread.a(findfp.o) value=0x4; +file /usr/lib/libc.so value=0x140); + /my/local/pthreads/lib/libpthread.a(findfp.o) definition taken +ld: warning: symbol `__iob' has differing sizes: + (file /my/local/pthreads/lib/libpthread.a(findfp.o) value=0x4; +file /usr/lib/libc.so value=0x140); + /my/local/pthreads/lib/libpthread.a(findfp.o) definition taken +@end example -@item Concurrent reads -@strong{MySQL} depends on the @code{pread()} and @code{pwrite()} calls to be -able to mix @code{INSERT} and @code{SELECT}. Currently we use mutexes -to emulate @code{pread()}/@code{pwrite()}. We will, in the long run, -replace the file level interface with a virtual interface so that we can -use the @code{readfile()}/@code{writefile()} interface on NT to get more speed. -The current implementation limits the number of open files @strong{MySQL} -can use to 1024, which means that you will not be able to run as many -concurrent threads on NT as on Unix. +@item +Some other warnings also can be ignored: -@item Blocking read -@strong{MySQL} uses a blocking read for each connection. -This means that: +@example +implicit declaration of function `int strtoll(...)' +implicit declaration of function `int strtoul(...)' +@end example -@itemize @bullet @item -A connection will not be disconnected automatically after 8 hours, as happens -with the Unix version of @strong{MySQL}. +We haven't gotten @code{readline} to work with MIT-pthreads. (This isn't +needed, but may be interesting for someone.) +@end itemize + +@cindex Perl, installing +@cindex installing, Perl +@node Perl support, Source install system issues, MIT-pthreads, Installing +@section Perl Installation Comments + +@menu +* Perl installation:: Installing Perl on Unix +* ActiveState Perl:: Installing ActiveState Perl on Windows +* Windows Perl:: Installing the @strong{MySQL} Perl distribution on Windows +* Perl support problems:: Problems using the Perl @code{DBI}/@code{DBD} interface +@end menu + +@node Perl installation, ActiveState Perl, Perl support, Perl support +@subsection Installing Perl on Unix + +Perl support for @strong{MySQL} is provided by means of the +@code{DBI}/@code{DBD} client interface. @xref{Perl}. The Perl +@code{DBD}/@code{DBI} client code requires Perl Version 5.004 or later. The +interface @strong{will not work} if you have an older version of Perl. + +@strong{MySQL} Perl support also requires that you've installed +@strong{MySQL} client programming support. If you installed @strong{MySQL} +from RPM files, client programs are in the client RPM, but client programming +support is in the developer RPM. Make sure you've installed the latter RPM. + +As of Version 3.22.8, Perl support is distributed separately from the main +@strong{MySQL} distribution. If you want to install Perl support, the files +you will need can be obtained from +@uref{http://www.mysql.com/Downloads/Contrib/}. + +The Perl distributions are provided as compressed @code{tar} archives and +have names like @file{MODULE-VERSION.tar.gz}, where @code{MODULE} is the +module name and @code{VERSION} is the version number. You should get the +@code{Data-Dumper}, @code{DBI}, and @code{Msql-Mysql-modules} distributions +and install them in that order. The installation procedure is shown below. +The example shown is for the @code{Data-Dumper} module, but the procedure is +the same for all three distributions: +@enumerate @item -If a connection hangs, it's impossible to break it without killing -@strong{MySQL}. +Unpack the distribution into the current directory: +@example +shell> gunzip < Data-Dumper-VERSION.tar.gz | tar xvf - +@end example +This command creates a directory named @file{Data-Dumper-VERSION}. @item -@code{mysqladmin kill} will not work on a sleeping connection. +Change into the top-level directory of the unpacked distribution: +@example +shell> cd Data-Dumper-VERSION +@end example @item -@code{mysqladmin shutdown} can't abort as long as there are sleeping -connections. -@end itemize +Build the distribution and compile everything: +@example +shell> perl Makefile.PL +shell> make +shell> make test +shell> make install +@end example +@end enumerate -We plan to fix this problem when our Windows developers have figured out a -nice workaround. +The @code{make test} command is important because it verifies that the +module is working. Note that when you run that command during the +@code{Msql-Mysql-modules} installation to exercise the interface code, the +@strong{MySQL} server must be running or the test will fail. -@item UDF functions -For the moment, @strong{MySQL}-Windows does not support user-definable -functions. +It is a good idea to rebuild and reinstall the @code{Msql-Mysql-modules} +distribution whenever you install a new release of @strong{MySQL}, +particularly if you notice symptoms such as all your @code{DBI} scripts +dumping core after you upgrade @strong{MySQL}. -@item @code{DROP DATABASE} -You can't drop a database that is in use by some thread. +If you don't have the right to install Perl modules in the system directory +or if you to install local Perl modules, the following reference may help +you: -@item Killing @strong{MySQL} from the task manager -You can't kill @strong{MySQL} from the task manager or with the shutdown -utility in Win95. You must take it down with @code{mysqladmin shutdown}. +@example +@uref{http://www.iserver.com/support/contrib/perl5/modules.html} +@end example -@item Case-insensitive names -Filenames are case insensitive on Windows, so database and table names -are also case insensitive in @strong{MySQL} for Windows. The only -restriction is that database and table names must be specified using the same -case throughout a given statement. @xref{Name case sensitivity}. +Look under the heading +@code{Installing New Modules that Require Locally Installed Modules}. -@item The @samp{\} directory character -Pathname components in Win95 are separated by the @samp{\} character, which is -also the escape character in @strong{MySQL}. If you are using @code{LOAD -DATA INFILE} or @code{SELECT ... INTO OUTFILE}, you must double the @samp{\} -character: +@node ActiveState Perl, Windows Perl, Perl installation, Perl support +@subsection Installing ActiveState Perl on Windows +@cindex installing, Perl on Windows +@cindex Perl, installing on Windows +@cindex ActiveState Perl + +To install the @strong{MySQL} @code{DBD} module with ActiveState Perl on +Windows, you should do the following: + +@itemize @bullet +@item +Get ActiveState Perl from +@uref{http://www.activestate.com/Products/ActivePerl/index.html} +and install it. + +@item +Open a DOS shell. + +@item +If required, set the HTTP_proxy variable. For example, you might try: @example -mysql> LOAD DATA INFILE "C:\\tmp\\skr.txt" INTO TABLE skr; -mysql> SELECT * INTO OUTFILE 'C:\\tmp\\skr.txt' FROM skr; +set HTTP_proxy=my.proxy.com:3128 @end example -Alternatively, use Unix style filenames with @samp{/} characters: +@item +Start the PPM program: @example -mysql> LOAD DATA INFILE "C:/tmp/skr.txt" INTO TABLE skr; -mysql> SELECT * INTO OUTFILE 'C:/tmp/skr.txt' FROM skr; +C:\> c:\perl\bin\ppm.pl @end example -@item @code{Can't open named pipe} error -If you use a @strong{MySQL} 3.22 version on NT with the newest mysql-clients -you will get the following error: +@item +If you have not already done so, install @code{DBI}: @example -error 2017: can't open named pipe to host: . pipe... +ppm> install DBI @end example -@tindex .my.cnf file -This is because the release version of @strong{MySQL} uses named pipes on NT -by default. You can avoid this error by using the @code{--host=localhost} -option to the new @strong{MySQL} clients or create an option file -@file{C:\my.cnf} that contains the following information: +@item +If this succeeds, run the following command: @example -[client] -host = localhost +install ftp://ftp.de.uu.net/pub/CPAN/authors/id/JWIED/DBD-mysql-1.2212.x86.ppd @end example +@end itemize -@item @code{Access denied for user} error -If you get the error @code{Access denied for user: 'some-user@@unknown' -to database 'mysql'} when accessing a @strong{MySQL} server on the same -machine, this means that @strong{MySQL} can't resolve your host name -properly. +The above should work at least with ActiveState Perl Version 5.6. -To fix this, you should create a file @file{\windows\hosts} with the -following information: +If you can't get the above to work, you should instead install the +@strong{MyODBC} driver and connect to @strong{MySQL} server through +ODBC: @example -127.0.0.1 localhost +use DBI; +$dbh= DBI->connect("DBI:ODBC:$dsn","$user","$password") || + die "Got error $DBI::errstr when connecting to $dsn\n"; @end example -@item @code{ALTER TABLE} -While you are executing an @code{ALTER TABLE} statement, the table is locked -from usage by other threads. This has to do with the fact that on Windows, -you can't delete a file that is in use by another threads. (In the future, -we may find some way to work around this problem.) - -@item @code{DROP TABLE} on a table that is in use by a @code{MERGE} table will not work -The @code{MERGE} handler does its table mapping hidden from @strong{MySQL}. -Because Windows doesn't allow you to drop files that are open, you first -must flush all @code{MERGE} tables (with @code{FLUSH TABLES}) or drop the -@code{MERGE} table before dropping the table. We will fix this at the same -time we introduce @code{VIEW}s. -@end table - -Here are some open issues for anyone who might want to help us with the Windows -release: +@node Windows Perl, Perl support problems, ActiveState Perl, Perl support +@subsection Installing the MySQL Perl Distribution on Windows -@cindex Windows, open issues +The @strong{MySQL} Perl distribution contains @code{DBI}, +@code{DBD:MySQL} and @code{DBD:ODBC}. @itemize @bullet -@item -Make a single-user @code{MYSQL.DLL} server. This should include everything in -a standard @strong{MySQL} server, except thread creation. This will make -@strong{MySQL} much easier to use in applications that don't need a true -client/server and don't need to access the server from other hosts. +@item +Get the Perl distribution for Windows from +@uref{http://www.mysql.com/download.html}. @item -Add some nice start and shutdown icons to the @strong{MySQL} installation. +Unzip the distribution in @code{C:} so that you get a @file{C:\PERL} directory. @item -Create a tool to manage registry entries for the @strong{MySQL} startup -options. The registry entry reading is already coded into @file{mysqld.cc}, -but it should be recoded to be more parameter oriented. The tool should -also be able to update the @file{C:\my.cnf} option file if the user prefers -to use that instead of the registry. +Add the directory @file{C:\PERL\BIN} to your path. @item -When registering @code{mysqld} as a service with @code{--install} (on NT) -it would be nice if you could also add default options on the command line. -For the moment, the workaround is to list the parameters in the -@file{C:\my.cnf} file instead. +Add the directory @file{C:\PERL\BIN\MSWIN32-x86-thread} or +@file{C:\PERL\BIN\MSWIN32-x86} to your path. @item -It would be real nice to be able to kill @code{mysqld} from the task manager. -For the moment, you must use @code{mysqladmin shutdown}. +Test that @code{perl} works by executing @code{perl -v} in a DOS shell. +@end itemize -@item -Port @code{readline} to Windows for use in the @code{mysql} command line tool. +@cindex problems, installing Perl +@cindex Perl DBI/DBD, installation problems +@node Perl support problems, , Windows Perl, Perl support +@subsection Problems Using the Perl @code{DBI}/@code{DBD} Interface -@item -GUI versions of the standard @strong{MySQL} clients (@code{mysql}, -@code{mysqlshow}, @code{mysqladmin}, and @code{mysqldump}) would be nice. +If Perl reports that it can't find the @file{../mysql/mysql.so} module, +then the problem is probably that Perl can't locate the shared library +@file{libmysqlclient.so}. -@item -It would be nice if the socket read and write functions in @file{net.c} were -interruptible. This would make it possible to kill open threads with -@code{mysqladmin kill} on Windows. +You can fix this by any of the following methods: +@itemize @bullet @item -@code{mysqld} always starts in the "C" locale and not in the default locale. -We would like to have @code{mysqld} use the current locale for the sort order. +Compile the @code{Msql-Mysql-modules} distribution with @code{perl +Makefile.PL -static -config} rather than @code{perl Makefile.PL}. @item -Implement UDF functions with @code{.DLL}s. +Copy @code{libmysqlclient.so} to the directory where your other shared +libraries are located (probably @file{/usr/lib} or @file{/lib}). @item -Add macros to use the faster thread-safe increment/decrement methods -provided by Windows. +On Linux you can add the pathname of the directory where +@file{libmysqlclient.so} is located to the @file{/etc/ld.so.conf} file. +@tindex LD_RUN_PATH environment variable +@tindex Environment variable, LD_RUN_PATH +@item +Add the pathname of the directory where @file{libmysqlclient.so} is located +to the @code{LD_RUN_PATH} environment variable. @end itemize -Other Windows-specific issues are described in the @file{README} file that -comes with the @strong{MySQL}-Windows distribution. +If you get the following errors from @code{DBD-mysql}, +you are probably using @code{gcc} (or using an old binary compiled with +@code{gcc}): -@node OS/2, MySQL binaries, Windows, Installing -@section OS/2 Notes +@example +/usr/bin/perl: can't resolve symbol '__moddi3' +/usr/bin/perl: can't resolve symbol '__divdi3' +@end example -@strong{MySQL} uses quite a few open files. Because of this, you should add -something like the following to your @file{CONFIG.SYS} file: +Add @code{-L/usr/lib/gcc-lib/... -lgcc} to the link command when the +@file{mysql.so} library gets built (check the output from @code{make} for +@file{mysql.so} when you compile the Perl client). The @code{-L} option +should specify the pathname of the directory where @file{libgcc.a} is located +on your system. + +Another cause of this problem may be that Perl and @strong{MySQL} aren't both +compiled with @code{gcc}. In this case, you can solve the mismatch by +compiling both with @code{gcc}. + +If you get the following error from @code{Msql-Mysql-modules} +when you run the tests: @example -SET EMXOPT=-c -n -h1024 +t/00base............install_driver(mysql) failed: Can't load '../blib/arch/auto/DBD/mysql/mysql.so' for module DBD::mysql: ../blib/arch/auto/DBD/mysql/mysql.so: undefined symbol: uncompress at /usr/lib/perl5/5.00503/i586-linux/DynaLoader.pm line 169. @end example -If you don't do this, you will probably run into the following error: +it means that you need to include the compression library, -lz, to the +link line. This can be doing the following change in the file +@file{lib/DBD/mysql/Install.pm}: @example -File 'xxxx' not found (Errcode: 24) -@end example +$sysliblist .= " -lm"; -When using @strong{MySQL} with OS/2 Warp 3, FixPack 29 or above is -required. With OS/2 Warp 4, FixPack 4 or above is required. This is a -requirement of the Pthreads library. @strong{MySQL} must be installed -in a partition that supports long filenames such as HPFS, FAT32, etc. +to -The @file{INSTALL.CMD} script must be run from OS/2's own @file{CMD.EXE} -and may not work with replacement shells such as @file{4OS2.EXE}. +$sysliblist .= " -lm -lz"; +@end example -The @file{scripts/mysql-install-db} script has been renamed. It is now called -@file{install.cmd} and is a REXX script, which will set up the default -@strong{MySQL} security settings and create the WorkPlace Shell icons -for @strong{MySQL}. +After this, you MUST run 'make realclean' and then proceed with the +installation from the beginning. -Dynamic module support is compiled in but not fully tested. Dynamic -modules should be compiled using the Pthreads run-time library. +If you want to use the Perl module on a system that doesn't support dynamic +linking (like SCO) you can generate a static version of Perl that includes +@code{DBI} and @code{DBD-mysql}. The way this works is that you generate a +version of Perl with the @code{DBI} code linked in and install it on top of +your current Perl. Then you use that to build a version of Perl that +additionally has the @code{DBD} code linked in, and install that. + +On SCO, you must have the following environment variables set: @example -gcc -Zdll -Zmt -Zcrtdll=pthrdrtl -I../include -I../regex -I.. \ - -o example udf_example.cc -L../lib -lmysqlclient udf_example.def -mv example.dll example.udf +shell> LD_LIBRARY_PATH=/lib:/usr/lib:/usr/local/lib:/usr/progressive/lib +or +shell> LD_LIBRARY_PATH=/usr/lib:/lib:/usr/local/lib:/usr/ccs/lib:/usr/progressive/lib:/usr/skunk/lib +shell> LIBPATH=/usr/lib:/lib:/usr/local/lib:/usr/ccs/lib:/usr/progressive/lib:/usr/skunk/lib +shell> MANPATH=scohelp:/usr/man:/usr/local1/man:/usr/local/man:/usr/skunk/man: @end example -@strong{Note:} Due to limitations in OS/2, UDF module name stems must not -exceed 8 characters. Modules are stored in the @file{/mysql2/udf} -directory; the @code{safe-mysqld.cmd} script will put this directory in -the @code{BEGINLIBPATH} environment variable. When using UDF modules, -specified extensions are ignored --- it is assumed to be @file{.udf}. -For example, in Unix, the shared module might be named @file{example.so} -and you would load a function from it like this: +First, create a Perl that includes a statically linked @code{DBI} by running +these commands in the directory where your @code{DBI} distribution is +located: @example -mysql> CREATE FUNCTION metaphon RETURNS STRING SONAME "example.so"; +shell> perl Makefile.PL -static -config +shell> make +shell> make install +shell> make perl @end example -Is OS/2, the module would be named @file{example.udf}, but you would not -specify the module extension: +Then you must install the new Perl. The output of @code{make perl} will +indicate the exact @code{make} command you will need to execute to perform +the installation. On SCO, this is @code{make -f Makefile.aperl inst_perl +MAP_TARGET=perl}. + +Next, use the just-created Perl to create another Perl that also includes a +statically-linked @code{DBD::mysql} by running these commands in the +directory where your @code{Msql-Mysql-modules} distribution is located: @example -mysql> CREATE FUNCTION metaphon RETURNS STRING SONAME "example"; +shell> perl Makefile.PL -static -config +shell> make +shell> make install +shell> make perl @end example -@cindex binary distributions -@node MySQL binaries, Post-installation, OS/2, Installing -@section MySQL Binaries +Finally, you should install this new Perl. Again, the output of @code{make +perl} indicates the command to use. -As a service, we at @strong{MySQL AB} provide a set of binary distributions -of @strong{MySQL} that are compiled at our site or at sites where customers -kindly have given us access to their machines. +@node Source install system issues, Windows, Perl support, Installing +@section System-specific Issues -These distributions are generated with @code{scripts/make_binary_distribution} -and are configured with the following compilers and options: +The following sections indicate some of the issues that have been observed to +occur on particular systems when installing @strong{MySQL} from a source +distribution. -@table @asis -@item SunOS 4.1.4 2 sun4c with @code{gcc} 2.7.2.1 -@code{CC=gcc CXX=gcc CXXFLAGS="-O3 -felide-constructors" ./configure --prefix=/usr/local/mysql --disable-shared --with-extra-charsets=complex --enable-assembler} +@menu +* Solaris:: Solaris notes +* Solaris 2.7:: Solaris 2.7 / 2.8 notes +* Solaris x86:: Solaris x86 notes +* SunOS:: SunOS 4 notes +* Linux:: Linux notes (all Linux versions) +* Alpha-DEC-UNIX:: Alpha-DEC-UNIX notes +* Alpha-DEC-OSF1:: Alpha-DEC-OSF1 notes +* SGI-Irix:: SGI-Irix notes +* FreeBSD:: FreeBSD notes +* NetBSD:: NetBSD notes +* OpenBSD:: OpenBSD 2.5 notes +* BSDI:: BSD/OS notes +* SCO:: SCO notes +* SCO Unixware:: SCO Unixware 7.0 notes +* IBM-AIX:: IBM-AIX notes +* HP-UX 10.20:: HP-UX 10.20 notes +* HP-UX 11.x:: HP-UX 11.x notes +* Mac OS X:: Mac OS X notes +* BEOS:: BeOS Notes +@end menu -@item SunOS 5.5.1 (and above) sun4u with @code{egcs} 1.0.3a or 2.90.27 or gcc 2.95.2 and newer -@code{CC=gcc CFLAGS="-O3" CXX=gcc CXXFLAGS="-O3 -felide-constructors -fno-exceptions -fno-rtti" ./configure --prefix=/usr/local/mysql --with-low-memory --with-extra-charsets=complex --enable-assembler} -@item SunOS 5.6 i86pc with @code{gcc} 2.8.1 -@code{CC=gcc CXX=gcc CXXFLAGS=-O3 ./configure --prefix=/usr/local/mysql --with-low-memory --with-extra-charsets=complex} +@cindex Solaris installation problems +@cindex problems, installing on Solaris +@cindex tar, problems on Solaris +@cindex errors, directory checksum +@cindex checksum errors +@node Solaris, Solaris 2.7, Source install system issues, Source install system issues +@subsection Solaris Notes -@item Linux 2.0.33 i386 with @code{pgcc} 2.90.29 (@code{egcs} 1.0.3a) -@code{CFLAGS="-O3 -mpentium -mstack-align-double" CXX=gcc CXXFLAGS="-O3 -mpentium -mstack-align-double -felide-constructors -fno-exceptions -fno-rtti" ./configure --prefix=/usr/local/mysql --enable-assembler --with-mysqld-ldflags=-all-static --with-extra-charsets=complex} +On Solaris, you may run into trouble even before you get the @strong{MySQL} +distribution unpacked! Solaris @code{tar} can't handle long file names, so +you may see an error like this when you unpack @strong{MySQL}: -@item Linux 2.2.x with x686 with @code{gcc} 2.95.2 -@code{CFLAGS="-O3 -mpentiumpro" CXX=gcc CXXFLAGS="-O3 -mpentiumpro -felide-constructors -fno-exceptions -fno-rtti" ./configure --prefix=/usr/local/mysql --enable-assembler --with-mysqld-ldflags=-all-static --disable-shared --with-extra-charset=complex} +@example +x mysql-3.22.12-beta/bench/Results/ATIS-mysql_odbc-NT_4.0-cmp-db2,informix,ms-sql,mysql,oracle,solid,sybase, 0 bytes, 0 tape blocks +tar: directory checksum error +@end example -@item SCO 3.2v5.0.4 i386 with @code{gcc} 2.7-95q4 -@code{CC=gcc CXX=gcc CXXFLAGS=-O3 ./configure --prefix=/usr/local/mysql --with-extra-charsets=complex} +In this case, you must use GNU @code{tar} (@code{gtar}) to unpack the +distribution. You can find a precompiled copy for Solaris at +@uref{http://www.mysql.com/Downloads/}. -@item AIX 2 4 with @code{gcc} 2.7.2.2 -@code{CC=gcc CXX=gcc CXXFLAGS=-O3 ./configure --prefix=/usr/local/mysql --with-extra-charsets=complex} +Sun native threads work only on Solaris 2.5 and higher. For Version 2.4 and +earlier, @strong{MySQL} will automatically use MIT-pthreads. +@xref{MIT-pthreads}. -@item OSF1 V4.0 564 alpha with @code{gcc} 2.8.1 -@code{CC=gcc CFLAGS=-O CXX=gcc CXXFLAGS=-O3 ./configure --prefix=/usr/local/mysql --with-low-memory --with-extra-charsets=complex} +If you get the following error from configure: -@item Irix 6.3 IP32 with @code{gcc} 2.8.0 -@code{CC=gcc CXX=gcc CXXFLAGS=-O3 ./configure --prefix=/usr/local/mysql --with-extra-charsets=complex} +@example +checking for restartable system calls... configure: error can not run test +programs while cross compiling +@end example -@item BSDI BSD/OS 3.1 i386 with @code{gcc} 2.7.2.1 -@code{CC=gcc CXX=gcc CXXFLAGS=-O ./configure --prefix=/usr/local/mysql --with-extra-charsets=complex} +This means that you have something wrong with your compiler installation! +In this case you should upgrade your compiler to a newer version. You may +also be able to solve this problem by inserting the following row into the +@file{config.cache} file: -@item BSDI BSD/OS 2.1 i386 with @code{gcc} 2.7.2 -@code{CC=gcc CXX=gcc CXXFLAGS=-O3 ./configure --prefix=/usr/local/mysql --with-extra-charsets=complex} -@end table +@example +ac_cv_sys_restartable_syscalls=$@{ac_cv_sys_restartable_syscalls='no'@} +@end example -Anyone who has more optimal options for any of the configurations listed -above can always mail them to the developer's mailing list at -@email{internals@@lists.mysql.com}. +If you are using Solaris on a SPARC, the recommended compiler is +@code{gcc} 2.95.2. You can find this at @uref{http://gcc.gnu.org/}. +Note that @code{egcs} 1.1.1 and @code{gcc} 2.8.1 don't work reliably on +SPARC! -RPM distributions prior to @strong{MySQL} Version 3.22 are user-contributed. -Beginning with Version 3.22, the RPMs are generated by us at -@strong{MySQL AB}. +The recommended @code{configure} line when using @code{gcc} 2.95.2 is: -If you want to compile a debug version of @strong{MySQL}, you should add -@code{--with-debug} or @code{--with-debug=full} to the above configure lines -and remove any @code{-fomit-frame-pointer} options. +@example +CC=gcc CFLAGS="-O3" \ +CXX=gcc CXXFLAGS="-O3 -felide-constructors -fno-exceptions -fno-rtti" \ +./configure --prefix=/usr/local/mysql --with-low-memory --enable-assembler +@end example -@cindex post-installation, setup and testing -@cindex testing, post-installation -@cindex setup, post-installation -@node Post-installation, Installing many servers, MySQL binaries, Installing -@section Post-installation Setup and Testing +If you have a ultra sparc, you can get 4 % more performance by adding +"-mcpu=v8 -Wa,-xarch=v8plusa" to CFLAGS and CXXFLAGS. -@menu -* mysql_install_db:: Problems running @code{mysql_install_db} -* Starting server:: Problems starting the @strong{MySQL} server -* Automatic start:: Starting and stopping @strong{MySQL} automatically -* Command-line options:: Command-line options -* Option files:: Option files -@end menu +If you have the Sun Workshop (SunPro) 4.2 (or newer) compiler, you can +run @code{configure} like this: -Once you've installed @strong{MySQL} (from either a binary or source -distribution), you need to initialize the grant tables, start the server, -and make sure that the server works okay. You may also wish to arrange -for the server to be started and stopped automatically when your system -starts up and shuts down. +@example +CC=cc CFLAGS="-Xa -fast -xO4 -native -xstrconst -mt" \ +CXX=CC CXXFLAGS="-noex -xO4 -mt" \ +./configure --prefix=/usr/local/mysql --enable-assembler +@end example -Normally you install the grant tables and start the server like this -for installation from a source distribution: -@cindex starting, the server -@cindex server, starting +You may also have to edit the @code{configure} script to change this line: @example -shell> ./scripts/mysql_install_db -shell> cd mysql_installation_directory -shell> ./bin/safe_mysqld --user=mysql & +#if !defined(__STDC__) || __STDC__ != 1 @end example -For a binary distribution (not RPM or pkg packages), do this: +to this: @example -shell> cd mysql_installation_directory -shell> ./bin/mysql_install_db -shell> ./bin/safe_mysqld --user=mysql & +#if !defined(__STDC__) @end example -This creates the @code{mysql} database which will hold all database -privileges, the @code{test} database which you can use to test -@strong{MySQL} and also privilege entries for the user that run -@code{mysql_install_db} and a @code{root} user (without any passwords). -This also starts the @code{mysqld} server. - -@code{mysql_install_db} will not overwrite any old privilege tables, so -it should be safe to run in any circumstances. If you don't want to -have the @code{test} database you can remove it with @code{mysqladmin -u -root drop test}. +If you turn on @code{__STDC__} with the @code{-Xc} option, the Sun compiler +can't compile with the Solaris @file{pthread.h} header file. This is a Sun +bug (broken compiler or broken include file). -Testing is most easily done from the top-level directory of the @strong{MySQL} -distribution. For a binary distribution, this is your installation directory -(typically something like @file{/usr/local/mysql}). For a source -distribution, this is the main directory of your @strong{MySQL} source tree. -@cindex testing, the server +If @code{mysqld} issues the error message shown below when you run it, you have +tried to compile @strong{MySQL} with the Sun compiler without enabling the +multi-thread option (@code{-mt}): -In the commands shown below in this section and in the following -subsections, @code{BINDIR} is the path to the location in which programs -like @code{mysqladmin} and @code{safe_mysqld} are installed. For a -binary distribution, this is the @file{bin} directory within the -distribution. For a source distribution, @code{BINDIR} is probably -@file{/usr/local/bin}, unless you specified an installation directory -other than @file{/usr/local} when you ran @code{configure}. -@code{EXECDIR} is the location in which the @code{mysqld} server is -installed. For a binary distribution, this is the same as -@code{BINDIR}. For a source distribution, @code{EXECDIR} is probably -@file{/usr/local/libexec}. +@example +libc internal error: _rmutex_unlock: rmutex not held +@end example -Testing is described in detail below: -@cindex testing, installation +Add @code{-mt} to @code{CFLAGS} and @code{CXXFLAGS} and try again. -@enumerate -@item -If necessary, start the @code{mysqld} server and set up the initial -@strong{MySQL} grant tables containing the privileges that determine how -users are allowed to connect to the server. This is normally done with the -@code{mysql_install_db} script: +If you get the following error when compiling @strong{MySQL} with @code{gcc}, +it means that your @code{gcc} is not configured for your version of Solaris: @example -shell> scripts/mysql_install_db +shell> gcc -O3 -g -O2 -DDBUG_OFF -o thr_alarm ... +./thr_alarm.c: In function `signal_hand': +./thr_alarm.c:556: too many arguments to function `sigwait' @end example -Typically, @code{mysql_install_db} needs to be run only the first time you -install @strong{MySQL}. Therefore, if you are upgrading an existing -installation, you can skip this step. (However, @code{mysql_install_db} is -quite safe to use and will not update any tables that already exist, so if -you are unsure of what to do, you can always run @code{mysql_install_db}.) +The proper thing to do in this case is to get the newest version of +@code{gcc} and compile it with your current @code{gcc} compiler! At +least for Solaris 2.5, almost all binary versions of @code{gcc} have +old, unusable include files that will break all programs that use +threads (and possibly other programs)! -@code{mysql_install_db} creates six tables (@code{user}, @code{db}, -@code{host}, @code{tables_priv}, @code{columns_priv}, and @code{func}) in the -@code{mysql} database. A description of the initial privileges is given in -@ref{Default privileges}. Briefly, these privileges allow the @strong{MySQL} -@code{root} user to do anything, and allow anybody to create or use databases -with a name of @code{'test'} or starting with @code{'test_'}. +Solaris doesn't provide static versions of all system libraries +(@code{libpthreads} and @code{libdl}), so you can't compile @strong{MySQL} +with @code{--static}. If you try to do so, you will get the error: -If you don't set up the grant tables, the following error will appear in the -log file when you start the server: +@example +ld: fatal: library -ldl: not found + +or + +undefined reference to `dlopen' +@end example + +If too many processes try to connect very rapidly to @code{mysqld}, you will +see this error in the @strong{MySQL} log: -@tindex host.frm, problems finding @example -mysqld: Can't find file: 'host.frm' +Error in accept: Protocol error @end example -The above may also happen with a binary @strong{MySQL} distribution if you -don't start @strong{MySQL} by executing exactly @code{./bin/safe_mysqld}! -@xref{safe_mysqld, , @code{safe_mysqld}}. +You might try starting the server with the @code{--set-variable back_log=50} +option as a workaround for this. @xref{Command-line options}. -You might need to run @code{mysql_install_db} as @code{root}. However, -if you prefer, you can run the @strong{MySQL} server as an unprivileged -(non-@code{root}) user, provided that user can read and write files in -the database directory. Instructions for running @strong{MySQL} as an -unprivileged user are given in @ref{Changing MySQL user, , Changing -@strong{MySQL} user}. +If you are linking your own @strong{MySQL} client, you might get the +following error when you try to execute it: -If you have problems with @code{mysql_install_db}, see -@ref{mysql_install_db, , @code{mysql_install_db}}. +@example +ld.so.1: ./my: fatal: libmysqlclient.so.#: open failed: No such file or directory +@end example -There are some alternatives to running the @code{mysql_install_db} -script as it is provided in the @strong{MySQL} distribution: +The problem can be avoided by one of the following methods: @itemize @bullet @item -You may want to edit @code{mysql_install_db} before running it, to change -the initial privileges that are installed into the grant tables. This is -useful if you want to install @strong{MySQL} on a lot of machines with the -same privileges. In this case you probably should need only to add a few -extra @code{INSERT} statements to the @code{mysql.user} and @code{mysql.db} -tables! +Link the client with the following flag (instead of @code{-Lpath}): +@code{-Wl,r/full-path-to-libmysqlclient.so}. @item -If you want to change things in the grant tables after installing them, you -can run @code{mysql_install_db}, then use @code{mysql -u root mysql} to -connect to the grant tables as the @strong{MySQL} @code{root} user and issue -SQL statements to modify the grant tables directly. +Copy @file{libmysqclient.so} to @file{/usr/lib}. +@tindex LD_RUN_PATH environment variable +@tindex Environment variable, LD_RUN_PATH @item -It is possible to re-create the grant tables completely after they have -already been created. You might want to do this if you've already installed -the tables but then want to re-create them after editing -@code{mysql_install_db}. +Add the pathname of the directory where @file{libmysqlclient.so} is located +to the @code{LD_RUN_PATH} environment variable before running your client. @end itemize -For more information about these alternatives, see @ref{Default privileges}. - -@item -Start the @strong{MySQL} server like this: +When using the @code{--with-libwrap} configure option, you must also +include the libraries that @file{libwrap.a} needs: @example -shell> cd mysql_installation_directory -shell> bin/safe_mysqld & +--with-libwrap="/opt/NUtcpwrapper-7.6/lib/libwrap.a -lnsl -lsocket @end example -If you have problems starting the server, see @ref{Starting server}. +If you have problems with configure trying to link with @code{-lz} and +you don't have @code{zlib} installed, you have two options: +@itemize @bullet @item -Use @code{mysqladmin} to verify that the server is running. The following -commands provide a simple test to check that the server is up and responding -to connections: - -@example -shell> BINDIR/mysqladmin version -shell> BINDIR/mysqladmin variables -@end example - -The output from @code{mysqladmin version} varies slightly depending on your -platform and version of @strong{MySQL}, but should be similar to that shown -below: - -@example -shell> BINDIR/mysqladmin version -mysqladmin Ver 8.14 Distrib 3.23.32, for linux on i586 -Copyright (C) 2000 MySQL AB & MySQL Finland AB & TCX DataKonsult AB -This software comes with ABSOLUTELY NO WARRANTY. This is free software, -and you are welcome to modify and redistribute it under the GPL license +If you want to be able to use the compressed communication protocol, +you need to get and install zlib from ftp.gnu.org. -Server version 3.23.32-debug -Protocol version 10 -Connection Localhost via Unix socket -TCP port 3306 -UNIX socket /tmp/mysql.sock -Uptime: 16 sec +@item +Configure with @code{--with-named-z-libs=no}. +@end itemize -Threads: 1 Questions: 9 Slow queries: 0 Opens: 7 Flush tables: 2 Open tables: 0 Queries per second avg: 0.000 Memory in use: 132K Max memory used: 16773K -@end example +If you are using gcc and have problems with loading @code{UDF} functions +into @strong{MySQL}, try adding @code{-lgcc} to the link line for the +@code{UDF} function. -To get a feeling for what else you can do with @code{BINDIR/mysqladmin}, -invoke it with the @code{--help} option. +If you would like @strong{MySQL} to start automatically, you can copy +@file{support-files/mysql.server} to @file{/etc/init.d} and create a +symbolic link to it named @file{/etc/rc3.d/S99mysql.server}. -@item -Verify that you can shut down the server: -@cindex server, shutdown -@cindex shutting down, the server +@node Solaris 2.7, Solaris x86, Solaris, Source install system issues +@subsection Solaris 2.7/2.8 Notes -@example -shell> BINDIR/mysqladmin -u root shutdown -@end example +You can normally use a Solaris 2.6 binary on Solaris 2.7 and 2.8. Most +of the Solaris 2.6 issues also apply for Solaris 2.7 and 2.8. -@item -Verify that you can restart the server. Do this using @code{safe_mysqld} or -by invoking @code{mysqld} directly. For example: +Note that @strong{MySQL} Version 3.23.4 and above should be able to autodetect +new versions of Solaris and enable workarounds for the following problems! -@cindex server, restart -@cindex restarting, the server +Solaris 2.7 / 2.8 has some bugs in the include files. You may see the +following error when you use @code{gcc}: @example -shell> BINDIR/safe_mysqld --log & +/usr/include/widec.h:42: warning: `getwc' redefined +/usr/include/wchar.h:326: warning: this is the location of the previous +definition @end example -If @code{safe_mysqld} fails, try running it from the @strong{MySQL} -installation directory (if you are not already there). If that doesn't work, -see @ref{Starting server}. +If this occurs, you can do the following to fix the problem: -@item -Run some simple tests to verify that the server is working. -The output should be similar to what is shown below: +Copy @code{/usr/include/widec.h} to +@code{.../lib/gcc-lib/os/gcc-version/include} and change line 41 from: @example -shell> BINDIR/mysqlshow -+-----------+ -| Databases | -+-----------+ -| mysql | -+-----------+ +#if !defined(lint) && !defined(__lint) -shell> BINDIR/mysqlshow mysql -Database: mysql -+--------------+ -| Tables | -+--------------+ -| columns_priv | -| db | -| func | -| host | -| tables_priv | -| user | -+--------------+ +to -shell> BINDIR/mysql -e "select host,db,user from db" mysql -+------+--------+------+ -| host | db | user | -+------+--------+------+ -| % | test | | -| % | test_% | | -+------+--------+------+ +#if !defined(lint) && !defined(__lint) && !defined(getwc) @end example -There is also a benchmark suite in the @file{sql-bench} directory (under the -@strong{MySQL} installation directory) that you can use to compare how -@strong{MySQL} performs on different platforms. The @file{sql-bench/Results} -directory contains the results from many runs against different databases and -platforms. To run all tests, execute these commands: +Alternatively, you can edit @file{/usr/include/widec.h} directly. Either +way, after you make the fix, you should remove @file{config.cache} and run +@code{configure} again! + +If you get errors like this when you run @code{make}, it's because +@code{configure} didn't detect the @file{curses.h} file (probably +because of the error in @file{/usr/include/widec.h}): @example -shell> cd sql-bench -shell> run-all-tests +In file included from mysql.cc:50: +/usr/include/term.h:1060: syntax error before `,' +/usr/include/term.h:1081: syntax error before `;' @end example -If you don't have the @file{sql-bench} directory, you are probably using an -RPM for a binary distribution. (Source distribution RPMs include the -benchmark directory.) In this case, you must first install the benchmark -suite before you can use it. Beginning with @strong{MySQL} Version 3.22, -there are benchmark RPM files named @file{mysql-bench-VERSION-i386.rpm} that -contain benchmark code and data. +The solution to this is to do one of the following: -If you have a source distribution, you can also run the tests in the -@file{tests} subdirectory. For example, to run @file{auto_increment.tst}, do -this: +@itemize @bullet +@item +Configure with @code{CFLAGS=-DHAVE_CURSES_H CXXFLAGS=-DHAVE_CURSES_H ./configure}. -@example -shell> BINDIR/mysql -vvf test < ./tests/auto_increment.tst -@end example +@item +Edit @file{/usr/include/widec.h} as indicted above and rerun configure. -The expected results are shown in the @file{./tests/auto_increment.res} file. -@end enumerate +@item +Remove the @code{#define HAVE_TERM} line from @file{config.h} file and +run @code{make} again. +@end itemize -@node mysql_install_db, Starting server, Post-installation, Post-installation -@subsection Problems Running @code{mysql_install_db} -@cindex @code{mysql_install_db} script -@cindex scripts, @code{mysql_install_db} +If you get a problem that your linker can't find @code{-lz} when linking +your client program, the problem is probably that your @file{libz.so} file is +installed in @file{/usr/local/lib}. You can fix this by one of the +following methods: -The purpose of the @code{mysql_install_db} script is to generate new -@strong{MySQL} privilege tables. It will not affect any other data! -It will also not do anything if you already have @strong{MySQL} privilege -tables installed! +@itemize @bullet +@item +Add @file{/usr/local/lib} to @code{LD_LIBRARY_PATH}. -If you want to re-create your privilege tables, you should take down -the @code{mysqld} server, if it's running, and then do something like: +@item +Add a link to @file{libz.so} from @file{/lib}. -@example -mv mysql-data-directory/mysql mysql-data-directory/mysql-old -mysql_install_db -@end example +@item +If you are using Solaris 8, you can install the optional zlib from your +Solaris 8 CD distribution. -This section lists problems you might encounter when you run -@code{mysql_install_db}: +@item +Configure @strong{MySQL} with the @code{--with-named-z-libs=no} option. +@end itemize -@table @strong -@item @code{mysql_install_db} doesn't install the grant tables +@node Solaris x86, SunOS, Solaris 2.7, Source install system issues +@subsection Solaris x86 Notes -You may find that @code{mysql_install_db} fails to install the grant -tables and terminates after displaying the following messages: +On Solaris 2.8 on x86, @strong{mysqld} will core dump if you run +'strip' in. + +If you are using @code{gcc} or @code{egcs} on Solaris x86 and you +experience problems with core dumps under load, you should use the +following @code{configure} command: @example -starting mysqld daemon with databases from XXXXXX -mysql daemon ended +CC=gcc CFLAGS="-O3 -fomit-frame-pointer -DHAVE_CURSES_H" \ +CXX=gcc \ +CXXFLAGS="-O3 -fomit-frame-pointer -felide-constructors -fno-exceptions -fno-rtti -DHAVE_CURSES_H" \ +./configure --prefix=/usr/local/mysql @end example -In this case, you should examine the log file very carefully! The log -should be located in the directory @file{XXXXXX} named by the error message, -and should indicate why @code{mysqld} didn't start. If you don't understand -what happened, include the log when you post a bug report using -@code{mysqlbug}! -@xref{Bug reports}. - -@item There is already a @code{mysqld} daemon running - -In this case, you probably don't have to run @code{mysql_install_db} at -all. You have to run @code{mysql_install_db} only once, when you install -@strong{MySQL} the first time. - -@item Installing a second @code{mysqld} daemon doesn't work when one daemon is running +This will avoid problems with the @code{libstdc++} library and with C++ +exceptions. -This can happen when you already have an existing @strong{MySQL} -installation, but want to put a new installation in a different place (for -example, for testing, or perhaps you simply want to run two installations at -the same time). Generally the problem that occurs when you try to run the -second server is that it tries to use the same socket and port as the old one. -In this case you will get the error message: @code{Can't start server: Bind on -TCP/IP port: Address already in use} or @code{Can't start server : Bind on -unix socket...}. @xref{Installing many servers}. +If this doesn't help, you should compile a debug version and run +it with a trace file or under @code{gdb}. @xref{Using gdb on mysqld}. -@item You don't have write access to @file{/tmp} -@cindex write access, tmp -@cindex temporary file, write access -@cindex files, @code{tmp} +@node SunOS, Linux, Solaris x86, Source install system issues +@subsection SunOS 4 Notes -If you don't have write access to create a socket file at the default place -(in @file{/tmp}) or permission to create temporary files in @file{/tmp,} -you will get an error when running @code{mysql_install_db} or when -starting or using @code{mysqld}. +On SunOS 4, MIT-pthreads is needed to compile @strong{MySQL}, which in turn +means you will need GNU @code{make}. -You can specify a different socket and temporary directory as follows: +Some SunOS 4 systems have problems with dynamic libraries and @code{libtool}. +You can use the following @code{configure} line to avoid this problem: -@tindex TMPDIR environment variable -@tindex MYSQL_UNIX_PORT environment variable -@tindex Environment variable, TMPDIR -@tindex Environment variable, MYSQL_UNIX_PORT @example -shell> TMPDIR=/some_tmp_dir/ -shell> MYSQL_UNIX_PORT=/some_tmp_dir/mysqld.sock -shell> export TMPDIR MYSQL_UNIX_PORT +shell> ./configure --disable-shared --with-mysqld-ldflags=-all-static @end example -@file{some_tmp_dir} should be the path to some directory for which you -have write permission. @xref{Environment variables}. - -After this you should be able to run @code{mysql_install_db} and start -the server with these commands: +When compiling @code{readline}, you may get warnings about duplicate defines. +These may be ignored. -@example -shell> scripts/mysql_install_db -shell> BINDIR/safe_mysqld & -@end example +When compiling @code{mysqld}, there will be some @code{implicit declaration +of function} warnings. These may be ignored. -@item @code{mysqld} crashes immediately +@node Linux, Alpha-DEC-UNIX, SunOS, Source install system issues +@subsection Linux Notes (All Linux Versions) -If you are running RedHat Version 5.0 with a version of @code{glibc} older than -2.0.7-5, you should make sure you have installed all @code{glibc} patches! -There is a lot of information about this in the @strong{MySQL} mail -archives. Links to the mail archives are available online at -@uref{http://www.mysql.com/documentation/}. -Also, see @ref{Linux}. +The notes below regarding @strong{glibc} apply only to the situation +when you build @strong{MySQL} +yourself. If you are running Linux on an x86 machine, in most cases it is +much better for you to just use our binary. We link our binaries against +the best patched version of @strong{glibc} we can come up with and with the +best compiler options, in an attempt to make it suitable for a high-load +server. So if you read the text below, and are in doubt about +what you should do, try our binary first to see if it meets your needs, and +worry about your own build only after you have discovered that our binary is +not good enough. In that case, we would appreciate a note about it, so we +can build a better binary next time. For a typical user, even for setups with +a lot of concurrent connections and/or tables exceeding 2GB limit, our +binary in most cases is the best choice. -You can also start @code{mysqld} manually using the @code{--skip-grant-tables} -option and add the privilege information yourself using @code{mysql}: +@strong{MySQL} uses LinuxThreads on Linux. If you are using an old +Linux version that doesn't have @code{glibc2}, you must install +LinuxThreads before trying to compile @strong{MySQL}. You can get +LinuxThreads at @uref{http://www.mysql.com/Downloads/Linux}. -@example -shell> BINDIR/safe_mysqld --skip-grant-tables & -shell> BINDIR/mysql -u root mysql -@end example +@strong{NOTE:} We have seen some strange problems with Linux 2.2.14 and +@strong{MySQL} on SMP systems; If you have a SMP system, we recommend +you to upgrade to Linux 2.4 ASAP! Your system will be faster and more +stable by doing this! -From @code{mysql}, manually execute the SQL commands in -@code{mysql_install_db}. Make sure you run @code{mysqladmin -flush-privileges} or @code{mysqladmin reload} afterward to tell the server to -reload the grant tables. -@end table +Note that @code{glibc} versions before and including Version 2.1.1 have +a fatal bug in @code{pthread_mutex_timedwait} handling, which is used +when you do @code{INSERT DELAYED}. We recommend you to not use +@code{INSERT DELAYED} before upgrading glibc. -@node Starting server, Automatic start, mysql_install_db, Post-installation -@subsection Problems Starting the MySQL Server -@cindex server, starting problems -@cindex problems, starting the server +If you plan to have 1000+ concurrent connections, you will need to make +some changes to LinuxThreads, recompile it, and relink @strong{MySQL} against +the new @file{libpthread.a}. Increase @code{PTHREAD_THREADS_MAX} in +@file{sysdeps/unix/sysv/linux/bits/local_lim.h} to 4096 and decrease +@code{STACK_SIZE} in @file{linuxthreads/internals.h} to 256 KB. The paths are +relative to the root of @code{glibc} Note that @strong{MySQL} will not be +stable with around 600-1000 connections if @code{STACK_SIZE} is the default +of 2 MB. -If you are going to use tables that support transactions (BDB, InnoDB), -you should first create a my.cnf file and set startup options -for the table types you plan to use. @xref{Table types}. +If you have a problem with that @strong{MySQL} can't open enough files, +or connections, it may be that you haven't configured Linux to handle +enough files. -Generally, you start the @code{mysqld} server in one of three ways: +In Linux 2.2 and forwards, you can check the number of allocated +file handlers by doing: -@itemize @bullet -@item -By invoking @code{mysql.server}. This script is used primarily at -system startup and shutdown, and is described more fully in -@ref{Automatic start}. +@example +cat /proc/sys/fs/file-max +cat /proc/sys/fs/dquot-max +cat /proc/sys/fs/super-max +@end example -@item -By invoking @code{safe_mysqld}, which tries to determine the proper options -for @code{mysqld} and then runs it with those options. @xref{safe_mysqld, , -@code{safe_mysqld}}. +If you have more than 16M of memory, you should add something like the +following in your boot script (@file{/etc/rc/boot.local} on SuSE): -@item -On NT you should install @code{mysqld} as a service as follows: @example -bin\mysqld-nt --install # Install MySQL as a service +echo 65536 > /proc/sys/fs/file-max +echo 8192 > /proc/sys/fs/dquot-max +echo 1024 > /proc/sys/fs/super-max @end example -You can now start/stop @code{mysqld} as follows: +You can also run the above from the command line as root, but in this case +your old limits will be used next time your computer reboots. + +You should also add /etc/my.cnf: + @example -NET START mysql -NET STOP mysql +[safe_mysqld] +open_files_limit=8192 @end example -Note that in this case you can't use any other options for @code{mysqld}! +The above should allow @strong{MySQL} to create up to 8192 connections/files. -You can remove the service as follows: -@example -bin\mysqld-nt --remove # remove MySQL as a service -@end example +The @code{STACK_SIZE} constant in LinuxThreads controls the spacing of thread +stacks in the address space. It needs to be large enough so that there will +be plenty of room for the stack of each individual thread, but small enough +to keep the stack of some thread from running into the global @code{mysqld} +data. Unfortunately, the Linux implementation of @code{mmap()}, as we have +experimentally discovered, will successfully unmap an already mapped region +if you ask it to map out an address already in use, zeroing out the data +on the entire page, instead of returning an error. So, the safety of +@code{mysqld} or any other threaded application depends on the "gentleman" +behavior of the code that creates threads. The user must take measures to +make sure the number of running threads at any time is sufficiently low for +thread stacks to stay away from the global heap. With @code{mysqld}, you +should enforce this "gentleman" behavior by setting a reasonable value for +the @code{max_connections} variable. -@item -By invoking @code{mysqld} directly. -@end itemize +If you build @strong{MySQL} yourself and do not what to mess with patching +LinuxThreads, you should set @code{max_connections} to a value no higher +than 500. It should be even less if you have a large key buffer, large +heap tables, or some other things that make @strong{mysqld} allocate a lot +of memory or if you are running a 2.2 kernel with a 2GB patch. If you are +using our binary or RPM version 3.23.25 or later, you can safely set +@code{max_connections} at 1500, assuming no large key buffer or heap tables +with lots of data. The more you reduce @code{STACK_SIZE} in LinuxThreads +the more threads you can safely create. We recommend the values between +128K and 256K. -When the @code{mysqld} daemon starts up, it changes directory to the -data directory. This is where it expects to write log files and the pid -(process ID) file, and where it expects to find databases. +If you use a lot of concurrent connections, you may suffer from a "feature" +in the 2.2 kernel that penalizes a process for forking or cloning a child +in an attempt to prevent a fork bomb attack. This will cause @strong{MySQL} +not to scale well as you increase the number of concurrent clients. On +single CPU systems, we have seen this manifested in a very slow thread +creation, which means it may take a long time to connect to @strong{MySQL} +(as long as 1 minute), and it may take just as long to shut it down. On +multiple CPU systems, we have observed a gradual drop in query speed as +the number of clients increases. In the process of trying to find a +solution, we have received a kernel patch from one of our users, who +claimed it made a lot of difference for his site. The patch is available here +(@uref{http://www.mysql.com/Downloads/Patches/linux-fork.patch}). We have +now done rather extensive testing of this patch on both development and +production systems. It has significantly +improved @code{MySQL} performance without causing any problems and we now +recommend it to our users who are still running high-load servers on +2.2 kernels. This issue has been fixed in the 2.4 kernel, so if you are not +satisfied with +the current performance of your system, rather than patching your 2.2 kernel, +it might be easier to just upgrade to 2.4, which will also give you a nice +SMP boost in addition to fixing this fairness bug. -The data directory location is hardwired in when the distribution is -compiled. However, if @code{mysqld} expects to find the data directory -somewhere other than where it really is on your system, it will not work -properly. If you have problems with incorrect paths, you can find out -what options @code{mysqld} allows and what the default path settings are by -invoking @code{mysqld} with the @code{--help} option. You can override the -defaults by specifying the correct pathnames as command-line arguments to -@code{mysqld}. (These options can be used with @code{safe_mysqld} as well.) +We have tested @strong{MySQL} on the 2.4 kernel on a 2 CPU machine and +found @strong{MySQL} scales MUCH better - there was virtually no slowdown +on query throughput all the way up +to 1000 clients, and @strong{MySQL} scaling factor ( computed as the ratio of +maximum throughput to the throughput with one client) was 180%. +We have observed similar results on a 4-CPU system - virtually no +slowdown as the number of +clients was increased up to 1000, and 300% scaling factor. So for a high-load +SMP server we would definitely recommend the 2.4 kernel at this point. We +have discovered that it is essential to run @code{mysqld} process with the +highest possible priority on the 2.4 kernel to achieve maximum performance. +This can be done by adding +@code{renice -20 $$} command to @code{safe_mysqld}. In our testing on a +4-CPU machine, increasing the priority gave 60% increase in throughput with +400 clients. -Normally you should need to tell @code{mysqld} only the base directory under -which @strong{MySQL} is installed. You can do this with the @code{--basedir} -option. You can also use @code{--help} to check the effect of changing path -options (note that @code{--help} @emph{must} be the final option of the -@code{mysqld} command). For example: +We are currently also trying to collect +more info on how well @code{MySQL} performs on 2.4 kernel on 4-way and 8-way +systems. If you have access such a system and have done some benchmarks, +please send a mail to @email{docs@@mysql.com} with the results - we will +include them in the manual. -@example -shell> EXECDIR/mysqld --basedir=/usr/local --help -@end example +There is another issue that greatly hurts @strong{MySQL} performance, +especially on SMP systems. The implementation of mutex in +LinuxThreads in @strong{glibc-2.1} is very bad for programs with many +threads that only +hold the mutex for a short time. On an SMP system, ironic as it is, if +you link @strong{MySQL} against unmodified @strong{LinuxThreads}, +removing processors from the machine improves @strong{MySQL} performance +in many cases. We have made a patch available for @strong{glibc 2.1.3}, +@uref{http://www.mysql.com/Downloads/Linux/linuxthreads-2.1-patch,linuxthreads-2.1-patch} +to correct this behavior. -Once you determine the path settings you want, start the server without -the @code{--help} option. +With @strong{glibc-2.2.2} +@strong{MySQL} version 3.23.36 will use the adaptive mutex, which is much +better than even the patched one in @strong{glibc-2.1.3}. Be warned, however, +that under some conditions, the current mutex code in @strong{glibc-2.2.2} +overspins, which hurts @strong{MySQL} performance. The chance of this +condition can be reduced by renicing @code{mysqld} process to the highest +priority. We have also been able to correct the overspin behavior with +a patch, available @uref{http://www.mysql.com/Downloads/Linux/linuxthreads-2.2.2.patch,here}. It combines the correction of overspin, maximum number of +threads, and stack spacing all in one. You will need to apply it in the +@code{linuxthreads} directory with +@code{patch -p0 </tmp/linuxthreads-2.2.2.patch}. +We hope it will be included in +some form in to the future releases of @code{glibc-2.2}. In any case, if +you link against @code{glibc-2.2.2} you still need to correct +@code{STACK_SIZE} and @code{PTHREAD_THREADS_MAX}. We hope that the defaults +will be corrected to some more acceptable values for high-load +@strong{MySQL} setup in the future, so that your own build can be reduced +to @code{./configure; make; make install}. -Whichever method you use to start the server, if it fails to start up -correctly, check the log file to see if you can find out why. Log files -are located in the data directory (typically -@file{/usr/local/mysql/data} for a binary distribution, -@file{/usr/local/var} for a source distribution, -@file{\mysql\data\mysql.err} on Windows.) Look in the data directory for -files with names of the form @file{host_name.err} and -@file{host_name.log} where @code{host_name} is the name of your server -host. Then check the last few lines of these files: +We recommend that you use the above patches to build a special static +version of @code{libpthread.a} and use it only for statically linking +against @code{MySQL}. We know that the patches are safe for @code{MySQL} +and significantly improve its performance, but we cannot say anything +about other applications. If you link other applications against the +patched version of the library, or build a patched shared version and +install it on your system, you are doing it at your own risk with regard +to other applications that depend on @code{LinuxThreads}. -@example -shell> tail host_name.err -shell> tail host_name.log -@end example +If you experience any strange problems during the installation of +@strong{MySQL}, or with some common utilties hanging, it is very likely that +they are either library or compiler related. If this is the case, using our +binary will resolve them. -If you find something like the following in the log file: -@example -000729 14:50:10 bdb: Recovery function for LSN 1 27595 failed -000729 14:50:10 bdb: warning: ./test/t1.db: No such file or directory -000729 14:50:10 Can't init databases -@end example +One known problem with the binary distribution is that with older Linux +systems that use @code{libc} (like RedHat 4.x or Slackware), you will get +some non-fatal problems with hostname resolution. +@xref{Binary notes-Linux}. -This means that you didn't start @code{mysqld} with @code{--bdb-no-recover} -and Berkeley DB found something wrong with its log files when it -tried to recover your databases. To be able to continue, you should -move away the old Berkeley DB log file from the database directory to -some other place, where you can later examine these. The log files are -named @file{log.0000000001}, where the number will increase over time. +When using LinuxThreads you will see a minimum of three processes +running. These are in fact threads. There will be one thread for the +LinuxThreads manager, one thread to handle connections, and one thread +to handle alarms and signals. -If you are running @code{mysqld} with BDB table support and @code{mysqld} core -dumps at start this could be because of some problems with the BDB -recover log. In this case you can try starting @code{mysqld} with -@code{--bdb-no-recover}. If this helps, then you should remove all -@file{log.*} files from the data directory and try starting @code{mysqld} -again. +Note that the Linux kernel and the LinuxThread library can by default +only have 1024 threads. This means that you can only have up to 1021 +connections to @strong{MySQL} on an unpatched system. The page +@uref{http://www.volano.com/linuxnotes.html} contains information how to +go around this limit. -If you get the following error, it means that some other program (or another -@code{mysqld} server) is already using the TCP/IP port or socket -@code{mysqld} is trying to use: +If you see a dead @code{mysqld} daemon process with @code{ps}, this usually +means that you have found a bug in @strong{MySQL} or you have a corrupted +table. @xref{Crashing}. -@example -Can't start server: Bind on TCP/IP port: Address already in use - or -Can't start server : Bind on unix socket... -@end example +To get a core dump on Linux if @code{mysqld} dies with a SIGSEGV +signal, you can start @code{mysqld} with the @code{--core-file} option. Note +that you also probably need to raise the @code{core file size} by adding +@code{ulimit -c 1000000} to @code{safe_mysqld} or starting @code{safe_mysqld} +with @code{--core-file-sizes=1000000}. @xref{safe_mysqld, , @code{safe_mysqld}}. -Use @code{ps} to make sure that you don't have another @code{mysqld} server -running. If you can't find another server running, you can try to execute -the command @code{telnet your-host-name tcp-ip-port-number} and press -@code{RETURN} a couple of times. If you don't get an error message like -@code{telnet: Unable to connect to remote host: Connection refused}, -something is using the TCP/IP port @code{mysqld} is trying to use. -See @ref{mysql_install_db} and @ref{Multiple servers}. +To get a core dump on Linux if @code{mysqld} dies with a SIGSEGV signal, you can +start @code{mysqld} with the @code{--core-file} option. Note that you also probably +need to raise the @code{core file size} by adding @code{ulimit -c 1000000} to +@code{safe_mysqld} or starting @code{safe_mysqld} with +@code{--core-file-sizes=1000000}. @xref{safe_mysqld, , @code{safe_mysqld}}. -If @code{mysqld} is currently running, you can find out what path settings -it is using by executing this command: +If you are linking your own @strong{MySQL} client and get the error: @example -shell> mysqladmin variables +ld.so.1: ./my: fatal: libmysqlclient.so.4: open failed: No such file or directory @end example -or +When executing them, the problem can be avoided by one of the following +methods: -@example -shell> mysqladmin -h 'your-host-name' variables -@end example +@itemize @bullet +@item +Link the client with the following flag (instead of @code{-Lpath}): +@code{-Wl,r/path-libmysqlclient.so}. -If @code{safe_mysqld} starts the server but you can't connect to it, -you should make sure you have an entry in @file{/etc/hosts} that looks like -this: +@item +Copy @code{libmysqclient.so} to @file{/usr/lib}. + +@tindex LD_RUN_PATH environment variable +@tindex environment variable, LD_RUN_PATH +@item +Add the pathname of the directory where @code{libmysqlclient.so} is located +to the @code{LD_RUN_PATH} environment variable before running your client. +@end itemize + +If you are using the Fujitsu compiler @code{(fcc / FCC)} you will have +some problems compiling @strong{MySQL} because the Linux header files are very +@code{gcc} oriented. + +The following @code{configure} line should work with @code{fcc/FCC}: @example -127.0.0.1 localhost +CC=fcc CFLAGS="-O -K fast -K lib -K omitfp -Kpreex -D_GNU_SOURCE -DCONST=const -DNO_STRTOLL_PROTO" CXX=FCC CXXFLAGS="-O -K fast -K lib -K omitfp -K preex --no_exceptions --no_rtti -D_GNU_SOURCE -DCONST=const -Dalloca=__builtin_alloca -DNO_STRTOLL_PROTO '-D_EXTERN_INLINE=static __inline'" ./configure --prefix=/usr/local/mysql --enable-assembler --with-mysqld-ldflags=-all-static --disable-shared --with-low-memory @end example -This problem occurs only on systems that don't have a working thread -library and for which @strong{MySQL} must be configured to use MIT-pthreads. +@menu +* Linux-x86:: Linux-x86 notes +* Linux-RedHat50:: RedHat 5.0 notes +* Linux-RedHat51:: RedHat 5.1 notes +* Linux-SPARC:: Linux-SPARC notes +* Linux-Alpha:: Linux-Alpha notes +* MKLinux:: MkLinux notes +* Qube2:: Qube2 Linux notes +* Linux-Ia64:: Linux-Ia64 notes +@end menu -If you can't get @code{mysqld} to start you can try to make a trace file -to find the problem. @xref{Making trace files}. +@node Linux-x86, Linux-RedHat50, Linux, Linux +@subsubsection Linux-x86 Notes -If you are using InnoDB tables, refer to the InnoDB-specific startup -options. @xref{InnoDB start}. +@strong{MySQL} requires @code{libc} Version 5.4.12 or newer. It's known to +work with @code{libc} 5.4.46. @code{glibc} Version 2.0.6 and later should +also work. There have been some problems with the @code{glibc} RPMs from +RedHat, so if you have problems, check whether or not there are any updates! +The @code{glibc} 2.0.7-19 and 2.0.7-29 RPMs are known to work. -If you are using BDB (Berkeley DB) tables, you should familiarize -yourself with the different BDB specific startup options. @xref{BDB start}. +On some older Linux distributions, @code{configure} may produce an error +like this: -@node Automatic start, Command-line options, Starting server, Post-installation -@subsection Starting and Stopping MySQL Automatically -@cindex starting, the server automatically -@cindex stopping, the server -@cindex server, starting and stopping +@example +Syntax error in sched.h. Change _P to __P in the /usr/include/sched.h file. +See the Installation chapter in the Reference Manual. +@end example -The @code{mysql.server} and @code{safe_mysqld} scripts can be used to start -the server automatically at system startup time. @code{mysql.server} can also -be used to stop the server. +Just do what the error message says and add an extra underscore to the +@code{_P} macro that has only one underscore, then try again. -The @code{mysql.server} script can be used to start or stop the server -by invoking it with @code{start} or @code{stop} arguments: +You may get some warnings when compiling; those shown below can be ignored: @example -shell> mysql.server start -shell> mysql.server stop +mysqld.cc -o objs-thread/mysqld.o +mysqld.cc: In function `void init_signals()': +mysqld.cc:315: warning: assignment of negative value `-1' to `long unsigned int' +mysqld.cc: In function `void * signal_hand(void *)': +mysqld.cc:346: warning: assignment of negative value `-1' to `long unsigned int' @end example -@code{mysql.server} can be found in the @file{share/mysql} directory -under the @strong{MySQL} installation directory or in the @file{support-files} -directory of the @strong{MySQL} source tree. +In Debian GNU/Linux, if you want @strong{MySQL} to start automatically when +the system boots, do the following: -Before @code{mysql.server} starts the server, it changes directory to -the @strong{MySQL} installation directory, then invokes @code{safe_mysqld}. -You might need to edit @code{mysql.server} if you have a binary distribution -that you've installed in a non-standard location. Modify it to @code{cd} -into the proper directory before it runs @code{safe_mysqld}. If you want the -server to run as some specific user, add an appropriate @code{user} line -to the @file{/etc/my.cnf} file, as shown later in this section. +@example +shell> cp support-files/mysql.server /etc/init.d/mysql.server +shell> /usr/sbin/update-rc.d mysql.server defaults 99 +@end example -@code{mysql.server stop} brings down the server by sending a signal to it. -You can take down the server manually by executing @code{mysqladmin shutdown}. +@code{mysql.server} can be found in the @file{share/mysql} directory +under the @strong{MySQL} installation directory or in the +@file{support-files} directory of the @strong{MySQL} source tree. -You might want to add these start and stop commands to the appropriate places -in your @file{/etc/rc*} files when you start using @strong{MySQL} for -production applications. Note that if you modify @code{mysql.server}, then -upgrade @strong{MySQL} sometime, your modified version will be overwritten, -so you should make a copy of your edited version that you can reinstall. +If @code{mysqld} always core dumps when it starts up, the problem may be that +you have an old @file{/lib/libc.a}. Try renaming it, then remove +@file{sql/mysqld} and do a new @code{make install} and try again. This +problem has been reported on some Slackware installations. RedHat Version 5.0 +also has a similar problem with some new @code{glibc} versions. +@xref{Linux-RedHat50}. + +If you get the following error when linking @code{mysqld}, +it means that your @file{libg++.a} is not installed correctly: -If your system uses @file{/etc/rc.local} to start external scripts, you -should append the following to it: @example -/bin/sh -c 'cd /usr/local/mysql ; ./bin/safe_mysqld --user=mysql &' +/usr/lib/libc.a(putc.o): In function `_IO_putc': +putc.o(.text+0x0): multiple definition of `_IO_putc' @end example -You can also add options for @code{mysql.server} in a global -@file{/etc/my.cnf} file. A typical @file{/etc/my.cnf} file might look like -this: +You can avoid using @file{libg++.a} by running @code{configure} like this: @example -[mysqld] -datadir=/usr/local/mysql/var -socket=/tmp/mysqld.sock -port=3306 -user=mysql - -[mysql.server] -basedir=/usr/local/mysql +shell> CXX=gcc ./configure @end example -The @code{mysql.server} script understands the following options: -@code{datadir}, @code{basedir}, and @code{pid-file}. - -The following table shows which option groups each of the startup scripts -read from option files: +@node Linux-RedHat50, Linux-RedHat51, Linux-x86, Linux +@subsubsection RedHat Version 5.0 Notes -@multitable @columnfractions .20 .80 -@item @strong{Script} @tab @strong{Option groups} -@item @code{mysqld} @tab @code{mysqld} and @code{server} -@item @code{mysql.server} @tab @code{mysql.server}, @code{mysqld}, and @code{server} -@item @code{safe_mysqld} @tab @code{mysql.server}, @code{mysqld}, and @code{server} -@end multitable +If you have any problems with @strong{MySQL} on RedHat, you should start by +upgrading @code{glibc} to the newest possible version! -@xref{Option files}. +If you install all the official RedHat patches (including +@code{glibc-2.0.7-19} and @code{glibc-devel-2.0.7-19}), both the +binary and source distributions of @strong{MySQL} should work without +any trouble! -@findex command-line options -@cindex options, command-line -@cindex mysqld options -@node Command-line options, Option files, Automatic start, Post-installation -@subsection mysqld Command-line Options +The updates are needed because there is a bug in @code{glibc} 2.0.5 in how +@code{pthread_key_create} variables are freed. With @code{glibc} 2.0.5, you +must use a statically linked @strong{MySQL} binary distribution. If you +want to compile from source, you must install the corrected version of +LinuxThreads from @uref{http://www.mysql.com/Downloads/Linux} or upgrade your +@code{glibc}. -@code{mysqld} accepts the following command-line options: +If you have an incorrect version of @code{glibc} or LinuxThreads, the symptom +is that @code{mysqld} crashes after each connection. For example, +@code{mysqladmin version} will crash @code{mysqld} when it finishes! -@table @code -@item --ansi -Use ANSI SQL syntax instead of MySQL syntax. @xref{ANSI mode}. +Another symptom of incorrect libraries is that @code{mysqld} crashes at +once when it starts. On some Linux systems, this can be fixed by configuring +like this: -@item -b, --basedir=path -Path to installation directory. All paths are usually resolved relative to -this. +@example +shell> ./configure --with-mysqld-ldflags=-all-static +@end example -@item --big-tables -Allow big result sets by saving all temporary sets on file. It solves -most 'table full' errors, but also slows down the queries where -in-memory tables would suffice. Since Version 3.23.2, @strong{MySQL} is -able to solve it automatically by using memory for small temporary -tables and switching to disk tables where necessary. +On Redhat Version 5.0, the easy way out is to install the @code{glibc} +2.0.7-19 RPM and run @code{configure} @strong{without} the +@code{--with-mysqld-ldflags=-all-static} option. -@item --bind-address=IP -IP address to bind to. +For the source distribution of @code{glibc} 2.0.7, a patch that is easy to +apply and is tested with @strong{MySQL} may be found at: -@item --character-sets-dir=path -Directory where character sets are. @xref{Character sets}. +@example +@uref{http://www.mysql.com/Downloads/Linux/glibc-2.0.7-total-patch.tar.gz} +@end example -@item --chroot=path -Chroot @code{mysqld} daemon during startup. Recommended security measure. It will -somewhat limit @code{LOAD DATA INFILE} and @code{SELECT ... INTO OUTFILE} -though. +If you experience crashes like these when you build @strong{MySQL}, you can +always download the newest binary version of @strong{MySQL}. This is +statically-linked to avoid library conflicts and should work on all Linux +systems! -@item --core-file -Write a core file if @code{mysqld} dies. For some systems you must also -specify @code{--core-file-size} to @code{safe_mysqld}. @xref{safe_mysqld, , -@code{safe_mysqld}}. +@strong{MySQL} comes with an internal debugger that can generate +trace files with a lot of information that can be used to find and solve a +wide range of different problems. +@xref{Debugging server}. -@item -h, --datadir=path -Path to the database root. +@node Linux-RedHat51, Linux-SPARC, Linux-RedHat50, Linux +@subsubsection RedHat Version 5.1 notes -@item --default-character-set=charset -Set the default character set. @xref{Character sets}. +The @code{glibc} of RedHat Version 5.1 (@code{glibc} 2.0.7-13) has a memory +leak, so to get a stable @strong{MySQL} version, you must upgrade @code{glibc}, +to 2.0.7-19, downgrade @code{glibc} or use a binary version of @code{mysqld}. +If you don't do this, you will encounter memory problems (out of memory, etc.). +The most common error in this case is: -@item --default-table-type=type -Set the default table type for tables. @xref{Table types}. +@example +Can't create a new thread (errno 11). If you are not out of available +memory, you can consult the manual for any possible OS dependent bug +@end example -@item --debug[...]= -If @strong{MySQL} is configured with @code{--with-debug}, you can use this -option to get a trace file of what @code{mysqld} is doing. -@xref{Making trace files}. +After you have upgraded to @code{glibc} 2.0.7-19, you can configure +@strong{MySQL} with dynamic linking (the default), but you @strong{cannot} +run @code{configure} with the @code{--with-mysqld-ldflags=-all-static} option +until you have installed @code{glibc} 2.0.7-19 from source! -@item --delay-key-write-for-all-tables -Don't flush key buffers between writes for any @code{MyISAM} table. -@xref{Server parameters}. +You can check which version of @code{glibc} you have with @code{rpm -q glibc}. -@item --enable-locking -Enable system locking. Note that if you use this option on a system -which a not fully working lockd() (as on Linux) you will easily get -mysqld to deadlock. +Another reason for the above error is if you try to use more threads +than your Linux kernel is configured for. In this case you should raise +the limits in @file{include/linux/tasks.h} and recompile your kernel! -@item -T, --exit-info -This is a bit mask of different flags one can use for debugging the -mysqld server; One should not use this option if one doesn't know -exactly what it does! +@node Linux-SPARC, Linux-Alpha, Linux-RedHat51, Linux +@subsubsection Linux-SPARC Notes -@item --flush -Flush all changes to disk after each SQL command. Normally @strong{MySQL} -only does a write of all changes to disk after each SQL command and lets -the operating system handle the syncing to disk. -@xref{Crashing}. +In some implementations, @code{readdir_r()} is broken. The symptom is that +@code{SHOW DATABASES} always returns an empty set. This can be fixed by +removing @code{HAVE_READDIR_R} from @file{config.h} after configuring and +before compiling. -@item -?, --help -Display short help and exit. +Some problems will require patching your Linux installation. The patch can +be found at +@uref{http://www.mysql.com/Downloads/patches/Linux-sparc-2.0.30.diff}. +This patch is against the Linux distribution @file{sparclinux-2.0.30.tar.gz} +that is available at @code{vger.rutgers.edu} (a version of Linux that was +never merged with the official 2.0.30). You must also install LinuxThreads +Version 0.6 or newer. -@item --init-file=file -Read SQL commands from this file at startup. +@node Linux-Alpha, MKLinux, Linux-SPARC, Linux +@subsubsection Linux-Alpha Notes -@item -L, --language=... -Client error messages in given language. May be given as a full path. -@xref{Languages}. +@strong{MySQL} Version 3.23.12 is the first @strong{MySQL} version that is +tested on Linux-Alpha. If you plan to use @strong{MySQL} on Linux-Alpha, +you should ensure that you have this version or newer. -@item -l, --log[=file] -Log connections and queries to file. @xref{Query log}. +We have tested @strong{MySQL} on Alpha with our benchmarks and test suite, +and it appears to work nicely. The main thing we haven't yet had time to +test is how things works with many concurrent users. -@item --log-isam[=file] -Log all ISAM/MyISAM changes to file (only used when debugging ISAM/MyISAM). +When we compiled the standard @strong{MySQL} binary we are using SuSE 6.4, +kernel 2.2.13-SMP, Compaq C compiler (V6.2-504) and Compaq C++ compiler +(V6.3-005) on a Comaq DS20 machine with an Alpha EV6 processor. -@item --log-slow-queries[=file] -Log all queries that have taken more than @code{long_query_time} seconds to -execute to file. @xref{Slow query log}. +You can find the above compilers at +@uref{http://www.support.compaq.com/alpha-tools}). By using these compilers, +instead of gcc, we get about 9-14 % better performance with @strong{MySQL}. -@item --log-update[=file] -Log updates to @code{file.#} where @code{#} is a unique number if not given. -@xref{Update log}. - -@item --log-long-format -Log some extra information to update log. If you are using -@code{--log-slow-queries} then queries that are not using indexes are logged -to the slow query log. - -@item --low-priority-updates -Table-modifying operations (@code{INSERT}/@code{DELETE}/@code{UPDATE}) -will have lower priority than selects. It can also be done via -@code{@{INSERT | REPLACE | UPDATE | DELETE@} LOW_PRIORITY ...} to lower -the priority of only one query, or by -@code{SET OPTION SQL_LOW_PRIORITY_UPDATES=1} to change the priority in one -thread. @xref{Table locking}. +Note that the configure line optimized the binary for the current CPU; This +means you can only use our binary if you have an Alpha EV6 processor. We also +compile statically to avoid library problems. -@item --memlock -Lock the @code{mysqld} process in memory. This works only if your -system supports the @code{mlockall()} system call (like Solaris). This -may help if you have a problem where the operating system is causing -@code{mysqld} to swap on disk. +@example +CC=ccc CFLAGS="-fast" CXX=cxx CXXFLAGS="-fast -noexceptions -nortti" ./configure --prefix=/usr/local/mysql --disable-shared --with-extra-charsets=complex --enable-thread-safe-client --with-mysqld-ldflags=-non_shared --with-client-ldflags=-non_shared +@end example -@item --myisam-recover [=option[,option...]]] where option is one of DEFAULT, BACKUP, FORCE or QUICK. -If this option is used, @code{mysqld} will on open check if the table is -marked as crashed or if if the table wasn't closed properly. (The last -option only works if you are running with @code{--skip-locking}). If this -is the case @code{mysqld} will run check on the table. If the table was corrupted, -@code{mysqld} will attempt to repair it. +If you want to use egcs the following configure line worked for us: -The following options affects how the repair works. +@example +CFLAGS="-O3 -fomit-frame-pointer" CXX=gcc CXXFLAGS="-O3 -fomit-frame-pointer -felide-constructors -fno-exceptions -fno-rtti" ./configure --prefix=/usr/local/mysql --disable-shared +@end example -@multitable @columnfractions .3 .7 -@item DEFAULT @tab The same as not giving any option to - @code{--myisam-recover}. -@item BACKUP @tab If the data table was changed during recover, save a - backup of the @file{table_name.MYD} data file as - @file{table_name-datetime.BAK}. -@item FORCE @tab Run recover even if we will loose more than one row - from the .MYD file. -@item QUICK @tab Don't check the rows in the table if there isn't any - delete blocks. -@end multitable +Some known problems when running @strong{MySQL} on Linux-Alpha: -Before a table is automatically repaired, @strong{MySQL} will add a note -about this in the error log. If you want to be able to recover from most -things without user intervention, you should use the options -@code{BACKUP,FORCE}. This will force a repair of a table even if some rows -would be deleted, but it will keep the old data file as a backup so that -you can later examine what happened. +@itemize @bullet +@item +Debugging threaded applications like @strong{MySQL} will not work with +@code{gdb 4.18}. You should download and use gdb 5.0 instead! -@item --pid-file=path -Path to pid file used by @code{safe_mysqld}. +@item +If you try linking @code{mysqld} statically when using @code{gcc}, the +resulting image will core dump at start. In other words, @strong{DON'T} +use @code{--with-mysqld-ldflags=-all-static} with @code{gcc}. +@end itemize -@item -P, --port=... -Port number to listen for TCP/IP connections. +@node MKLinux, Qube2, Linux-Alpha, Linux +@subsubsection MkLinux Notes -@item -o, --old-protocol -Use the 3.20 protocol for compatibility with some very old clients. -@xref{Upgrading-from-3.20}. +@strong{MySQL} should work on MkLinux with the newest @code{glibc} package +(tested with @code{glibc} 2.0.7). -@item --one-thread -Only use one thread (for debugging under Linux). @xref{Debugging server}. +@node Qube2, Linux-Ia64, MKLinux, Linux +@subsubsection Qube2 Linux Notes -@item -O, --set-variable var=option -Give a variable a value. @code{--help} lists variables. You can find a full -description for all variables in the @code{SHOW VARIABLES} section in this -manual. @xref{SHOW VARIABLES}. The tuning server parameters section includes -information of how to optimize these. @xref{Server parameters}. +To get @strong{MySQL} to work on Qube2, (Linux Mips), you need the +newest @code{glibc} libraries (@code{glibc-2.0.7-29C2} is known to +work). You must also use the @code{egcs} C++ compiler +(@code{egcs-1.0.2-9}, @code{gcc 2.95.2} or newer). -@item --safe-mode -Skip some optimize stages. Implies @code{--skip-delay-key-write}. +@node Linux-Ia64, , Qube2, Linux +@subsubsection Linux IA64 Notes -@item --safe-show-database -Don't show databases for which the user doesn't have any privileges. +To get @strong{MySQL} to compile on Linux Ia64, we had to do the following +(we assume that this will be easier when next gcc version for ia64 is +released). -@item --secure -IP numbers returned by the @code{gethostbyname()} system call are -checked to make sure they resolve back to the original hostname. This -makes it harder for someone on the outside to get access by pretending -to be another host. This option also adds some sanity checks of -hostnames. The option is turned off by default in @strong{MySQL} Version 3.21 -because sometimes it takes a long time to perform backward resolutions. -@strong{MySQL} Version 3.22 caches hostnames (unless @code{--skip-host-cache} -is used) and has this option enabled by default. +Using @code{gcc-2.9-final}: -@item --skip-concurrent-insert -Turn off the ability to select and insert at the same time on @code{MyISAM} -tables. (This is only to be used if you think you have found a bug in this -feature). +@example +CFLAGS="-O2" CXX=gcc CXXFLAGS="-O2 -felide-constructors -fno-exceptions -fno-rtti" ./configure --prefix=/usr/local/mysql --enable-assembler --with-mysqld-ldflags=-all-static --disable-shared --with-extra-charsets=complex +@end example -@item --skip-delay-key-write -Ignore the @code{delay_key_write} option for all tables. -@xref{Server parameters}. +After @code{make} you will get an error that @code{sql/opt_range.cc} +will not compile (internal compiler error). To fix this, go to the sql +directory and type @code{make} again. Copy the compile line, but change +-O2 to -O0. The file should now compile. -@item --skip-grant-tables -This option causes the server not to use the privilege system at all. This -gives everyone @emph{full access} to all databases! (You can tell a running -server to start using the grant tables again by executing @code{mysqladmin -flush-privileges} or @code{mysqladmin reload}.) +Now you can do: -@item --skip-host-cache -Never use host name cache for faster name-ip resolution, but query DNS server -on every connect instead. @xref{DNS}. +@example +cd .. +make +make_install +@end example -@item --skip-locking -Don't use system locking. To use @code{isamchk} or @code{myisamchk} you must -shut down the server. @xref{Stability}. Note that in @strong{MySQL} Version -3.23 you can use @code{REPAIR} and @code{CHECK} to repair/check @code{MyISAM} -tables. +and @strong{mysqld} should be ready to run. -@item --skip-name-resolve -Hostnames are not resolved. All @code{Host} column values in the grant tables -must be IP numbers or @code{localhost}. @xref{DNS}. +@node Alpha-DEC-UNIX, Alpha-DEC-OSF1, Linux, Source install system issues +@subsection Alpha-DEC-UNIX Notes (Tru64) -@item --skip-networking -Don't listen for TCP/IP connections at all. All interaction with -@code{mysqld} must be made via Unix sockets. This option is highly -recommended for systems where only local requests are allowed. @xref{DNS}. +If you are using egcs 1.1.2 on Digital Unix, you should upgrade to gcc +2.95.2, as egcs on DEC has some serious bugs! -@item --skip-new -Don't use new, possible wrong routines. Implies @code{--skip-delay-key-write}. -This will also set default table type to @code{ISAM}. @xref{ISAM}. +When compiling threaded programs under Digital Unix, the documentation +recommends using the @code{-pthread} option for @code{cc} and @code{cxx} and +the libraries @code{-lmach -lexc} (in addition to @code{-lpthread}). You +should run @code{configure} something like this: -@item --skip-symlink -Don't delete or rename files that a symlinked file in the data directory -points to. +@example +CC="cc -pthread" CXX="cxx -pthread -O" \ +./configure --with-named-thread-libs="-lpthread -lmach -lexc -lc" +@end example -@item --skip-safemalloc -If @strong{MySQL} is configured with @code{--with-debug=full}, all programs -will check the memory for overruns for every memory allocation and memory -freeing. As this checking is very slow, you can avoid this, when you don't -need memory checking, by using this option. +When compiling @code{mysqld}, you may see a couple of warnings like this: -@item --skip-show-database -Don't allow 'SHOW DATABASE' commands, unless the user has @strong{process} -privilege. +@example +mysqld.cc: In function void handle_connections()': +mysqld.cc:626: passing long unsigned int *' as argument 3 of +accept(int,sockadddr *, int *)' +@end example -@item --skip-stack-trace -Don't write stack traces. This option is useful when you are running -@code{mysqld} under a debugger. @xref{Debugging server}. +You can safely ignore these warnings. They occur because @code{configure} +can detect only errors, not warnings. -@item --skip-thread-priority -Disable using thread priorities for faster response time. +If you start the server directly from the command line, you may have problems +with it dying when you log out. (When you log out, your outstanding processes +receive a @code{SIGHUP} signal.) If so, try starting the server like this: -@item --socket=path -Socket file to use for local connections instead of default -@code{/tmp/mysql.sock}. +@example +shell> nohup mysqld [options] & +@end example -@item transaction-isolation= @{ READ-UNCOMMITTED | READ-COMMITTED | REPEATABLE-READ | SERIALIZABLE @} -Sets the default transaction isolation level. @xref{SET TRANSACTION}. +@code{nohup} causes the command following it to ignore any @code{SIGHUP} +signal sent from the terminal. Alternatively, start the server by running +@code{safe_mysqld}, which invokes @code{mysqld} using @code{nohup} for you. +@xref{safe_mysqld, , @code{safe_mysqld}}. -@item -t, --tmpdir=path -Path for temporary files. It may be useful if your default @code{/tmp} -directory resides on a partition too small to hold temporary tables. +If you get a problem when compiling mysys/get_opt.c, just remove the +line #define _NO_PROTO from the start of that file! -@item -u, --user=user_name -Run @code{mysqld} daemon as user @code{user_name}. This option is -@emph{mandatory} when starting @code{mysqld} as root. +If you are using Compac's CC compiler, the following configure line should +work: -@item -V, --version -Output version information and exit. +@example +CC="cc -pthread" +CFLAGS="-O4 -ansi_alias -ansi_args -fast -inline speed -speculate all -arch host" +CXX="cxx -pthread" +CXXFLAGS="-O4 -ansi_alias -ansi_args -fast -inline speed -speculate all -arch host" +export CC CFLAGS CXX CXXFLAGS +./configure \ +--prefix=/usr/local/mysql \ +--with-low-memory \ +--enable-large-files \ +--enable-shared=yes \ +--with-named-thread-libs="-lpthread -lmach -lexc -lc" +gnumake +@end example -@item -W, --warnings -Print out warnings like @code{Aborted connection...} to the @code{.err} file. -@xref{Communication errors}. -@end table +If you get a problem with libtool, when compiling with shared libraries +as above, when linking @code{mysql}, you should be able to get around +this by issuing: -@cindex default options -@cindex option files -@cindex creating, default startup options -@cindex startup options, default -@node Option files, , Command-line options, Post-installation -@subsection Option Files +@example +cd mysql +/bin/sh ../libtool --mode=link cxx -pthread -O3 -DDBUG_OFF \ +-O4 -ansi_alias -ansi_args -fast -inline speed \ +-speculate all \ -arch host -DUNDEF_HAVE_GETHOSTBYNAME_R \ +-o mysql mysql.o readline.o sql_string.o completion_hash.o \ +../readline/libreadline.a -lcurses \ +../libmysql/.libs/libmysqlclient.so -lm +cd .. +gnumake +gnumake install +scripts/mysql_install_db +@end example -@strong{MySQL} can, since Version 3.22, read default startup options for the -server and for clients from option files. +@node Alpha-DEC-OSF1, SGI-Irix, Alpha-DEC-UNIX, Source install system issues +@subsection Alpha-DEC-OSF1 Notes -@strong{MySQL} reads default options from the following files on Unix: +If you have problems compiling and have DEC @code{CC} and @code{gcc} +installed, try running @code{configure} like this: -@tindex .my.cnf file -@multitable @columnfractions .3 .7 -@item @strong{Filename} @tab @strong{Purpose} -@item @code{/etc/my.cnf} @tab Global options -@item @code{DATADIR/my.cnf} @tab Server-specific options -@item @code{defaults-extra-file} @tab The file specified with --defaults-extra-file=# -@item @code{~/.my.cnf} @tab User-specific options -@end multitable +@example +CC=cc CFLAGS=-O CXX=gcc CXXFLAGS=-O3 \ +./configure --prefix=/usr/local/mysql +@end example -@code{DATADIR} is the @strong{MySQL} data directory (typically -@file{/usr/local/mysql/data} for a binary installation or -@file{/usr/local/var} for a source installation). Note that this is the -directory that was specified at configuration time, not the one specified -with @code{--datadir} when @code{mysqld} starts up! (@code{--datadir} has no -effect on where the server looks for option files, because it looks for them -before it processes any command-line arguments.) +If you get problems with the @file{c_asm.h} file, you can create and use +a 'dummy' @file{c_asm.h} file with: -@strong{MySQL} reads default options from the following files on Windows: +@example +touch include/c_asm.h +CC=gcc CFLAGS=-I./include \ +CXX=gcc CXXFLAGS=-O3 \ +./configure --prefix=/usr/local/mysql +@end example -@multitable @columnfractions .3 .7 -@item @strong{Filename} @tab @strong{Purpose} -@item @code{windows-system-directory\my.ini} @tab Global options -@item @code{C:\my.cnf} @tab Global options -@item @code{C:\mysql\data\my.cnf} @tab Server-specific options -@end multitable +Note that the following problems with the @code{ld} program can be fixed +by downloading the latest DEC (Compaq) patch kit from: +@uref{http://ftp.support.compaq.com/public/unix/}. -Note that on Windows, you should specify all paths with @code{/} instead of -@code{\}. If you use @code{\}, you need to specify this twice, as -@code{\} is the escape character in @strong{MySQL}. +On OSF1 V4.0D and compiler "DEC C V5.6-071 on Digital Unix V4.0 (Rev. 878)" +the compiler had some strange behavior (undefined @code{asm} symbols). +@code{/bin/ld} also appears to be broken (problems with @code{_exit +undefined} errors occuring while linking @code{mysqld}). On this system, we +have managed to compile @strong{MySQL} with the following @code{configure} +line, after replacing @code{/bin/ld} with the version from OSF 4.0C: -@cindex Environment variables -@strong{MySQL} tries to read option files in the order listed above. If -multiple option files exist, an option specified in a file read later takes -precedence over the same option specified in a file read earlier. Options -specified on the command line take precedence over options specified in any -option file. Some options can be specified using environment variables. -Options specified on the command line or in option files take precedence over -environment variable values. @xref{Environment variables}. +@example +CC=gcc CXX=gcc CXXFLAGS=-O3 ./configure --prefix=/usr/local/mysql +@end example -The following programs support option files: @code{mysql}, -@code{mysqladmin}, @code{mysqld}, @code{mysqldump}, @code{mysqlimport}, -@code{mysql.server}, @code{myisamchk}, and @code{myisampack}. +With the Digital compiler "C++ V6.1-029", the following should work: -You can use option files to specify any long option that a program supports! -Run the program with @code{--help} to get a list of available options. +@example +CC=cc -pthread +CFLAGS=-O4 -ansi_alias -ansi_args -fast -inline speed -speculate all -arch host +CXX=cxx -pthread +CXXFLAGS=-O4 -ansi_alias -ansi_args -fast -inline speed -speculate all -arch host -noexceptions -nortti +export CC CFLAGS CXX CXXFLAGS +./configure --prefix=/usr/mysql/mysql --with-mysqld-ldflags=-all-static --disable-shared --with-named-thread-libs="-lmach -lexc -lc" +@end example -An option file can contain lines of the following forms: +In some versions of OSF1, the @code{alloca()} function is broken. Fix +this by removing the line in @file{config.h} that defines @code{'HAVE_ALLOCA'}. -@table @code -@item #comment -Comment lines start with @samp{#} or @samp{;}. Empty lines are ignored. +The @code{alloca()} function also may have an incorrect prototype in +@code{/usr/include/alloca.h}. This warning resulting from this can be ignored. -@item [group] -@code{group} is the name of the program or group for which you want to set -options. After a group line, any @code{option} or @code{set-variable} lines -apply to the named group until the end of the option file or another group -line is given. +@code{configure} will use the following thread libraries automatically: +@code{--with-named-thread-libs="-lpthread -lmach -lexc -lc"}. -@item option -This is equivalent to @code{--option} on the command line. +When using @code{gcc}, you can also try running @code{configure} like this: -@item option=value -This is equivalent to @code{--option=value} on the command line. +@example +shell> CFLAGS=-D_PTHREAD_USE_D4 CXX=gcc CXXFLAGS=-O3 ./configure .... +@end example -@item set-variable = variable=value -This is equivalent to @code{--set-variable variable=value} on the command line. -This syntax must be used to set a @code{mysqld} variable. -@end table +If you have problems with signals (@strong{MySQL} dies unexpectedly +under high load), you may have found an OS bug with threads and +signals. In this case you can tell @strong{MySQL} not to use signals by +configuring with: -The @code{client} group allows you to specify options that apply to all -@strong{MySQL} clients (not @code{mysqld}). This is the perfect group to use -to specify the password you use to connect to the server. (But make -sure the option file is readable and writable only by yourself.) +@example +shell> CFLAGS=-DDONT_USE_THR_ALARM \ + CXXFLAGS=-DDONT_USE_THR_ALARM \ + ./configure ... +@end example -Note that for options and values, all leading and trailing blanks are -automatically deleted. You may use the escape sequences @samp{\b}, -@samp{\t}, @samp{\n}, @samp{\r}, @samp{\\}, and @samp{\s} in your value string -(@samp{\s} == blank). +This doesn't affect the performance of @strong{MySQL}, but has the side +effect that you can't kill clients that are ``sleeping'' on a connection with +@code{mysqladmin kill} or @code{mysqladmin shutdown}. Instead, the client +will die when it issues its next command. -Here is a typical global option file: +With @code{gcc} 2.95.2, you will probably run into the following compile error: @example -[client] -port=3306 -socket=/tmp/mysql.sock - -[mysqld] -port=3306 -socket=/tmp/mysql.sock -set-variable = key_buffer_size=16M -set-variable = max_allowed_packet=1M - -[mysqldump] -quick +sql_acl.cc:1456: Internal compiler error in `scan_region', at except.c:2566 +Please submit a full bug report. @end example -Here is typical user option file: - -@example -[client] -# The following password will be sent to all standard MySQL clients -password=my_password +To fix this you should change to the @code{sql} directory and do a ``cut +and paste'' of the last @code{gcc} line, but change @code{-O3} to +@code{-O0} (or add @code{-O0} immediately after @code{gcc} if you don't +have any @code{-O} option on your compile line.) After this is done you +can just change back to the top-level directly and run @code{make} +again. -[mysql] -no-auto-rehash -set-variable = connect_timeout=2 +@node SGI-Irix, FreeBSD, Alpha-DEC-OSF1, Source install system issues +@subsection SGI-Irix Notes -[mysqlhotcopy] -interactive-timeout +If you are using Irix Version 6.5.3 or newer @code{mysqld} will only be able to +create threads if you run it as a user with @code{CAP_SCHED_MGT} +privileges (like @code{root}) or give the @code{mysqld} server this privilege +with the following shell command: +@example +shell> chcap "CAP_SCHED_MGT+epi" /opt/mysql/libexec/mysqld @end example -@tindex .my.cnf file -If you have a source distribution, you will find sample configuration -files named @file{my-xxxx.cnf} in the @file{support-files} directory. -If you have a binary distribution, look in the @file{DIR/support-files} -directory, where @code{DIR} is the pathname to the @strong{MySQL} -installation directory (typically @file{/usr/local/mysql}). Currently -there are sample configuration files for small, medium, large, and very -large systems. You can copy @file{my-xxxx.cnf} to your home directory -(rename the copy to @file{.my.cnf}) to experiment with this. +You may have to undefine some things in @file{config.h} after running +@code{configure} and before compiling. -All @strong{MySQL} clients that support option files support the -following options: +In some Irix implementations, the @code{alloca()} function is broken. If the +@code{mysqld} server dies on some @code{SELECT} statements, remove the lines +from @file{config.h} that define @code{HAVE_ALLOC} and @code{HAVE_ALLOCA_H}. +If @code{mysqladmin create} doesn't work, remove the line from @file{config.h} +that defines @code{HAVE_READDIR_R}. You may have to remove the +@code{HAVE_TERM_H} line as well. -@multitable @columnfractions .40 .60 -@item --no-defaults @tab Don't read any option files. -@item --print-defaults @tab Print the program name and all options that it will get. -@item --defaults-file=full-path-to-default-file @tab Only use the given configuration file. -@item --defaults-extra-file=full-path-to-default-file @tab Read this configuration file after the global configuration file but before the user configuration file. -@end multitable +SGI recommends that you install all of the patches on this page as a set: +http://support.sgi.com/surfzone/patches/patchset/6.2_indigo.rps.html -Note that the above options must be first on the command line to work! -@code{--print-defaults} may however be used directly after the -@code{--defaults-xxx-file} commands. +At the very minimum, you should install the latest kernel rollup, the +latest @code{rld} rollup, and the latest @code{libc} rollup. -Note for developers: Option file handling is implemented simply by -processing all matching options (that is, options in the appropriate group) -before any command-line arguments. This works nicely for programs that use -the last instance of an option that is specified multiple times. If you have -an old program that handles multiply-specified options this way but doesn't -read option files, you need add only two lines to give it that capability. -Check the source code of any of the standard @strong{MySQL} clients to see -how to do this. +You definitely need all the POSIX patches on this page, for pthreads support: -In shell scripts you can use the @file{my_print_defaults} command to parse the -config files: +@uref{http://support.sgi.com/surfzone/patches/patchset/6.2_posix.rps.html} -@example +If you get the something like the following error when compiling +@file{mysql.cc}: -shell> my_print_defaults client mysql ---port=3306 ---socket=/tmp/mysql.sock ---no-auto-rehash +@example +"/usr/include/curses.h", line 82: error(1084): invalid combination of type @end example -The above output contains all options for the groups 'client' and 'mysql'. +Type the following in the top-level directory of your @strong{MySQL} source +tree: -@node Installing many servers, Upgrade, Post-installation, Installing -@section Installing Many Servers on the Same Machine -@cindex post-install, many servers -@cindex Installing many servers -@cindex Starting many servers +@example +shell> extra/replace bool curses_bool < /usr/include/curses.h > include/curses.h +shell> make +@end example -In some cases you may want to have many different @code{mysqld} daemons -(servers) running on the same machine. You may for example want to run -a new version of @strong{MySQL} for testing together with an old version -that is in production. Another case is when you want to give different -users access to different @code{mysqld} servers that they manage themselves. +There have also been reports of scheduling problems. If only one thread is +running, things go slow. Avoid this by starting another client. This may +lead to a 2-to-10-fold increase in execution speed thereafter for the other +thread. This is a poorly understood problem with Irix threads; you may have +to improvise to find solutions until this can be fixed. -One way to get a new server running is by starting it with a different -socket and port as follows: +If you are compiling with @code{gcc}, you can use the following +@code{configure} command: -@tindex MYSQL_UNIX_PORT environment variable -@tindex MYSQL_TCP_PORT environment variable -@tindex Environment variable, MYSQL_UNIX_PORT -@tindex Environment variable, MYSQL_TCP_PORT @example -shell> MYSQL_UNIX_PORT=/tmp/mysqld-new.sock -shell> MYSQL_TCP_PORT=3307 -shell> export MYSQL_UNIX_PORT MYSQL_TCP_PORT -shell> scripts/mysql_install_db -shell> bin/safe_mysqld & +CC=gcc CXX=gcc CXXFLAGS=-O3 \ +./configure --prefix=/usr/local/mysql --with-thread-safe-client --with-named-thread-libs=-lpthread @end example -The environment variables appendix includes a list of other environment -variables you can use to affect @code{mysqld}. @xref{Environment variables}. +On Irix 6.5.11 with native Irix C and C++ compilers ver. 7.3.1.2, the +following is reported to work -The above is the quick and dirty way that one commonly uses for testing. -The nice thing with this is that all connections you do in the above shell -will automatically be directed to the new running server! +@example +CC=cc CXX=CC CFLAGS='-O3 -n32 -TARG:platform=IP22 -I/usr/local/include \ +-L/usr/local/lib' CXXFLAGS='-O3 -n32 -TARG:platform=IP22 \ +-I/usr/local/include -L/usr/local/lib' ./configure --prefix=/usr/local/mysql \ +--with-berkeley-db --with-innodb \ +--with-libwrap=/usr/local --with-named-curses-libs=/usr/local/lib/libncurses.a +@end example -If you need to do this more permanently, you should create an option -file for each server. @xref{Option files}. In your startup script that -is executed at boot time (mysql.server?) you should specify for both -servers: +@node FreeBSD, NetBSD, SGI-Irix, Source install system issues +@subsection FreeBSD Notes -@code{safe_mysqld --default-file=path-to-option-file} +FreeBSD 3.x is recommended for running @strong{MySQL} since the thread package +is much more integrated. -At least the following options should be different per server: +The easiest and therefor the preferred way to install is to use the +mysql-server and mysql-client ports available on @uref{http://www.freebsd.org}. -@table @code -@item port=# -@item socket=path -@item pid-file=path -@end table +Using these gives you: +@itemize @bullet +@item +A working @strong{MySQL} with all optimizations known to work on your version +of FreeBSD enabled. -The following options should be different, if they are used: +@item +Automatic configuration and build. -@table @code -@item log=path -@item log-bin=path -@item log-update=path -@item log-isam=path -@item bdb-logdir=path -@end table +@item +Startup scripts installed in /usr/local/etc/rc.d. -If you want more performance, you can also specify the following differently: +@item +Ability to see which files that are installed with pkg_info -L. And to +remove them all with pkg_delete if you no longer want @strong{MySQL} on that +machine. +@end itemize -@table @code -@item tmpdir=path -@item bdb-tmpdir=path -@end table +It is recommended you use MIT-pthreads on FreeBSD 2.x and native threads on +Versions 3 and up. It is possible to run with native threads on some late +2.2.x versions but you may encounter problems shutting down @code{mysqld}. -@xref{Command-line options}. +The @strong{MYSQL} Makefiles require GNU make (@code{gmake}) to work. +If you want to compile @strong{MYSQL} you need to install GNU make +first. -If you are installing binary @strong{MySQL} versions (.tar files) and -start them with @code{./bin/safe_mysqld} then in most cases the only -option you need to add/change is the @code{socket} and @code{port} -argument to @code{safe_mysqld}. +Be sure to have your name resolver setup correct. Otherwise you may +experience resolver delays or failures when connecting to @code{mysqld}. -@node Upgrade, , Installing many servers, Installing -@section Upgrading/Downgrading MySQL -@cindex upgrading -@cindex downgrading +Make sure that the @code{localhost} entry in the @file{/etc/hosts} file is +correct (otherwise you will have problems connecting to the database). The +@file{/etc/hosts} file should start with a line: -You can always move the @strong{MySQL} form and data files between -different versions on the same architecture as long as you have the same -base version of @strong{MySQL}. The current base version is -3. If you change the character set when running @strong{MySQL} (which may -also change the sort order), you must run @code{myisamchk -r -q} on all -tables. Otherwise your indexes may not be ordered correctly. +@example +127.0.0.1 localhost localhost.your.domain +@end example -If you are afraid of new versions, you can always rename your old -@code{mysqld} to something like @code{mysqld}-'old-version-number'. If -your new @code{mysqld} then does something unexpected, you can simply shut it -down and restart with your old @code{mysqld}! +If you notice that @code{configure} will use MIT-pthreads, you should read +the MIT-pthreads notes. @xref{MIT-pthreads}. -When you do an upgrade you should also back up your old databases, of course. +If you get an error from @code{make install} that it can't find +@file{/usr/include/pthreads}, @code{configure} didn't detect that you need +MIT-pthreads. This is fixed by executing these commands: -If after an upgrade, you experience problems with recompiled client programs, -like @code{Commands out of sync} or unexpected core dumps, you probably have -used an old header or library file when compiling your programs. In this -case you should check the date for your @file{mysql.h} file and -@file{libmysqlclient.a} library to verify that they are from the new -@strong{MySQL} distribution. If not, please recompile your programs! +@example +shell> rm config.cache +shell> ./configure --with-mit-threads +@end example -If you get some problems that the new @code{mysqld} server doesn't want to -start or that you can't connect without a password, check that you don't -have some old @file{my.cnf} file from your old installation! You can -check this with: @code{program-name --print-defaults}. If this outputs -anything other than the program name, you have an active @code{my.cnf} -file that will affect things! +FreeBSD is also known to have a very low default file handle limit. +@xref{Not enough file handles}. Uncomment the ulimit -n section in +safe_mysqld or raise the limits for the @code{mysqld} user in /etc/login.conf +(and rebuild it with cap_mkdb /etc/login.conf). Also be sure you set the +appropriate class for this user in the password file if you are not +using the default (use: chpass mysqld-user-name). @xref{safe_mysqld, , +@code{safe_mysqld}}. -It is a good idea to rebuild and reinstall the @code{Msql-Mysql-modules} -distribution whenever you install a new release of @strong{MySQL}, -particularly if you notice symptoms such as all your @code{DBI} scripts -dumping core after you upgrade @strong{MySQL}. +If you get problems with the current date in @strong{MySQL}, setting the +@code{TZ} variable will probably help. @xref{Environment variables}. + +To get a secure and stable system you should only use FreeBSD kernels +that are marked @code{-STABLE}. + +@node NetBSD, OpenBSD, FreeBSD, Source install system issues +@subsection NetBSD notes + +To compile on NetBSD you need GNU @code{make}. Otherwise the compile will +crash when @code{make} tries to run @code{lint} on C++ files. + +@node OpenBSD, BSDI, NetBSD, Source install system issues +@subsection OpenBSD Notes @menu -* Upgrading-from-3.22:: Upgrading from a 3.22 version to 3.23 -* Upgrading-from-3.21:: Upgrading from a 3.21 version to 3.22 -* Upgrading-from-3.20:: Upgrading from a 3.20 version to 3.21 -* Upgrading-to-arch:: Upgrading to another architecture +* OpenBSD 2.5:: OpenBSD 2.5 Notes +* OpenBSD 2.8:: OpenBSD 2.8 Notes @end menu -@cindex compatibility, between MySQL versions -@cindex upgrading, 3.22 to 3.23 -@node Upgrading-from-3.22, Upgrading-from-3.21, Upgrade, Upgrade -@subsection Upgrading From Version 3.22 to Version 3.23 +@node OpenBSD 2.5, OpenBSD 2.8, OpenBSD, OpenBSD +@subsubsection OpenBSD 2.5 Notes -@strong{MySQL} Version 3.23 supports tables of the new @code{MyISAM} type and -the old @code{ISAM} type. You don't have to convert your old tables to -use these with Version 3.23. By default, all new tables will be created with -type @code{MyISAM} (unless you start @code{mysqld} with the -@code{--default-table-type=isam} option). You can change an @code{ISAM} -table to a @code{MyISAM} table with @code{ALTER TABLE table_name TYPE=MyISAM} -or the Perl script @code{mysql_convert_table_format}. +On OpenBSD Version 2.5, you can compile @strong{MySQL} with native threads +with the following options: -Version 3.22 and 3.21 clients will work without any problems with a Version -3.23 server. +@example +CFLAGS=-pthread CXXFLAGS=-pthread ./configure --with-mit-threads=no +@end example -The following lists tell what you have to watch out for when upgrading to -Version 3.23: +@node OpenBSD 2.8, , OpenBSD 2.5, OpenBSD +@subsubsection OpenBSD 2.8 Notes -@itemize @bullet -@item -All tables that uses the @code{tis620} character set must be fixed -with @code{myisamchk -r} or @code{REPAIR TABLE}. -@item -If you do a @code{DROP DATABASE} on a symbolic linked database, both the -link and the original database is deleted. (This didn't happen in 3.22 -because configure didn't detect the @code{readlink} system call). -@item -@code{OPTIMIZE TABLE} now only works for @strong{MyISAM} tables. -For other table types, you can use @code{ALTER TABLE} to optimize the table. -During @code{OPTIMIZE TABLE} the table is now locked from other threads. -@item -The @strong{MySQL} client @code{mysql} is now by default started with the -option @code{--no-named-commands (-g)}. This option can be disabled with -@code{--enable-named-commands (-G)}. This may cause incompatibility problems in -some cases, for example in SQL scripts that use named commands without a -semicolon! Long format commands still work from the first line. -@item -If you are using the @code{german} character sort order, you must repair -all your tables with @code{isamchk -r}, as we have made some changes in -the sort order! -@item The default return type of @code{IF} will now depend on both arguments -and not only the first argument. -@item @code{AUTO_INCREMENT} will not work with negative numbers. The reason -for this is that negative numbers caused problems when wrapping from -1 to 0. -@code{AUTO_INCREMENT} is now for MyISAM tables handled at a lower level and -is much faster than before. For MyISAM tables old numbers are also not reused -anymore, even if you delete some rows from the table. -@item @code{CASE}, @code{DELAYED}, @code{ELSE}, @code{END}, @code{FULLTEXT}, @code{INNER}, @code{RIGHT}, @code{THEN} and @code{WHEN} are now reserved words. -@item @code{FLOAT(X)} is now a true floating-point type and not a value with -a fixed number of decimals. -@item When declaring @code{DECIMAL(length,dec)} the length argument no -longer includes a place for the sign or the decimal point. -@item A @code{TIME} string must now be of one of the following formats: -@code{[[[DAYS] [H]H:]MM:]SS[.fraction]} or -@code{[[[[[H]H]H]H]MM]SS[.fraction]} -@item @code{LIKE} now compares strings using the same character -comparison rules as @code{'='}. If you require the old behavior, you -can compile @strong{MySQL} with the @code{CXXFLAGS=-DLIKE_CMP_TOUPPER} -flag. -@item @code{REGEXP} is now case insensitive for normal (not binary) strings. -@item When you check/repair tables you should use @code{CHECK TABLE} -or @code{myisamchk} for @code{MyISAM} tables (@code{.MYI}) and -@code{isamchk} for ISAM (@code{.ISM}) tables. -@item If you want your @code{mysqldump} files to be compatible between -@strong{MySQL} Version 3.22 and Version 3.23, you should not use the -@code{--opt} or @code{--full} option to @code{mysqldump}. -@item Check all your calls to @code{DATE_FORMAT()} to make sure there is a -@samp{%} before each format character. (Later @strong{MySQL} Version 3.22 -did allow this syntax.) -@item -@code{mysql_fetch_fields_direct} is now a function (it was a macro) and -it returns a pointer to a @code{MYSQL_FIELD} instead of a -@code{MYSQL_FIELD}. -@item -@code{mysql_num_fields()} can no longer be used on a @code{MYSQL*} object (it's -now a function that takes @code{MYSQL_RES*} as an argument. You should now -use @code{mysql_field_count()} instead. -@item -In @strong{MySQL} Version 3.22, the output of @code{SELECT DISTINCT ...} was -almost always sorted. In Version 3.23, you must use @code{GROUP BY} or -@code{ORDER BY} to obtain sorted output. -@item -@code{SUM()} now returns @code{NULL}, instead of 0, if there is no matching -rows. This is according to ANSI SQL. -@item An @code{AND} or @code{OR} with @code{NULL} values will now return -@code{NULL} instead of 0. This mostly affects queries that use @code{NOT} -on an @code{AND/OR} expression as @code{NOT NULL} = @code{NULL}. -@code{LPAD()} and @code{RPAD()} will shorten the result string if it's longer -than the length argument. -@end itemize +Our users have reported that OpenBSD 2.8 has a threading bug which causes +problems with @strong{MySQL}. The OpenBSD Developers have fixed the problem, +but as of January 25th, 2001, it's only available in the ``-current'' branch. +The symptoms of this threading bug are: slow response, high load, high CPU +usage, and crashes. -@cindex compatibility, between MySQL versions -@node Upgrading-from-3.21, Upgrading-from-3.20, Upgrading-from-3.22, Upgrade -@subsection Upgrading from Version 3.21 to Version 3.22 -@cindex upgrading, 3.21 to 3.22 +@node BSDI, SCO, OpenBSD, Source install system issues +@subsection BSD/OS Notes -Nothing that affects compatibility has changed between Version 3.21 and 3.22. -The only pitfall is that new tables that are created with @code{DATE} type -columns will use the new way to store the date. You can't access these new -fields from an old version of @code{mysqld}. +@menu +* BSDI2:: BSD/OS 2.x notes +* BSDI3:: BSD/OS 3.x notes +* BSDI4:: BSD/OS 4.x notes +@end menu -After installing @strong{MySQL} Version 3.22, you should start the new server -and then run the @code{mysql_fix_privilege_tables} script. This will add the -new privileges that you need to use the @code{GRANT} command. If you forget -this, you will get @code{Access denied} when you try to use @code{ALTER -TABLE}, @code{CREATE INDEX}, or @code{DROP INDEX}. If your @strong{MySQL} root -user requires a password, you should give this as an argument to -@code{mysql_fix_privilege_tables}. +@node BSDI2, BSDI3, BSDI, BSDI +@subsubsection BSD/OS Version 2.x Notes -The C API interface to @code{mysql_real_connect()} has changed. If you have -an old client program that calls this function, you must place a @code{0} for -the new @code{db} argument (or recode the client to send the @code{db} -element for faster connections). You must also call @code{mysql_init()} -before calling @code{mysql_real_connect()}! This change was done to allow -the new @code{mysql_options()} function to save options in the @code{MYSQL} -handler structure. +If you get the following error when compiling @strong{MySQL}, your +@code{ulimit} value for virtual memory is too low: -The @code{mysqld} variable @code{key_buffer} has changed names to -@code{key_buffer_size}, but you can still use the old name in your -startup files. +@example +item_func.h: In method `Item_func_ge::Item_func_ge(const Item_func_ge &)': +item_func.h:28: virtual memory exhausted +make[2]: *** [item_func.o] Error 1 +@end example -@node Upgrading-from-3.20, Upgrading-to-arch, Upgrading-from-3.21, Upgrade -@subsection Upgrading from Version 3.20 to Version 3.21 -@cindex upgrading, 3.20 to 3.21 +Try using @code{ulimit -v 80000} and run @code{make} again. If this +doesn't work and you are using @code{bash}, try switching to @code{csh} +or @code{sh}; some BSDI users have reported problems with @code{bash} +and @code{ulimit}. -If you are running a version older than Version 3.20.28 and want to -switch to Version 3.21, you need to do the following: +If you are using @code{gcc}, you may also use have to use the +@code{--with-low-memory} flag for @code{configure} to be able to compile +@file{sql_yacc.cc}. -You can start the @code{mysqld} Version 3.21 server with @code{safe_mysqld ---old-protocol} to use it with clients from a Version 3.20 distribution. -In this case, the new client function @code{mysql_errno()} will not -return any server error, only @code{CR_UNKNOWN_ERROR} (but it -works for client errors), and the server uses the old @code{password()} -checking rather than the new one. +If you get problems with the current date in @strong{MySQL}, setting the +@code{TZ} variable will probably help. @xref{Environment variables}. -If you are @strong{NOT} using the @code{--old-protocol} option to -@code{mysqld}, you will need to make the following changes: +@node BSDI3, BSDI4, BSDI2, BSDI +@subsubsection BSD/OS Version 3.x Notes -@itemize @bullet -@item -All client code must be recompiled. If you are using ODBC, you must get -the new @strong{MyODBC} 2.x driver. -@item -The script @code{scripts/add_long_password} must be run to convert the -@code{Password} field in the @code{mysql.user} table to @code{CHAR(16)}. -@item -All passwords must be reassigned in the @code{mysql.user} table (to get 62-bit -rather than 31-bit passwords). -@item -The table format hasn't changed, so you don't have to convert any tables. -@end itemize +Upgrade to BSD/OS Version 3.1. If that is not possible, install +BSDIpatch M300-038. -@strong{MySQL} Version 3.20.28 and above can handle the new @code{user} table -format without affecting clients. If you have a @strong{MySQL} version earlier -than Version 3.20.28, passwords will no longer work with it if you convert the -@code{user} table. So to be safe, you should first upgrade to at least Version -3.20.28 and then upgrade to Version 3.21. +Use the following command when configuring @strong{MySQL}: -@cindex Protocol mismatch -The new client code works with a 3.20.x @code{mysqld} server, so -if you experience problems with 3.21.x, you can use the old 3.20.x server -without having to recompile the clients again. +@example +shell> env CXX=shlicc++ CC=shlicc2 \ + ./configure \ + --prefix=/usr/local/mysql \ + --localstatedir=/var/mysql \ + --without-perl \ + --with-unix-socket-path=/var/mysql/mysql.sock +@end example -If you are not using the @code{--old-protocol} option to @code{mysqld}, -old clients will issue the error message: +The following is also known to work: @example -ERROR: Protocol mismatch. Server Version = 10 Client Version = 9 +shell> env CC=gcc CXX=gcc CXXFLAGS=-O3 \ + ./configure \ + --prefix=/usr/local/mysql \ + --with-unix-socket-path=/var/mysql/mysql.sock @end example -The new Perl @code{DBI}/@code{DBD} interface also supports the old -@code{mysqlperl} interface. The only change you have to make if you use -@code{mysqlperl} is to change the arguments to the @code{connect()} function. -The new arguments are: @code{host}, @code{database}, @code{user}, -@code{password} (the @code{user} and @code{password} arguments have changed -places). -@xref{Perl DBI Class, , Perl @code{DBI} Class}. - -The following changes may affect queries in old applications: +You can change the directory locations if you wish, or just use the +defaults by not specifying any locations. -@itemize @bullet -@item -@code{HAVING} must now be specified before any @code{ORDER BY} clause. -@item -The parameters to @code{LOCATE()} have been swapped. -@item -There are some new reserved words. The most notable are @code{DATE}, -@code{TIME}, and @code{TIMESTAMP}. -@end itemize +If you have problems with performance under heavy load, try using the +@code{--skip-thread-priority} option to @code{mysqld}! This will run +all threads with the same priority; on BSDI Version 3.1, this gives better +performance (at least until BSDI fixes their thread scheduler). -@cindex upgrading, different architecture -@node Upgrading-to-arch, , Upgrading-from-3.20, Upgrade -@subsection Upgrading to Another Architecture +If you get the error @code{virtual memory exhausted} while compiling, +you should try using @code{ulimit -v 80000} and run @code{make} again. +If this doesn't work and you are using @code{bash}, try switching to +@code{csh} or @code{sh}; some BSDI users have reported problems with +@code{bash} and @code{ulimit}. -If you are using @strong{MySQL} Version 3.23, you can copy the @code{.frm}, -@code{.MYI}, and @code{.MYD} files between different architectures that -support the same floating-point format. (@strong{MySQL} takes care of any -byte swapping issues.) +@node BSDI4, , BSDI3, BSDI +@subsubsection BSD/OS Version 4.x Notes -The @strong{MySQL} @code{ISAM} data and index files (@file{.ISD} and -@file{*.ISM}, respectively) are architecture-dependent and in some cases -OS-dependent. If you want to move your applications to another machine -that has a different architecture or OS than your current machine, you -should not try to move a database by simply copying the files to the -other machine. Use @code{mysqldump} instead. +BSDI Version 4.x has some thread-related bugs. If you want to use +@strong{MySQL} on this, you should install all thread-related patches. At +least M400-023 should be installed. -By default, @code{mysqldump} will create a file full of SQL statements. -You can then transfer the file to the other machine and feed it as input -to the @code{mysql} client. +On some BSDI Version 4.x systems, you may get problems with shared libraries. +The symptom is that you can't execute any client programs, for example, +@code{mysqladmin}. In this case you need to reconfigure not to use +shared libraries with the @code{--disable-shared} option to configure. -Try @code{mysqldump --help} to see what options are available. -If you are moving the data to a newer version of @strong{MySQL}, you should use -@code{mysqldump --opt} with the newer version to get a fast, compact dump. +Some customers have had problems on BSDI 4.0.1 that the @code{mysqld} +binary after a while can't open tables. This is because some +library/system related bug causes @code{mysqld} to change current +directory without asking for this! -The easiest (although not the fastest) way to move a database between two -machines is to run the following commands on the machine on which the -database is located: +The fix is to either upgrade to 3.23.34 or after running @code{configure} +remove the line @code{#define HAVE_REALPATH} from @code{config.h} +before running make. -@example -shell> mysqladmin -h 'other hostname' create db_name -shell> mysqldump --opt db_name \ - | mysql -h 'other hostname' db_name -@end example +Note that the above means that you can't symbolic link a database directories +to another database directory or symbolic link a table to another database +on BSDI! (Making a symbolic link to another disk is ok). -If you want to copy a database from a remote machine over a slow network, -you can use: +@node SCO, SCO Unixware, BSDI, Source install system issues +@subsection SCO Notes -@example -shell> mysqladmin create db_name -shell> mysqldump -h 'other hostname' --opt --compress db_name \ - | mysql db_name -@end example +The current port is tested only on a ``sco3.2v5.0.4'' and +``sco3.2v5.0.5'' system. There has also been a lot of progress on a +port to ``sco 3.2v4.2''. -You can also store the result in a file, then transfer the file to the -target machine and load the file into the database there. For example, -you can dump a database to a file on the source machine like this: +For the moment the recommended compiler on OpenServer is gcc 2.95.2. With this +you should be able to compile @strong{MySQL} with just: @example -shell> mysqldump --quick db_name | gzip > db_name.contents.gz +CC=gcc CXX=gcc ./configure ... (options) @end example -(The file created in this example is compressed.) Transfer the file -containing the database contents to the target machine and run these commands -there: +@enumerate +@item +For OpenServer 5.0.X you need to use GDS in Skunkware 95 (95q4c). This +is necessary because GNU @code{gcc} 2.7.2 in Skunkware 97 does not have +GNU @code{as}. You can also use @code{egcs} 1.1.2 or newer +@uref{http://www.egcs.com/}. If you are using @code{egcs} 1.1.2 you have +to execute the following command: @example -shell> mysqladmin create db_name -shell> gunzip < db_name.contents.gz | mysql db_name +shell> cp -p /usr/include/pthread/stdtypes.h /usr/local/lib/gcc-lib/i386-pc-sco3.2v5.0.5/egcs-2.91.66/include/pthread/ @end example -@cindex @code{mysqldump} -@cindex @code{mysqlimport} -You can also use @code{mysqldump} and @code{mysqlimport} to accomplish -the database transfer. -For big tables, this is much faster than simply using @code{mysqldump}. -In the commands shown below, @code{DUMPDIR} represents the full pathname -of the directory you use to store the output from @code{mysqldump}. - -First, create the directory for the output files and dump the database: +@item +You need the port of GCC 2.5.x for this product and the Development +system. They are required on this version of SCO Unix. You cannot +just use the GCC Dev system. -@example -shell> mkdir DUMPDIR -shell> mysqldump --tab=DUMPDIR db_name -@end example +@item +You should get the FSU Pthreads package and install it first. This can be +found at +@uref{http://www.cs.wustl.edu/~schmidt/ACE_wrappers/FSU-threads.tar.gz}. +You can also get a precompiled package from +@uref{http://www.mysql.com/Downloads/SCO/FSU-threads-3.5c.tar.gz}. -Then transfer the files in the @code{DUMPDIR} directory to some corresponding -directory on the target machine and load the files into @strong{MySQL} -there: +@item +FSU Pthreads can be compiled with SCO Unix 4.2 with tcpip. Or +OpenServer 3.0 or Open Desktop 3.0 (OS 3.0 ODT 3.0), with the SCO +Development System installed using a good port of GCC 2.5.x ODT or OS +3.0 you will need a good port of GCC 2.5.x There are a lot of problems +without a good port. The port for this product requires the SCO Unix +Development system. Without it, you are missing the libraries and the +linker that is needed. -@example -shell> mysqladmin create db_name # create database -shell> cat DUMPDIR/*.sql | mysql db_name # create tables in database -shell> mysqlimport db_name DUMPDIR/*.txt # load data into tables -@end example +@item +To build FSU Pthreads on your system, do the following: -Also, don't forget to copy the @code{mysql} database, because that's where the -grant tables (@code{user}, @code{db}, @code{host}) are stored. You may have -to run commands as the @strong{MySQL} @code{root} user on the new machine -until you have the @code{mysql} database in place. +@enumerate a +@item +Run @code{./configure} in the @file{threads/src} directory and select +the SCO OpenServer option. This command copies @file{Makefile.SCO5} to +@file{Makefile}. -After you import the @code{mysql} database on the new machine, execute -@code{mysqladmin flush-privileges} so that the server reloads the grant table -information. +@item +Run @code{make}. -@cindex compatibility, with ANSI SQL -@cindex standards compatibility -@cindex extensions, to ANSI SQL -@cindex ANSI SQL92, extensions to -@node Compatibility, Privilege system, Installing, Top -@chapter How Standards-compatible Is MySQL? +@item +To install in the default @file{/usr/include} directory, login as root, +then @code{cd} to the @file{thread/src} directory, and run @code{make +install}. +@end enumerate -@menu -* Extensions to ANSI:: @strong{MySQL} extensions to ANSI SQL92 -* ANSI mode:: Running @strong{MySQL} in ANSI mode -* Differences from ANSI:: @strong{MySQL} differences compared to ANSI SQL92 -* Missing functions:: Functionality missing from @strong{MySQL} -* Standards:: What standards does @strong{MySQL} follow? -* Commit-rollback:: How to cope without @code{COMMIT}-@code{ROLLBACK} -@end menu +@item +Remember to use GNU @code{make} when making @strong{MySQL}. -This chapter describes how @strong{MySQL} relates to the ANSI SQL standards. -@strong{MySQL} has many extensions to the ANSI SQL standards, and here you -will find out what they are, and how to use them. You will also find -information about functionality missing from @strong{MySQL}, and how to work -around some differences. +@item +If you don't start @code{safe_mysqld} as root, you probably will get only the +default 110 open files per process. @code{mysqld} will write a note about this +in the log file. -@node Extensions to ANSI, ANSI mode, Compatibility, Compatibility -@section MySQL Extensions to ANSI SQL92 +@item +With SCO 3.2V5.0.5, you should use FSU Pthreads version 3.5c or newer. +You should also use gcc 2.95.2 or newer! -@strong{MySQL} includes some extensions that you probably will not find in -other SQL databases. Be warned that if you use them, your code will not be -portable to other SQL servers. In some cases, you can write code that -includes @strong{MySQL} extensions, but is still portable, by using comments -of the form @code{/*! ... */}. In this case, @strong{MySQL} will parse and -execute the code within the comment as it would any other @strong{MySQL} -statement, but other SQL servers will ignore the extensions. For example: +The following @code{configure} command should work: @example -SELECT /*! STRAIGHT_JOIN */ col_name FROM table1,table2 WHERE ... +shell> ./configure --prefix=/usr/local/mysql --disable-shared @end example -If you add a version number after the @code{'!'}, the syntax will only be -executed if the @strong{MySQL} version is equal to or newer than the used -version number: +@item +With SCO 3.2V4.2, you should use FSU Pthreads version 3.5c or newer. +The following @code{configure} command should work: @example -CREATE /*!32302 TEMPORARY */ TABLE (a int); +shell> CFLAGS="-D_XOPEN_XPG4" CXX=gcc CXXFLAGS="-D_XOPEN_XPG4" \ + ./configure \ + --prefix=/usr/local/mysql \ + --with-named-thread-libs="-lgthreads -lsocket -lgen -lgthreads" \ + --with-named-curses-libs="-lcurses" @end example -The above means that if you have Version 3.23.02 or newer, then @strong{MySQL} -will use the @code{TEMPORARY} keyword. +You may get some problems with some include files. In this case, you can +find new SCO-specific include files at +@uref{http://www.mysql.com/Downloads/SCO/SCO-3.2v4.2-includes.tar.gz}. +You should unpack this file in the @file{include} directory of your +@strong{MySQL} source tree. +@end enumerate -@strong{MySQL} extensions are listed below: +SCO development notes: @itemize @bullet @item -The field types @code{MEDIUMINT}, @code{SET}, @code{ENUM}, and the -different @code{BLOB} and @code{TEXT} types. - -@item -The field attributes @code{AUTO_INCREMENT}, @code{BINARY}, @code{NULL}, -@code{UNSIGNED}, and @code{ZEROFILL}. - -@item -All string comparisons are case insensitive by default, with sort -ordering determined by the current character set (ISO-8859-1 Latin1 by -default). If you don't like this, you should declare your columns with -the @code{BINARY} attribute or use the @code{BINARY} cast, which causes -comparisons to be done according to the ASCII order used on the -@strong{MySQL} server host. - -@item -@strong{MySQL} maps each database to a directory under the @strong{MySQL} -data directory, and tables within a database to filenames in the database -directory. - -This has a few implications: - -@cindex database names, case sensitivity -@cindex table names, case sensitivity -@cindex case sensitivity, of database names -@cindex case sensitivity, of table names - -@itemize @minus +@strong{MySQL} should automatically detect FSU Pthreads and link @code{mysqld} +with @code{-lgthreads -lsocket -lgthreads}. @item -Database names and table names are case sensitive in @strong{MySQL} on -operating systems that have case-sensitive filenames (like most Unix -systems). @xref{Name case sensitivity}. +The SCO development libraries are re-entrant in FSU Pthreads. SCO claims +that its libraries' functions are re-entrant, so they must be reentrant with +FSU Pthreads. FSU Pthreads on OpenServer tries to use the SCO scheme to +make re-entrant libraries. @item -Database, table, index, column, or alias names may begin with a digit -(but may not consist solely of digits). +FSU Pthreads (at least the version at @uref{http://www.mysql.com/}) comes +linked with GNU @code{malloc}. If you encounter problems with memory usage, +make sure that @file{gmalloc.o} is included in @file{libgthreads.a} and +@file{libgthreads.so}. @item -You can use standard system commands to backup, rename, move, delete, and copy -tables. For example, to rename a table, rename the @file{.MYD}, @file{.MYI}, -and @file{.frm} files to which the table corresponds. +In FSU Pthreads, the following system calls are pthreads-aware: @code{read()}, +@code{write()}, @code{getmsg()}, @code{connect()}, @code{accept()}, +@code{select()}, and @code{wait()}. @end itemize -@item -In SQL statements, you can access tables from different databases -with the @code{db_name.tbl_name} syntax. Some SQL servers provide -the same functionality but call this @code{User space}. -@strong{MySQL} doesn't support tablespaces as in: -@code{create table ralph.my_table...IN my_tablespace}. - -@item -@code{LIKE} is allowed on numeric columns. - -@item -Use of @code{INTO OUTFILE} and @code{STRAIGHT_JOIN} in a @code{SELECT} -statement. @xref{SELECT, , @code{SELECT}}. +If you want to install DBI on SCO, you have to edit the @file{Makefile} in +DBI-xxx and each subdirectory. -@item -The @code{SQL_SMALL_RESULT} option in a @code{SELECT} statement. +Note that the following assumes gcc 2.95.2 or newer: -@item -@code{EXPLAIN SELECT} to get a description on how tables are joined. +@example +OLD: NEW: +CC = cc CC = gcc +CCCDLFLAGS = -KPIC -W1,-Bexport CCCDLFLAGS = -fpic +CCDLFLAGS = -wl,-Bexport CCDLFLAGS = -@item -Use of index names, indexes on a prefix of a field, and use of -@code{INDEX} or @code{KEY} in a @code{CREATE TABLE} -statement. @xref{CREATE TABLE, , @code{CREATE TABLE}}. +LD = ld LD = gcc -G -fpic +LDDLFLAGS = -G -L/usr/local/lib LDDLFLAGS = -L/usr/local/lib +LDFLAGS = -belf -L/usr/local/lib LDFLAGS = -L/usr/local/lib -@item -Use of @code{TEMPORARY} or @code{IF NOT EXISTS} with @code{CREATE TABLE}. +LD = ld LD = gcc -G -fpic +OPTIMISE = -Od OPTIMISE = -O1 -@item -Use of @code{COUNT(DISTINCT list)} where 'list' is more than one element. +OLD: +CCCFLAGS = -belf -dy -w0 -U M_XENIX -DPERL_SCO5 -I/usr/local/include -@item -Use of @code{CHANGE col_name}, @code{DROP col_name}, or @code{DROP -INDEX}, @code{IGNORE} or @code{RENAME} in an @code{ALTER TABLE} -statement. @xref{ALTER TABLE, , @code{ALTER TABLE}}. +NEW: +CCFLAGS = -U M_XENIX -DPERL_SCO5 -I/usr/local/include +@end example -@item -Use of @code{RENAME TABLE}. @xref{RENAME TABLE, , @code{RENAME TABLE}}. +This is because the Perl dynaloader will not load the @code{DBI} modules +if they were compiled with @code{icc} or @code{cc}. -@item -Use of multiple @code{ADD}, @code{ALTER}, @code{DROP}, or @code{CHANGE} -clauses in an @code{ALTER TABLE} statement. +Perl works best when compiled with @code{cc}. -@item -Use of @code{DROP TABLE} with the keywords @code{IF EXISTS}. +@node SCO Unixware, IBM-AIX, SCO, Source install system issues +@subsection SCO Unixware Version 7.0 Notes -@item -You can drop multiple tables with a single @code{DROP TABLE} statement. +You must use a version of @strong{MySQL} at least as recent as Version 3.22.13 +because that version fixes some portability problems under Unixware. -@item -The @code{LIMIT} clause of the @code{DELETE} statement. +We have been able to compile @strong{MySQL} with the following @code{configure} +command on Unixware Version 7.0.1: -@item -The @code{DELAYED} clause of the @code{INSERT} and @code{REPLACE} -statements. +@example +CC=cc CXX=CC ./configure --prefix=/usr/local/mysql +@end example -@item -The @code{LOW_PRIORITY} clause of the @code{INSERT}, @code{REPLACE}, -@code{DELETE}, and @code{UPDATE} statements. +If you want to use @code{gcc}, you must use @code{gcc} 2.95.2 or newer. -@cindex Oracle compatibility -@cindex compatibility, with Oracle -@item -Use of @code{LOAD DATA INFILE}. In many cases, this syntax is compatible with -Oracle's @code{LOAD DATA INFILE}. @xref{LOAD DATA, , @code{LOAD DATA}}. +@node IBM-AIX, HP-UX 10.20, SCO Unixware, Source install system issues +@subsection IBM-AIX notes +@cindex problems, installing on IBM-AIX -@item -The @code{ANALYZE TABLE}, @code{CHECK TABLE}, @code{OPTIMIZE TABLE}, and -@code{REPAIR TABLE} statements. +Automatic detection of @code{xlC} is missing from Autoconf, so a +@code{configure} command something like this is needed when compiling +@strong{MySQL} (This example uses the IBM compiler): -@item -The @code{SHOW} statement. -@xref{SHOW, , @code{SHOW}}. +@example +export CC="xlc_r -ma -O3 -qstrict -qoptimize=3 -qmaxmem=8192 " +export CXX="xlC_r -ma -O3 -qstrict -qoptimize=3 -qmaxmem=8192" +export CFLAGS="-I /usr/local/include" +export LDLFAGS="-L /usr/local/lib" +export CPPFLAGS=$CFLAGS +export CXXFLAGS=$CFLAGS -@item -Strings may be enclosed by either @samp{"} or @samp{'}, not just by @samp{'}. +./configure --prefix=/usr/local \ + --localstatedir=/var/mysql \ + --sysconfdir=/etc/mysql \ + --sbindir='/usr/local/bin' \ + --libexecdir='/usr/local/bin' \ + --enable-thread-safe-client \ + --enable-large-files +@end example -@item -Use of the escape @samp{\} character. +Above are the options used to compile the @strong{MySQL} distribution that +can be found at @uref{http://www-frec.bull.com/}. -@item -The @code{SET OPTION} statement. @xref{SET OPTION, , @code{SET OPTION}}. +If you change the @code{-O3} to @code{-O2} in the above configure line, +you must also remove the @code{-qstrict} option (this is a limitation in +the IBM C compiler). -@item -You don't need to name all selected columns in the @code{GROUP BY} part. -This gives better performance for some very specific, but quite normal -queries. -@xref{Group by functions}. +If you are using @code{gcc} or @code{egcs} to compile @strong{MySQL}, you +@strong{MUST} use the @code{-fno-exceptions} flag, as the exception +handling in @code{gcc}/@code{egcs} is not thread safe! (This is tested with +@code{egcs} 1.1.). There are also some known problems with IBM's assembler, +which may cause it to generate bad code when used with gcc. -@item -One can specify @code{ASC} and @code{DESC} with @code{GROUP BY}. +We recommend the following @code{configure} line with @code{egcs} and +@code{gcc 2.95} on AIX: -@item -To make it easier for users who come from other SQL environments, -@strong{MySQL} supports aliases for many functions. For example, all -string functions support both ANSI SQL syntax and ODBC syntax. +@example +CC="gcc -pipe -mcpu=power -Wa,-many" \ +CXX="gcc -pipe -mcpu=power -Wa,-many" \ +CXXFLAGS="-felide-constructors -fno-exceptions -fno-rtti" \ +./configure --prefix=/usr/local/mysql --with-low-memory +@end example -@item -@strong{MySQL} understands the @code{||} and @code{&&} operators to mean -logical OR and AND, as in the C programming language. In @strong{MySQL}, -@code{||} and @code{OR} are synonyms, as are @code{&&} and @code{AND}. -Because of this nice syntax, @strong{MySQL} doesn't support -the ANSI SQL @code{||} operator for string concatenation; use -@code{CONCAT()} instead. Because @code{CONCAT()} takes any number -of arguments, it's easy to convert use of the @code{||} operator to -@strong{MySQL}. +The @code{-Wa,-many} is necessary for the compile to be successful. IBM is +aware of this problem but is in to hurry to fix it because of the workaround +available. We don't know if the @code{-fno-exceptions} is required with +@code{gcc 2.95}, but as @strong{MySQL} doesn't use exceptions and the above +option generates faster code, we recommend that you should always use this +option with @code{egcs / gcc}. -@item -@code{CREATE DATABASE} or @code{DROP DATABASE}. -@xref{CREATE DATABASE, , @code{CREATE DATABASE}}. +If you get a problem with assembler code try changing the -mcpu=xxx to +match your cpu. Typically power2, power, or powerpc may need to be used, +alternatively you might need to use 604 or 604e. I'm not positive but I +would think using "power" would likely be safe most of the time, even on +a power2 machine. -@cindex PostgreSQL compatibility -@cindex compatibility, with PostgreSQL -@item -The @code{%} operator is a synonym for @code{MOD()}. That is, -@code{N % M} is equivalent to @code{MOD(N,M)}. @code{%} is supported -for C programmers and for compatibility with PostgreSQL. +If you don't know what your cpu is then do a "uname -m", this will give +you back a string that looks like "000514676700", with a format of +xxyyyyyymmss where xx and ss are always 0's, yyyyyy is a unique system +id and mm is the id of the CPU Planar. A chart of these values can be +found at +@uref{http://www.rs6000.ibm.com/doc_link/en_US/a_doc_lib/cmds/aixcmds5/uname.htm}. +This will give you a machine type and a machine model you can use to +determine what type of cpu you have. -@item -The @code{=}, @code{<>}, @code{<=} ,@code{<}, @code{>=},@code{>}, -@code{<<}, @code{>>}, @code{<=>}, @code{AND}, @code{OR}, or @code{LIKE} -operators may be used in column comparisons to the left of the -@code{FROM} in @code{SELECT} statements. For example: +If you have problems with signals (@strong{MySQL} dies unexpectedly +under high load) you may have found an OS bug with threads and +signals. In this case you can tell @strong{MySQL} not to use signals by +configuring with: @example -mysql> SELECT col1=1 AND col2=2 FROM tbl_name; +shell> CFLAGS=-DDONT_USE_THR_ALARM CXX=gcc \ + CXXFLAGS="-felide-constructors -fno-exceptions -fno-rtti -DDONT_USE_THR_ALARM" \ + ./configure --prefix=/usr/local/mysql --with-debug --with-low-memory @end example -@item -The @code{LAST_INSERT_ID()} function. -@xref{mysql_insert_id, , @code{mysql_insert_id()}}. - -@item -The @code{REGEXP} and @code{NOT REGEXP} extended regular expression -operators. - -@item -@code{CONCAT()} or @code{CHAR()} with one argument or more than two -arguments. (In @strong{MySQL}, these functions can take any number of -arguments.) +This doesn't affect the performance of @strong{MySQL}, but has the side +effect that you can't kill clients that are ``sleeping'' on a connection with +@code{mysqladmin kill} or @code{mysqladmin shutdown}. Instead, the client +will die when it issues its next command. -@item The @code{BIT_COUNT()}, @code{CASE}, @code{ELT()}, -@code{FROM_DAYS()}, @code{FORMAT()}, @code{IF()}, @code{PASSWORD()}, -@code{ENCRYPT()}, @code{md5()}, @code{ENCODE()}, @code{DECODE()}, -@code{PERIOD_ADD()}, @code{PERIOD_DIFF()}, @code{TO_DAYS()}, or -@code{WEEKDAY()} functions. +On some versions of AIX, linking with @code{libbind.a} makes +@code{getservbyname} core dump. This is an AIX bug and should be reported +to IBM. -@item -Use of @code{TRIM()} to trim substrings. ANSI SQL only supports removal -of single characters. +For AIX 4.2.1 and gcc you have to do the following changes. -@item -The @code{GROUP BY} functions @code{STD()}, @code{BIT_OR()}, and -@code{BIT_AND()}. +After configuring, edit @file{config.h} and @file{include/my_config.h} +and change the line that says -@item -Use of @code{REPLACE} instead of @code{DELETE} + @code{INSERT}. -@xref{REPLACE, , @code{REPLACE}}. +@example +#define HAVE_SNPRINTF 1 +@end example -@item -The @code{FLUSH flush_option} statement. +to -@item -The possibility to set variables in a statement with @code{:=}: @example -SELECT @@a:=SUM(total),@@b=COUNT(*),@@a/@@b AS avg FROM test_table; -SELECT @@t1:=(@@t2:=1)+@@t3:=4,@@t1,@@t2,@@t3; +#undef HAVE_SNPRINTF @end example -@end itemize +And finally, in @file{mysqld.cc} you need to add a prototype for initgoups. -@node ANSI mode, Differences from ANSI, Extensions to ANSI, Compatibility -@section Running MySQL in ANSI Mode -@cindex running, ANSI mode -@cindex ANSI mode, running +@example +#ifdef _AIX41 +extern "C" int initgroups(const char *,int); +#endif +@end example -If you start @code{mysqld} with the @code{--ansi} option, the following behavior -of @strong{MySQL} changes: +@node HP-UX 10.20, HP-UX 11.x, IBM-AIX, Source install system issues +@subsection HP-UX Version 10.20 Notes -@itemize @bullet -@item -@code{||} is string concatenation instead of @code{OR}. -@item -You can have any number of spaces between a function name and the @samp{(}. -This forces all function names to be treated as reserved words. -@item -@samp{"} will be an identifier quote character (like the @strong{MySQL} -@samp{`} quote character) and not a string quote character. -@item -@code{REAL} will be a synonym for @code{FLOAT} instead of a synonym of -@code{DOUBLE}. -@item -The default transaction isolation level is @code{SERIALIZABLE}. -@xref{SET TRANSACTION}. -@end itemize +There are a couple of small problems when compiling @strong{MySQL} on +HP-UX. We recommend that you use @code{gcc} instead of the HP-UX native +compiler, because @code{gcc} produces better code! -@node Differences from ANSI, Missing functions, ANSI mode, Compatibility -@section MySQL Differences Compared to ANSI SQL92 +We recommend using gcc 2.95 on HP-UX. Don't use high optimization +flags (like -O6) as this may not be safe on HP-UX. -We try to make @strong{MySQL} follow the ANSI SQL standard and the -ODBC SQL standard, but in some cases @strong{MySQL} does some things -differently: +Note that MIT-pthreads can't be compiled with the HP-UX compiler +because it can't compile @code{.S} (assembler) files. -@itemize @bullet -@item -@code{--} is only a comment if followed by a white space. @xref{Missing -comments}. -@item -For @code{VARCHAR} columns, trailing spaces are removed when the value is -stored. @xref{Bugs}. -@item -In some cases, @code{CHAR} columns are silently changed to @code{VARCHAR} -columns. @xref{Silent column changes}. -@item -Privileges for a table are not automatically revoked when you delete a -table. You must explicitly issue a @code{REVOKE} to revoke privileges for -a table. @xref{GRANT, , @code{GRANT}}. -@item -@code{NULL AND FALSE} will evaluate to @code{NULL} and not to @code{FALSE}. -This is because we don't think it's good to have to evaluate a lot of -extra conditions in this case. -@end itemize +The following configure line should work: -@node Missing functions, Standards, Differences from ANSI, Compatibility -@section Functionality Missing from MySQL -@cindex missing functionality -@cindex functionality, missing +@example +CFLAGS="-DHPUX -I/opt/dce/include" CXXFLAGS="-DHPUX -I/opt/dce/include -felide-constructors -fno-exceptions -fno-rtti" CXX=gcc ./configure --with-pthread --with-named-thread-libs='-ldce' --prefix=/usr/local/mysql --disable-shared +@end example -The following functionality is missing in the current version of -@strong{MySQL}. For a prioritized list indicating when new extensions -may be added to @strong{MySQL}, you should consult -@uref{http://www.mysql.com/documentation/manual.php?section=TODO, the -online @strong{MySQL} TODO list}. That is the latest version of the TODO -list in this manual. @xref{TODO}. +If you are compiling @code{gcc} 2.95 yourself, you should NOT link it with +the DCE libraries (@code{libdce.a} or @code{libcma.a}) if you want to compile +@strong{MySQL} with MIT-pthreads. If you mix the DCE and MIT-pthreads +packages you will get a @code{mysqld} to which you cannot connect. Remove +the DCE libraries while you compile @code{gcc} 2.95! -@menu -* Missing Sub-selects:: Sub-selects -* Missing SELECT INTO TABLE:: @code{SELECT INTO TABLE} -* Missing Transactions:: Transactions -* Missing Triggers:: Triggers -* Missing Foreign Keys:: Foreign Keys -* Missing Views:: Views -* Missing comments:: @samp{--} as the start of a comment -@end menu +@node HP-UX 11.x, Mac OS X, HP-UX 10.20, Source install system issues +@subsection HP-UX Version 11.x Notes -@node Missing Sub-selects, Missing SELECT INTO TABLE, Missing functions, Missing functions -@subsection Sub-selects -@cindex sub-selects +For HP-UX Version 11.x we recommend @strong{MySQL} Version 3.23.15 or later. -The following will not yet work in @strong{MySQL}: +Because of some critical bugs in the standard HP-UX libraries, you should +install the following patches before trying to run @strong{MySQL} on HP-UX 11.0: @example -SELECT * FROM table1 WHERE id IN (SELECT id FROM table2); -SELECT * FROM table1 WHERE id NOT IN (SELECT id FROM table2); -SELECT * FROM table1 WHERE NOT EXISTS (SELECT id FROM table2 where table1.id=table2.id); +PHKL_22840 Streams cumulative +PHNE_22397 ARPA cumulative @end example -However, in many cases you can rewrite the query without a sub-select: +This will solve a problem that one gets @code{EWOULDBLOCK} from @code{recv()} +and @code{EBADF} from @code{accept()} in threaded applications. + +If you are using @code{gcc} 2.95.1 on an unpatched HP-UX 11.x system, +you will get the error: @example -SELECT table1.* FROM table1,table2 WHERE table1.id=table2.id; -SELECT table1.* FROM table1 LEFT JOIN table2 ON table1.id=table2.id where table2.id IS NULL +In file included from /usr/include/unistd.h:11, + from ../include/global.h:125, + from mysql_priv.h:15, + from item.cc:19: +/usr/include/sys/unistd.h:184: declaration of C function ... +/usr/include/sys/pthread.h:440: previous declaration ... +In file included from item.h:306, + from mysql_priv.h:158, + from item.cc:19: @end example -For more complicated subqueries you can often create temporary tables -to hold the subquery. In some cases, however this option will not -work. The most frequently encountered of these cases arises with -@code{DELETE} statements, for which standard SQL does not support joins -(except in sub-selects). For this situation there are two options -available until subqueries are supported by @strong{MySQL}. - -The first option is to use a procedural programming language (such as -Perl or PHP) to submit a @code{SELECT} query to obtain the primary keys -for the records to be deleted, and then use these values to construct -the @code{DELETE} statement (@code{DELETE FROM ... WHERE ... IN (key1, -key2, ...)}). +The problem is that HP-UX doesn't define @code{pthreads_atfork()} consistently. +It has conflicting prototypes in +@file{/usr/include/sys/unistd.h}:184 and +@file{/usr/include/sys/pthread.h}:440 (details below). -The second option is to use interactive SQL to contruct a set of -@code{DELETE} statements automatically, using the @strong{MySQL} -extension @code{CONCAT()} (in lieu of the standard @code{||} operator). -For example: +One solution is to copy @file{/usr/include/sys/unistd.h} into +@file{mysql/include} and edit @file{unistd.h} and change it to match +the definition in @file{pthread.h}. Here's the diff: @example -SELECT CONCAT('DELETE FROM tab1 WHERE pkid = ', tab1.pkid, ';') - FROM tab1, tab2 - WHERE tab1.col1 = tab2.col2; +183,184c183,184 +< extern int pthread_atfork(void (*prepare)(), void (*parent)(), +< void (*child)()); +--- +> extern int pthread_atfork(void (*prepare)(void), void (*parent)(void), +> void (*child)(void)); @end example -You can place this query in a script file and redirect input from it to -the @code{mysql} command-line interpreter, piping its output back to a -second instance of the interpreter: +After this, the following configure line should work: @example -prompt> mysql --skip-column-names mydb < myscript.sql | mysql mydb +CFLAGS="-fomit-frame-pointer -O3 -fpic" CXX=gcc CXXFLAGS="-felide-constructors -fno-exceptions -fno-rtti -O3" ./configure --prefix=/usr/local/mysql --disable-shared @end example -@strong{MySQL} only supports @code{INSERT ... SELECT ...} and -@code{REPLACE ... SELECT ...} Independent sub-selects will probably -be available in Version 4.0. You can now use the function @code{IN()} in -other contexts, however. +Here is some information that a HP-UX Version 11.x user sent us about compiling +@strong{MySQL} with HP-UX:x compiler: -@node Missing SELECT INTO TABLE, Missing Transactions, Missing Sub-selects, Missing functions -@subsection @code{SELECT INTO TABLE} -@findex SELECT INTO TABLE +@example + Environment: + proper compilers. + setenv CC cc + setenv CXX aCC + flags + setenv CFLAGS -D_REENTRANT + setenv CXXFLAGS -D_REENTRANT + setenv CPPFLAGS -D_REENTRANT + % aCC -V + aCC: HP ANSI C++ B3910B X.03.14.06 + % cc -V /tmp/empty.c + cpp.ansi: HP92453-01 A.11.02.00 HP C Preprocessor (ANSI) + ccom: HP92453-01 A.11.01.00 HP C Compiler + cc: "/tmp/empty.c", line 1: warning 501: Empty source file. -@strong{MySQL} doesn't yet support the Oracle SQL extension: -@code{SELECT ... INTO TABLE ...}. @strong{MySQL} supports instead the -ANSI SQL syntax @code{INSERT INTO ... SELECT ...}, which is basically -the same thing. @xref{INSERT SELECT}. + configuration: + ./configure --with-pthread \ + --prefix=/source-control/mysql \ + --with-named-thread-libs=-lpthread \ + --with-low-memory -@example -INSERT INTO tblTemp2 (fldID) SELECT tblTemp1.fldOrder_ID FROM tblTemp1 WHERE -tblTemp1.fldOrder_ID > 100; + added '#define _CTYPE_INCLUDED' to include/m_ctype.h. This + symbol is the one defined in HP's /usr/include/ctype.h: + + /* Don't include std ctype.h when this is included */ + #define _CTYPE_H + #define __CTYPE_INCLUDED + #define _CTYPE_INCLUDED + #define _CTYPE_USING /* Don't put names in global namespace. */ @end example -Alternatively, you can use @code{SELECT INTO OUTFILE...} or @code{CREATE -TABLE ... SELECT} to solve your problem. +@itemize @bullet +@item +I had to use the compile-time flag @code{-D_REENTRANT} to get the compiler +to recognize the prototype for @code{localtime_r}. Alternatively I could have +supplied the prototype for @code{localtime_r}. But I wanted to catch other +bugs without needing to run into them. I wasn't sure where I needed it, so I +added it to all flags. +@item +The optimization flags used by @strong{MySQL} (-O3) are not recognized by HP's +compilers. I did not change the flags. +@end itemize -@node Missing Transactions, Missing Triggers, Missing SELECT INTO TABLE, Missing functions -@subsection Transactions -@cindex transactions, support +If you get the following error from @code{configure} -As @strong{MySQL} does nowadays support transactions, the following -discussion is only valid if you are only using the non-transaction-safe -table types. @xref{COMMIT}. +@example +checking for cc option to accept ANSI C... no +configure: error: MySQL requires a ANSI C compiler (and a C++ compiler). Try gcc. See the Installation chapter in the Reference Manual. +@end example -The question is often asked, by the curious and the critical, ``Why is -@strong{MySQL} not a transactional database?'' or ``Why does @strong{MySQL} -not support transactions?'' +Check that you don't have the path to the K&R compiler before the path +to the HP-UX C and C++ compiler. -@strong{MySQL} has made a conscious decision to support another paradigm -for data integrity, ``atomic operations.'' It is our thinking and -experience that atomic operations offer equal or even better integrity -with much better performance. We, nonetheless, appreciate and understand -the transactional database paradigm and plan, within the next few releases, -to introduce transaction-safe tables on a per table basis. We will be -giving our users the possibility to decide if they need the speed of -atomic operations or if they need to use transactional features in their -applications. +@node Mac OS X, BEOS, HP-UX 11.x, Source install system issues +@subsection Mac OS X Notes -How does one use the features of @strong{MySQL} to maintain rigorous integrity -and how do these features compare with the transactional paradigm? +@menu +* Mac OS X Public Data:: +* Mac OS X Server:: +@end menu -First, in the transactional paradigm, if your applications are written -in a way that is dependent on the calling of ``rollback'' instead of -``commit'' in critical situations, then transactions are more -convenient. Moreover, transactions ensure that unfinished updates or -corrupting activities are not committed to the database; the server is -given the opportunity to do an automatic rollback and your database is -saved. +@node Mac OS X Public Data, Mac OS X Server, Mac OS X, Mac OS X +@subsubsection Mac OS X Public beta -@strong{MySQL}, in almost all cases, allows you to solve for potential -problems by including simple checks before updates and by running simple -scripts that check the databases for inconsistencies and automatically -repair or warn if such occurs. Note that just by using the -@strong{MySQL} log or even adding one extra log, one can normally fix -tables perfectly with no data integrity loss. +@strong{MySQL} should work without any problems on Mac OS X Public Beta +(Darwin). You don't need the pthread patches for this OS! -Moreover, fatal transactional updates can be rewritten to be -atomic. In fact,we will go so far as to say that all integrity problems -that transactions solve can be done with @code{LOCK TABLES} or atomic updates, -ensuring that you never will get an automatic abort from the database, -which is a common problem with transactional databases. +@node Mac OS X Server, , Mac OS X Public Data, Mac OS X +@subsubsection Mac OS X Server -Not even transactions can prevent all loss if the server goes down. In -such cases even a transactional system can lose data. The difference -between different systems lies in just how small the time-lap is where -they could lose data. No system is 100% secure, only ``secure -enough.'' Even Oracle, reputed to be the safest of transactional -databases, is reported to sometimes lose data in such situations. +Before trying to configure @strong{MySQL} on Mac OS X server you must +first install the pthread package from +@uref{http://www.prnet.de/RegEx/mysql.html}. -To be safe with @strong{MySQL}, you only need to have backups and have -the update logging turned on. With this you can recover from any -situation that you could with any transactional database. It is, of -course, always good to have backups, independent of which database you -use. +Our binary for Mac OS X is compiled on Rhapsody 5.5 with the following +configure line: -The transactional paradigm has its benefits and its drawbacks. Many -users and application developers depend on the ease with which they can -code around problems where an abort appears to be, or is necessary, and they -may have to do a little more work with @strong{MySQL} to either think -differently or write more. If you are new to the atomic operations -paradigm, or more familiar or more comfortable with transactions, do not -jump to the conclusion that @strong{MySQL} has not addressed these -issues. Reliability and integrity are foremost in our minds. Recent -estimates indicate that there are more than 1,000,000 @code{mysqld} servers -currently running, many of which are in production environments. We -hear very, very seldom from our users that they have lost any data, and -in almost all of those cases user error is involved. This is, in our -opinion, the best proof of @strong{MySQL}'s stability and reliability. +@example +CC=gcc CFLAGS="-O2 -fomit-frame-pointer" CXX=gcc CXXFLAGS="-O2 -fomit-frame-pointer" ./configure --prefix=/usr/local/mysql "--with-comment=Official MySQL binary" --with-extra-charsets=complex --disable-shared +@end example -Lastly, in situations where integrity is of highest importance, -@strong{MySQL}'s current features allow for transaction-level or better -reliability and integrity. If you lock tables with @code{LOCK TABLES}, all -updates will stall until any integrity checks are made. If you only obtain -a read lock (as opposed to a write lock), then reads and inserts are -still allowed to happen. The new inserted records will not be seen by -any of the clients that have a @code{READ} lock until they release their read -locks. With @code{INSERT DELAYED} you can queue inserts into a local queue, -until the locks are released, without having the client wait for the insert -to complete. @xref{INSERT DELAYED}. +You might want to also add aliases to your shell's resource file to +access @code{mysql} and @code{mysqladmin} from the command line: -``Atomic,'' in the sense that we mean it, is nothing magical. It only means -that you can be sure that while each specific update is running, no other -user can interfere with it, and there will never be an automatic -rollback (which can happen on transaction based systems if you are not -very careful). @strong{MySQL} also guarantees that there will not be -any dirty reads. You can find some example of how to write atomic updates -in the commit-rollback section. @xref{Commit-rollback}. +@example +alias mysql '/usr/local/mysql/bin/mysql' +alias mysqladmin '/usr/local/mysql/bin/mysqladmin' +@end example -We have thought quite a bit about integrity and performance, and we -believe that our atomic operations paradigm allows for both high -reliability and extremely high performance, on the order of three to -five times the speed of the fastest and most optimally tuned of -transactional databases. We didn't leave out transactions because they -are hard to do. The main reason we went with atomic operations as -opposed to transactions is that by doing this we could apply many speed -optimizations that would not otherwise have been possible. +@node BEOS, , Mac OS X, Source install system issues +@subsection BeOS Notes -Many of our users who have speed foremost in their minds are not at all -concerned about transactions. For them transactions are not an -issue. For those of our users who are concerned with or have wondered -about transactions vis-a-vis @strong{MySQL}, there is a ``@strong{MySQL} -way'' as we have outlined above. For those where safety is more -important than speed, we recommend them to use the @code{BDB}, -or @code{InnoDB} tables for all their critical data. @xref{Table types}. +We are really interested in getting @strong{MySQL} to work on BeOS, but +unfortunately we don't have any person who knows BeOS or has time to do +a port. -One final note: We are currently working on a safe replication schema -that we believe to be better than any commercial replication system we -know of. This system will work most reliably under the atomic -operations, non-transactional, paradigm. Stay tuned. +We are interested in finding someone to do a port, and we will help them +with any technical questions they may have while doing the port. -@node Missing Triggers, Missing Foreign Keys, Missing Transactions, Missing functions -@subsection Stored Procedures and Triggers -@cindex stored procedures and triggers, defined -@cindex procedures, stored -@cindex triggers, stored +We have previously talked with some BeOS developers that have said that +@strong{MySQL} is 80% ported to BeOS, but we haven't heard from them +in a while. +@node Windows, OS/2, Source install system issues, Installing +@section Windows Notes -A stored procedure is a set of SQL commands that can be compiled and stored -in the server. Once this has been done, clients don't need to keep reissuing -the entire query but can refer to the stored procedure. This provides better -performance because the query has to be parsed only once, and less information -needs to be sent between the server and the client. You can also raise the -conceptual level by having libraries of functions in the server. +This section describes installation and use of @strong{MySQL} on Windows. +This information is also provided in the @file{README} file that comes +with the @strong{MySQL} Windows distribution. -A trigger is a stored procedure that is invoked when a particular event -occurs. For example, you can install a stored procedure that is triggered -each time a record is deleted from a transaction table and that automatically -deletes the corresponding customer from a customer table when all his -transactions are deleted. +@menu +* Windows installation:: Installing @strong{MySQL} on Windows +* Win95 start:: Starting @strong{MySQL} on Win95 / Win98 +* NT start:: Starting @strong{MySQL} on NT / Win2000 +* Windows running:: Running @strong{MySQL} on Windows +* Windows and SSH:: Connecting to a remote @strong{MySQL} from Windows with SSH +* Windows symbolic links:: Splitting data across different disks under Win32 +* Windows compiling:: Compiling MySQL clients on Windows. +* Windows vs Unix:: @strong{MySQL}-Windows compared to Unix @strong{MySQL} +@end menu -The planned update language will be able to -handle stored procedures, but without triggers. Triggers usually slow -down everything, even queries for which they are not needed. +@node Windows installation, Win95 start, Windows, Windows +@subsection Installing MySQL on Windows -To see when @strong{MySQL} might get stored procedures, see @ref{TODO}. +The following instructions apply to precompiled binary distributions. +If you download a source distribution, you will have to compile and install +it yourself. -@node Missing Foreign Keys, Missing Views, Missing Triggers, Missing functions -@subsection Foreign Keys -@cindex foreign keys -@cindex keys, foreign +If you don't have a copy of the @strong{MySQL} distribution, you should +first download one from @uref{http://www.mysql.com/downloads/mysql-3.23.html}. -Note that foreign keys in SQL are not used to join tables, but are used -mostly for checking referential integrity (foreign key constraints). If -you want to get results from multiple tables from a @code{SELECT} -statement, you do this by joining tables: +If you plan to connect to @strong{MySQL} from some other program, you will +probably also need the @strong{MyODBC} driver. You can find this at the +@strong{MyODBC} download page +(@uref{http://www.mysql.com/downloads/api-myodbc.html}). + +To install either distribution, unzip it in some empty directory and run the +@code{Setup.exe} program. + +By default, @strong{MySQL}-Windows is configured to be installed in +@file{C:\mysql}. If you want to install @strong{MySQL} elsewhere, +install it in @file{C:\mysql} first, then move the installation to +where you want it. If you do move @strong{MySQL}, you must indicate +where everything is located by supplying a @code{--basedir} option when +you start the server. For example, if you have moved the @strong{MySQL} +distribution to @file{D:\programs\mysql}, you must start @code{mysqld} +like this: @example -SELECT * from table1,table2 where table1.id = table2.id; +C:\> D:\programs\mysql\bin\mysqld --basedir D:\programs\mysql @end example -@xref{JOIN, , @code{JOIN}}. @xref{example-Foreign keys}. - -The @code{FOREIGN KEY} syntax in @strong{MySQL} exists only for compatibility -with other SQL vendors' @code{CREATE TABLE} commands; it doesn't do -anything. The @code{FOREIGN KEY} syntax without @code{ON DELETE ...} is -mostly used for documentation purposes. Some ODBC applications may use this -to produce automatic @code{WHERE} clauses, but this is usually easy to -override. @code{FOREIGN KEY} is sometimes used as a constraint check, but -this check is unnecessary in practice if rows are inserted into the tables in -the right order. @strong{MySQL} only supports these clauses because some -applications require them to exist (regardless of whether or not they -work). +Use @code{mysqld --help} to display all the options that @code{mysqld} +understands! -In @strong{MySQL}, you can work around the problem of @code{ON DELETE -...} not being implemented by adding the appropriate @code{DELETE} statement to -an application when you delete records from a table that has a foreign key. -In practice this is as quick (in some cases quicker) and much more portable -than using foreign keys. +With all newer @strong{MySQL} versions, you can also create a +@file{C:\my.cnf} file that holds any default options for the +@strong{MySQL} server. Copy the file @file{\mysql\my-xxxxx.cnf} to +@file{C:\my.cnf} and edit it to suit your setup. Note that you should +specify all paths with @samp{/} instead of @samp{\}. If you use +@samp{\}, you need to specify it twice, because @samp{\} is the escape +character in @strong{MySQL}. @xref{Option files}. -In the near future we will extend the @code{FOREIGN KEY} implementation so -that at least the information will be saved in the table specification file -and may be retrieved by @code{mysqldump} and ODBC. At a later stage we will -implement the foreign key constraints for application that can't easily be -coded to avoid them. +Starting with @strong{MySQL} 3.23.38, the Windows distribution includes +both the normal and the @strong{MySQL-Max} binaries. The main benefit +of using the normal @code{mysqld.exe} binary is that it's a little +faster and uses less resources. -@menu -* Broken Foreign KEY:: Reasons NOT to use foreign keys constraints -@end menu +Here is a list of the different @strong{MySQL} servers you can use: -@node Broken Foreign KEY, , Missing Foreign Keys, Missing Foreign Keys -@subsubsection Reasons NOT to Use Foreign Keys constraints -@cindex foreign keys, reasons not to use +@multitable @columnfractions .25 .75 +@item @code{mysqld} @tab +Compiled with full debugging and automatic memory allocation checking, +symbolic links, BDB and InnoDB tables. +@item @code{mysqld-opt} @tab +Optimized binary with no support for transactional tables. +@item @code{mysqld-nt} @tab +Optimized binary for NT with support for named pipes. You can run this +version on Win98, but in this case no named pipes are created and you must +have TCP/IP installed. +@item @code{mysqld-max} @tab +Optimized binary with support for symbolic links, BDB and InnoDB tables. +@item @code{mysqld-max-nt} @tab +Like @code{mysqld-max}, but compiled with support for named pipes. +@end multitable -There are so many problems with foreign key constraints that we don't -know where to start: +All of the above binaries are optimized for the Pentium Pro processor but +should work on any Intel processor >= i386. -@itemize @bullet -@item -Foreign key constraints make life very complicated, because the foreign -key definitions must be stored in a database and implementing them would -destroy the whole ``nice approach'' of using files that can be moved, -copied, and removed. +NOTE: If you want to use InnoDB tables, there are certain startup +options that must be specified in your @file{my.ini} file! @xref{InnoDB start}. -@item -The speed impact is terrible for @code{INSERT} and @code{UPDATE} -statements, and in this case almost all @code{FOREIGN KEY} constraint -checks are useless because you usually insert records in the right -tables in the right order, anyway. +@node Win95 start, NT start, Windows installation, Windows +@subsection Starting MySQL on Windows 95 or Windows 98 -@item -There is also a need to hold locks on many more tables when updating one -table, because the side effects can cascade through the entire database. It's -MUCH faster to delete records from one table first and subsequently delete -them from the other tables. +@strong{MySQL} uses TCP/IP to connect a client to a server. (This will +allow any machine on your network to connect to your @strong{MySQL} +server.) Because of this, you must install TCP/IP on your machine before +starting @strong{MySQL}. You can find TCP/IP on your Windows CD-ROM. -@item -You can no longer restore a table by doing a full delete from the table -and then restoring all records (from a new source or from a backup). +Note that if you are using an old Win95 release (for example OSR2), it's +likely that you have an old Winsock package! @strong{MySQL} requires +Winsock 2! You can get the newest Winsock from +@uref{http://www.microsoft.com/}. Win98 has the new Winsock 2 library, so +the above doesn't apply for Win98. -@item -If you use foreign key constraints you can't dump and restore tables -unless you do so in a very specific order. +To start the @code{mysqld} server, you should start an MS-DOS window and type: -@item -It's very easy to do ``allowed'' circular definitions that make the -tables impossible to re-create each table with a single create statement, -even if the definition works and is usable. +@example +C:\> C:\mysql\bin\mysqld +@end example -@item -It's very easy to overlook @code{FOREIGN KEY ... ON DELETE} rules when -one codes an application. It's not unusual that one loses a lot of -important information just because a wrong or misused @code{ON DELETE} rule. -@end itemize +This will start @code{mysqld} in the background without a window. -The only nice aspect of @code{FOREIGN KEY} is that it gives ODBC and some -other client programs the ability to see how a table is connected and to use -this to show connection diagrams and to help in building applications. +You can kill the @strong{MySQL} server by executing: -@strong{MySQL} will soon store @code{FOREIGN KEY} definitions so that a -client can ask for and receive an answer about how the original -connection was made. The current @file{.frm} file format does not have -any place for it. At a later stage we will implement the foreign key -constraints for application that can't easily be coded to avoid them. +@example +C:\> C:\mysql\bin\mysqladmin -u root shutdown +@end example -@node Missing Views, Missing comments, Missing Foreign Keys, Missing functions -@subsection Views -@cindex views +Note that Win95 and Win98 don't support creation of named pipes. +On Win95 and Win98, you can only use named pipes to connect to a +remote @strong{MySQL} server running on a Windows NT server host. +(The @strong{MySQL} server must also support named pipes, of +course. For example, using @code{mysqld-opt} under NT will not allow +named pipe connections. You should use either @code{mysqld-nt} or +@code{mysqld-max-nt}.) -@strong{MySQL} doesn't yet support views, but we plan to implement these -to about 4.1. +If @code{mysqld} doesn't start, please check the +@file{\mysql\data\mysql.err} file to see if the server wrote any message +there to indicate the cause of the problem. You can also try to start +the server with @code{mysqld --standalone}; In this case, you may get +some useful information on the screen that may help solve the problem. -Views are mostly useful for letting users access a set of relations as one -table (in read-only mode). Many SQL databases don't allow one to update -any rows in a view, but you have to do the updates in the separate tables. +The last option is to start @code{mysqld} with @code{--standalone +--debug}. In this case @code{mysqld} will write a log file +@file{C:\mysqld.trace} that should contain the reason why @code{mysqld} +doesn't start. @xref{Making trace files}. -As @strong{MySQL} is mostly used in applications and on web system where -the application writer has full control on the database usage, most of -our users haven't regarded views to be very important. (At least no one -has been interested enough in this to be prepared to finance the -implementation of views). +@node NT start, Windows running, Win95 start, Windows +@subsection Starting MySQL on Windows NT or Windows 2000 -One doesn't need views in @strong{MySQL} to restrict access to columns -as @strong{MySQL} has a very sophisticated privilege -system. @xref{Privilege system}. +The Win95/Win98 section also applies to @strong{MySQL} on NT/Win2000, with +the following differences: -@node Missing comments, , Missing Views, Missing functions -@subsection @samp{--} as the Start of a Comment -@cindex comments, starting -@cindex starting, comments +To get @strong{MySQL} to work with TCP/IP on NT, you must install +service pack 3 (or newer)! -Some other SQL databases use @samp{--} to start comments. @strong{MySQL} -has @samp{#} as the start comment character, even if the @code{mysql} -command-line tool removes all lines that start with @samp{--}. -You can also use the C comment style @code{/* this is a comment */} with -@strong{MySQL}. -@xref{Comments}. +Note that everything in the following that applies for NT also applies +for Win2000! -@strong{MySQL} Version 3.23.3 and above supports the @samp{--} comment style -only if the comment is followed by a space. This is because this -degenerate comment style has caused many problems with automatically -generated SQL queries that have used something like the following code, -where we automatically insert the value of the payment for -@code{!payment!}: +For NT/Win2000, the server name is @code{mysqld-nt}. Normally you +should install @strong{MySQL} as a service on NT/Win2000: @example -UPDATE tbl_name SET credit=credit-!payment! +C:\> C:\mysql\bin\mysqld-nt --install @end example -What do you think will happen when the value of @code{payment} is negative? - -Because @code{1--1} is legal in SQL, we think it is terrible that -@samp{--} means start comment. +or -In @strong{MySQL} Version 3.23 you can, however, use: -@code{1-- This is a comment} +@example +C:\> C:\mysql\bin\mysqld-max-nt --install +@end example -The following discussion only concerns you if you are running a @strong{MySQL} -version earlier than Version 3.23: +(Under Windows NT, you can actually install any of the server binaries +as a service, but only those having names that end with @code{-nt.exe} +provide support for named pipes.) -If you have a SQL program in a text file that contains @samp{--} comments -you should use: +You can start and stop the @strong{MySQL} service with these commands: @example -shell> replace " --" " #" < text-file-with-funny-comments.sql \ - | mysql database +C:\> NET START mysql +C:\> NET STOP mysql @end example -instead of the usual: +Note that in this case you can't use any other options for @code{mysqld-nt}! + +You can also run @code{mysqld-nt} as a stand-alone program on NT if you need +to start @code{mysqld-nt} with any options! If you start @code{mysqld-nt} +without options on NT, @code{mysqld-nt} tries to start itself as a service +with the default service options. If you have stopped @code{mysqld-nt}, you +have to start it with @code{NET START mysql}. + +The service is installed with the name @code{MySQL}. Once installed, it must +be started using the Services Control Manager (SCM) Utility found in the +Control Panel, or by using the @code{NET START MySQL} command. If any options +are desired, they must be specified as ``Startup parameters'' in the SCM utility +before you start the @strong{MySQL} service. Once running, @code{mysqld-nt} +can be stopped using @code{mysqladmin}, or from the SCM utility or by using +the command @code{NET STOP MySQL}. If you use SCM to stop @code{mysqld-nt}, +there is a strange message from SCM about @code{mysqld shutdown normally}. +When run as a service, @code{mysqld-nt} has no access to a console and so no +messages can be seen. + +On NT you can get the following service error messages: + +@multitable @columnfractions .3 .7 +@item Permission Denied @tab Means that it cannot find @code{mysqld-nt.exe}. +@item Cannot Register @tab Means that the path is incorrect. +@item Failed to install service. @tab Means that the service is already installed or that the Service Control Manager is in bad state. +@end multitable + +If you have problems installing @code{mysqld-nt} as a service, try starting +it with the full path: @example -shell> mysql database < text-file-with-funny-comments.sql +C:\> C:\mysql\bin\mysqld-nt --install @end example -You can also edit the command file ``in place'' to change the @samp{--} -comments to @samp{#} comments: +If this doesn't work, you can get @code{mysqld-nt} to start properly by fixing +the path in the registry! + +If you don't want to start @code{mysqld-nt} as a service, you can start it as +follows: @example -shell> replace " --" " #" -- text-file-with-funny-comments.sql +C:\> C:\mysql\bin\mysqld-nt --standalone @end example -Change them back with this command: +or @example -shell> replace " #" " --" -- text-file-with-funny-comments.sql +C:\> C:\mysql\bin\mysqld --standalone --debug @end example -@node Standards, Commit-rollback, Missing functions, Compatibility -@section What Standards Does MySQL Follow? - -Entry level SQL92. ODBC levels 0-2. +The last version gives you a debug trace in @file{C:\mysqld.trace}. +@xref{Making trace files}. -@node Commit-rollback, , Standards, Compatibility -@section How to Cope Without @code{COMMIT}/@code{ROLLBACK} -@findex COMMIT -@findex ROLLBACK -@cindex transaction-safe tables -@cindex tables, updating -@cindex updating, tables -@cindex @code{BDB} tables -@cindex @code{InnoDB} tables +@node Windows running, Windows and SSH, NT start, Windows +@subsection Running MySQL on Windows -The following mostly applies only for @code{ISAM}, @code{MyISAM}, and -@code{HEAP} tables. If you only use transaction-safe tables (@code{BDB}, -or @code{InnoDB} tables) in an an update, you can do -@code{COMMIT} and @code{ROLLBACK} also with @strong{MySQL}. -@xref{COMMIT}. +@cindex TCP/IP +@cindex named pipes -The problem with handling @code{COMMIT}-@code{ROLLBACK} efficiently with -the above table types would require a completely different table layout -than @strong{MySQL} uses today. The table type would also need extra -threads that do automatic cleanups on the tables, and the disk usage -would be much higher. This would make these table types about 2-4 times -slower than they are today. +@strong{MySQL} supports TCP/IP on all Windows platforms and named pipes on NT. +The default is to use named pipes for local connections on NT and TCP/IP for +all other cases if the client has TCP/IP installed. The host name specifies +which protocol is used: -For the moment, we prefer implementing the SQL server language (something -like stored procedures). With this you would very seldom really need -@code{COMMIT}-@code{ROLLBACK.} This would also give much better performance. +@multitable @columnfractions .3 .7 +@strong{Host name} @tab @strong{Protocol} +@item NULL (none) @tab On NT, try named pipes first; if that doesn't work, use TCP/IP. On Win95/Win98, TCP/IP is used. +@item . @tab Named pipes +@item localhost @tab TCP/IP to current host +@item hostname @tab TCP/IP +@end multitable -Loops that need transactions normally can be coded with the help of -@code{LOCK TABLES}, and you don't need cursors when you can update records -on the fly. +You can force a @strong{MySQL} client to use named pipes by specifying the +@code{--pipe} option or by specifying @code{.} as the host name. Use the +@code{--socket} option to specify the name of the pipe. -We at TcX had a greater need for a real fast database than a 100% -general database. Whenever we find a way to implement these features without -any speed loss, we will probably do it. For the moment, there are many more -important things to do. Check the TODO for how we prioritize things at -the moment. (Customers with higher levels of support can alter this, so -things may be reprioritized.) +You can test whether or not @strong{MySQL} is working by executing the +following commands: -The current problem is actually @code{ROLLBACK}. Without -@code{ROLLBACK}, you can do any kind of @code{COMMIT} action with -@code{LOCK TABLES}. To support @code{ROLLBACK} with the above table -types, @strong{MySQL} would have to be changed to store all old records -that were updated and revert everything back to the starting point if -@code{ROLLBACK} was issued. For simple cases, this isn't that hard to do -(the current @code{isamlog} could be used for this purpose), but it -would be much more difficult to implement @code{ROLLBACK} for -@code{ALTER/DROP/CREATE TABLE}. +@example +C:\> C:\mysql\bin\mysqlshow +C:\> C:\mysql\bin\mysqlshow -u root mysql +C:\> C:\mysql\bin\mysqladmin version status proc +C:\> C:\mysql\bin\mysql test +@end example -To avoid using @code{ROLLBACK}, you can use the following strategy: +If @code{mysqld} is slow to answer to connections on Win95/Win98, there is +probably a problem with your DNS. In this case, start @code{mysqld} with +@code{--skip-name-resolve} and use only @code{localhost} and IP numbers in +the @strong{MySQL} grant tables. You can also avoid DNS when connecting to a +@code{mysqld-nt} @strong{MySQL} server running on NT by using the +@code{--pipe} argument to specify use of named pipes. This works for most +@strong{MySQL} clients. -@enumerate +There are two versions of the @strong{MySQL} command-line tool: +@multitable @columnfractions .25 .75 +@item @code{mysql} @tab Compiled on native Windows, which offers very limited text editing capabilities. +@item @code{mysqlc} @tab Compiled with the Cygnus GNU compiler and libraries, which offers @code{readline} editing. +@end multitable + +If you want to use @code{mysqlc.exe}, you must copy +@file{C:\mysql\lib\cygwinb19.dll} to your Windows system directory +(@file{\windows\system} or similar place). + +The default privileges on Windows give all local users full privileges +to all databases without specifying a password. To make @strong{MySQL} +more secure, you should set a password for all users and remove the row in +the @code{mysql.user} table that has @code{Host='localhost'} and +@code{User=''}. + +You should also add a password for the @code{root} user. The following +example starts by removing the anonymous user that can be used by anyone +to access the @code{test} database, then sets a @code{root} user password: + +@example +C:\> C:\mysql\bin\mysql mysql +mysql> DELETE FROM user WHERE Host='localhost' AND User=''; +mysql> QUIT +C:\> C:\mysql\bin\mysqladmin reload +C:\> C:\mysql\bin\mysqladmin -u root password your_password +@end example + +After you've set the password, if you want to take down the @code{mysqld} +server, you can do so using this command: + +@example +C:\> mysqladmin --user=root --password=your_password shutdown +@end example + +If you are using the old shareware version of @strong{MySQL} Version +3.21 under Windows, the above command will fail with an error: +@code{parse error near 'SET OPTION password'}. The fix is in to upgrade +to the current @strong{MySQL} version, which is freely available. + +With the current @strong{MySQL} versions you can easily add new users +and change privileges with @code{GRANT} and @code{REVOKE} commands. +@xref{GRANT}. + +@c FIX this is ugly, real ugly. +@cindex SSH +@cindex connecting, remotely with SSH +@node Windows and SSH, Windows symbolic links, Windows running, Windows +@subsection Connecting to a Remote MySQL from Windows with SSH + +Here is a note about how to connect to get a secure connection to remote +@strong{MySQL} server with SSH (by David Carlson @email{dcarlson@@mplcomm.com}): + +@itemize @bullet @item -Use @code{LOCK TABLES ...} to lock all the tables you want to access. +Install an SSH client on your Windows machine. As a user, the best non-free +one I've found is from @code{SecureCRT} from @uref{http://www.vandyke.com/}. +Another option is @code{f-secure} from @uref{http://www.f-secure.com/}. You +can also find some free ones on @strong{Google} at +@uref{http://directory.google.com/Top/Computers/Security/Products_and_Tools/Cryptography/SSH/Clients/Windows/}. + @item -Test conditions. +Start your Windows SSH client. +Set @code{Host_Name = yourmysqlserver_URL_or_IP}. +Set @code{userid=your_userid} to log in to your server (probably not the same +as your @strong{MySQL} login/password. + @item -Update if everything is okay. +Set up port forwarding. Either do a remote forward (Set @code{local_port: 3306}, @code{remote_host: yourmysqlservername_or_ip}, @code{remote_port: 3306} ) +or a local forward (Set @code{port: 3306}, @code{host: localhost}, @code{remote port: 3306}). + @item -Use @code{UNLOCK TABLES} to release your locks. -@end enumerate +Save everything, otherwise you'll have to redo it the next time. -This is usually a much faster method than using transactions with possible -@code{ROLLBACK}s, although not always. The only situation this solution -doesn't handle is when someone kills the threads in the middle of an -update. In this case, all locks will be released but some of the updates may -not have been executed. +@item +Log in to your server with SSH session you just created. -You can also use functions to update records in a single operation. -You can get a very efficient application by using the following techniques: +@item +On your Windows machine, start some ODBC application (such as Access). -@itemize @bullet -@item Modify fields relative to their current value. -@item Update only those fields that actually have changed. +@item +Create a new file in Windows and link to @strong{MySQL} using the ODBC +driver the same way you normally do, EXCEPT type in @code{localhost} +for the @strong{MySQL} host server --- not @code{yourmysqlservername}. @end itemize -For example, when we are doing updates to some customer information, we -update only the customer data that has changed and test only that none of -the changed data, or data that depend on the changed data, has changed -compared to the original row. The test for changed data is done with the -@code{WHERE} clause in the @code{UPDATE} statement. If the record wasn't -updated, we give the client a message: "Some of the data you have changed -have been changed by another user". Then we show the old row versus the new -row in a window, so the user can decide which version of the customer record -he should use. - -This gives us something that is similar to column locking but is actually -even better, because we only update some of the columns, using values that -are relative to their current values. This means that typical @code{UPDATE} -statements look something like these: +You should now have an ODBC connection to @strong{MySQL}, encrypted using SSH. -@example -UPDATE tablename SET pay_back=pay_back+'relative change'; +@cindex symbolic links +@cindex using multiple disks to start data +@cindex disks, splitting data across +@node Windows symbolic links, Windows compiling, Windows and SSH, Windows +@subsection Splitting Data Across Different Disks on Windows -UPDATE customer - SET - customer_date='current_date', - address='new address', - phone='new phone', - money_he_owes_us=money_he_owes_us+'new_money' - WHERE - customer_id=id AND address='old address' AND phone='old phone'; -@end example +Beginning with @strong{MySQL} Version 3.23.16, the @code{mysqld-max} +and @code{mysql-max-nt} servers in the @strong{MySQL} distribution are +compiled with the @code{-DUSE_SYMDIR} option. This allows you to put a +database on different disk by adding a symbolic link to it +(in a manner similar to the way that symbolic links work on Unix). -As you can see, this is very efficient and works even if another client has -changed the values in the @code{pay_back} or @code{money_he_owes_us} columns. +On Windows, you make a symbolic link to a database by creating a file +that contains the path to the destination directory and saving this in +the @file{mysql_data} directory under the filename @file{database.sym}. +Note that the symbolic link will be used only if the directory +@file{mysql_data_dir\database} doesn't exist. -@findex mysql_insert_id() -@findex LAST_INSERT_ID() -In many cases, users have wanted @code{ROLLBACK} and/or @code{LOCK -TABLES} for the purpose of managing unique identifiers for some tables. This -can be handled much more efficiently by using an @code{AUTO_INCREMENT} column -and either the SQL function @code{LAST_INSERT_ID()} or the C API function -@code{mysql_insert_id()}. @xref{mysql_insert_id, , @code{mysql_insert_id()}}. +For example, if the @strong{MySQL} data directory is @file{C:\mysql\data} +and you want to have database @code{foo} located at @file{D:\data\foo}, you +should create the file @file{C:\mysql\data\foo.sym} that contains the +text @code{D:\data\foo\}. After that, all tables created in the database +@code{foo} will be created in @file{D:\data\foo}. -@cindex rows, locking -At @strong{MySQL AB}, we have never had any need for row-level locking -because we have always been able to code around it. Some cases really need -row locking, but they are very few. If you want row-level locking, you -can use a flag column in the table and do something like this: +Note that because of the speed penalty you get when opening every table, +we have not enabled this by default even if you have compiled +@strong{MySQL} with support for this. To enable symlinks you should put +in your @code{my.cnf} or @code{my.ini} file the following entry: @example -UPDATE tbl_name SET row_flag=1 WHERE id=ID; +[mysqld] +use-symbolic-links @end example -@strong{MySQL} returns 1 for the number of affected rows if the row was -found and @code{row_flag} wasn't already 1 in the original row. +In @strong{MySQL} 4.0 we will enable symlinks by default. Then you +should instead use the @code{skip-symlink} option if you want to +disable this. -You can think of it as @strong{MySQL} changed the above query to: +@cindex compiling, on Windows +@cindex Windows, compiling on +@node Windows compiling, Windows vs Unix, Windows symbolic links, Windows +@subsection Compiling MySQL Clients on Windows + +In your source files, you should include @file{windows.h} before you include +@file{mysql.h}: @example -UPDATE tbl_name SET row_flag=1 WHERE id=ID and row_flag <> 1; +#if defined(_WIN32) || defined(_WIN64) +#include <windows.h> +#endif +#include <mysql.h> @end example -@node Privilege system, Reference, Compatibility, Top -@chapter The MySQL Access Privilege System -@cindex system, security -@cindex access privileges -@cindex privileges, access -@cindex security system -@cindex ACLs - - -@strong{MySQL} has an advanced but non-standard security/privilege -system. This section describes how it works. +You can either link your code with the dynamic @file{libmysql.lib} library, +which is just a wrapper to load in @file{libmysql.dll} on demand, or link +with the static @file{mysqlclient.lib} library. -@menu -* General security:: General security -* Security:: How to make @strong{MySQL} secure against crackers -* Privileges options:: -* What Privileges:: What the privilege system does -* User names:: @strong{MySQL} user names and passwords -* Connecting:: Connecting to the @strong{MySQL} server -* Password security:: Keeping your password secure -* Privileges provided:: Privileges provided by @strong{MySQL} -* Privileges:: How the privilege system works -* Connection access:: Access control, stage 1: Connection verification -* Request access:: Access control, stage 2: Request verification -* Privilege changes:: When privilege changes take effect -* Default privileges:: Setting up the initial @strong{MySQL} privileges -* Adding users:: Adding new users to @strong{MySQL} -* Passwords:: How to set up passwords -* Access denied:: Causes of @code{Access denied} errors -@end menu +Note that as the mysqlclient libraries are compiled as threaded libraries, +you should also compile your code to be multi-threaded! -@node General security, Security, Privilege system, Privilege system -@section General Security +@cindex Windows, versus Unix +@cindex operating systems, Windows versus Unix +@node Windows vs Unix, , Windows compiling, Windows +@subsection MySQL-Windows Compared to Unix MySQL -Anyone using @strong{MySQL} on a computer connected to the Internet -should read this section to avoid the most common security mistakes. +@strong{MySQL}-Windows has by now proven itself to be very stable. This version +of @strong{MySQL} has the same features as the corresponding Unix version +with the following exceptions: -In discussing security, we emphasize the necessity of fully protecting the -entire server host (not simply the @strong{MySQL} server) against all types -of applicable attacks: eavesdropping, altering, playback, and denial of -service. We do not cover all aspects of availability and fault tolerance -here. +@table @strong +@item Win95 and threads +Win95 leaks about 200 bytes of main memory for each thread creation. +Each connection in @strong{MySQL} creates a new thread, so you shouldn't +run @code{mysqld} for an extended time on Win95 if your server handles +many connections! WinNT and Win98 don't suffer from this bug. -@strong{MySQL} uses security based on Access Control Lists (ACLs) for all -connections, queries, and other operations that a user may attempt to -perform. There is also some support for SSL-encrypted connections -between @strong{MySQL} clients and servers. Many of the concepts -discussed here are not specific to @strong{MySQL} at all; the same -general ideas apply to almost all applications. +@item Concurrent reads +@strong{MySQL} depends on the @code{pread()} and @code{pwrite()} calls to be +able to mix @code{INSERT} and @code{SELECT}. Currently we use mutexes +to emulate @code{pread()}/@code{pwrite()}. We will, in the long run, +replace the file level interface with a virtual interface so that we can +use the @code{readfile()}/@code{writefile()} interface on NT to get more speed. +The current implementation limits the number of open files @strong{MySQL} +can use to 1024, which means that you will not be able to run as many +concurrent threads on NT as on Unix. -When running @strong{MySQL}, follow these guidelines whenever possible: +@item Blocking read +@strong{MySQL} uses a blocking read for each connection. +This means that: @itemize @bullet @item -DON'T EVER GIVE ANYONE (EXCEPT THE @strong{MySQL} ROOT USER) ACCESS TO THE -@code{user} TABLE IN THE @code{mysql} DATABASE! The encrypted password -is the real password in @strong{MySQL}. If you know the password listed in -the @code{user} table for a given user, you can easily log in as that -user if you have access to the host listed for that account. +A connection will not be disconnected automatically after 8 hours, as happens +with the Unix version of @strong{MySQL}. @item -Learn the @strong{MySQL} access privilege system. The @code{GRANT} and -@code{REVOKE} commands are used for controlling access to @strong{MySQL}. Do -not grant any more privileges than necessary. Never grant privileges to all -hosts. +If a connection hangs, it's impossible to break it without killing +@strong{MySQL}. -Checklist: -@itemize @minus @item -Try @code{mysql -u root}. If you are able to connect successfully to the -server without being asked for a password, you have problems. Anyone -can connect to your @strong{MySQL} server as the @strong{MySQL} -@code{root} user with full privileges! -Review the @strong{MySQL} installation instructions, paying particular -attention to the item about setting a @code{root} password. +@code{mysqladmin kill} will not work on a sleeping connection. + @item -Use the command @code{SHOW GRANTS} and check to see who has access to -what. Remove those privileges that are not necessary using the @code{REVOKE} -command. +@code{mysqladmin shutdown} can't abort as long as there are sleeping +connections. @end itemize -@item -Do not keep any plain-text passwords in your database. When your -computer becomes compromised, the intruder can take the full list of -passwords and use them. Instead use @code{MD5()} or another one-way -hashing function. -@item -Do not choose passwords from dictionaries. There are special programs to -break them. Even passwords like ``xfish98'' are very bad. Much better is -``duag98'' which contains the same word ``fish'' but typed one key to the -left on a standard QWERTY keyboard. Another method is to use ``Mhall'' which -is taken from the first characters of each word in the sentence ``Mary had -a little lamb.'' This is easy to remember and type, but difficult to guess -for someone who does not know it. -@item -Invest in a firewall. This protects you from at least 50% of all types of -exploits in any software. Put @strong{MySQL} behind the firewall or in -a demilitarized zone (DMZ). +We plan to fix this problem when our Windows developers have figured out a +nice workaround. -Checklist: -@itemize @minus -@item -Try to scan your ports from the Internet using a tool such as -@code{nmap}. @strong{MySQL} uses port 3306 by default. This port should -be inaccessible from untrusted hosts. Another simple way to check whether -or not your @strong{MySQL} port is open is to try the following command -from some remote machine, where @code{server_host} is the hostname of -your @strong{MySQL} server: +@item UDF functions +For the moment, @strong{MySQL}-Windows does not support user-definable +functions. + +@item @code{DROP DATABASE} +You can't drop a database that is in use by some thread. + +@item Killing @strong{MySQL} from the task manager +You can't kill @strong{MySQL} from the task manager or with the shutdown +utility in Win95. You must take it down with @code{mysqladmin shutdown}. + +@item Case-insensitive names +Filenames are case insensitive on Windows, so database and table names +are also case insensitive in @strong{MySQL} for Windows. The only +restriction is that database and table names must be specified using the same +case throughout a given statement. @xref{Name case sensitivity}. + +@item The @samp{\} directory character +Pathname components in Win95 are separated by the @samp{\} character, which is +also the escape character in @strong{MySQL}. If you are using @code{LOAD +DATA INFILE} or @code{SELECT ... INTO OUTFILE}, you must double the @samp{\} +character: @example -shell> telnet server_host 3306 +mysql> LOAD DATA INFILE "C:\\tmp\\skr.txt" INTO TABLE skr; +mysql> SELECT * INTO OUTFILE 'C:\\tmp\\skr.txt' FROM skr; @end example -If you get a connection and some garbage characters, the port is -open, and should be closed on your firewall or router, unless you really -have a good reason to keep it open. If @code{telnet} just hangs or the -connection is refused, everything is OK; the port is blocked. -@end itemize +Alternatively, use Unix style filenames with @samp{/} characters: -@item -Do not trust any data entered by your users. They can try to trick your -code by entering special or escaped character sequences in Web forms, -URLs, or whatever application you have built. Be sure that your -application remains secure if a user enters something like ``@code{; DROP -DATABASE mysql;}''. This is an extreme example, but large security leaks -and data loss may occur as a result of hackers using similar techniques, -if you do not prepare for them. +@example +mysql> LOAD DATA INFILE "C:/tmp/skr.txt" INTO TABLE skr; +mysql> SELECT * INTO OUTFILE 'C:/tmp/skr.txt' FROM skr; +@end example -Also remember to check numeric data. A common mistake is to protect only -strings. Sometimes people think that if a database contains only publicly -available data that it need not be protected. This is incorrect. At least -denial-of-service type attacks can be performed on such -databases. The simplest way to protect from this type of attack is to use -apostrophes around the numeric constants: @code{SELECT * FROM table -WHERE ID='234'} rather than @code{SELECT * FROM table WHERE ID=234}. -@strong{MySQL} automatically converts this string to a number and -strips all non-numeric symbols from it. +@item @code{Can't open named pipe} error +If you use a @strong{MySQL} 3.22 version on NT with the newest mysql-clients +you will get the following error: + +@example +error 2017: can't open named pipe to host: . pipe... +@end example + +@tindex .my.cnf file +This is because the release version of @strong{MySQL} uses named pipes on NT +by default. You can avoid this error by using the @code{--host=localhost} +option to the new @strong{MySQL} clients or create an option file +@file{C:\my.cnf} that contains the following information: + +@example +[client] +host = localhost +@end example + +@item @code{Access denied for user} error +If you get the error @code{Access denied for user: 'some-user@@unknown' +to database 'mysql'} when accessing a @strong{MySQL} server on the same +machine, this means that @strong{MySQL} can't resolve your host name +properly. + +To fix this, you should create a file @file{\windows\hosts} with the +following information: + +@example +127.0.0.1 localhost +@end example + +@item @code{ALTER TABLE} +While you are executing an @code{ALTER TABLE} statement, the table is locked +from usage by other threads. This has to do with the fact that on Windows, +you can't delete a file that is in use by another threads. (In the future, +we may find some way to work around this problem.) + +@item @code{DROP TABLE} on a table that is in use by a @code{MERGE} table will not work +The @code{MERGE} handler does its table mapping hidden from @strong{MySQL}. +Because Windows doesn't allow you to drop files that are open, you first +must flush all @code{MERGE} tables (with @code{FLUSH TABLES}) or drop the +@code{MERGE} table before dropping the table. We will fix this at the same +time we introduce @code{VIEW}s. +@end table + +Here are some open issues for anyone who might want to help us with the Windows +release: + +@cindex Windows, open issues -Checklist: -@itemize @minus -@item -All Web applications: @itemize @bullet @item -Try to enter @samp{'} and @samp{"} in all your Web forms. If you get any kind -of @strong{MySQL} error, investigate the problem right away. +Make a single-user @code{MYSQL.DLL} server. This should include everything in +a standard @strong{MySQL} server, except thread creation. This will make +@strong{MySQL} much easier to use in applications that don't need a true +client/server and don't need to access the server from other hosts. + @item -Try to modify any dynamic URLs by adding @code{%22} (@samp{"}), @code{%23} -(@samp{#}), and @code{%27} (@samp{'}) in the URL. +Add some nice start and shutdown icons to the @strong{MySQL} installation. + @item -Try to modify datatypes in dynamic URLs from numeric ones to character -ones containing characters from previous examples. Your application -should be safe against this and similar attacks. +Create a tool to manage registry entries for the @strong{MySQL} startup +options. The registry entry reading is already coded into @file{mysqld.cc}, +but it should be recoded to be more parameter oriented. The tool should +also be able to update the @file{C:\my.cnf} option file if the user prefers +to use that instead of the registry. + @item -Try to enter characters, spaces, and special symbols instead of numbers in -numeric fields. Your application should remove them before passing them to -@strong{MySQL} or your application should generate an error. Passing -unchecked values to @strong{MySQL} is very dangerous! +When registering @code{mysqld} as a service with @code{--install} (on NT) +it would be nice if you could also add default options on the command line. +For the moment, the workaround is to list the parameters in the +@file{C:\my.cnf} file instead. + @item -Check data sizes before passing them to @strong{MySQL}. +It would be real nice to be able to kill @code{mysqld} from the task manager. +For the moment, you must use @code{mysqladmin shutdown}. + @item -Consider having your application connect to the database using a -different user name than the one you use for administrative purposes. Do -not give your applications any more access privileges than they need. -@end itemize -@item -Users of PHP: -@itemize @bullet -@item Check out the @code{addslashes()} function. -As of PHP 4.0.3, a @code{mysql_escape_string()} function is available -that is based on the function of the same name in the @strong{MySQL} C API. -@end itemize -@item -Users of @strong{MySQL} C API: -@itemize @bullet -@item Check out the @code{mysql_escape_string()} API call. -@end itemize -@item -Users of @strong{MySQL}++: -@itemize @bullet -@item Check out the @code{escape} and @code{quote} modifiers for query streams. -@end itemize +Port @code{readline} to Windows for use in the @code{mysql} command line tool. + @item -Users of Perl DBI: -@itemize @bullet -@item Check out the @code{quote()} method or use placeholders. -@end itemize +GUI versions of the standard @strong{MySQL} clients (@code{mysql}, +@code{mysqlshow}, @code{mysqladmin}, and @code{mysqldump}) would be nice. + @item -Users of Java JDBC: -@itemize @bullet -@item Use a @code{PreparedStatement} object and placeholders. -@end itemize -@end itemize +It would be nice if the socket read and write functions in @file{net.c} were +interruptible. This would make it possible to kill open threads with +@code{mysqladmin kill} on Windows. @item -Do not transmit plain (unencrypted) data over the Internet. These data are -accessible to everyone who has the time and ability to intercept it and use -it for their own purposes. Instead, use an encrypted protocol such as SSL or -SSH. @strong{MySQL} supports internal SSL connections as of Version 3.23.9. -SSH port-forwarding can be used to create an encrypted (and compressed) -tunnel for the communication. +@code{mysqld} always starts in the "C" locale and not in the default locale. +We would like to have @code{mysqld} use the current locale for the sort order. + @item -Learn to use the @code{tcpdump} and @code{strings} utilities. For most cases, -you can check whether or not @strong{MySQL} data streams are unencrypted -by issuing a command like the following: +Implement UDF functions with @code{.DLL}s. -@example -shell> tcpdump -l -i eth0 -w - src or dst port 3306 | strings -@end example +@item +Add macros to use the faster thread-safe increment/decrement methods +provided by Windows. -(This works under Linux and should work with small modifications under -other systems). Warning: If you do not see data this doesn't always -actually mean that it is encrypted. If you need high security, you should -consult with a security expert. @end itemize -@node Security, Privileges options, General security, Privilege system -@section How to Make MySQL Secure Against Crackers -@cindex crackers, security against -@cindex security, against crackers - -When you connect to a @strong{MySQL} server, you normally should use a -password. The password is not transmitted in clear text over the -connection, however the encryption algorithm is not very strong, and -with some effort a clever attacker can crack the password if he is able -to sniff the traffic between the client and the server. If the -connection between the client and the server goes through an untrusted -network, you should use an SSH tunnel to encrypt the -communication. - -All other information is transferred as text that can be read by anyone -who is able to watch the connection. If you are concerned about this, -you can use the compressed protocol (in @strong{MySQL} Version 3.22 and above) -to make things much harder. To make things even more secure you should use -@code{ssh}. You can find an open source @code{ssh} client at -@uref{http://www.openssh.org}, and a commercial @code{ssh} client at -@uref{http://www.ssh.com}. With this, you can get an encrypted TCP/IP -connection between a @strong{MySQL} server and a @strong{MySQL} client. +Other Windows-specific issues are described in the @file{README} file that +comes with the @strong{MySQL}-Windows distribution. -To make a @strong{MySQL} system secure, you should strongly consider the -following suggestions: +@node OS/2, MySQL binaries, Windows, Installing +@section OS/2 Notes -@itemize @bullet -@item -Use passwords for all @strong{MySQL} users. Remember that anyone can log in -as any other person as simply as @code{mysql -u other_user db_name} if -@code{other_user} has no password. It is common behavior with client/server -applications that the client may specify any user name. You can change the -password of all users by editing the @code{mysql_install_db} script before -you run it, or only the password for the @strong{MySQL} @code{root} user like -this: +@strong{MySQL} uses quite a few open files. Because of this, you should add +something like the following to your @file{CONFIG.SYS} file: @example -shell> mysql -u root mysql -mysql> UPDATE user SET Password=PASSWORD('new_password') - WHERE user='root'; -mysql> FLUSH PRIVILEGES; +SET EMXOPT=-c -n -h1024 @end example -@item -Don't run the @strong{MySQL} daemon as the Unix @code{root} user. This is -very dangerous, because any user with @code{FILE} privileges will be able -to create files as @code{root} (for example, @code{~root/.bashrc}). To -prevent this, @code{mysqld} will refuse to run as @code{root} unless it -is specified directly using a @code{--user=root} option. - -@code{mysqld} can be run as an ordinary unprivileged user instead. -You can also create a new Unix user @code{mysql} to make everything -even more secure. If you run @code{mysqld} as another Unix user, -you don't need to change the @code{root} user name in the @code{user} -table, because @strong{MySQL} user names have nothing to do with Unix -user names. To start @code{mysqld} as another Unix user, add a @code{user} -line that specifies the user name to the @code{[mysqld]} group of the -@file{/etc/my.cnf} option file or the @file{my.cnf} option file in the -server's data directory. For example: +If you don't do this, you will probably run into the following error: @example -[mysqld] -user=mysql +File 'xxxx' not found (Errcode: 24) @end example -This will cause the server to start as the designated user whether you -start it manually or by using @code{safe_mysqld} or @code{mysql.server}. -For more details, see @ref{Changing MySQL user, , Changing @strong{MySQL} -user}. +When using @strong{MySQL} with OS/2 Warp 3, FixPack 29 or above is +required. With OS/2 Warp 4, FixPack 4 or above is required. This is a +requirement of the Pthreads library. @strong{MySQL} must be installed +in a partition that supports long filenames such as HPFS, FAT32, etc. -@item -Don't support symlinks to tables (This can be disabled with the -@code{--skip-symlink} option. This is especially important if you run -@code{mysqld} as root as anyone that has write access to the mysqld data -directories could then delete any file in the system! -@xref{Symbolic links to tables}. +The @file{INSTALL.CMD} script must be run from OS/2's own @file{CMD.EXE} +and may not work with replacement shells such as @file{4OS2.EXE}. -@item -Check that the Unix user that @code{mysqld} runs as is the only user with -read/write privileges in the database directories. +The @file{scripts/mysql-install-db} script has been renamed. It is now called +@file{install.cmd} and is a REXX script, which will set up the default +@strong{MySQL} security settings and create the WorkPlace Shell icons +for @strong{MySQL}. -@item -Don't give the @strong{process} privilege to all users. The output of -@code{mysqladmin processlist} shows the text of the currently executing -queries, so any user who is allowed to execute that command might be able to -see if another user issues an @code{UPDATE user SET -password=PASSWORD('not_secure')} query. +Dynamic module support is compiled in but not fully tested. Dynamic +modules should be compiled using the Pthreads run-time library. -@code{mysqld} reserves an extra connection for users who have the -@strong{process} privilege, so that a @strong{MySQL} @code{root} user can log -in and check things even if all normal connections are in use. +@example +gcc -Zdll -Zmt -Zcrtdll=pthrdrtl -I../include -I../regex -I.. \ + -o example udf_example.cc -L../lib -lmysqlclient udf_example.def +mv example.dll example.udf +@end example -@item -Don't give the @strong{file} privilege to all users. Any user that has this -privilege can write a file anywhere in the file system with the privileges of -the @code{mysqld} daemon! To make this a bit safer, all files generated with -@code{SELECT ... INTO OUTFILE} are readable to everyone, and you cannot -overwrite existing files. +@strong{Note:} Due to limitations in OS/2, UDF module name stems must not +exceed 8 characters. Modules are stored in the @file{/mysql2/udf} +directory; the @code{safe-mysqld.cmd} script will put this directory in +the @code{BEGINLIBPATH} environment variable. When using UDF modules, +specified extensions are ignored --- it is assumed to be @file{.udf}. +For example, in Unix, the shared module might be named @file{example.so} +and you would load a function from it like this: -@tindex /etc/passwd -The @strong{file} privilege may also be used to read any file accessible -to the Unix user that the server runs as. This could be abused, for example, -by using @code{LOAD DATA} to load @file{/etc/passwd} into a table, which -can then be read with @code{SELECT}. +@example +mysql> CREATE FUNCTION metaphon RETURNS STRING SONAME "example.so"; +@end example -@item -If you don't trust your DNS, you should use IP numbers instead of hostnames -in the grant tables. In principle, the @code{--secure} option to -@code{mysqld} should make hostnames safe. In any case, you should be very -careful about creating grant table entries using hostname values that -contain wild cards! +Is OS/2, the module would be named @file{example.udf}, but you would not +specify the module extension: -@item -If you want to restrict the number of connections for a single user, you -can do this by setting the @code{max_user_connections} variable in -@code{mysqld}. -@end itemize +@example +mysql> CREATE FUNCTION metaphon RETURNS STRING SONAME "example"; +@end example -@node Privileges options, What Privileges, Security, Privilege system -@section Startup Options for @code{mysqld} Concerning Security +@cindex binary distributions +@node MySQL binaries, Post-installation, OS/2, Installing +@section MySQL Binaries -The following @code{mysqld} options affect networking security: +As a service, we at @strong{MySQL AB} provide a set of binary distributions +of @strong{MySQL} that are compiled at our site or at sites where customers +kindly have given us access to their machines. -@table @code -@item --secure -IP numbers returned by the @code{gethostbyname()} system call are -checked to make sure they resolve back to the original hostname. This -makes it harder for someone on the outside to get access by pretending -to be another host. This option also adds some sanity checks of -hostnames. The option is turned off by default in @strong{MySQL} Version -3.21 because sometimes it takes a long time to perform backward resolutions. -@strong{MySQL} Version 3.22 caches hostnames and has this option enabled by -default. +These distributions are generated with @code{scripts/make_binary_distribution} +and are configured with the following compilers and options: -@item --skip-grant-tables -This option causes the server not to use the privilege system at all. This -gives everyone @emph{full access} to all databases! (You can tell a running -server to start using the grant tables again by executing @code{mysqladmin -flush-privileges} or @code{mysqladmin reload}.) +@table @asis +@item SunOS 4.1.4 2 sun4c with @code{gcc} 2.7.2.1 +@code{CC=gcc CXX=gcc CXXFLAGS="-O3 -felide-constructors" ./configure --prefix=/usr/local/mysql --disable-shared --with-extra-charsets=complex --enable-assembler} -@item --skip-name-resolve -Hostnames are not resolved. All @code{Host} column values in the grant -tables must be IP numbers or @code{localhost}. +@item SunOS 5.5.1 (and above) sun4u with @code{egcs} 1.0.3a or 2.90.27 or gcc 2.95.2 and newer +@code{CC=gcc CFLAGS="-O3" CXX=gcc CXXFLAGS="-O3 -felide-constructors -fno-exceptions -fno-rtti" ./configure --prefix=/usr/local/mysql --with-low-memory --with-extra-charsets=complex --enable-assembler} -@item --skip-networking -Don't allow TCP/IP connections over the network. All connections to -@code{mysqld} must be made via Unix sockets. This option is unsuitable for -systems that use MIT-pthreads, because the MIT-pthreads package doesn't -support Unix sockets. +@item SunOS 5.6 i86pc with @code{gcc} 2.8.1 +@code{CC=gcc CXX=gcc CXXFLAGS=-O3 ./configure --prefix=/usr/local/mysql --with-low-memory --with-extra-charsets=complex} -@item --skip-show-database -With this option, the -@code{SHOW DATABASES} statement doesn't return anything. +@item Linux 2.0.33 i386 with @code{pgcc} 2.90.29 (@code{egcs} 1.0.3a) +@code{CFLAGS="-O3 -mpentium -mstack-align-double" CXX=gcc CXXFLAGS="-O3 -mpentium -mstack-align-double -felide-constructors -fno-exceptions -fno-rtti" ./configure --prefix=/usr/local/mysql --enable-assembler --with-mysqld-ldflags=-all-static --with-extra-charsets=complex} -@item --safe-show-database -With this option, -@code{SHOW DATABASES} returns only those databases for which the user has -some kind of privilege. +@item Linux 2.2.x with x686 with @code{gcc} 2.95.2 +@code{CFLAGS="-O3 -mpentiumpro" CXX=gcc CXXFLAGS="-O3 -mpentiumpro -felide-constructors -fno-exceptions -fno-rtti" ./configure --prefix=/usr/local/mysql --enable-assembler --with-mysqld-ldflags=-all-static --disable-shared --with-extra-charset=complex} -@end table +@item SCO 3.2v5.0.4 i386 with @code{gcc} 2.7-95q4 +@code{CC=gcc CXX=gcc CXXFLAGS=-O3 ./configure --prefix=/usr/local/mysql --with-extra-charsets=complex} -@node What Privileges, User names, Privileges options, Privilege system -@section What the Privilege System Does -@cindex system, privilege -@cindex privilege system -@cindex passwords, security +@item AIX 2 4 with @code{gcc} 2.7.2.2 +@code{CC=gcc CXX=gcc CXXFLAGS=-O3 ./configure --prefix=/usr/local/mysql --with-extra-charsets=complex} -The primary function of the @strong{MySQL} privilege system is to -authenticate a user connecting from a given host, and to associate that user -with privileges on a database such as -@strong{select}, @strong{insert}, @strong{update} and @strong{delete}. +@item OSF1 V4.0 564 alpha with @code{gcc} 2.8.1 +@code{CC=gcc CFLAGS=-O CXX=gcc CXXFLAGS=-O3 ./configure --prefix=/usr/local/mysql --with-low-memory --with-extra-charsets=complex} -Additional functionality includes the ability to have an anonymous user and -to grant privileges for @strong{MySQL}-specific functions such as @code{LOAD -DATA INFILE} and administrative operations. +@item Irix 6.3 IP32 with @code{gcc} 2.8.0 +@code{CC=gcc CXX=gcc CXXFLAGS=-O3 ./configure --prefix=/usr/local/mysql --with-extra-charsets=complex} -@node User names, Connecting, What Privileges, Privilege system -@section MySQL User Names and Passwords -@cindex user names, and passwords -@cindex passwords, for users +@item BSDI BSD/OS 3.1 i386 with @code{gcc} 2.7.2.1 +@code{CC=gcc CXX=gcc CXXFLAGS=-O ./configure --prefix=/usr/local/mysql --with-extra-charsets=complex} -There are several distinctions between the way user names and passwords are -used by @strong{MySQL} and the way they are used by Unix or Windows: +@item BSDI BSD/OS 2.1 i386 with @code{gcc} 2.7.2 +@code{CC=gcc CXX=gcc CXXFLAGS=-O3 ./configure --prefix=/usr/local/mysql --with-extra-charsets=complex} +@end table -@itemize @bullet -@item -User names, as used by @strong{MySQL} for authentication purposes, have -nothing to do with Unix user names (login names) or Windows user names. Most -@strong{MySQL} clients by default try to log in using the current Unix user -name as the @strong{MySQL} user name, but that is for convenience only. -Client programs allow a different name to be specified with the @code{-u} or -@code{--user} options. This means that you can't make a database secure in -any way unless all @strong{MySQL} user names have passwords. Anyone may -attempt to connect to the server using any name, and they will succeed if -they specify any name that doesn't have a password. +Anyone who has more optimal options for any of the configurations listed +above can always mail them to the developer's mailing list at +@email{internals@@lists.mysql.com}. -@item -@strong{MySQL} user names can be up to 16 characters long; Unix user names -typically are limited to 8 characters. +RPM distributions prior to @strong{MySQL} Version 3.22 are user-contributed. +Beginning with Version 3.22, the RPMs are generated by us at +@strong{MySQL AB}. -@item -@strong{MySQL} passwords have nothing to do with Unix passwords. There is no -necessary connection between the password you use to log in to a Unix machine -and the password you use to access a database on that machine. +If you want to compile a debug version of @strong{MySQL}, you should add +@code{--with-debug} or @code{--with-debug=full} to the above configure lines +and remove any @code{-fomit-frame-pointer} options. -@item -@strong{MySQL} encrypts passwords using a different algorithm than the -one used during the Unix login process. See the descriptions of the -@code{PASSWORD()} and @code{ENCRYPT()} functions in @ref{Miscellaneous -functions}. Note that even if the password is stored 'scrambled', and -knowing your 'scrambled' password is enough to be able to connect to -the @strong{MySQL} server! -@end itemize +@cindex post-installation, setup and testing +@cindex testing, post-installation +@cindex setup, post-installation +@node Post-installation, Installing many servers, MySQL binaries, Installing +@section Post-installation Setup and Testing -@strong{MySQL} users and they privileges are normally created with the -@code{GRANT} command. @xref{GRANT}. +@menu +* mysql_install_db:: Problems running @code{mysql_install_db} +* Starting server:: Problems starting the @strong{MySQL} server +* Automatic start:: Starting and stopping @strong{MySQL} automatically +* Command-line options:: Command-line options +* Option files:: Option files +@end menu -When you login to a @strong{MySQL} server with a command line client you -should specify the password with @code{--password=your-password}. -@xref{Connecting}. +Once you've installed @strong{MySQL} (from either a binary or source +distribution), you need to initialize the grant tables, start the server, +and make sure that the server works okay. You may also wish to arrange +for the server to be started and stopped automatically when your system +starts up and shuts down. + +Normally you install the grant tables and start the server like this +for installation from a source distribution: +@cindex starting, the server +@cindex server, starting @example -mysql --user=monty --password=guess database_name +shell> ./scripts/mysql_install_db +shell> cd mysql_installation_directory +shell> ./bin/safe_mysqld --user=mysql & @end example -If you want the client to prompt for a password, you should use -@code{--password} without any argument +For a binary distribution (not RPM or pkg packages), do this: @example -mysql --user=monty --password database_name +shell> cd mysql_installation_directory +shell> ./bin/mysql_install_db +shell> ./bin/safe_mysqld --user=mysql & @end example -or the short form: +This creates the @code{mysql} database which will hold all database +privileges, the @code{test} database which you can use to test +@strong{MySQL} and also privilege entries for the user that run +@code{mysql_install_db} and a @code{root} user (without any passwords). +This also starts the @code{mysqld} server. -@example -mysql -u monty -p database_name -@end example +@code{mysql_install_db} will not overwrite any old privilege tables, so +it should be safe to run in any circumstances. If you don't want to +have the @code{test} database you can remove it with @code{mysqladmin -u +root drop test}. -Note that in the last example the password is @strong{NOT} 'database_name'. +Testing is most easily done from the top-level directory of the @strong{MySQL} +distribution. For a binary distribution, this is your installation directory +(typically something like @file{/usr/local/mysql}). For a source +distribution, this is the main directory of your @strong{MySQL} source tree. +@cindex testing, the server -If you want to use the @code{-p} option to supply a password you should do like this: +In the commands shown below in this section and in the following +subsections, @code{BINDIR} is the path to the location in which programs +like @code{mysqladmin} and @code{safe_mysqld} are installed. For a +binary distribution, this is the @file{bin} directory within the +distribution. For a source distribution, @code{BINDIR} is probably +@file{/usr/local/bin}, unless you specified an installation directory +other than @file{/usr/local} when you ran @code{configure}. +@code{EXECDIR} is the location in which the @code{mysqld} server is +installed. For a binary distribution, this is the same as +@code{BINDIR}. For a source distribution, @code{EXECDIR} is probably +@file{/usr/local/libexec}. + +Testing is described in detail below: +@cindex testing, installation + +@enumerate +@item +If necessary, start the @code{mysqld} server and set up the initial +@strong{MySQL} grant tables containing the privileges that determine how +users are allowed to connect to the server. This is normally done with the +@code{mysql_install_db} script: @example -mysql -u monty -pguess database_name +shell> scripts/mysql_install_db @end example -On some system the library call that @strong{MySQL} uses to prompt for a -password will automatically cut the password to 8 characters. Internally -@strong{MySQL} doesn't have any limit for the length of the password. +Typically, @code{mysql_install_db} needs to be run only the first time you +install @strong{MySQL}. Therefore, if you are upgrading an existing +installation, you can skip this step. (However, @code{mysql_install_db} is +quite safe to use and will not update any tables that already exist, so if +you are unsure of what to do, you can always run @code{mysql_install_db}.) -@node Connecting, Password security, User names, Privilege system -@section Connecting to the MySQL Server -@cindex connecting, to the server -@cindex default hostname -@cindex hostname, default -@cindex server, connecting +@code{mysql_install_db} creates six tables (@code{user}, @code{db}, +@code{host}, @code{tables_priv}, @code{columns_priv}, and @code{func}) in the +@code{mysql} database. A description of the initial privileges is given in +@ref{Default privileges}. Briefly, these privileges allow the @strong{MySQL} +@code{root} user to do anything, and allow anybody to create or use databases +with a name of @code{'test'} or starting with @code{'test_'}. -@strong{MySQL} client programs generally require that you specify connection -parameters when you want to access a @strong{MySQL} server: the host you want -to connect to, your user name, and your password. For example, the -@code{mysql} client can be started like this (optional arguments are enclosed -between @samp{[} and @samp{]}): +If you don't set up the grant tables, the following error will appear in the +log file when you start the server: +@tindex host.frm, problems finding @example -shell> mysql [-h host_name] [-u user_name] [-pyour_pass] +mysqld: Can't find file: 'host.frm' @end example -Alternate forms of the @code{-h}, @code{-u}, and @code{-p} options are -@code{--host=host_name}, @code{--user=user_name}, and -@code{--password=your_pass}. Note that there is @emph{no space} between -@code{-p} or @code{--password=} and the password following it. +The above may also happen with a binary @strong{MySQL} distribution if you +don't start @strong{MySQL} by executing exactly @code{./bin/safe_mysqld}! +@xref{safe_mysqld, , @code{safe_mysqld}}. -@strong{NOTE:} Specifying a password on the command line is not secure! -Any user on your system may then find out your password by typing a command -like: @code{ps auxww}. @xref{Option files}. +You might need to run @code{mysql_install_db} as @code{root}. However, +if you prefer, you can run the @strong{MySQL} server as an unprivileged +(non-@code{root}) user, provided that user can read and write files in +the database directory. Instructions for running @strong{MySQL} as an +unprivileged user are given in @ref{Changing MySQL user, , Changing +@strong{MySQL} user}. -@code{mysql} uses default values for connection parameters that are missing -from the command line: +If you have problems with @code{mysql_install_db}, see +@ref{mysql_install_db, , @code{mysql_install_db}}. + +There are some alternatives to running the @code{mysql_install_db} +script as it is provided in the @strong{MySQL} distribution: @itemize @bullet @item -The default hostname is @code{localhost}. +You may want to edit @code{mysql_install_db} before running it, to change +the initial privileges that are installed into the grant tables. This is +useful if you want to install @strong{MySQL} on a lot of machines with the +same privileges. In this case you probably should need only to add a few +extra @code{INSERT} statements to the @code{mysql.user} and @code{mysql.db} +tables! @item -The default user name is your Unix login name. +If you want to change things in the grant tables after installing them, you +can run @code{mysql_install_db}, then use @code{mysql -u root mysql} to +connect to the grant tables as the @strong{MySQL} @code{root} user and issue +SQL statements to modify the grant tables directly. @item -No password is supplied if @code{-p} is missing. +It is possible to re-create the grant tables completely after they have +already been created. You might want to do this if you've already installed +the tables but then want to re-create them after editing +@code{mysql_install_db}. @end itemize -Thus, for a Unix user @code{joe}, the following commands are equivalent: +For more information about these alternatives, see @ref{Default privileges}. + +@item +Start the @strong{MySQL} server like this: @example -shell> mysql -h localhost -u joe -shell> mysql -h localhost -shell> mysql -u joe -shell> mysql +shell> cd mysql_installation_directory +shell> bin/safe_mysqld & @end example -Other @strong{MySQL} clients behave similarly. - -On Unix systems, you can specify different default values to be used when you -make a connection, so that you need not enter them on the command line each -time you invoke a client program. This can be done in a couple of ways: +If you have problems starting the server, see @ref{Starting server}. -@itemize @bullet @item -@tindex .my.cnf file -You can specify connection parameters in the @code{[client]} section of the -@file{.my.cnf} configuration file in your home directory. The relevant -section of the file might look like this: +Use @code{mysqladmin} to verify that the server is running. The following +commands provide a simple test to check that the server is up and responding +to connections: @example -[client] -host=host_name -user=user_name -password=your_pass +shell> BINDIR/mysqladmin version +shell> BINDIR/mysqladmin variables @end example -@xref{Option files}. +The output from @code{mysqladmin version} varies slightly depending on your +platform and version of @strong{MySQL}, but should be similar to that shown +below: -@item -@tindex MYSQL_HOST environment variable -@tindex Environment variable, MYSQL_HOST -@tindex MYSQL_PWD environment variable -@tindex Environment variable, MYSQL_PWD -@tindex USER environment variable -@tindex Environment variable, USER -You can specify connection parameters using environment variables. The -host can be specified for @code{mysql} using @code{MYSQL_HOST}. The -@strong{MySQL} user name can be specified using @code{USER} (this is for -Windows only). The password can be specified using @code{MYSQL_PWD} -(but this is insecure; see the next section). @xref{Environment variables}. -@end itemize +@example +shell> BINDIR/mysqladmin version +mysqladmin Ver 8.14 Distrib 3.23.32, for linux on i586 +Copyright (C) 2000 MySQL AB & MySQL Finland AB & TCX DataKonsult AB +This software comes with ABSOLUTELY NO WARRANTY. This is free software, +and you are welcome to modify and redistribute it under the GPL license -@node Password security, Privileges provided, Connecting, Privilege system -@section Keeping Your Password Secure +Server version 3.23.32-debug +Protocol version 10 +Connection Localhost via Unix socket +TCP port 3306 +UNIX socket /tmp/mysql.sock +Uptime: 16 sec -It is inadvisable to specify your password in a way that exposes it to -discovery by other users. The methods you can use to specify your password -when you run client programs are listed below, along with an assessment of -the risks of each method: +Threads: 1 Questions: 9 Slow queries: 0 Opens: 7 Flush tables: 2 Open tables: 0 Queries per second avg: 0.000 Memory in use: 132K Max memory used: 16773K +@end example -@itemize @bullet -@item -Never give a normal user access to the @code{mysql.user} table. Knowing -the encrypted password for a user makes it possible to login as this -user. The passwords are only scrambled so that one shouldn't be able to -see the real password you used (if you happen to use a similar password -with your other applications). +To get a feeling for what else you can do with @code{BINDIR/mysqladmin}, +invoke it with the @code{--help} option. @item -Use a @code{-pyour_pass} or @code{--password=your_pass} option on the command -line. This is convenient but insecure, because your password becomes visible -to system status programs (such as @code{ps}) that may be invoked by other -users to display command lines. (@strong{MySQL} clients typically overwrite -the command-line argument with zeroes during their initialization sequence, -but there is still a brief interval during which the value is visible.) +Verify that you can shut down the server: +@cindex server, shutdown +@cindex shutting down, the server + +@example +shell> BINDIR/mysqladmin -u root shutdown +@end example @item -Use a @code{-p} or @code{--password} option (with no @code{your_pass} value -specified). In this case, the client program solicits the password from -the terminal: -@findex -p option -@findex -password option +Verify that you can restart the server. Do this using @code{safe_mysqld} or +by invoking @code{mysqld} directly. For example: + +@cindex server, restart +@cindex restarting, the server @example -shell> mysql -u user_name -p -Enter password: ******** +shell> BINDIR/safe_mysqld --log & @end example -The @samp{*} characters represent your password. - -It is more secure to enter your password this way than to specify it on the -command line because it is not visible to other users. However, this method -of entering a password is suitable only for programs that you run -interactively. If you want to invoke a client from a script that runs -non-interactively, there is no opportunity to enter the password from the -terminal. On some systems, you may even find that the first line of your -script is read and interpreted (incorrectly) as your password! +If @code{safe_mysqld} fails, try running it from the @strong{MySQL} +installation directory (if you are not already there). If that doesn't work, +see @ref{Starting server}. @item -@tindex .my.cnf file -Store your password in a configuration file. For example, you can list your -password in the @code{[client]} section of the @file{.my.cnf} file in your -home directory: +Run some simple tests to verify that the server is working. +The output should be similar to what is shown below: @example -[client] -password=your_pass +shell> BINDIR/mysqlshow ++-----------+ +| Databases | ++-----------+ +| mysql | ++-----------+ + +shell> BINDIR/mysqlshow mysql +Database: mysql ++--------------+ +| Tables | ++--------------+ +| columns_priv | +| db | +| func | +| host | +| tables_priv | +| user | ++--------------+ + +shell> BINDIR/mysql -e "select host,db,user from db" mysql ++------+--------+------+ +| host | db | user | ++------+--------+------+ +| % | test | | +| % | test_% | | ++------+--------+------+ @end example -If you store your password in @file{.my.cnf}, the file should not be group or -world readable or writable. Make sure the file's access mode is @code{400} -or @code{600}. +There is also a benchmark suite in the @file{sql-bench} directory (under the +@strong{MySQL} installation directory) that you can use to compare how +@strong{MySQL} performs on different platforms. The @file{sql-bench/Results} +directory contains the results from many runs against different databases and +platforms. To run all tests, execute these commands: -@xref{Option files}. -@item -You can store your password in the @code{MYSQL_PWD} environment variable, but -this method must be considered extremely insecure and should not be used. -Some versions of @code{ps} include an option to display the environment of -running processes; your password will be in plain sight for all to see if -you set @code{MYSQL_PWD}. Even on systems without such a version of -@code{ps}, it is unwise to assume there is no other method to observe process -environments. @xref{Environment variables}. -@end itemize +@example +shell> cd sql-bench +shell> run-all-tests +@end example -All in all, the safest methods are to have the client program prompt for the -password or to specify the password in a properly protected @file{.my.cnf} -file. +If you don't have the @file{sql-bench} directory, you are probably using an +RPM for a binary distribution. (Source distribution RPMs include the +benchmark directory.) In this case, you must first install the benchmark +suite before you can use it. Beginning with @strong{MySQL} Version 3.22, +there are benchmark RPM files named @file{mysql-bench-VERSION-i386.rpm} that +contain benchmark code and data. -@node Privileges provided, Privileges, Password security, Privilege system -@section Privileges Provided by MySQL -@cindex privilege information, location +If you have a source distribution, you can also run the tests in the +@file{tests} subdirectory. For example, to run @file{auto_increment.tst}, do +this: -Information about user privileges is stored in the @code{user}, @code{db}, -@code{host}, @code{tables_priv}, and @code{columns_priv} tables in the -@code{mysql} database (that is, in the database named @code{mysql}). The -@strong{MySQL} server reads the contents of these tables when it starts up -and under the circumstances indicated in @ref{Privilege changes}. +@example +shell> BINDIR/mysql -vvf test < ./tests/auto_increment.tst +@end example -The names used in this manual to refer to the privileges provided by -@strong{MySQL} are shown below, along with the table column name associated -with each privilege in the grant tables and the context in which the -privilege applies: +The expected results are shown in the @file{./tests/auto_increment.res} file. +@end enumerate -@multitable @columnfractions .15 .25 .6 -@item @strong{Privilege} @tab @strong{Column} @tab @strong{Context} -@item @strong{select} @tab @code{Select_priv} @tab tables -@item @strong{insert} @tab @code{Insert_priv} @tab tables -@item @strong{update} @tab @code{Update_priv} @tab tables -@item @strong{delete} @tab @code{Delete_priv} @tab tables -@item @strong{index} @tab @code{Index_priv} @tab tables -@item @strong{alter} @tab @code{Alter_priv} @tab tables -@item @strong{create} @tab @code{Create_priv} @tab databases, tables, or indexes -@item @strong{drop} @tab @code{Drop_priv} @tab databases or tables -@item @strong{grant} @tab @code{Grant_priv} @tab databases or tables -@item @strong{references} @tab @code{References_priv} @tab databases or tables -@item @strong{reload} @tab @code{Reload_priv} @tab server administration -@item @strong{shutdown} @tab @code{Shutdown_priv} @tab server administration -@item @strong{process} @tab @code{Process_priv} @tab server administration -@item @strong{file} @tab @code{File_priv} @tab file access on server -@end multitable +@node mysql_install_db, Starting server, Post-installation, Post-installation +@subsection Problems Running @code{mysql_install_db} +@cindex @code{mysql_install_db} script +@cindex scripts, @code{mysql_install_db} -The @strong{select}, @strong{insert}, @strong{update}, and @strong{delete} -privileges allow you to perform operations on rows in existing tables in -a database. +The purpose of the @code{mysql_install_db} script is to generate new +@strong{MySQL} privilege tables. It will not affect any other data! +It will also not do anything if you already have @strong{MySQL} privilege +tables installed! -@code{SELECT} statements require the @strong{select} privilege only if they -actually retrieve rows from a table. You can execute certain @code{SELECT} -statements even without permission to access any of the databases on the -server. For example, you could use the @code{mysql} client as a simple -calculator: +If you want to re-create your privilege tables, you should take down +the @code{mysqld} server, if it's running, and then do something like: @example -mysql> SELECT 1+1; -mysql> SELECT PI()*2; +mv mysql-data-directory/mysql mysql-data-directory/mysql-old +mysql_install_db @end example -The @strong{index} privilege allows you to create or drop (remove) indexes. - -The @strong{alter} privilege allows you to use @code{ALTER TABLE}. +This section lists problems you might encounter when you run +@code{mysql_install_db}: -The @strong{create} and @strong{drop} privileges allow you to create new -databases and tables, or to drop (remove) existing databases and tables. +@table @strong +@item @code{mysql_install_db} doesn't install the grant tables -Note that if you grant the @strong{drop} privilege for the @code{mysql} -database to a user, that user can drop the database in which the -@strong{MySQL} access privileges are stored! +You may find that @code{mysql_install_db} fails to install the grant +tables and terminates after displaying the following messages: -The @strong{grant} privilege allows you to give to other users those -privileges you yourself possess. +@example +starting mysqld daemon with databases from XXXXXX +mysql daemon ended +@end example -The @strong{file} privilege gives you permission to read and write files on -the server using the @code{LOAD DATA INFILE} and @code{SELECT ... INTO -OUTFILE} statements. Any user to whom this privilege is granted can read or -write any file that the @strong{MySQL} server can read or write. +In this case, you should examine the log file very carefully! The log +should be located in the directory @file{XXXXXX} named by the error message, +and should indicate why @code{mysqld} didn't start. If you don't understand +what happened, include the log when you post a bug report using +@code{mysqlbug}! +@xref{Bug reports}. -The remaining privileges are used for administrative operations, which are -performed using the @code{mysqladmin} program. The table below shows which -@code{mysqladmin} commands each administrative privilege allows you to -execute: +@item There is already a @code{mysqld} daemon running -@multitable @columnfractions .15 .85 -@item @strong{Privilege} @tab @strong{Commands permitted to privilege holders} -@item @strong{reload} @tab @code{reload}, @code{refresh}, -@code{flush-privileges}, @code{flush-hosts}, @code{flush-logs}, and -@code{flush-tables} -@item @strong{shutdown} @tab @code{shutdown} -@item @strong{process} @tab @code{processlist}, @code{kill} -@end multitable +In this case, you probably don't have to run @code{mysql_install_db} at +all. You have to run @code{mysql_install_db} only once, when you install +@strong{MySQL} the first time. -The @code{reload} command tells the server to re-read the grant tables. The -@code{refresh} command flushes all tables and opens and closes the log -files. @code{flush-privileges} is a synonym for @code{reload}. The other -@code{flush-*} commands perform functions similar to @code{refresh} but are -more limited in scope, and may be preferable in some instances. For example, -if you want to flush just the log files, @code{flush-logs} is a better choice -than @code{refresh}. +@item Installing a second @code{mysqld} daemon doesn't work when one daemon is running -The @code{shutdown} command shuts down the server. +This can happen when you already have an existing @strong{MySQL} +installation, but want to put a new installation in a different place (for +example, for testing, or perhaps you simply want to run two installations at +the same time). Generally the problem that occurs when you try to run the +second server is that it tries to use the same socket and port as the old one. +In this case you will get the error message: @code{Can't start server: Bind on +TCP/IP port: Address already in use} or @code{Can't start server : Bind on +unix socket...}. @xref{Installing many servers}. -The @code{processlist} command displays information about the threads -executing within the server. The @code{kill} command kills server threads. -You can always display or kill your own threads, but you need the -@strong{process} privilege to display or kill threads initiated by other -users. @xref{KILL}. +@item You don't have write access to @file{/tmp} +@cindex write access, tmp +@cindex temporary file, write access +@cindex files, @code{tmp} -It is a good idea in general to grant privileges only to those users who need -them, but you should exercise particular caution in granting certain -privileges: +If you don't have write access to create a socket file at the default place +(in @file{/tmp}) or permission to create temporary files in @file{/tmp,} +you will get an error when running @code{mysql_install_db} or when +starting or using @code{mysqld}. -@itemize @bullet -@item -The @strong{grant} privilege allows users to give away their privileges to -other users. Two users with different privileges and with the @strong{grant} -privilege are able to combine privileges. +You can specify a different socket and temporary directory as follows: -@item -The @strong{alter} privilege may be used to subvert the privilege system -by renaming tables. +@tindex TMPDIR environment variable +@tindex MYSQL_UNIX_PORT environment variable +@tindex Environment variable, TMPDIR +@tindex Environment variable, MYSQL_UNIX_PORT +@example +shell> TMPDIR=/some_tmp_dir/ +shell> MYSQL_UNIX_PORT=/some_tmp_dir/mysqld.sock +shell> export TMPDIR MYSQL_UNIX_PORT +@end example -@item -The @strong{file} privilege can be abused to read any world-readable file on -the server into a database table, the contents of which can then be -accessed using @code{SELECT}. This includes the contents of all databases -hosted by the server! +@file{some_tmp_dir} should be the path to some directory for which you +have write permission. @xref{Environment variables}. -@item -The @strong{shutdown} privilege can be abused to deny service to other -users entirely, by terminating the server. +After this you should be able to run @code{mysql_install_db} and start +the server with these commands: -@item -The @strong{process} privilege can be used to view the plain text of -currently executing queries, including queries that set or change passwords. +@example +shell> scripts/mysql_install_db +shell> BINDIR/safe_mysqld & +@end example -@item -Privileges on the @code{mysql} database can be used to change passwords -and other access privilege information. (Passwords are stored -encrypted, so a malicious user cannot simply read them to know the plain -text password). If they can access the @code{mysql.user} password -column, they can use it to log into the @strong{MySQL} server -for the given user. (With sufficient privileges, the same user can -replace a password with a different one.) -@end itemize +@item @code{mysqld} crashes immediately -There are some things that you cannot do with the @strong{MySQL} -privilege system: +If you are running RedHat Version 5.0 with a version of @code{glibc} older than +2.0.7-5, you should make sure you have installed all @code{glibc} patches! +There is a lot of information about this in the @strong{MySQL} mail +archives. Links to the mail archives are available online at +@uref{http://www.mysql.com/documentation/}. +Also, see @ref{Linux}. -@itemize @bullet -@item -You cannot explicitly specify that a given user should be denied access. -That is, you cannot explicitly match a user and then refuse the connection. +You can also start @code{mysqld} manually using the @code{--skip-grant-tables} +option and add the privilege information yourself using @code{mysql}: -@item -You cannot specify that a user has privileges to create or drop tables -in a database but not to create or drop the database itself. -@end itemize +@example +shell> BINDIR/safe_mysqld --skip-grant-tables & +shell> BINDIR/mysql -u root mysql +@end example -@node Privileges, Connection access, Privileges provided, Privilege system -@section How the Privilege System Works -@cindex privilege system, described +From @code{mysql}, manually execute the SQL commands in +@code{mysql_install_db}. Make sure you run @code{mysqladmin +flush-privileges} or @code{mysqladmin reload} afterward to tell the server to +reload the grant tables. +@end table -The @strong{MySQL} privilege system ensures that all users may do exactly the -things that they are supposed to be allowed to do. When you connect to a -@strong{MySQL} server, your identity is determined by @strong{the host from -which you connect} and @strong{the user name you specify}. The system grants -privileges according to your identity and @strong{what you want to do}. +@node Starting server, Automatic start, mysql_install_db, Post-installation +@subsection Problems Starting the MySQL Server +@cindex server, starting problems +@cindex problems, starting the server -@strong{MySQL} considers both your hostname and user name in identifying you -because there is little reason to assume that a given user name belongs to -the same person everywhere on the Internet. For example, the user -@code{bill} who connects from @code{whitehouse.gov} need not be the same -person as the user @code{bill} who connects from @code{microsoft.com}. -@strong{MySQL} handles this by allowing you to distinguish users on different -hosts that happen to have the same name: you can grant @code{bill} one set -of privileges for connections from @code{whitehouse.gov}, and a different set -of privileges for connections from @code{microsoft.com}. +If you are going to use tables that support transactions (BDB, InnoDB), +you should first create a my.cnf file and set startup options +for the table types you plan to use. @xref{Table types}. -@strong{MySQL} access control involves two stages: +Generally, you start the @code{mysqld} server in one of three ways: @itemize @bullet @item -Stage 1: The server checks whether or not you are even allowed to connect. +By invoking @code{mysql.server}. This script is used primarily at +system startup and shutdown, and is described more fully in +@ref{Automatic start}. @item -Stage 2: Assuming you can connect, the server checks each request you issue -to see whether or not you have sufficient privileges to perform it. For -example, if you try to select rows from a table in a database or drop a table -from the database, the server makes sure you have the @strong{select} -privilege for the table or the @strong{drop} privilege for the database. -@end itemize - -The server uses the @code{user}, @code{db}, and @code{host} tables in the -@code{mysql} database at both stages of access control. The fields in these -grant tables are shown below: +By invoking @code{safe_mysqld}, which tries to determine the proper options +for @code{mysqld} and then runs it with those options. @xref{safe_mysqld, , +@code{safe_mysqld}}. -@multitable @columnfractions .2 .25 .25 .25 -@item @strong{Table name} @tab @code{user} @tab @code{db} @tab @code{host} +@item +On NT you should install @code{mysqld} as a service as follows: +@example +bin\mysqld-nt --install # Install MySQL as a service +@end example -@item @strong{Scope fields} @tab @code{Host} @tab @code{Host} @tab @code{Host} -@item @tab @code{User} @tab @code{Db} @tab @code{Db} -@item @tab @code{Password} @tab @code{User} @tab +You can now start/stop @code{mysqld} as follows: +@example +NET START mysql +NET STOP mysql +@end example -@item @strong{Privilege fields} @tab @code{Select_priv} @tab @code{Select_priv} @tab @code{Select_priv} -@item @tab @code{Insert_priv} @tab @code{Insert_priv} @tab @code{Insert_priv} -@item @tab @code{Update_priv} @tab @code{Update_priv} @tab @code{Update_priv} -@item @tab @code{Delete_priv} @tab @code{Delete_priv} @tab @code{Delete_priv} -@item @tab @code{Index_priv} @tab @code{Index_priv} @tab @code{Index_priv} -@item @tab @code{Alter_priv} @tab @code{Alter_priv} @tab @code{Alter_priv} -@item @tab @code{Create_priv} @tab @code{Create_priv} @tab @code{Create_priv} -@item @tab @code{Drop_priv} @tab @code{Drop_priv} @tab @code{Drop_priv} -@item @tab @code{Grant_priv} @tab @code{Grant_priv} @tab @code{Grant_priv} -@item @tab @code{References_priv} @tab @tab -@item @tab @code{Reload_priv} @tab @tab -@item @tab @code{Shutdown_priv} @tab @tab -@item @tab @code{Process_priv} @tab @tab -@item @tab @code{File_priv} @tab @tab -@end multitable +Note that in this case you can't use any other options for @code{mysqld}! -For the second stage of access control (request verification), the server -may, if the request involves tables, additionally consult the -@code{tables_priv} and @code{columns_priv} tables. The fields in these -tables are shown below: +You can remove the service as follows: +@example +bin\mysqld-nt --remove # remove MySQL as a service +@end example -@multitable @columnfractions .2 .25 .25 -@item @strong{Table name} @tab @code{tables_priv} @tab @code{columns_priv} +@item +By invoking @code{mysqld} directly. +@end itemize -@item @strong{Scope fields} @tab @code{Host} @tab @code{Host} -@item @tab @code{Db} @tab @code{Db} -@item @tab @code{User} @tab @code{User} -@item @tab @code{Table_name} @tab @code{Table_name} -@item @tab @tab @code{Column_name} +When the @code{mysqld} daemon starts up, it changes directory to the +data directory. This is where it expects to write log files and the pid +(process ID) file, and where it expects to find databases. -@item @strong{Privilege fields} @tab @code{Table_priv} @tab @code{Column_priv} -@item @tab @code{Column_priv} @tab +The data directory location is hardwired in when the distribution is +compiled. However, if @code{mysqld} expects to find the data directory +somewhere other than where it really is on your system, it will not work +properly. If you have problems with incorrect paths, you can find out +what options @code{mysqld} allows and what the default path settings are by +invoking @code{mysqld} with the @code{--help} option. You can override the +defaults by specifying the correct pathnames as command-line arguments to +@code{mysqld}. (These options can be used with @code{safe_mysqld} as well.) -@item @strong{Other fields} @tab @code{Timestamp} @tab @code{Timestamp} -@item @tab @code{Grantor} @tab -@end multitable +Normally you should need to tell @code{mysqld} only the base directory under +which @strong{MySQL} is installed. You can do this with the @code{--basedir} +option. You can also use @code{--help} to check the effect of changing path +options (note that @code{--help} @emph{must} be the final option of the +@code{mysqld} command). For example: -Each grant table contains scope fields and privilege fields. +@example +shell> EXECDIR/mysqld --basedir=/usr/local --help +@end example -Scope fields determine the scope of each entry in the tables, that is, the -context in which the entry applies. For example, a @code{user} table entry -with @code{Host} and @code{User} values of @code{'thomas.loc.gov'} and -@code{'bob'} would be used for authenticating connections made to the server -by @code{bob} from the host @code{thomas.loc.gov}. Similarly, a @code{db} -table entry with @code{Host}, @code{User}, and @code{Db} fields of -@code{'thomas.loc.gov'}, @code{'bob'} and @code{'reports'} would be used when -@code{bob} connects from the host @code{thomas.loc.gov} to access the -@code{reports} database. The @code{tables_priv} and @code{columns_priv} -tables contain scope fields indicating tables or table/column combinations -to which each entry applies. +Once you determine the path settings you want, start the server without +the @code{--help} option. -@cindex case sensitivity, in access checking -For access-checking purposes, comparisons of @code{Host} values are -case insensitive. @code{User}, @code{Password}, @code{Db}, and -@code{Table_name} values are case sensitive. -@code{Column_name} values are case insensitive in @strong{MySQL} Version -3.22.12 or later. +Whichever method you use to start the server, if it fails to start up +correctly, check the log file to see if you can find out why. Log files +are located in the data directory (typically +@file{/usr/local/mysql/data} for a binary distribution, +@file{/usr/local/var} for a source distribution, +@file{\mysql\data\mysql.err} on Windows.) Look in the data directory for +files with names of the form @file{host_name.err} and +@file{host_name.log} where @code{host_name} is the name of your server +host. Then check the last few lines of these files: -Privilege fields indicate the privileges granted by a table entry, that is, -what operations can be performed. The server combines the information in the -various grant tables to form a complete description of a user's privileges. -The rules used to do this are described in @ref{Request access}. +@example +shell> tail host_name.err +shell> tail host_name.log +@end example -Scope fields are strings, declared as shown below; the default value for -each is the empty string: +If you find something like the following in the log file: +@example +000729 14:50:10 bdb: Recovery function for LSN 1 27595 failed +000729 14:50:10 bdb: warning: ./test/t1.db: No such file or directory +000729 14:50:10 Can't init databases +@end example -@multitable @columnfractions .15 .15 .7 -@item @strong{Field name} @tab @strong{Type} -@item @code{Host} @tab @code{CHAR(60)} -@item @code{User} @tab @code{CHAR(16)} -@item @code{Password} @tab @code{CHAR(16)} -@item @code{Db} @tab @code{CHAR(64)} @tab (@code{CHAR(60)} for the -@code{tables_priv} and @code{columns_priv} tables) -@item @code{Table_name} @tab @code{CHAR(60)} -@item @code{Column_name} @tab @code{CHAR(60)} -@end multitable - -In the @code{user}, @code{db} and @code{host} tables, -all privilege fields are declared as @code{ENUM('N','Y')} --- each can have a -value of @code{'N'} or @code{'Y'}, and the default value is @code{'N'}. - -In the @code{tables_priv} and @code{columns_priv} tables, the privilege -fields are declared as @code{SET} fields: - -@multitable @columnfractions .2 .2 .6 -@item @strong{Table name} @tab @strong{Field name} @tab @strong{Possible set elements} -@item @code{tables_priv} @tab @code{Table_priv} @tab @code{'Select', 'Insert', -'Update', 'Delete', 'Create', 'Drop', 'Grant', 'References', 'Index', 'Alter'} -@item @code{tables_priv} @tab @code{Column_priv} @tab @code{'Select', 'Insert', -'Update', 'References'} -@item @code{columns_priv} @tab @code{Column_priv} @tab @code{'Select', 'Insert', -'Update', 'References'} -@end multitable - -Briefly, the server uses the grant tables like this: +This means that you didn't start @code{mysqld} with @code{--bdb-no-recover} +and Berkeley DB found something wrong with its log files when it +tried to recover your databases. To be able to continue, you should +move away the old Berkeley DB log file from the database directory to +some other place, where you can later examine these. The log files are +named @file{log.0000000001}, where the number will increase over time. -@itemize @bullet -@item -The @code{user} table scope fields determine whether to allow or reject -incoming connections. For allowed connections, any privileges granted in -the @code{user} table indicate the user's global (superuser) privileges. -These privileges apply to @strong{all} databases on the server. +If you are running @code{mysqld} with BDB table support and @code{mysqld} core +dumps at start this could be because of some problems with the BDB +recover log. In this case you can try starting @code{mysqld} with +@code{--bdb-no-recover}. If this helps, then you should remove all +@file{log.*} files from the data directory and try starting @code{mysqld} +again. -@item -The @code{db} and @code{host} tables are used together: +If you get the following error, it means that some other program (or another +@code{mysqld} server) is already using the TCP/IP port or socket +@code{mysqld} is trying to use: -@itemize @minus -@item -The @code{db} table scope fields determine which users can access which -databases from which hosts. The privilege fields determine which operations -are allowed. +@example +Can't start server: Bind on TCP/IP port: Address already in use + or +Can't start server : Bind on unix socket... +@end example -@item -The @code{host} table is used as an extension of the @code{db} table when you -want a given @code{db} table entry to apply to several hosts. For example, -if you want a user to be able to use a database from several hosts in -your network, leave the @code{Host} value empty in the user's @code{db} table -entry, then populate the @code{host} table with an entry for each of those -hosts. This mechanism is described more detail in @ref{Request access}. -@end itemize +Use @code{ps} to make sure that you don't have another @code{mysqld} server +running. If you can't find another server running, you can try to execute +the command @code{telnet your-host-name tcp-ip-port-number} and press +@code{RETURN} a couple of times. If you don't get an error message like +@code{telnet: Unable to connect to remote host: Connection refused}, +something is using the TCP/IP port @code{mysqld} is trying to use. +See @ref{mysql_install_db} and @ref{Multiple servers}. -@item -The @code{tables_priv} and @code{columns_priv} tables are similar to -the @code{db} table, but are more fine-grained: they apply at the -table and column levels rather than at the database level. -@end itemize +If @code{mysqld} is currently running, you can find out what path settings +it is using by executing this command: -Note that administrative privileges (@strong{reload}, @strong{shutdown}, -etc.) are specified only in the @code{user} table. This is because -administrative operations are operations on the server itself and are not -database-specific, so there is no reason to list such privileges in the -other grant tables. In fact, only the @code{user} table need -be consulted to determine whether or not you can perform an administrative -operation. +@example +shell> mysqladmin variables +@end example -The @strong{file} privilege is specified only in the @code{user} table, too. -It is not an administrative privilege as such, but your ability to read or -write files on the server host is independent of the database you are -accessing. +or -The @code{mysqld} server reads the contents of the grant tables once, when it -starts up. Changes to the grant tables take effect as indicated in -@ref{Privilege changes}. +@example +shell> mysqladmin -h 'your-host-name' variables +@end example -When you modify the contents of the grant tables, it is a good idea to make -sure that your changes set up privileges the way you want. For help in -diagnosing problems, see @ref{Access denied}. For advice on security issues, -@pxref{Security}. +If @code{safe_mysqld} starts the server but you can't connect to it, +you should make sure you have an entry in @file{/etc/hosts} that looks like +this: -A useful -diagnostic tool is the @code{mysqlaccess} script, which Yves Carlier has -provided for the @strong{MySQL} distribution. Invoke @code{mysqlaccess} with -the @code{--help} option to find out how it works. -Note that @code{mysqlaccess} checks access using only the @code{user}, -@code{db} and @code{host} tables. It does not check table- or column-level -privileges. +@example +127.0.0.1 localhost +@end example -@node Connection access, Request access, Privileges, Privilege system -@section Access Control, Stage 1: Connection Verification -@cindex access control -@cindex control access -@cindex connecting, verification -@cindex testing, connection to the server +This problem occurs only on systems that don't have a working thread +library and for which @strong{MySQL} must be configured to use MIT-pthreads. -When you attempt to connect to a @strong{MySQL} server, the server accepts or -rejects the connection based on your identity and whether or not you can -verify your identity by supplying the correct password. If not, the server -denies access to you completely. Otherwise, the server accepts the -connection, then enters Stage 2 and waits for requests. +If you can't get @code{mysqld} to start you can try to make a trace file +to find the problem. @xref{Making trace files}. -Your identity is based on two pieces of information: +If you are using InnoDB tables, refer to the InnoDB-specific startup +options. @xref{InnoDB start}. -@itemize @bullet -@item -The host from which you connect +If you are using BDB (Berkeley DB) tables, you should familiarize +yourself with the different BDB specific startup options. @xref{BDB start}. -@item -Your @strong{MySQL} user name -@end itemize +@node Automatic start, Command-line options, Starting server, Post-installation +@subsection Starting and Stopping MySQL Automatically +@cindex starting, the server automatically +@cindex stopping, the server +@cindex server, starting and stopping -Identity checking is performed using the three @code{user} table scope fields -(@code{Host}, @code{User}, and @code{Password}). The server accepts the -connection only if a @code{user} table entry matches your hostname and user -name, and you supply the correct password. +The @code{mysql.server} and @code{safe_mysqld} scripts can be used to start +the server automatically at system startup time. @code{mysql.server} can also +be used to stop the server. -Values in the @code{user} table scope fields may be specified as follows: +The @code{mysql.server} script can be used to start or stop the server +by invoking it with @code{start} or @code{stop} arguments: -@itemize @bullet -@item -A @code{Host} value may be a hostname or an IP number, or @code{'localhost'} -to indicate the local host. +@example +shell> mysql.server start +shell> mysql.server stop +@end example -@item -@cindex wildcards, in @code{mysql.user} table -You can use the wild-card characters @samp{%} and @samp{_} in the @code{Host} -field. +@code{mysql.server} can be found in the @file{share/mysql} directory +under the @strong{MySQL} installation directory or in the @file{support-files} +directory of the @strong{MySQL} source tree. -@item -A @code{Host} value of @code{'%'} matches any hostname. +Before @code{mysql.server} starts the server, it changes directory to +the @strong{MySQL} installation directory, then invokes @code{safe_mysqld}. +You might need to edit @code{mysql.server} if you have a binary distribution +that you've installed in a non-standard location. Modify it to @code{cd} +into the proper directory before it runs @code{safe_mysqld}. If you want the +server to run as some specific user, add an appropriate @code{user} line +to the @file{/etc/my.cnf} file, as shown later in this section. -@item -A blank @code{Host} value means that the privilege should be anded -with the entry in the @code{host} table that matches the given host name. -You can find more information about this in the next chapter. +@code{mysql.server stop} brings down the server by sending a signal to it. +You can take down the server manually by executing @code{mysqladmin shutdown}. -@cindex netmask notation, in @code{mysql.user} table -@item -As of @strong{MySQL} Version 3.23, for @code{Host} values specified as -IP numbers, you can specify a netmask indicating how many address bits to -use for the network number. For example: +You might want to add these start and stop commands to the appropriate places +in your @file{/etc/rc*} files when you start using @strong{MySQL} for +production applications. Note that if you modify @code{mysql.server}, then +upgrade @strong{MySQL} sometime, your modified version will be overwritten, +so you should make a copy of your edited version that you can reinstall. +If your system uses @file{/etc/rc.local} to start external scripts, you +should append the following to it: @example -GRANT ALL PRIVILEGES on db.* to david@@'192.58.197.0/255.255.255.0'; +/bin/sh -c 'cd /usr/local/mysql ; ./bin/safe_mysqld --user=mysql &' @end example -This will allow everyone to connect from an IP where the following is true: +You can also add options for @code{mysql.server} in a global +@file{/etc/my.cnf} file. A typical @file{/etc/my.cnf} file might look like +this: @example -user_ip & netmask = host_ip. +[mysqld] +datadir=/usr/local/mysql/var +socket=/tmp/mysqld.sock +port=3306 +user=mysql + +[mysql.server] +basedir=/usr/local/mysql @end example -In the above example all IP:s in the interval 192.58.197.0 - -192.58.197.255 can connect to the @strong{MySQL} server. +The @code{mysql.server} script understands the following options: +@code{datadir}, @code{basedir}, and @code{pid-file}. -@item -@cindex anonymous user -Wild-card characters are not allowed in the @code{User} field, but you can -specify a blank value, which matches any name. If the @code{user} table -entry that matches an incoming connection has a blank user name, the user is -considered to be the anonymous user (the user with no name), rather than the -name that the client actually specified. This means that a blank user name -is used for all further access checking for the duration of the connection -(that is, during Stage 2). +The following table shows which option groups each of the startup scripts +read from option files: -@item -The @code{Password} field can be blank. This does not mean that any password -matches, it means the user must connect without specifying a password. -@end itemize +@multitable @columnfractions .20 .80 +@item @strong{Script} @tab @strong{Option groups} +@item @code{mysqld} @tab @code{mysqld} and @code{server} +@item @code{mysql.server} @tab @code{mysql.server}, @code{mysqld}, and @code{server} +@item @code{safe_mysqld} @tab @code{mysql.server}, @code{mysqld}, and @code{server} +@end multitable -@findex PASSWORD() -Non-blank @code{Password} values represent encrypted passwords. -@strong{MySQL} does not store passwords in plaintext form for anyone to -see. Rather, the password supplied by a user who is attempting to -connect is encrypted (using the @code{PASSWORD()} function). The -encrypted password is then used when the client/server is checking if -the password is correct (This is done without the encrypted password -ever traveling over the connection.) Note that from @strong{MySQL}'s -point of view the encrypted password is the REAL password, so you should -not give anyone access to it! In particular, don't give normal users -read access to the tables in the @code{mysql} database! +@xref{Option files}. -The examples below show how various combinations of @code{Host} and -@code{User} values in @code{user} table entries apply to incoming -connections: +@findex command-line options +@cindex options, command-line +@cindex mysqld options +@node Command-line options, Option files, Automatic start, Post-installation +@subsection mysqld Command-line Options -@multitable @columnfractions .25 .15 .60 -@item @code{Host} @strong{value} @tab @code{User} @strong{value} @tab @strong{Connections matched by entry} -@item @code{'thomas.loc.gov'} @tab @code{'fred'} @tab @code{fred}, connecting from @code{thomas.loc.gov} -@item @code{'thomas.loc.gov'} @tab @code{''} @tab Any user, connecting from @code{thomas.loc.gov} -@item @code{'%'} @tab @code{'fred'} @tab @code{fred}, connecting from any host -@item @code{'%'} @tab @code{''} @tab Any user, connecting from any host -@item @code{'%.loc.gov'} @tab @code{'fred'} @tab @code{fred}, connecting from any host in the @code{loc.gov} domain -@item @code{'x.y.%'} @tab @code{'fred'} @tab @code{fred}, connecting from @code{x.y.net}, @code{x.y.com},@code{x.y.edu}, etc. (this is probably not useful) -@item @code{'144.155.166.177'} @tab @code{'fred'} @tab @code{fred}, connecting from the host with IP address @code{144.155.166.177} -@item @code{'144.155.166.%'} @tab @code{'fred'} @tab @code{fred}, connecting from any host in the @code{144.155.166} class C subnet -@item @code{'144.155.166.0/255.255.255.0'} @tab @code{'fred'} @tab Same as previous example -@end multitable +@code{mysqld} accepts the following command-line options: -Because you can use IP wild-card values in the @code{Host} field (for example, -@code{'144.155.166.%'} to match every host on a subnet), there is the -possibility that someone might try to exploit this capability by naming a -host @code{144.155.166.somewhere.com}. To foil such attempts, @strong{MySQL} -disallows matching on hostnames that start with digits and a dot. Thus, if -you have a host named something like @code{1.2.foo.com}, its name will never -match the @code{Host} column of the grant tables. Only an IP number can -match an IP wild-card value. +@table @code +@item --ansi +Use ANSI SQL syntax instead of MySQL syntax. @xref{ANSI mode}. -An incoming connection may be matched by more than one entry in the -@code{user} table. For example, a connection from @code{thomas.loc.gov} by -@code{fred} would be matched by several of the entries just shown above. How -does the server choose which entry to use if more than one matches? The -server resolves this question by sorting the @code{user} table after reading -it at startup time, then looking through the entries in sorted order when a -user attempts to connect. The first matching entry is the one that is used. +@item -b, --basedir=path +Path to installation directory. All paths are usually resolved relative to +this. -@code{user} table sorting works as follows. Suppose the @code{user} table -looks like this: +@item --big-tables +Allow big result sets by saving all temporary sets on file. It solves +most 'table full' errors, but also slows down the queries where +in-memory tables would suffice. Since Version 3.23.2, @strong{MySQL} is +able to solve it automatically by using memory for small temporary +tables and switching to disk tables where necessary. -@example -+-----------+----------+- -| Host | User | ... -+-----------+----------+- -| % | root | ... -| % | jeffrey | ... -| localhost | root | ... -| localhost | | ... -+-----------+----------+- -@end example +@item --bind-address=IP +IP address to bind to. -When the server reads in the table, it orders the entries with the -most-specific @code{Host} values first (@code{'%'} in the @code{Host} column -means ``any host'' and is least specific). Entries with the same @code{Host} -value are ordered with the most-specific @code{User} values first (a blank -@code{User} value means ``any user'' and is least specific). The resulting -sorted @code{user} table looks like this: +@item --character-sets-dir=path +Directory where character sets are. @xref{Character sets}. -@example -+-----------+----------+- -| Host | User | ... -+-----------+----------+- -| localhost | root | ... -| localhost | | ... -| % | jeffrey | ... -| % | root | ... -+-----------+----------+- -@end example +@item --chroot=path +Chroot @code{mysqld} daemon during startup. Recommended security measure. It will +somewhat limit @code{LOAD DATA INFILE} and @code{SELECT ... INTO OUTFILE} +though. -@cindex grant tables, sorting -@cindex sorting, grant tables -@cindex @code{user} table, sorting -When a connection is attempted, the server looks through the sorted entries -and uses the first match found. For a connection from @code{localhost} by -@code{jeffrey}, the entries with @code{'localhost'} in the @code{Host} column -match first. Of those, the entry with the blank user name matches both the -connecting hostname and user name. (The @code{'%'/'jeffrey'} entry would -have matched, too, but it is not the first match in the table.) +@item --core-file +Write a core file if @code{mysqld} dies. For some systems you must also +specify @code{--core-file-size} to @code{safe_mysqld}. @xref{safe_mysqld, , +@code{safe_mysqld}}. -Here is another example. Suppose the @code{user} table looks like this: +@item -h, --datadir=path +Path to the database root. -@example -+----------------+----------+- -| Host | User | ... -+----------------+----------+- -| % | jeffrey | ... -| thomas.loc.gov | | ... -+----------------+----------+- -@end example +@item --default-character-set=charset +Set the default character set. @xref{Character sets}. -The sorted table looks like this: +@item --default-table-type=type +Set the default table type for tables. @xref{Table types}. -@example -+----------------+----------+- -| Host | User | ... -+----------------+----------+- -| thomas.loc.gov | | ... -| % | jeffrey | ... -+----------------+----------+- -@end example +@item --debug[...]= +If @strong{MySQL} is configured with @code{--with-debug}, you can use this +option to get a trace file of what @code{mysqld} is doing. +@xref{Making trace files}. -A connection from @code{thomas.loc.gov} by @code{jeffrey} is matched by the -first entry, whereas a connection from @code{whitehouse.gov} by -@code{jeffrey} is matched by the second. +@item --delay-key-write-for-all-tables +Don't flush key buffers between writes for any @code{MyISAM} table. +@xref{Server parameters}. -A common misconception is to think that for a given user name, all entries -that explicitly name that user will be used first when the server attempts to -find a match for the connection. This is simply not true. The previous -example illustrates this, where a connection from @code{thomas.loc.gov} by -@code{jeffrey} is first matched not by the entry containing @code{'jeffrey'} -as the @code{User} field value, but by the entry with no user name! +@item --enable-locking +Enable system locking. Note that if you use this option on a system +which a not fully working lockd() (as on Linux) you will easily get +mysqld to deadlock. -If you have problems connecting to the server, print out the @code{user} -table and sort it by hand to see where the first match is being made. +@item -T, --exit-info +This is a bit mask of different flags one can use for debugging the +mysqld server; One should not use this option if one doesn't know +exactly what it does! -@node Request access, Privilege changes, Connection access, Privilege system -@section Access Control, Stage 2: Request Verification +@item --flush +Flush all changes to disk after each SQL command. Normally @strong{MySQL} +only does a write of all changes to disk after each SQL command and lets +the operating system handle the syncing to disk. +@xref{Crashing}. -Once you establish a connection, the server enters Stage 2. For each request -that comes in on the connection, the server checks whether you have -sufficient privileges to perform it, based on the type of operation you wish -to perform. This is where the privilege fields in the grant tables come into -play. These privileges can come from any of the @code{user}, @code{db}, -@code{host}, @code{tables_priv}, or @code{columns_priv} tables. The grant -tables are manipulated with @code{GRANT} and @code{REVOKE} commands. -@xref{GRANT, , @code{GRANT}}. (You may find it helpful to refer to -@ref{Privileges}, which lists the fields present in each of the grant -tables.) +@item -?, --help +Display short help and exit. -The @code{user} table grants privileges that are assigned to you on a global -basis and that apply no matter what the current database is. For example, if -the @code{user} table grants you the @strong{delete} privilege, you can -delete rows from any database on the server host! In other words, -@code{user} table privileges are superuser privileges. It is wise to grant -privileges in the @code{user} table only to superusers such as server or -database administrators. For other users, you should leave the privileges -in the @code{user} table set to @code{'N'} and grant privileges on a -database-specific basis only, using the @code{db} and @code{host} tables. +@item --init-file=file +Read SQL commands from this file at startup. -@cindex Anonymous user -@cindex wild cards, in @code{mysql.db} table -@cindex wild cards, in @code{mysql.host} table -The @code{db} and @code{host} tables grant database-specific privileges. -Values in the scope fields may be specified as follows: +@item -L, --language=... +Client error messages in given language. May be given as a full path. +@xref{Languages}. -@itemize @bullet -@item -The wild-card characters @samp{%} and @samp{_} can be used in the @code{Host} -and @code{Db} fields of either table. +@item -l, --log[=file] +Log connections and queries to file. @xref{Query log}. -@item -A @code{'%'} @code{Host} value in the @code{db} table means ``any host.'' A -blank @code{Host} value in the @code{db} table means ``consult the -@code{host} table for further information.'' +@item --log-isam[=file] +Log all ISAM/MyISAM changes to file (only used when debugging ISAM/MyISAM). -@item -A @code{'%'} or blank @code{Host} value in the @code{host} table means ``any -host.'' +@item --log-slow-queries[=file] +Log all queries that have taken more than @code{long_query_time} seconds to +execute to file. @xref{Slow query log}. -@item -A @code{'%'} or blank @code{Db} value in either table means ``any database.'' +@item --log-update[=file] +Log updates to @code{file.#} where @code{#} is a unique number if not given. +@xref{Update log}. -@item -A blank @code{User} value in either table matches the anonymous user. -@end itemize +@item --log-long-format +Log some extra information to update log. If you are using +@code{--log-slow-queries} then queries that are not using indexes are logged +to the slow query log. -@cindex grant tables, sorting -@cindex sorting, grant tables -@cindex @code{db} table, sorting -@cindex @code{host} table, sorting -The @code{db} and @code{host} tables are read in and sorted when the server -starts up (at the same time that it reads the @code{user} table). The -@code{db} table is sorted on the @code{Host}, @code{Db}, and @code{User} scope -fields, and the @code{host} table is sorted on the @code{Host} and @code{Db} -scope fields. As with the @code{user} table, sorting puts the most-specific -values first and least-specific values last, and when the server looks for -matching entries, it uses the first match that it finds. +@item --low-priority-updates +Table-modifying operations (@code{INSERT}/@code{DELETE}/@code{UPDATE}) +will have lower priority than selects. It can also be done via +@code{@{INSERT | REPLACE | UPDATE | DELETE@} LOW_PRIORITY ...} to lower +the priority of only one query, or by +@code{SET OPTION SQL_LOW_PRIORITY_UPDATES=1} to change the priority in one +thread. @xref{Table locking}. +@item --memlock +Lock the @code{mysqld} process in memory. This works only if your +system supports the @code{mlockall()} system call (like Solaris). This +may help if you have a problem where the operating system is causing +@code{mysqld} to swap on disk. -@cindex wild cards, in @code{mysql.tables_priv} table -@cindex wild cards, in @code{mysql.columns_priv} table -The @code{tables_priv} and @code{columns_priv} tables grant table- and -column-specific privileges. Values in the scope fields may be specified as -follows: +@item --myisam-recover [=option[,option...]]] where option is one of DEFAULT, BACKUP, FORCE or QUICK. +If this option is used, @code{mysqld} will on open check if the table is +marked as crashed or if if the table wasn't closed properly. (The last +option only works if you are running with @code{--skip-locking}). If this +is the case @code{mysqld} will run check on the table. If the table was corrupted, +@code{mysqld} will attempt to repair it. -@itemize @bullet -@item -The wild-card characters @samp{%} and @samp{_} -can be used in the @code{Host} field of either table. +The following options affects how the repair works. -@item -A @code{'%'} or blank @code{Host} value in either table means ``any host.'' - -@item -The @code{Db}, @code{Table_name} and @code{Column_name} fields cannot contain -wild cards or be blank in either table. -@end itemize +@multitable @columnfractions .3 .7 +@item DEFAULT @tab The same as not giving any option to + @code{--myisam-recover}. +@item BACKUP @tab If the data table was changed during recover, save a + backup of the @file{table_name.MYD} data file as + @file{table_name-datetime.BAK}. +@item FORCE @tab Run recover even if we will loose more than one row + from the .MYD file. +@item QUICK @tab Don't check the rows in the table if there isn't any + delete blocks. +@end multitable -The @code{tables_priv} and @code{columns_priv} tables are sorted on -the @code{Host}, @code{Db}, and @code{User} fields. This is similar to -@code{db} table sorting, although the sorting is simpler because -only the @code{Host} field may contain wild cards. +Before a table is automatically repaired, @strong{MySQL} will add a note +about this in the error log. If you want to be able to recover from most +things without user intervention, you should use the options +@code{BACKUP,FORCE}. This will force a repair of a table even if some rows +would be deleted, but it will keep the old data file as a backup so that +you can later examine what happened. -The request verification process is described below. (If you are familiar -with the access-checking source code, you will notice that the description -here differs slightly from the algorithm used in the code. The description -is equivalent to what the code actually does; it differs only to make the -explanation simpler.) +@item --pid-file=path +Path to pid file used by @code{safe_mysqld}. -For administrative requests (@strong{shutdown}, @strong{reload}, etc.), the -server checks only the @code{user} table entry, because that is the only table -that specifies administrative privileges. Access is granted if the entry -allows the requested operation and denied otherwise. For example, if you -want to execute @code{mysqladmin shutdown} but your @code{user} table entry -doesn't grant the @strong{shutdown} privilege to you, access is denied -without even checking the @code{db} or @code{host} tables. (They -contain no @code{Shutdown_priv} column, so there is no need to do so.) +@item -P, --port=... +Port number to listen for TCP/IP connections. -For database-related requests (@strong{insert}, @strong{update}, etc.), the -server first checks the user's global (superuser) privileges by looking in -the @code{user} table entry. If the entry allows the requested operation, -access is granted. If the global privileges in the @code{user} table are -insufficient, the server determines the user's database-specific privileges -by checking the @code{db} and @code{host} tables: +@item -o, --old-protocol +Use the 3.20 protocol for compatibility with some very old clients. +@xref{Upgrading-from-3.20}. -@enumerate -@item -The server looks in the @code{db} table for a match on the @code{Host}, -@code{Db}, and @code{User} fields. The @code{Host} and @code{User} fields are -matched to the connecting user's hostname and @strong{MySQL} user name. The -@code{Db} field is matched to the database the user wants to access. If -there is no entry for the @code{Host} and @code{User}, access is denied. +@item --one-thread +Only use one thread (for debugging under Linux). @xref{Debugging server}. -@item -If there is a matching @code{db} table entry and its @code{Host} field is -not blank, that entry defines the user's database-specific privileges. +@item -O, --set-variable var=option +Give a variable a value. @code{--help} lists variables. You can find a full +description for all variables in the @code{SHOW VARIABLES} section in this +manual. @xref{SHOW VARIABLES}. The tuning server parameters section includes +information of how to optimize these. @xref{Server parameters}. -@item -If the matching @code{db} table entry's @code{Host} field is blank, it -signifies that the @code{host} table enumerates which hosts should be allowed -access to the database. In this case, a further lookup is done in the -@code{host} table to find a match on the @code{Host} and @code{Db} fields. -If no @code{host} table entry matches, access is denied. If there is a -match, the user's database-specific privileges are computed as the -intersection (@emph{not} the union!) of the privileges in the @code{db} and -@code{host} table entries, that is, the privileges that are @code{'Y'} in both -entries. (This way you can grant general privileges in the @code{db} table -entry and then selectively restrict them on a host-by-host basis using the -@code{host} table entries.) -@end enumerate +@item --safe-mode +Skip some optimize stages. Implies @code{--skip-delay-key-write}. -After determining the database-specific privileges granted by the @code{db} -and @code{host} table entries, the server adds them to the global privileges -granted by the @code{user} table. If the result allows the requested -operation, access is granted. Otherwise, the server checks the user's -table and column privileges in the @code{tables_priv} and @code{columns_priv} -tables and adds those to the user's privileges. Access is allowed or denied -based on the result. +@item --safe-show-database +Don't show databases for which the user doesn't have any privileges. -Expressed in boolean terms, the preceding description of how a user's -privileges are calculated may be summarized like this: +@item --secure +IP numbers returned by the @code{gethostbyname()} system call are +checked to make sure they resolve back to the original hostname. This +makes it harder for someone on the outside to get access by pretending +to be another host. This option also adds some sanity checks of +hostnames. The option is turned off by default in @strong{MySQL} Version 3.21 +because sometimes it takes a long time to perform backward resolutions. +@strong{MySQL} Version 3.22 caches hostnames (unless @code{--skip-host-cache} +is used) and has this option enabled by default. -@example -global privileges -OR (database privileges AND host privileges) -OR table privileges -OR column privileges -@end example +@item --skip-concurrent-insert +Turn off the ability to select and insert at the same time on @code{MyISAM} +tables. (This is only to be used if you think you have found a bug in this +feature). -It may not be apparent why, if the global @code{user} entry privileges are -initially found to be insufficient for the requested operation, the server -adds those privileges to the database-, table-, and column-specific privileges -later. The reason is that a request might require more than one type of -privilege. For example, if you execute an @code{INSERT ... SELECT} -statement, you need both @strong{insert} and @strong{select} privileges. -Your privileges might be such that the @code{user} table entry grants one -privilege and the @code{db} table entry grants the other. In this case, you -have the necessary privileges to perform the request, but the server cannot -tell that from either table by itself; the privileges granted by the entries -in both tables must be combined. +@item --skip-delay-key-write +Ignore the @code{delay_key_write} option for all tables. +@xref{Server parameters}. -@cindex @code{host} table -@cindex tables, @code{host} +@item --skip-grant-tables +This option causes the server not to use the privilege system at all. This +gives everyone @emph{full access} to all databases! (You can tell a running +server to start using the grant tables again by executing @code{mysqladmin +flush-privileges} or @code{mysqladmin reload}.) -The @code{host} table can be used to maintain a list of secure servers. +@item --skip-host-cache +Never use host name cache for faster name-ip resolution, but query DNS server +on every connect instead. @xref{DNS}. -At TcX, the @code{host} table contains a list of all machines on the local -network. These are granted all privileges. +@item --skip-locking +Don't use system locking. To use @code{isamchk} or @code{myisamchk} you must +shut down the server. @xref{Stability}. Note that in @strong{MySQL} Version +3.23 you can use @code{REPAIR} and @code{CHECK} to repair/check @code{MyISAM} +tables. -You can also use the @code{host} table to indicate hosts that are @emph{not} -secure. Suppose you have a machine @code{public.your.domain} that is located -in a public area that you do not consider secure. You can allow access to -all hosts on your network except that machine by using @code{host} table -entries -like this: +@item --skip-name-resolve +Hostnames are not resolved. All @code{Host} column values in the grant tables +must be IP numbers or @code{localhost}. @xref{DNS}. -@example -+--------------------+----+- -| Host | Db | ... -+--------------------+----+- -| public.your.domain | % | ... (all privileges set to 'N') -| %.your.domain | % | ... (all privileges set to 'Y') -+--------------------+----+- -@end example +@item --skip-networking +Don't listen for TCP/IP connections at all. All interaction with +@code{mysqld} must be made via Unix sockets. This option is highly +recommended for systems where only local requests are allowed. @xref{DNS}. -@cindex privilege, changes -@cindex changes to privileges -@cindex tables, grant -@cindex grant tables +@item --skip-new +Don't use new, possible wrong routines. Implies @code{--skip-delay-key-write}. +This will also set default table type to @code{ISAM}. @xref{ISAM}. -Naturally, you should always test your entries in the grant tables (for -example, using @code{mysqlaccess}) to make sure your access privileges are -actually set up the way you think they are. +@item --skip-symlink +Don't delete or rename files that a symlinked file in the data directory +points to. -@node Privilege changes, Default privileges, Request access, Privilege system -@section When Privilege Changes Take Effect +@item --skip-safemalloc +If @strong{MySQL} is configured with @code{--with-debug=full}, all programs +will check the memory for overruns for every memory allocation and memory +freeing. As this checking is very slow, you can avoid this, when you don't +need memory checking, by using this option. -When @code{mysqld} starts, all grant table contents are read into memory and -become effective at that point. +@item --skip-show-database +Don't allow 'SHOW DATABASE' commands, unless the user has @strong{process} +privilege. -Modifications to the grant tables that you perform using @code{GRANT}, -@code{REVOKE}, or @code{SET PASSWORD} are noticed by the server immediately. +@item --skip-stack-trace +Don't write stack traces. This option is useful when you are running +@code{mysqld} under a debugger. @xref{Debugging server}. -If you modify the grant tables manually (using @code{INSERT}, @code{UPDATE}, -etc.), you should execute a @code{FLUSH PRIVILEGES} statement or run -@code{mysqladmin flush-privileges} or @code{mysqladmin reload} to tell the -server to reload the grant tables. Otherwise your changes will have @emph{no -effect} until you restart the server. If you change the grant tables manually -but forget to reload the privileges, you will be wondering why your changes -don't seem to make any difference! +@item --skip-thread-priority +Disable using thread priorities for faster response time. -When the server notices that the grant tables have been changed, -existing client connections are affected as follows: +@item --socket=path +Socket file to use for local connections instead of default +@code{/tmp/mysql.sock}. -@itemize @bullet -@item -Table and column privilege changes take effect with the client's next -request. +@item transaction-isolation= @{ READ-UNCOMMITTED | READ-COMMITTED | REPEATABLE-READ | SERIALIZABLE @} +Sets the default transaction isolation level. @xref{SET TRANSACTION}. -@item -Database privilege changes take effect at the next @code{USE db_name} -command. -@end itemize +@item -t, --tmpdir=path +Path for temporary files. It may be useful if your default @code{/tmp} +directory resides on a partition too small to hold temporary tables. -Global privilege changes and password changes take effect the next time the -client connects. +@item -u, --user=user_name +Run @code{mysqld} daemon as user @code{user_name}. This option is +@emph{mandatory} when starting @code{mysqld} as root. -@cindex privileges, default -@cindex default, privileges -@cindex root password -@cindex superuser -@cindex users, root -@cindex anonymous user -@cindex password, root user -@node Default privileges, Adding users, Privilege changes, Privilege system -@section Setting Up the Initial MySQL Privileges +@item -V, --version +Output version information and exit. -After installing @strong{MySQL}, you set up the initial access privileges by -running @code{scripts/mysql_install_db}. -@xref{Quick install}. -The @code{mysql_install_db} script starts up the @code{mysqld} -server, then initializes the grant tables to contain the following set -of privileges: +@item -W, --warnings +Print out warnings like @code{Aborted connection...} to the @code{.err} file. +@xref{Communication errors}. +@end table -@itemize @bullet -@item -The @strong{MySQL} @code{root} user is created as a superuser who can do -anything. Connections must be made from the local host. +@cindex default options +@cindex option files +@cindex creating, default startup options +@cindex startup options, default +@node Option files, , Command-line options, Post-installation +@subsection Option Files -@strong{NOTE:} -The initial @code{root} password is empty, so anyone can connect as @code{root} -@emph{without a password} and be granted all privileges. +@strong{MySQL} can, since Version 3.22, read default startup options for the +server and for clients from option files. -@item -@cindex anonymous user -An anonymous user is created that can do anything with databases that have a -name of @code{'test'} or starting with @code{'test_'}. Connections must be -made from the local host. This means any local user can connect without a -password and be treated as the anonymous user. +@strong{MySQL} reads default options from the following files on Unix: -@item -Other privileges are denied. For example, normal users can't use -@code{mysqladmin shutdown} or @code{mysqladmin processlist}. -@end itemize +@tindex .my.cnf file +@multitable @columnfractions .3 .7 +@item @strong{Filename} @tab @strong{Purpose} +@item @code{/etc/my.cnf} @tab Global options +@item @code{DATADIR/my.cnf} @tab Server-specific options +@item @code{defaults-extra-file} @tab The file specified with --defaults-extra-file=# +@item @code{~/.my.cnf} @tab User-specific options +@end multitable -@strong{NOTE:} The default privileges are different for Windows. -@xref{Windows running}. +@code{DATADIR} is the @strong{MySQL} data directory (typically +@file{/usr/local/mysql/data} for a binary installation or +@file{/usr/local/var} for a source installation). Note that this is the +directory that was specified at configuration time, not the one specified +with @code{--datadir} when @code{mysqld} starts up! (@code{--datadir} has no +effect on where the server looks for option files, because it looks for them +before it processes any command-line arguments.) -Because your installation is initially wide open, one of the first things you -should do is specify a password for the @strong{MySQL} -@code{root} user. You can do this as follows (note that you specify the -password using the @code{PASSWORD()} function): +@strong{MySQL} reads default options from the following files on Windows: -@example -shell> mysql -u root mysql -mysql> UPDATE user SET Password=PASSWORD('new_password') - WHERE user='root'; -mysql> FLUSH PRIVILEGES; -@end example +@multitable @columnfractions .3 .7 +@item @strong{Filename} @tab @strong{Purpose} +@item @code{windows-system-directory\my.ini} @tab Global options +@item @code{C:\my.cnf} @tab Global options +@item @code{C:\mysql\data\my.cnf} @tab Server-specific options +@end multitable -You can, in @strong{MySQL} Version 3.22 and above, use the @code{SET PASSWORD} -statement: +Note that on Windows, you should specify all paths with @code{/} instead of +@code{\}. If you use @code{\}, you need to specify this twice, as +@code{\} is the escape character in @strong{MySQL}. -@example -shell> mysql -u root mysql -mysql> SET PASSWORD FOR root=PASSWORD('new_password'); -@end example +@cindex Environment variables +@strong{MySQL} tries to read option files in the order listed above. If +multiple option files exist, an option specified in a file read later takes +precedence over the same option specified in a file read earlier. Options +specified on the command line take precedence over options specified in any +option file. Some options can be specified using environment variables. +Options specified on the command line or in option files take precedence over +environment variable values. @xref{Environment variables}. -Another way to set the password is by using the @code{mysqladmin} command: +The following programs support option files: @code{mysql}, +@code{mysqladmin}, @code{mysqld}, @code{mysqldump}, @code{mysqlimport}, +@code{mysql.server}, @code{myisamchk}, and @code{myisampack}. -@example -shell> mysqladmin -u root password new_password -@end example +You can use option files to specify any long option that a program supports! +Run the program with @code{--help} to get a list of available options. -Only users with write/update access to the @code{mysql} database can change the -password for others users. All normal users (not anonymous ones) can only -change their own password with either of the above commands or with -@code{SET PASSWORD=PASSWORD('new password')}. +An option file can contain lines of the following forms: -Note that if you update the password in the @code{user} table directly using -the first method, you must tell the server to re-read the grant tables (with -@code{FLUSH PRIVILEGES}), because the change will go unnoticed otherwise. +@table @code +@item #comment +Comment lines start with @samp{#} or @samp{;}. Empty lines are ignored. -Once the @code{root} password has been set, thereafter you must supply that -password when you connect to the server as @code{root}. +@item [group] +@code{group} is the name of the program or group for which you want to set +options. After a group line, any @code{option} or @code{set-variable} lines +apply to the named group until the end of the option file or another group +line is given. -You may wish to leave the @code{root} password blank so that you don't need -to specify it while you perform additional setup or testing. However, be sure -to set it before using your installation for any real production work. +@item option +This is equivalent to @code{--option} on the command line. -See the @code{scripts/mysql_install_db} script to see how it sets up -the default privileges. You can use this as a basis to see how to -add other users. +@item option=value +This is equivalent to @code{--option=value} on the command line. -If you want the initial privileges to be different than those just described -above, you can modify @code{mysql_install_db} before you run it. +@item set-variable = variable=value +This is equivalent to @code{--set-variable variable=value} on the command line. +This syntax must be used to set a @code{mysqld} variable. +@end table -@cindex grant tables, re-creating -@cindex re-creating, grant tables -To re-create the grant tables completely, remove all the @file{.frm}, -@file{.MYI}, and @file{.MYD} files in the directory containing the -@code{mysql} database. (This is the directory named @file{mysql} under -the database directory, which is listed when you run @code{mysqld ---help}.) Then run the @code{mysql_install_db} script, possibly after -editing it first to have the privileges you want. +The @code{client} group allows you to specify options that apply to all +@strong{MySQL} clients (not @code{mysqld}). This is the perfect group to use +to specify the password you use to connect to the server. (But make +sure the option file is readable and writable only by yourself.) -@strong{NOTE:} For @strong{MySQL} versions older than Version 3.22.10, -you should NOT delete the @file{.frm} files. If you accidentally do this, -you should copy them back from your @strong{MySQL} distribution before -running @code{mysql_install_db}. +Note that for options and values, all leading and trailing blanks are +automatically deleted. You may use the escape sequences @samp{\b}, +@samp{\t}, @samp{\n}, @samp{\r}, @samp{\\}, and @samp{\s} in your value string +(@samp{\s} == blank). -@cindex privileges, adding -@cindex adding, new user privileges -@cindex user privileges, adding -@findex GRANT statement -@findex statements, GRANT -@node Adding users, Passwords, Default privileges, Privilege system -@section Adding New Users to MySQL +Here is a typical global option file: -You can add users two different ways: by using @code{GRANT} statements -or by manipulating the @strong{MySQL} grant tables directly. The -preferred method is to use @code{GRANT} statements, because they are -more concise and less error-prone. @xref{GRANT}. +@example +[client] +port=3306 +socket=/tmp/mysql.sock -There is also a lot of contributed programs like @code{phpmyadmin} that -can be used to create and administrate users. @xref{Contrib}. +[mysqld] +port=3306 +socket=/tmp/mysql.sock +set-variable = key_buffer_size=16M +set-variable = max_allowed_packet=1M -The examples below show how to use the @code{mysql} client to set up new -users. These examples assume that privileges are set up according to the -defaults described in the previous section. This means that to make changes, -you must be on the same machine where @code{mysqld} is running, you must -connect as the @strong{MySQL} @code{root} user, and the @code{root} user must -have the @strong{insert} privilege for the @code{mysql} database and the -@strong{reload} administrative privilege. Also, if you have changed the -@code{root} user password, you must specify it for the @code{mysql} commands -below. +[mysqldump] +quick +@end example -You can add new users by issuing @code{GRANT} statements: +Here is typical user option file: @example -shell> mysql --user=root mysql -mysql> GRANT ALL PRIVILEGES ON *.* TO monty@@localhost - IDENTIFIED BY 'some_pass' WITH GRANT OPTION; -mysql> GRANT ALL PRIVILEGES ON *.* TO monty@@"%" - IDENTIFIED BY 'some_pass' WITH GRANT OPTION; -mysql> GRANT RELOAD,PROCESS ON *.* TO admin@@localhost; -mysql> GRANT USAGE ON *.* TO dummy@@localhost; -@end example +[client] +# The following password will be sent to all standard MySQL clients +password=my_password -These @code{GRANT} statements set up three new users: +[mysql] +no-auto-rehash +set-variable = connect_timeout=2 -@table @code -@item monty -A full superuser who can connect to the server from anywhere, but who must -use a password @code{'some_pass'} to do so. Note that we must issue -@code{GRANT} statements for both @code{monty@@localhost} and -@code{monty@@"%"}. If we don't add the entry with @code{localhost}, the -anonymous user entry for @code{localhost} that is created by -@code{mysql_install_db} will take precedence when we connect from the local -host, because it has a more specific @code{Host} field value and thus comes -earlier in the @code{user} table sort order. +[mysqlhotcopy] +interactive-timeout -@item admin -A user who can connect from @code{localhost} without a password and who is -granted the @strong{reload} and @strong{process} administrative privileges. -This allows the user to execute the @code{mysqladmin reload}, -@code{mysqladmin refresh}, and @code{mysqladmin flush-*} commands, as well as -@code{mysqladmin processlist} . No database-related privileges are granted. -(They can be granted later by issuing additional @code{GRANT} statements.) +@end example -@item dummy -A user who can connect without a password, but only from the local host. The -global privileges are all set to @code{'N'} --- the @code{USAGE} privilege -type allows you to create a user with no privileges. It is assumed that you -will grant database-specific privileges later. -@end table +@tindex .my.cnf file +If you have a source distribution, you will find sample configuration +files named @file{my-xxxx.cnf} in the @file{support-files} directory. +If you have a binary distribution, look in the @file{DIR/support-files} +directory, where @code{DIR} is the pathname to the @strong{MySQL} +installation directory (typically @file{/usr/local/mysql}). Currently +there are sample configuration files for small, medium, large, and very +large systems. You can copy @file{my-xxxx.cnf} to your home directory +(rename the copy to @file{.my.cnf}) to experiment with this. -@findex INSERT statement, grant privileges -@findex statements, INSERT -You can also add the same user access information directly by issuing -@code{INSERT} statements and then telling the server to reload the grant -tables: +All @strong{MySQL} clients that support option files support the +following options: -@example -shell> mysql --user=root mysql -mysql> INSERT INTO user VALUES('localhost','monty',PASSWORD('some_pass'), - 'Y','Y','Y','Y','Y','Y','Y','Y','Y','Y','Y','Y','Y','Y'); -mysql> INSERT INTO user VALUES('%','monty',PASSWORD('some_pass'), - 'Y','Y','Y','Y','Y','Y','Y','Y','Y','Y','Y','Y','Y','Y'); -mysql> INSERT INTO user SET Host='localhost',User='admin', - Reload_priv='Y', Process_priv='Y'; -mysql> INSERT INTO user (Host,User,Password) - VALUES('localhost','dummy',''); -mysql> FLUSH PRIVILEGES; -@end example +@multitable @columnfractions .40 .60 +@item --no-defaults @tab Don't read any option files. +@item --print-defaults @tab Print the program name and all options that it will get. +@item --defaults-file=full-path-to-default-file @tab Only use the given configuration file. +@item --defaults-extra-file=full-path-to-default-file @tab Read this configuration file after the global configuration file but before the user configuration file. +@end multitable -Depending on your @strong{MySQL} version, you may have to use a different -number of @code{'Y'} values above (versions prior to Version 3.22.11 had fewer -privilege columns). For the @code{admin} user, the more readable extended -@code{INSERT} syntax that is available starting with Version 3.22.11 is used. +Note that the above options must be first on the command line to work! +@code{--print-defaults} may however be used directly after the +@code{--defaults-xxx-file} commands. -Note that to set up a superuser, you need only create a @code{user} table -entry with the privilege fields set to @code{'Y'}. No @code{db} or -@code{host} table entries are necessary. - -The privilege columns in the @code{user} table were not set explicitly in the -last @code{INSERT} statement (for the @code{dummy} user), so those columns -are assigned the default value of @code{'N'}. This is the same thing that -@code{GRANT USAGE} does. - -The following example adds a user @code{custom} who can connect from hosts -@code{localhost}, @code{server.domain}, and @code{whitehouse.gov}. He wants -to access the @code{bankaccount} database only from @code{localhost}, -the @code{expenses} database only from @code{whitehouse.gov}, and -the @code{customer} database from all three hosts. He wants -to use the password @code{stupid} from all three hosts. +Note for developers: Option file handling is implemented simply by +processing all matching options (that is, options in the appropriate group) +before any command-line arguments. This works nicely for programs that use +the last instance of an option that is specified multiple times. If you have +an old program that handles multiply-specified options this way but doesn't +read option files, you need add only two lines to give it that capability. +Check the source code of any of the standard @strong{MySQL} clients to see +how to do this. -To set up this user's privileges using @code{GRANT} statements, run these -commands: +In shell scripts you can use the @file{my_print_defaults} command to parse the +config files: @example -shell> mysql --user=root mysql -mysql> GRANT SELECT,INSERT,UPDATE,DELETE,CREATE,DROP - ON bankaccount.* - TO custom@@localhost - IDENTIFIED BY 'stupid'; -mysql> GRANT SELECT,INSERT,UPDATE,DELETE,CREATE,DROP - ON expenses.* - TO custom@@whitehouse.gov - IDENTIFIED BY 'stupid'; -mysql> GRANT SELECT,INSERT,UPDATE,DELETE,CREATE,DROP - ON customer.* - TO custom@@'%' - IDENTIFIED BY 'stupid'; + +shell> my_print_defaults client mysql +--port=3306 +--socket=/tmp/mysql.sock +--no-auto-rehash @end example -The reason that we do to grant statements for the user 'custom' is that -we want the give the user access to @strong{MySQL} both from the local -machine with Unix sockets and from the remote machine 'whitehouse.gov' -over TCP/IP. +The above output contains all options for the groups 'client' and 'mysql'. -To set up the user's privileges by modifying the grant tables directly, -run these commands (note the @code{FLUSH PRIVILEGES} at the end): +@node Installing many servers, Upgrade, Post-installation, Installing +@section Installing Many Servers on the Same Machine +@cindex post-install, many servers +@cindex Installing many servers +@cindex Starting many servers + +In some cases you may want to have many different @code{mysqld} daemons +(servers) running on the same machine. You may for example want to run +a new version of @strong{MySQL} for testing together with an old version +that is in production. Another case is when you want to give different +users access to different @code{mysqld} servers that they manage themselves. + +One way to get a new server running is by starting it with a different +socket and port as follows: +@tindex MYSQL_UNIX_PORT environment variable +@tindex MYSQL_TCP_PORT environment variable +@tindex Environment variable, MYSQL_UNIX_PORT +@tindex Environment variable, MYSQL_TCP_PORT @example -shell> mysql --user=root mysql -mysql> INSERT INTO user (Host,User,Password) - VALUES('localhost','custom',PASSWORD('stupid')); -mysql> INSERT INTO user (Host,User,Password) - VALUES('server.domain','custom',PASSWORD('stupid')); -mysql> INSERT INTO user (Host,User,Password) - VALUES('whitehouse.gov','custom',PASSWORD('stupid')); -mysql> INSERT INTO db - (Host,Db,User,Select_priv,Insert_priv,Update_priv,Delete_priv, - Create_priv,Drop_priv) - VALUES - ('localhost','bankaccount','custom','Y','Y','Y','Y','Y','Y'); -mysql> INSERT INTO db - (Host,Db,User,Select_priv,Insert_priv,Update_priv,Delete_priv, - Create_priv,Drop_priv) - VALUES - ('whitehouse.gov','expenses','custom','Y','Y','Y','Y','Y','Y'); -mysql> INSERT INTO db - (Host,Db,User,Select_priv,Insert_priv,Update_priv,Delete_priv, - Create_priv,Drop_priv) - VALUES('%','customer','custom','Y','Y','Y','Y','Y','Y'); -mysql> FLUSH PRIVILEGES; +shell> MYSQL_UNIX_PORT=/tmp/mysqld-new.sock +shell> MYSQL_TCP_PORT=3307 +shell> export MYSQL_UNIX_PORT MYSQL_TCP_PORT +shell> scripts/mysql_install_db +shell> bin/safe_mysqld & @end example -The first three @code{INSERT} statements add @code{user} table entries that -allow user @code{custom} to connect from the various hosts with the given -password, but grant no permissions to him (all privileges are set to the -default value of @code{'N'}). The next three @code{INSERT} statements add -@code{db} table entries that grant privileges to @code{custom} for the -@code{bankaccount}, @code{expenses}, and @code{customer} databases, but only -when accessed from the proper hosts. As usual, when the grant tables are -modified directly, the server must be told to reload them (with -@code{FLUSH PRIVILEGES}) so that the privilege changes take effect. +The environment variables appendix includes a list of other environment +variables you can use to affect @code{mysqld}. @xref{Environment variables}. -If you want to give a specific user access from any machine in a given -domain, you can issue a @code{GRANT} statement like the following: +The above is the quick and dirty way that one commonly uses for testing. +The nice thing with this is that all connections you do in the above shell +will automatically be directed to the new running server! -@example -mysql> GRANT ... - ON *.* - TO myusername@@"%.mydomainname.com" - IDENTIFIED BY 'mypassword'; -@end example +If you need to do this more permanently, you should create an option +file for each server. @xref{Option files}. In your startup script that +is executed at boot time (mysql.server?) you should specify for both +servers: -To do the same thing by modifying the grant tables directly, do this: +@code{safe_mysqld --default-file=path-to-option-file} -@example -mysql> INSERT INTO user VALUES ('%.mydomainname.com', 'myusername', - PASSWORD('mypassword'),...); -mysql> FLUSH PRIVILEGES; -@end example +At least the following options should be different per server: -You can also use @code{xmysqladmin}, @code{mysql_webadmin}, and even -@code{xmysql} to insert, change, and update values in the grant tables. -You can find these utilities in the -@uref{http://www.mysql.com/Downloads/Contrib/,Contrib directory of the @strong{MySQL} -Website}. +@table @code +@item port=# +@item socket=path +@item pid-file=path +@end table -@cindex passwords, setting -@findex PASSWORD() -@findex SET PASSWORD statement -@cindex setting, passwords -@node Passwords, Access denied, Adding users, Privilege system -@section Setting Up Passwords +The following options should be different, if they are used: -In most cases you should use @code{GRANT} to set up your users/passwords, -so the following only applies for advanced users. @xref{GRANT, , @code{GRANT}}. +@table @code +@item log=path +@item log-bin=path +@item log-update=path +@item log-isam=path +@item bdb-logdir=path +@end table -The examples in the preceding sections illustrate an important principle: -when you store a non-empty password using @code{INSERT} or @code{UPDATE} -statements, you must use the @code{PASSWORD()} function to encrypt it. This -is because the @code{user} table stores passwords in encrypted form, not as -plaintext. If you forget that fact, you are likely to attempt to set -passwords like this: +If you want more performance, you can also specify the following differently: -@example -shell> mysql -u root mysql -mysql> INSERT INTO user (Host,User,Password) - VALUES('%','jeffrey','biscuit'); -mysql> FLUSH PRIVILEGES; -@end example +@table @code +@item tmpdir=path +@item bdb-tmpdir=path +@end table -The result is that the plaintext value @code{'biscuit'} is stored as the -password in the @code{user} table. When the user @code{jeffrey} attempts to -connect to the server using this password, the @code{mysql} client encrypts -it with @code{PASSWORD()} and sends the result to the server. The server -compares the value in the @code{user} table (the encrypted value of -@code{'biscuit'}) to the encrypted password (which is @emph{not} -@code{'biscuit'}). The comparison fails and the server rejects the -connection: +@xref{Command-line options}. -@example -shell> mysql -u jeffrey -pbiscuit test -Access denied -@end example +If you are installing binary @strong{MySQL} versions (.tar files) and +start them with @code{./bin/safe_mysqld} then in most cases the only +option you need to add/change is the @code{socket} and @code{port} +argument to @code{safe_mysqld}. -Passwords must be encrypted when they are inserted in the @code{user} -table, so the @code{INSERT} statement should have been specified like this -instead: +@node Upgrade, , Installing many servers, Installing +@section Upgrading/Downgrading MySQL +@cindex upgrading +@cindex downgrading -@example -mysql> INSERT INTO user (Host,User,Password) - VALUES('%','jeffrey',PASSWORD('biscuit')); -@end example +You can always move the @strong{MySQL} form and data files between +different versions on the same architecture as long as you have the same +base version of @strong{MySQL}. The current base version is +3. If you change the character set when running @strong{MySQL} (which may +also change the sort order), you must run @code{myisamchk -r -q} on all +tables. Otherwise your indexes may not be ordered correctly. -You must also use the @code{PASSWORD()} function when you use @code{SET -PASSWORD} statements: +If you are afraid of new versions, you can always rename your old +@code{mysqld} to something like @code{mysqld}-'old-version-number'. If +your new @code{mysqld} then does something unexpected, you can simply shut it +down and restart with your old @code{mysqld}! -@example -mysql> SET PASSWORD FOR jeffrey@@"%" = PASSWORD('biscuit'); -@end example +When you do an upgrade you should also back up your old databases, of course. -If you set passwords using the @code{GRANT ... IDENTIFIED BY} statement -or the @code{mysqladmin password} command, the @code{PASSWORD()} function -is unnecessary. They both take care of encrypting the password for you, -so you would specify a password of @code{'biscuit'} like this: +If after an upgrade, you experience problems with recompiled client programs, +like @code{Commands out of sync} or unexpected core dumps, you probably have +used an old header or library file when compiling your programs. In this +case you should check the date for your @file{mysql.h} file and +@file{libmysqlclient.a} library to verify that they are from the new +@strong{MySQL} distribution. If not, please recompile your programs! -@example -mysql> GRANT USAGE ON *.* TO jeffrey@@"%" IDENTIFIED BY 'biscuit'; -@end example +If you get some problems that the new @code{mysqld} server doesn't want to +start or that you can't connect without a password, check that you don't +have some old @file{my.cnf} file from your old installation! You can +check this with: @code{program-name --print-defaults}. If this outputs +anything other than the program name, you have an active @code{my.cnf} +file that will affect things! -or +It is a good idea to rebuild and reinstall the @code{Msql-Mysql-modules} +distribution whenever you install a new release of @strong{MySQL}, +particularly if you notice symptoms such as all your @code{DBI} scripts +dumping core after you upgrade @strong{MySQL}. -@example -shell> mysqladmin -u jeffrey password biscuit -@end example +@menu +* Upgrading-from-3.22:: Upgrading from a 3.22 version to 3.23 +* Upgrading-from-3.21:: Upgrading from a 3.21 version to 3.22 +* Upgrading-from-3.20:: Upgrading from a 3.20 version to 3.21 +* Upgrading-to-arch:: Upgrading to another architecture +@end menu -@strong{NOTE:} @code{PASSWORD()} does not perform password encryption in the -same way that Unix passwords are encrypted. You should not assume that if -your Unix password and your @strong{MySQL} password are the same, that -@code{PASSWORD()} will result in the same encrypted value as is stored in the -Unix password file. @xref{User names}. +@cindex compatibility, between MySQL versions +@cindex upgrading, 3.22 to 3.23 +@node Upgrading-from-3.22, Upgrading-from-3.21, Upgrade, Upgrade +@subsection Upgrading From Version 3.22 to Version 3.23 -@node Access denied, , Passwords, Privilege system -@section Causes of @code{Access denied} Errors +@strong{MySQL} Version 3.23 supports tables of the new @code{MyISAM} type and +the old @code{ISAM} type. You don't have to convert your old tables to +use these with Version 3.23. By default, all new tables will be created with +type @code{MyISAM} (unless you start @code{mysqld} with the +@code{--default-table-type=isam} option). You can change an @code{ISAM} +table to a @code{MyISAM} table with @code{ALTER TABLE table_name TYPE=MyISAM} +or the Perl script @code{mysql_convert_table_format}. -If you encounter @code{Access denied} errors when you try to connect to the -@strong{MySQL} server, the list below indicates some courses of -action you can take to correct the problem: +Version 3.22 and 3.21 clients will work without any problems with a Version +3.23 server. + +The following lists tell what you have to watch out for when upgrading to +Version 3.23: @itemize @bullet @item -After installing @strong{MySQL}, did you run the @code{mysql_install_db} -script to set up the initial grant table contents? If not, do so. -@xref{Default privileges}. Test the initial privileges by executing -this command: +All tables that uses the @code{tis620} character set must be fixed +with @code{myisamchk -r} or @code{REPAIR TABLE}. +@item +If you do a @code{DROP DATABASE} on a symbolic linked database, both the +link and the original database is deleted. (This didn't happen in 3.22 +because configure didn't detect the @code{readlink} system call). +@item +@code{OPTIMIZE TABLE} now only works for @strong{MyISAM} tables. +For other table types, you can use @code{ALTER TABLE} to optimize the table. +During @code{OPTIMIZE TABLE} the table is now locked from other threads. +@item +The @strong{MySQL} client @code{mysql} is now by default started with the +option @code{--no-named-commands (-g)}. This option can be disabled with +@code{--enable-named-commands (-G)}. This may cause incompatibility problems in +some cases, for example in SQL scripts that use named commands without a +semicolon! Long format commands still work from the first line. +@item +If you are using the @code{german} character sort order, you must repair +all your tables with @code{isamchk -r}, as we have made some changes in +the sort order! +@item The default return type of @code{IF} will now depend on both arguments +and not only the first argument. +@item @code{AUTO_INCREMENT} will not work with negative numbers. The reason +for this is that negative numbers caused problems when wrapping from -1 to 0. +@code{AUTO_INCREMENT} is now for MyISAM tables handled at a lower level and +is much faster than before. For MyISAM tables old numbers are also not reused +anymore, even if you delete some rows from the table. +@item @code{CASE}, @code{DELAYED}, @code{ELSE}, @code{END}, @code{FULLTEXT}, @code{INNER}, @code{RIGHT}, @code{THEN} and @code{WHEN} are now reserved words. +@item @code{FLOAT(X)} is now a true floating-point type and not a value with +a fixed number of decimals. +@item When declaring @code{DECIMAL(length,dec)} the length argument no +longer includes a place for the sign or the decimal point. +@item A @code{TIME} string must now be of one of the following formats: +@code{[[[DAYS] [H]H:]MM:]SS[.fraction]} or +@code{[[[[[H]H]H]H]MM]SS[.fraction]} +@item @code{LIKE} now compares strings using the same character +comparison rules as @code{'='}. If you require the old behavior, you +can compile @strong{MySQL} with the @code{CXXFLAGS=-DLIKE_CMP_TOUPPER} +flag. +@item @code{REGEXP} is now case insensitive for normal (not binary) strings. +@item When you check/repair tables you should use @code{CHECK TABLE} +or @code{myisamchk} for @code{MyISAM} tables (@code{.MYI}) and +@code{isamchk} for ISAM (@code{.ISM}) tables. +@item If you want your @code{mysqldump} files to be compatible between +@strong{MySQL} Version 3.22 and Version 3.23, you should not use the +@code{--opt} or @code{--full} option to @code{mysqldump}. +@item Check all your calls to @code{DATE_FORMAT()} to make sure there is a +@samp{%} before each format character. (Later @strong{MySQL} Version 3.22 +did allow this syntax.) +@item +@code{mysql_fetch_fields_direct} is now a function (it was a macro) and +it returns a pointer to a @code{MYSQL_FIELD} instead of a +@code{MYSQL_FIELD}. +@item +@code{mysql_num_fields()} can no longer be used on a @code{MYSQL*} object (it's +now a function that takes @code{MYSQL_RES*} as an argument. You should now +use @code{mysql_field_count()} instead. +@item +In @strong{MySQL} Version 3.22, the output of @code{SELECT DISTINCT ...} was +almost always sorted. In Version 3.23, you must use @code{GROUP BY} or +@code{ORDER BY} to obtain sorted output. +@item +@code{SUM()} now returns @code{NULL}, instead of 0, if there is no matching +rows. This is according to ANSI SQL. +@item An @code{AND} or @code{OR} with @code{NULL} values will now return +@code{NULL} instead of 0. This mostly affects queries that use @code{NOT} +on an @code{AND/OR} expression as @code{NOT NULL} = @code{NULL}. +@code{LPAD()} and @code{RPAD()} will shorten the result string if it's longer +than the length argument. +@end itemize -@example -shell> mysql -u root test -@end example +@cindex compatibility, between MySQL versions +@node Upgrading-from-3.21, Upgrading-from-3.20, Upgrading-from-3.22, Upgrade +@subsection Upgrading from Version 3.21 to Version 3.22 +@cindex upgrading, 3.21 to 3.22 -The server should let you connect without error. You should also make sure -you have a file @file{user.MYD} in the @strong{MySQL} database directory. -Ordinarily, this is @file{PATH/var/mysql/user.MYD}, where @code{PATH} is the -pathname to the @strong{MySQL} installation root. +Nothing that affects compatibility has changed between Version 3.21 and 3.22. +The only pitfall is that new tables that are created with @code{DATE} type +columns will use the new way to store the date. You can't access these new +fields from an old version of @code{mysqld}. -@item -After a fresh installation, you should connect to the server and set up -your users and their access permissions: +After installing @strong{MySQL} Version 3.22, you should start the new server +and then run the @code{mysql_fix_privilege_tables} script. This will add the +new privileges that you need to use the @code{GRANT} command. If you forget +this, you will get @code{Access denied} when you try to use @code{ALTER +TABLE}, @code{CREATE INDEX}, or @code{DROP INDEX}. If your @strong{MySQL} root +user requires a password, you should give this as an argument to +@code{mysql_fix_privilege_tables}. -@example -shell> mysql -u root mysql -@end example +The C API interface to @code{mysql_real_connect()} has changed. If you have +an old client program that calls this function, you must place a @code{0} for +the new @code{db} argument (or recode the client to send the @code{db} +element for faster connections). You must also call @code{mysql_init()} +before calling @code{mysql_real_connect()}! This change was done to allow +the new @code{mysql_options()} function to save options in the @code{MYSQL} +handler structure. -The server should let you connect because the @strong{MySQL} @code{root} user -has no password initially. That is also a security risk, so setting the -@code{root} password is something you should do while you're setting up -your other @strong{MySQL} users. +The @code{mysqld} variable @code{key_buffer} has changed names to +@code{key_buffer_size}, but you can still use the old name in your +startup files. -If you try to connect as @code{root} and get this error: +@node Upgrading-from-3.20, Upgrading-to-arch, Upgrading-from-3.21, Upgrade +@subsection Upgrading from Version 3.20 to Version 3.21 +@cindex upgrading, 3.20 to 3.21 -@example -Access denied for user: '@@unknown' to database mysql -@end example +If you are running a version older than Version 3.20.28 and want to +switch to Version 3.21, you need to do the following: -this means that you don't have an entry in the @code{user} table with a -@code{User} column value of @code{'root'} and that @code{mysqld} cannot -resolve the hostname for your client. In this case, you must restart the -server with the @code{--skip-grant-tables} option and edit your -@file{/etc/hosts} or @file{\windows\hosts} file to add an entry for your -host. +You can start the @code{mysqld} Version 3.21 server with @code{safe_mysqld +--old-protocol} to use it with clients from a Version 3.20 distribution. +In this case, the new client function @code{mysql_errno()} will not +return any server error, only @code{CR_UNKNOWN_ERROR} (but it +works for client errors), and the server uses the old @code{password()} +checking rather than the new one. -@item -If you get an error like the following: +If you are @strong{NOT} using the @code{--old-protocol} option to +@code{mysqld}, you will need to make the following changes: -@example -shell> mysqladmin -u root -pxxxx ver -Access denied for user: 'root@@localhost' (Using password: YES) -@end example +@itemize @bullet +@item +All client code must be recompiled. If you are using ODBC, you must get +the new @strong{MyODBC} 2.x driver. +@item +The script @code{scripts/add_long_password} must be run to convert the +@code{Password} field in the @code{mysql.user} table to @code{CHAR(16)}. +@item +All passwords must be reassigned in the @code{mysql.user} table (to get 62-bit +rather than 31-bit passwords). +@item +The table format hasn't changed, so you don't have to convert any tables. +@end itemize -It means that you are using a wrong password. @xref{Passwords}. +@strong{MySQL} Version 3.20.28 and above can handle the new @code{user} table +format without affecting clients. If you have a @strong{MySQL} version earlier +than Version 3.20.28, passwords will no longer work with it if you convert the +@code{user} table. So to be safe, you should first upgrade to at least Version +3.20.28 and then upgrade to Version 3.21. -If you have forgot the root password, you can restart @code{mysqld} with -@code{--skip-grant-tables} to change the password. You can find more -about this option later on in this manual section. +@cindex Protocol mismatch +The new client code works with a 3.20.x @code{mysqld} server, so +if you experience problems with 3.21.x, you can use the old 3.20.x server +without having to recompile the clients again. -If you get the above error even if you haven't specified a password, -this means that you a wrong password in some @code{my.ini} -file. @xref{Option files}. You can avoid using option files with the @code{--no-defaults} option, as follows: +If you are not using the @code{--old-protocol} option to @code{mysqld}, +old clients will issue the error message: @example -shell> mysqladmin --no-defaults -u root ver +ERROR: Protocol mismatch. Server Version = 10 Client Version = 9 @end example -@item -@cindex @code{mysql_fix_privilege_tables} -If you updated an existing @strong{MySQL} installation from a version earlier -than Version 3.22.11 to Version 3.22.11 or later, did you run the -@code{mysql_fix_privilege_tables} script? If not, do so. The structure of -the grant tables changed with @strong{MySQL} Version 3.22.11 when the -@code{GRANT} statement became functional. +The new Perl @code{DBI}/@code{DBD} interface also supports the old +@code{mysqlperl} interface. The only change you have to make if you use +@code{mysqlperl} is to change the arguments to the @code{connect()} function. +The new arguments are: @code{host}, @code{database}, @code{user}, +@code{password} (the @code{user} and @code{password} arguments have changed +places). +@xref{Perl DBI Class, , Perl @code{DBI} Class}. -@item -If your privileges seem to have changed in the middle of a session, it may be -that a superuser has changed them. Reloading the grant tables affects new -client connections, but it also affects existing connections as indicated in -@ref{Privilege changes}. +The following changes may affect queries in old applications: +@itemize @bullet @item -If you can't get your password to work, remember that you must use -the @code{PASSWORD()} function if you set the password with the -@code{INSERT}, @code{UPDATE}, or @code{SET PASSWORD} statements. The -@code{PASSWORD()} function is unnecessary if you specify the password using -the @code{GRANT ... INDENTIFIED BY} statement or the @code{mysqladmin -password} command. -@xref{Passwords}. - +@code{HAVING} must now be specified before any @code{ORDER BY} clause. @item -@code{localhost} is a synonym for your local hostname, and is also the -default host to which clients try to connect if you specify no host -explicitly. However, connections to @code{localhost} do not work if you are -running on a system that uses MIT-pthreads (@code{localhost} connections are -made using Unix sockets, which are not supported by MIT-pthreads). To avoid -this problem on such systems, you should use the @code{--host} option to name -the server host explicitly. This will make a TCP/IP connection to the -@code{mysqld} server. In this case, you must have your real hostname in -@code{user} table entries on the server host. (This is true even if you are -running a client program on the same host as the server.) - +The parameters to @code{LOCATE()} have been swapped. @item -If you get an @code{Access denied} error when trying to connect to the -database with @code{mysql -u user_name db_name}, you may have a problem -with the @code{user} table. Check this by executing @code{mysql -u root -mysql} and issuing this SQL statement: - -@example -mysql> SELECT * FROM user; -@end example +There are some new reserved words. The most notable are @code{DATE}, +@code{TIME}, and @code{TIMESTAMP}. +@end itemize -The result should include an entry with the @code{Host} and @code{User} -columns matching your computer's hostname and your @strong{MySQL} user name. +@cindex upgrading, different architecture +@node Upgrading-to-arch, , Upgrading-from-3.20, Upgrade +@subsection Upgrading to Another Architecture -@item -The @code{Access denied} error message will tell you who you are trying -to log in as, the host from which you are trying to connect, and whether -or not you were using a password. Normally, you should have one entry in -the @code{user} table that exactly matches the hostname and user name -that were given in the error message. For example if you get an error -message that contains @code{Using password: NO}, this means that you -tried to login without an password. - -@item -If you get the following error when you try to connect from a different host -than the one on which the @strong{MySQL} server is running, then there is no -row in the @code{user} table that matches that host: +If you are using @strong{MySQL} Version 3.23, you can copy the @code{.frm}, +@code{.MYI}, and @code{.MYD} files between different architectures that +support the same floating-point format. (@strong{MySQL} takes care of any +byte swapping issues.) -@example -Host ... is not allowed to connect to this MySQL server -@end example +The @strong{MySQL} @code{ISAM} data and index files (@file{.ISD} and +@file{*.ISM}, respectively) are architecture-dependent and in some cases +OS-dependent. If you want to move your applications to another machine +that has a different architecture or OS than your current machine, you +should not try to move a database by simply copying the files to the +other machine. Use @code{mysqldump} instead. -You can fix this by using the command-line tool @code{mysql} (on the -server host!) to add a row to the @code{user}, @code{db}, or @code{host} -table for the user/hostname combination from which you are trying to -connect and then execute @code{mysqladmin flush-privileges}. If you are -not running @strong{MySQL} Version 3.22 and you don't know the IP number or -hostname of the machine from which you are connecting, you should put an -entry with @code{'%'} as the @code{Host} column value in the @code{user} -table and restart @code{mysqld} with the @code{--log} option on the -server machine. After trying to connect from the client machine, the -information in the @strong{MySQL} log will indicate how you really did -connect. (Then replace the @code{'%'} in the @code{user} table entry -with the actual hostname that shows up in the log. Otherwise, you'll -have a system that is insecure.) +By default, @code{mysqldump} will create a file full of SQL statements. +You can then transfer the file to the other machine and feed it as input +to the @code{mysql} client. -Another reason for this error on Linux is that you are using a binary -@strong{MySQL} version that is compiled with a different glibc version -than the one you are using. In this case you should either upgrade your -OS/glibc or download the source @strong{MySQL} version and compile this -yourself. A source RPM is normally trivial to compile and install, so -this isn't a big problem. +Try @code{mysqldump --help} to see what options are available. +If you are moving the data to a newer version of @strong{MySQL}, you should use +@code{mysqldump --opt} with the newer version to get a fast, compact dump. -@item -If you get an error message where the hostname is not shown or where the -hostname is an IP, even if you try to connect with a hostname: +The easiest (although not the fastest) way to move a database between two +machines is to run the following commands on the machine on which the +database is located: @example -shell> mysqladmin -u root -pxxxx -h some-hostname ver -Access denied for user: 'root@' (Using password: YES) +shell> mysqladmin -h 'other hostname' create db_name +shell> mysqldump --opt db_name \ + | mysql -h 'other hostname' db_name @end example -This means that @strong{MySQL} got some error when trying to resolve the -IP to a hostname. In this case you can execute @code{mysqladmin -flush-hosts} to reset the internal DNS cache. @xref{DNS}. - -Some permanent solutions are: - -@itemize @minus -@item -Try to find out what is wrong with your DNS server and fix this. -@item -Specify IPs instead of hostnames in the @strong{MySQL} privilege tables. -@item -Start @code{mysqld} with @code{--skip-name-resolve}. -@item -Start @code{mysqld} with @code{--skip-host-cache}. -@item -Connect to @code{localhost} if you are running the server and the client -on the same machine. -@item -Put the client machine names in @code{/etc/hosts}. -@end itemize - -@item -If @code{mysql -u root test} works but @code{mysql -h your_hostname -u root -test} results in @code{Access denied}, then you may not have the correct name -for your host in the @code{user} table. A common problem here is that the -@code{Host} value in the user table entry specifies an unqualified hostname, -but your system's name resolution routines return a fully qualified domain -name (or vice-versa). For example, if you have an entry with host -@code{'tcx'} in the @code{user} table, but your DNS tells @strong{MySQL} that -your hostname is @code{'tcx.subnet.se'}, the entry will not work. Try adding -an entry to the @code{user} table that contains the IP number of your host as -the @code{Host} column value. (Alternatively, you could add an entry to the -@code{user} table with a @code{Host} value that contains a wild card---for -example, @code{'tcx.%'}. However, use of hostnames ending with @samp{%} is -@emph{insecure} and is @emph{not} recommended!) - -@item -If @code{mysql -u user_name test} works but @code{mysql -u user_name -other_db_name} doesn't work, you don't have an entry for @code{other_db_name} -listed in the @code{db} table. - -@item -If @code{mysql -u user_name db_name} works when executed on the server -machine, but @code{mysql -u host_name -u user_name db_name} doesn't work when -executed on another client machine, you don't have the client machine listed -in the @code{user} table or the @code{db} table. - -@item -If you can't figure out why you get @code{Access denied}, remove from the -@code{user} table all entries that have @code{Host} values containing -wild cards (entries that contain @samp{%} or @samp{_}). A very common error -is to insert a new entry with @code{Host}=@code{'%'} and -@code{User}=@code{'some user'}, thinking that this will allow you to specify -@code{localhost} to connect from the same machine. The reason that this -doesn't work is that the default privileges include an entry with -@code{Host}=@code{'localhost'} and @code{User}=@code{''}. Because that entry -has a @code{Host} value @code{'localhost'} that is more specific than -@code{'%'}, it is used in preference to the new entry when connecting from -@code{localhost}! The correct procedure is to insert a second entry with -@code{Host}=@code{'localhost'} and @code{User}=@code{'some_user'}, or to -remove the entry with @code{Host}=@code{'localhost'} and -@code{User}=@code{''}. - -@item -If you get the following error, you may have a problem with the @code{db} or -@code{host} table: +If you want to copy a database from a remote machine over a slow network, +you can use: @example -Access to database denied +shell> mysqladmin create db_name +shell> mysqldump -h 'other hostname' --opt --compress db_name \ + | mysql db_name @end example -If the entry selected from the @code{db} table has an empty value in the -@code{Host} column, make sure there are one or more corresponding entries in -the @code{host} table specifying which hosts the @code{db} table entry -applies to. - -If you get the error when using the SQL commands @code{SELECT ... -INTO OUTFILE} or @code{LOAD DATA INFILE}, your entry in the @code{user} table -probably doesn't have the @strong{file} privilege enabled. - -@item -@cindex configuration files -@cindex environment variables -@tindex .my.cnf file -Remember that client programs will use connection parameters specified -in configuration files or environment variables. @xref{Environment -variables}. If a client seems to be sending the wrong default -connection parameters when you don't specify them on the command line, -check your environment and the @file{.my.cnf} file in your home -directory. You might also check the system-wide @strong{MySQL} -configuration files, though it is far less likely that client connection -parameters will be specified there. @xref{Option files}. If you get -@code{Access denied} when you run a client without any options, make -sure you haven't specified an old password in any of your option files! -@xref{Option files}. - -@item -If you make changes to the grant tables directly (using an @code{INSERT} or -@code{UPDATE} statement) and your changes seem to be ignored, remember -that you must issue a @code{FLUSH PRIVILEGES} statement or execute a -@code{mysqladmin flush-privileges} command to cause the server to re-read -the privilege tables. Otherwise your changes have no effect until the -next time the server is restarted. Remember that after you set the -@code{root} password with an @code{UPDATE} command, you won't need to -specify it until after you flush the privileges, because the server -won't know you've changed the password yet! - -@item -If you have access problems with a Perl, PHP, Python, or ODBC program, try to -connect to the server with @code{mysql -u user_name db_name} or @code{mysql --u user_name -pyour_pass db_name}. If you are able to connect using the -@code{mysql} client, there is a problem with your program and not with the -access privileges. (Note that there is no space between @code{-p} and the -password; you can also use the @code{--password=your_pass} syntax to specify -the password. If you use the @code{-p} option alone, @strong{MySQL} will -prompt you for the password.) - -@item -For testing, start the @code{mysqld} daemon with the -@code{--skip-grant-tables} option. Then you can change the @strong{MySQL} -grant tables and use the @code{mysqlaccess} script to check whether or not -your modifications have the desired effect. When you are satisfied with your -changes, execute @code{mysqladmin flush-privileges} to tell the @code{mysqld} -server to start using the new grant tables. @strong{Note:} Reloading the -grant tables overrides the @code{--skip-grant-tables} option. This allows -you to tell the server to begin using the grant tables again without bringing -it down and restarting it. - -@item -If everything else fails, start the @code{mysqld} daemon with a debugging -option (for example, @code{--debug=d,general,query}). This will print host and -user information about attempted connections, as well as information about -each command issued. @xref{Making trace files}. - -@item -If you have any other problems with the @strong{MySQL} grant tables and -feel you must post the problem to the mailing list, always provide a -dump of the @strong{MySQL} grant tables. You can dump the tables with -the @code{mysqldump mysql} command. As always, post your problem using -the @code{mysqlbug} script. @xref{Bug reports}. In some cases you may need -to restart @code{mysqld} with @code{--skip-grant-tables} to run -@code{mysqldump}. -@end itemize - -@node Reference, Table types, Privilege system, Top -@chapter MySQL Language Reference - -@menu -* Literals:: Literals: How to write strings and numbers -* Variables:: User variables -* Column types:: Column types -* Functions:: Functions -* CREATE DATABASE:: @code{CREATE DATABASE} syntax -* DROP DATABASE:: @code{DROP DATABASE} syntax -* CREATE TABLE:: @code{CREATE TABLE} syntax -* ALTER TABLE:: @code{ALTER TABLE} syntax -* RENAME TABLE:: @code{RENAME TABLE} syntax -* DROP TABLE:: @code{DROP TABLE} syntax -* OPTIMIZE TABLE:: @code{OPTIMIZE TABLE} syntax -* CHECK TABLE:: @code{CHECK TABLE} syntax -* BACKUP TABLE:: @code{BACKUP TABLE} syntax -* RESTORE TABLE:: @code{RESTORE TABLE} syntax -* ANALYZE TABLE:: @code{ANALYZE TABLE} syntax -* REPAIR TABLE:: @code{REPAIR TABLE} syntax -* DELETE:: @code{DELETE} syntax -* TRUNCATE:: @code{TRUNCATE} syntax -* SELECT:: @code{SELECT} syntax -* JOIN:: @code{JOIN} syntax -* INSERT:: @code{INSERT} syntax -* REPLACE:: @code{REPLACE} syntax -* LOAD DATA:: @code{LOAD DATA INFILE} syntax -* UPDATE:: @code{UPDATE} syntax -* USE:: @code{USE} syntax -* FLUSH:: @code{FLUSH} syntax (clearing caches) -* KILL:: @code{KILL} syntax -* SHOW:: @code{SHOW} syntax (Get information about tables, columns, ...) -* EXPLAIN:: @code{EXPLAIN} syntax (Get information about a @code{SELECT}) -* DESCRIBE:: @code{DESCRIBE} syntax (Get information about names of columns) -* COMMIT:: @code{BEGIN/COMMIT/ROLLBACK} syntax -* LOCK TABLES:: @code{LOCK TABLES/UNLOCK TABLES} syntax -* SET OPTION:: @code{SET OPTION} syntax -* SET TRANSACTION:: @code{SET TRANSACTION} syntax -* GRANT:: @code{GRANT} and @code{REVOKE} syntax -* CREATE INDEX:: @code{CREATE INDEX} syntax -* DROP INDEX:: @code{DROP INDEX} syntax -* Comments:: Comment syntax -* CREATE FUNCTION:: @code{CREATE FUNCTION} syntax -* Reserved words:: Is @strong{MySQL} picky about reserved words? -@end menu - -@strong{MySQL} has a very complex, but intuitive and easy to learn SQL -interface. This chapter describes the various commands, types, and functions -you will need to know in order to use @strong{MySQL} efficiently and -effectively. This chapter also serves as a reference to all functionality -included in @strong{MySQL}. In order to use this chapter effectively, you -may find it useful to refer to the various indexes. +You can also store the result in a file, then transfer the file to the +target machine and load the file into the database there. For example, +you can dump a database to a file on the source machine like this: -@cindex strings, defined -@cindex strings, escaping characters -@cindex literals -@cindex escape characters -@cindex backslash, escape character -@node Literals, Variables, Reference, Reference -@section Literals: How to Write Strings and Numbers +@example +shell> mysqldump --quick db_name | gzip > db_name.contents.gz +@end example -@menu -* String syntax:: Strings -* Number syntax:: Numbers -* Hexadecimal values:: Hexadecimal values -* NULL values:: @code{NULL} values -* Legal names:: Database, Table, Index, Column, and Alias Names -@end menu +(The file created in this example is compressed.) Transfer the file +containing the database contents to the target machine and run these commands +there: -This section describes the various ways to write strings and numbers in -@strong{MySQL}. It also covers the various nuances and ``gotchas'' that -you may run into when dealing with these basic types in @strong{MySQL}. +@example +shell> mysqladmin create db_name +shell> gunzip < db_name.contents.gz | mysql db_name +@end example -@node String syntax, Number syntax, Literals, Literals -@subsection Strings +@cindex @code{mysqldump} +@cindex @code{mysqlimport} +You can also use @code{mysqldump} and @code{mysqlimport} to accomplish +the database transfer. +For big tables, this is much faster than simply using @code{mysqldump}. +In the commands shown below, @code{DUMPDIR} represents the full pathname +of the directory you use to store the output from @code{mysqldump}. -A string is a sequence of characters, surrounded by either single quote -(@samp{'}) or double quote (@samp{"}) characters (only the single quote -if you run in ANSI mode). Examples: +First, create the directory for the output files and dump the database: @example -'a string' -"another string" +shell> mkdir DUMPDIR +shell> mysqldump --tab=DUMPDIR db_name @end example -Within a string, certain sequences have special meaning. Each of these -sequences begins with a backslash (@samp{\}), known as the @emph{escape -character}. @strong{MySQL} recognizes the following escape sequences: +Then transfer the files in the @code{DUMPDIR} directory to some corresponding +directory on the target machine and load the files into @strong{MySQL} +there: -@c these aren't really functions, but that's probably the most reasonable index -@table @code -@findex \0 (ASCII 0) -@findex NUL -@item \0 -An ASCII 0 (@code{NUL}) character. +@example +shell> mysqladmin create db_name # create database +shell> cat DUMPDIR/*.sql | mysql db_name # create tables in database +shell> mysqlimport db_name DUMPDIR/*.txt # load data into tables +@end example -@findex \' (single quote) -@findex single quote (\') -@item \' -A single quote (@samp{'}) character. +Also, don't forget to copy the @code{mysql} database, because that's where the +grant tables (@code{user}, @code{db}, @code{host}) are stored. You may have +to run commands as the @strong{MySQL} @code{root} user on the new machine +until you have the @code{mysql} database in place. -@findex \" (double quote) -@findex double quote (\") -@item \" -A double quote (@samp{"}) character. +After you import the @code{mysql} database on the new machine, execute +@code{mysqladmin flush-privileges} so that the server reloads the grant table +information. -@findex \b (backspace) -@findex backspace (\b) -@item \b -A backspace character. -@findex \n (newline) -@findex newline (\n) -@item \n -A newline character. +@node Privilege system, Reference, Installing, Top +@chapter The MySQL Access Privilege System +@cindex system, security +@cindex access privileges +@cindex privileges, access +@cindex security system +@cindex ACLs -@findex \r (carriage return) -@findex return (\r) -@findex carriage return (\r) -@item \r -A carriage return character. -@findex \t (tab) -@findex tab (\t) -@item \t -A tab character. +@strong{MySQL} has an advanced but non-standard security/privilege +system. This section describes how it works. -@findex \z (Control-Z) ASCII(26) -@findex (Control-Z) \z -@item \z -ASCII(26) (Control-Z). This character can be encoded to allow you to -go around the problem that ASCII(26) stands for END-OF-FILE on Windows. -(ASCII(26) will cause problems if you try to use -@code{mysql database < filename}). +@menu +* General security:: General security +* Security:: How to make @strong{MySQL} secure against crackers +* Privileges options:: +* What Privileges:: What the privilege system does +* User names:: @strong{MySQL} user names and passwords +* Connecting:: Connecting to the @strong{MySQL} server +* Password security:: Keeping your password secure +* Privileges provided:: Privileges provided by @strong{MySQL} +* Privileges:: How the privilege system works +* Connection access:: Access control, stage 1: Connection verification +* Request access:: Access control, stage 2: Request verification +* Privilege changes:: When privilege changes take effect +* Default privileges:: Setting up the initial @strong{MySQL} privileges +* Adding users:: Adding new users to @strong{MySQL} +* Passwords:: How to set up passwords +* Access denied:: Causes of @code{Access denied} errors +@end menu -@findex \\ (escape) -@findex escape (\\) -@item \\ -A backslash (@samp{\}) character. +@node General security, Security, Privilege system, Privilege system +@section General Security -@findex % (wild card character) -@findex Wild card character (%) -@item \% -A @samp{%} character. This is used to search for literal instances of -@samp{%} in contexts where @samp{%} would otherwise be interpreted -as a wild-card character. @xref{String comparison functions}. +Anyone using @strong{MySQL} on a computer connected to the Internet +should read this section to avoid the most common security mistakes. -@findex _ (wild card character) -@findex Wild card character (_) -@item \_ -A @samp{_} character. This is used to search for literal instances of -@samp{_} in contexts where @samp{_} would otherwise be interpreted -as a wild-card character. @xref{String comparison functions}. -@end table +In discussing security, we emphasize the necessity of fully protecting the +entire server host (not simply the @strong{MySQL} server) against all types +of applicable attacks: eavesdropping, altering, playback, and denial of +service. We do not cover all aspects of availability and fault tolerance +here. -Note that if you use @samp{\%} or @samp{\_} in some string contexts, these -will return the strings @samp{\%} and @samp{\_} and not @samp{%} and -@samp{_}. +@strong{MySQL} uses security based on Access Control Lists (ACLs) for all +connections, queries, and other operations that a user may attempt to +perform. There is also some support for SSL-encrypted connections +between @strong{MySQL} clients and servers. Many of the concepts +discussed here are not specific to @strong{MySQL} at all; the same +general ideas apply to almost all applications. -@cindex quotes, in strings -@noindent -There are several ways to include quotes within a string: +When running @strong{MySQL}, follow these guidelines whenever possible: @itemize @bullet @item -A @samp{'} inside a string quoted with @samp{'} may be written as @samp{''}. +DON'T EVER GIVE ANYONE (EXCEPT THE @strong{MySQL} ROOT USER) ACCESS TO THE +@code{user} TABLE IN THE @code{mysql} DATABASE! The encrypted password +is the real password in @strong{MySQL}. If you know the password listed in +the @code{user} table for a given user, you can easily log in as that +user if you have access to the host listed for that account. @item -A @samp{"} inside a string quoted with @samp{"} may be written as @samp{""}. +Learn the @strong{MySQL} access privilege system. The @code{GRANT} and +@code{REVOKE} commands are used for controlling access to @strong{MySQL}. Do +not grant any more privileges than necessary. Never grant privileges to all +hosts. +Checklist: +@itemize @minus @item -You can precede the quote character with an escape character (@samp{\}). - +Try @code{mysql -u root}. If you are able to connect successfully to the +server without being asked for a password, you have problems. Anyone +can connect to your @strong{MySQL} server as the @strong{MySQL} +@code{root} user with full privileges! +Review the @strong{MySQL} installation instructions, paying particular +attention to the item about setting a @code{root} password. @item -A @samp{'} inside a string quoted with @samp{"} needs no special treatment -and need not be doubled or escaped. In the same way, @samp{"} inside a -string quoted with @samp{'} needs no special treatment. +Use the command @code{SHOW GRANTS} and check to see who has access to +what. Remove those privileges that are not necessary using the @code{REVOKE} +command. @end itemize -The @code{SELECT} statements shown below demonstrate how quoting and -escaping work: - -@example -mysql> SELECT 'hello', '"hello"', '""hello""', 'hel''lo', '\'hello'; -+-------+---------+-----------+--------+--------+ -| hello | "hello" | ""hello"" | hel'lo | 'hello | -+-------+---------+-----------+--------+--------+ +@item +Do not keep any plain-text passwords in your database. When your +computer becomes compromised, the intruder can take the full list of +passwords and use them. Instead use @code{MD5()} or another one-way +hashing function. +@item +Do not choose passwords from dictionaries. There are special programs to +break them. Even passwords like ``xfish98'' are very bad. Much better is +``duag98'' which contains the same word ``fish'' but typed one key to the +left on a standard QWERTY keyboard. Another method is to use ``Mhall'' which +is taken from the first characters of each word in the sentence ``Mary had +a little lamb.'' This is easy to remember and type, but difficult to guess +for someone who does not know it. +@item +Invest in a firewall. This protects you from at least 50% of all types of +exploits in any software. Put @strong{MySQL} behind the firewall or in +a demilitarized zone (DMZ). -mysql> SELECT "hello", "'hello'", "''hello''", "hel""lo", "\"hello"; -+-------+---------+-----------+--------+--------+ -| hello | 'hello' | ''hello'' | hel"lo | "hello | -+-------+---------+-----------+--------+--------+ +Checklist: +@itemize @minus +@item +Try to scan your ports from the Internet using a tool such as +@code{nmap}. @strong{MySQL} uses port 3306 by default. This port should +be inaccessible from untrusted hosts. Another simple way to check whether +or not your @strong{MySQL} port is open is to try the following command +from some remote machine, where @code{server_host} is the hostname of +your @strong{MySQL} server: -mysql> SELECT "This\nIs\nFour\nlines"; -+--------------------+ -| This -Is -Four -lines | -+--------------------+ +@example +shell> telnet server_host 3306 @end example -@cindex quoting binary data +If you get a connection and some garbage characters, the port is +open, and should be closed on your firewall or router, unless you really +have a good reason to keep it open. If @code{telnet} just hangs or the +connection is refused, everything is OK; the port is blocked. +@end itemize -If you want to insert binary data into a @code{BLOB} column, the following -characters must be represented by escape sequences: -@table @code -@item NUL -ASCII 0. You should represent this by @samp{\0} (a backslash and an ASCII @samp{0} character). -@item \ -ASCII 92, backslash. Represent this by @samp{\\}. -@item ' -ASCII 39, single quote. Represent this by @samp{\'}. -@item " -ASCII 34, double quote. Represent this by @samp{\"}. -@end table - -@cindex quoting -@cindex @code{BLOB}, inserting binary data -@findex mysql_escape_string() -@findex DBI->quote -If you write C code, you can use the C API function -@code{mysql_escape_string()} to escape characters for the @code{INSERT} -statement. @xref{C API function overview}. In Perl, you can use the -@code{quote} method of the @code{DBI} package to convert special -characters to the proper escape sequences. @xref{Perl DBI Class, , Perl -@code{DBI} Class}. - -You should use an escape function on any string that might contain any of the -special characters listed above! +@item +Do not trust any data entered by your users. They can try to trick your +code by entering special or escaped character sequences in Web forms, +URLs, or whatever application you have built. Be sure that your +application remains secure if a user enters something like ``@code{; DROP +DATABASE mysql;}''. This is an extreme example, but large security leaks +and data loss may occur as a result of hackers using similar techniques, +if you do not prepare for them. -@cindex numbers -@cindex valid numbers, examples -@cindex integers -@cindex floats -@cindex negative values -@node Number syntax, Hexadecimal values, String syntax, Literals -@subsection Numbers +Also remember to check numeric data. A common mistake is to protect only +strings. Sometimes people think that if a database contains only publicly +available data that it need not be protected. This is incorrect. At least +denial-of-service type attacks can be performed on such +databases. The simplest way to protect from this type of attack is to use +apostrophes around the numeric constants: @code{SELECT * FROM table +WHERE ID='234'} rather than @code{SELECT * FROM table WHERE ID=234}. +@strong{MySQL} automatically converts this string to a number and +strips all non-numeric symbols from it. -Integers are represented as a sequence of digits. Floats use @samp{.} as a -decimal separator. Either type of number may be preceded by @samp{-} to -indicate a negative value. +Checklist: +@itemize @minus +@item +All Web applications: +@itemize @bullet +@item +Try to enter @samp{'} and @samp{"} in all your Web forms. If you get any kind +of @strong{MySQL} error, investigate the problem right away. +@item +Try to modify any dynamic URLs by adding @code{%22} (@samp{"}), @code{%23} +(@samp{#}), and @code{%27} (@samp{'}) in the URL. +@item +Try to modify datatypes in dynamic URLs from numeric ones to character +ones containing characters from previous examples. Your application +should be safe against this and similar attacks. +@item +Try to enter characters, spaces, and special symbols instead of numbers in +numeric fields. Your application should remove them before passing them to +@strong{MySQL} or your application should generate an error. Passing +unchecked values to @strong{MySQL} is very dangerous! +@item +Check data sizes before passing them to @strong{MySQL}. +@item +Consider having your application connect to the database using a +different user name than the one you use for administrative purposes. Do +not give your applications any more access privileges than they need. +@end itemize +@item +Users of PHP: +@itemize @bullet +@item Check out the @code{addslashes()} function. +As of PHP 4.0.3, a @code{mysql_escape_string()} function is available +that is based on the function of the same name in the @strong{MySQL} C API. +@end itemize +@item +Users of @strong{MySQL} C API: +@itemize @bullet +@item Check out the @code{mysql_escape_string()} API call. +@end itemize +@item +Users of @strong{MySQL}++: +@itemize @bullet +@item Check out the @code{escape} and @code{quote} modifiers for query streams. +@end itemize +@item +Users of Perl DBI: +@itemize @bullet +@item Check out the @code{quote()} method or use placeholders. +@end itemize +@item +Users of Java JDBC: +@itemize @bullet +@item Use a @code{PreparedStatement} object and placeholders. +@end itemize +@end itemize -Examples of valid integers: +@item +Do not transmit plain (unencrypted) data over the Internet. These data are +accessible to everyone who has the time and ability to intercept it and use +it for their own purposes. Instead, use an encrypted protocol such as SSL or +SSH. @strong{MySQL} supports internal SSL connections as of Version 3.23.9. +SSH port-forwarding can be used to create an encrypted (and compressed) +tunnel for the communication. +@item +Learn to use the @code{tcpdump} and @code{strings} utilities. For most cases, +you can check whether or not @strong{MySQL} data streams are unencrypted +by issuing a command like the following: @example -1221 -0 --32 +shell> tcpdump -l -i eth0 -w - src or dst port 3306 | strings @end example -Examples of valid floating-point numbers: +(This works under Linux and should work with small modifications under +other systems). Warning: If you do not see data this doesn't always +actually mean that it is encrypted. If you need high security, you should +consult with a security expert. +@end itemize -@example -294.42 --32032.6809e+10 -148.00 -@end example +@node Security, Privileges options, General security, Privilege system +@section How to Make MySQL Secure Against Crackers +@cindex crackers, security against +@cindex security, against crackers -An integer may be used in a floating-point context; it is interpreted -as the equivalent floating-point number. +When you connect to a @strong{MySQL} server, you normally should use a +password. The password is not transmitted in clear text over the +connection, however the encryption algorithm is not very strong, and +with some effort a clever attacker can crack the password if he is able +to sniff the traffic between the client and the server. If the +connection between the client and the server goes through an untrusted +network, you should use an SSH tunnel to encrypt the +communication. -@tindex hexadecimal values -@node Hexadecimal values, NULL values, Number syntax, Literals -@subsection Hexadecimal Values +All other information is transferred as text that can be read by anyone +who is able to watch the connection. If you are concerned about this, +you can use the compressed protocol (in @strong{MySQL} Version 3.22 and above) +to make things much harder. To make things even more secure you should use +@code{ssh}. You can find an open source @code{ssh} client at +@uref{http://www.openssh.org}, and a commercial @code{ssh} client at +@uref{http://www.ssh.com}. With this, you can get an encrypted TCP/IP +connection between a @strong{MySQL} server and a @strong{MySQL} client. -@strong{MySQL} supports hexadecimal values. In number context these act -like an integer (64-bit precision). In string context these act like a binary -string where each pair of hex digits is converted to a character: +To make a @strong{MySQL} system secure, you should strongly consider the +following suggestions: + +@itemize @bullet +@item +Use passwords for all @strong{MySQL} users. Remember that anyone can log in +as any other person as simply as @code{mysql -u other_user db_name} if +@code{other_user} has no password. It is common behavior with client/server +applications that the client may specify any user name. You can change the +password of all users by editing the @code{mysql_install_db} script before +you run it, or only the password for the @strong{MySQL} @code{root} user like +this: @example -mysql> SELECT 0xa+0; - -> 10 -mysql> select 0x5061756c; - -> Paul +shell> mysql -u root mysql +mysql> UPDATE user SET Password=PASSWORD('new_password') + WHERE user='root'; +mysql> FLUSH PRIVILEGES; @end example -Hexadecimal strings are often used by ODBC to give values for BLOB columns. - -@tindex NULL value -@node NULL values, Legal names, Hexadecimal values, Literals -@subsection @code{NULL} Values - -The @code{NULL} value means ``no data'' and is different from values such -as @code{0} for numeric types or the empty string for string types. -@xref{Problems with NULL, , Problems with @code{NULL}}. - -@code{NULL} may be represented by @code{\N} when using the text file import -or export formats (@code{LOAD DATA INFILE}, @code{SELECT ... INTO OUTFILE}). -@xref{LOAD DATA, , @code{LOAD DATA}}. - -@cindex names -@cindex legal names -@cindex databases, names -@cindex tables, names -@cindex indexes, names -@cindex columns, names -@cindex aliases, names -@node Legal names, , NULL values, Literals -@subsection Database, Table, Index, Column, and Alias Names - -@menu -* Name case sensitivity:: Case sensitivity in names -@end menu +@item +Don't run the @strong{MySQL} daemon as the Unix @code{root} user. This is +very dangerous, because any user with @code{FILE} privileges will be able +to create files as @code{root} (for example, @code{~root/.bashrc}). To +prevent this, @code{mysqld} will refuse to run as @code{root} unless it +is specified directly using a @code{--user=root} option. -Database, table, index, column, and alias names all follow the same rules in -@strong{MySQL}. +@code{mysqld} can be run as an ordinary unprivileged user instead. +You can also create a new Unix user @code{mysql} to make everything +even more secure. If you run @code{mysqld} as another Unix user, +you don't need to change the @code{root} user name in the @code{user} +table, because @strong{MySQL} user names have nothing to do with Unix +user names. To start @code{mysqld} as another Unix user, add a @code{user} +line that specifies the user name to the @code{[mysqld]} group of the +@file{/etc/my.cnf} option file or the @file{my.cnf} option file in the +server's data directory. For example: -@tindex identifiers, quoting -@tindex quoting of identifiers -@tindex ` -@tindex " -Note that the rules changed starting with @strong{MySQL} Version 3.23.6 when we -introduced quoting of identifiers (database, table, and column names) -with @samp{`}. @samp{"} will also work to quote identifiers if you run -in ANSI mode. @xref{ANSI mode}. +@example +[mysqld] +user=mysql +@end example -@multitable @columnfractions .15 .15 .70 -@item @strong{Identifier} @tab @strong{Max length} @tab @strong{Allowed characters} -@item Database @tab 64 @tab Any character that is allowed in a directory name except @samp{/} or @samp{.}. -@item Table @tab 64 @tab Any character that is allowed in a file name, except @samp{/} or @samp{.}. -@item Column @tab 64 @tab All characters. -@item Alias @tab 255 @tab All characters. -@end multitable +This will cause the server to start as the designated user whether you +start it manually or by using @code{safe_mysqld} or @code{mysql.server}. +For more details, see @ref{Changing MySQL user, , Changing @strong{MySQL} +user}. -Note that in addition to the above, you can't have ASCII(0) or ASCII(255) or -the quoting character in an identifier. +@item +Don't support symlinks to tables (This can be disabled with the +@code{--skip-symlink} option. This is especially important if you run +@code{mysqld} as root as anyone that has write access to the mysqld data +directories could then delete any file in the system! +@xref{Symbolic links to tables}. -Note that if the identifier is a restricted word or contains special characters -you must always quote it with @code{`} when you use it: +@item +Check that the Unix user that @code{mysqld} runs as is the only user with +read/write privileges in the database directories. -@example -SELECT * from `select` where `select`.id > 100; -@end example +@item +Don't give the @strong{process} privilege to all users. The output of +@code{mysqladmin processlist} shows the text of the currently executing +queries, so any user who is allowed to execute that command might be able to +see if another user issues an @code{UPDATE user SET +password=PASSWORD('not_secure')} query. -In previous versions of @strong{MySQL}, the name rules are as follows: +@code{mysqld} reserves an extra connection for users who have the +@strong{process} privilege, so that a @strong{MySQL} @code{root} user can log +in and check things even if all normal connections are in use. -@itemize @bullet @item -A name may consist of alphanumeric characters from the current character set -and also @samp{_} and @samp{$}. The default character set is ISO-8859-1 -Latin1; this may be changed with the @code{--default-character-set} option -to @code{mysqld}. -@xref{Character sets}. +Don't give the @strong{file} privilege to all users. Any user that has this +privilege can write a file anywhere in the file system with the privileges of +the @code{mysqld} daemon! To make this a bit safer, all files generated with +@code{SELECT ... INTO OUTFILE} are readable to everyone, and you cannot +overwrite existing files. + +@tindex /etc/passwd +The @strong{file} privilege may also be used to read any file accessible +to the Unix user that the server runs as. This could be abused, for example, +by using @code{LOAD DATA} to load @file{/etc/passwd} into a table, which +can then be read with @code{SELECT}. @item -A name may start with any character that is legal in a name. In particular, -a name may start with a number (this differs from many other database -systems!). However, a name cannot consist @emph{only} of numbers. +If you don't trust your DNS, you should use IP numbers instead of hostnames +in the grant tables. In principle, the @code{--secure} option to +@code{mysqld} should make hostnames safe. In any case, you should be very +careful about creating grant table entries using hostname values that +contain wild cards! @item -You cannot use the @samp{.} character in names because it is used to extend the -format by which you can refer to columns (see immediately below). +If you want to restrict the number of connections for a single user, you +can do this by setting the @code{max_user_connections} variable in +@code{mysqld}. @end itemize -It is recommended that you do not use names like @code{1e}, because -an expression like @code{1e+1} is ambiguous. It may be interpreted as the -expression @code{1e + 1} or as the number @code{1e+1}. +@node Privileges options, What Privileges, Security, Privilege system +@section Startup Options for @code{mysqld} Concerning Security -In @strong{MySQL} you can refer to a column using any of the following forms: +The following @code{mysqld} options affect networking security: -@multitable @columnfractions .35 .65 -@item @strong{Column reference} @tab @strong{Meaning} -@item @code{col_name} @tab Column @code{col_name} -from whichever table used in the query contains a column of that name. -@item @code{tbl_name.col_name} @tab Column @code{col_name} from table -@code{tbl_name} of the current database. -@item @code{db_name.tbl_name.col_name} @tab Column @code{col_name} from table -@code{tbl_name} of the database @code{db_name}. This form is available in -@strong{MySQL} Version 3.22 or later. -@item -@code{`column_name`} @tab A column that is a keyword or contains special characters. -@end multitable +@table @code +@item --secure +IP numbers returned by the @code{gethostbyname()} system call are +checked to make sure they resolve back to the original hostname. This +makes it harder for someone on the outside to get access by pretending +to be another host. This option also adds some sanity checks of +hostnames. The option is turned off by default in @strong{MySQL} Version +3.21 because sometimes it takes a long time to perform backward resolutions. +@strong{MySQL} Version 3.22 caches hostnames and has this option enabled by +default. -You need not specify a @code{tbl_name} or @code{db_name.tbl_name} prefix for -a column reference in a statement unless the reference would be ambiguous. -For example, suppose tables @code{t1} and @code{t2} each contain a column -@code{c}, and you retrieve @code{c} in a @code{SELECT} statement that uses -both @code{t1} and @code{t2}. In this case, @code{c} is ambiguous because it -is not unique among the tables used in the statement, so you must indicate -which table you mean by writing @code{t1.c} or @code{t2.c}. Similarly, if -you are retrieving from a table @code{t} in database @code{db1} and from a -table @code{t} in database @code{db2}, you must refer to columns in those -tables as @code{db1.t.col_name} and @code{db2.t.col_name}. +@item --skip-grant-tables +This option causes the server not to use the privilege system at all. This +gives everyone @emph{full access} to all databases! (You can tell a running +server to start using the grant tables again by executing @code{mysqladmin +flush-privileges} or @code{mysqladmin reload}.) -@cindex ODBC compatibility -@cindex compatibility, with ODBC -The syntax @code{.tbl_name} means the table @code{tbl_name} in the current -database. This syntax is accepted for ODBC compatibility, because some ODBC -programs prefix table names with a @samp{.} character. +@item --skip-name-resolve +Hostnames are not resolved. All @code{Host} column values in the grant +tables must be IP numbers or @code{localhost}. -@cindex names, case-sensitivity -@cindex case-sensitivity, in names -@node Name case sensitivity, , Legal names, Legal names -@subsubsection Case Sensitivity in Names -@cindex database names, case sensitivity -@cindex table names, case sensitivity -@cindex column names, case sensitivity -@cindex alias names, case sensitivity +@item --skip-networking +Don't allow TCP/IP connections over the network. All connections to +@code{mysqld} must be made via Unix sockets. This option is unsuitable for +systems that use MIT-pthreads, because the MIT-pthreads package doesn't +support Unix sockets. -In @strong{MySQL}, databases and tables correspond to directories and files -within those directories. Consequently, the case sensitivity of the -underlying operating system determines the case sensitivity of database and -table names. This means database and table names are case sensitive in Unix -and case insensitive in Windows. @xref{Extensions to ANSI}. +@item --skip-show-database +With this option, the +@code{SHOW DATABASES} statement doesn't return anything. -@strong{NOTE:} Although database and table names are case insensitive for -Windows, you should not refer to a given database or table using different -cases within the same query. The following query would not work because it -refers to a table both as @code{my_table} and as @code{MY_TABLE}: +@item --safe-show-database +With this option, +@code{SHOW DATABASES} returns only those databases for which the user has +some kind of privilege. -@example -mysql> SELECT * FROM my_table WHERE MY_TABLE.col=1; -@end example +@end table -Column names are case insensitive in all cases. +@node What Privileges, User names, Privileges options, Privilege system +@section What the Privilege System Does +@cindex system, privilege +@cindex privilege system +@cindex passwords, security -Aliases on tables are case sensitive. The following query would not work -because it refers to the alias both as @code{a} and as @code{A}: +The primary function of the @strong{MySQL} privilege system is to +authenticate a user connecting from a given host, and to associate that user +with privileges on a database such as +@strong{select}, @strong{insert}, @strong{update} and @strong{delete}. -@example -mysql> SELECT col_name FROM tbl_name AS a - WHERE a.col_name = 1 OR A.col_name = 2; -@end example +Additional functionality includes the ability to have an anonymous user and +to grant privileges for @strong{MySQL}-specific functions such as @code{LOAD +DATA INFILE} and administrative operations. -Aliases on columns are case insensitive. +@node User names, Connecting, What Privileges, Privilege system +@section MySQL User Names and Passwords +@cindex user names, and passwords +@cindex passwords, for users -If you have a problem remembering the used cases for a table names, -adopt a consistent convention, such as always creating databases and -tables using lowercase names. +There are several distinctions between the way user names and passwords are +used by @strong{MySQL} and the way they are used by Unix or Windows: -One way to avoid this problem is to start @code{mysqld} with @code{-O -lower_case_table_names=1}. By default this option is 1 on Windows and 0 on -Unix. +@itemize @bullet +@item +User names, as used by @strong{MySQL} for authentication purposes, have +nothing to do with Unix user names (login names) or Windows user names. Most +@strong{MySQL} clients by default try to log in using the current Unix user +name as the @strong{MySQL} user name, but that is for convenience only. +Client programs allow a different name to be specified with the @code{-u} or +@code{--user} options. This means that you can't make a database secure in +any way unless all @strong{MySQL} user names have passwords. Anyone may +attempt to connect to the server using any name, and they will succeed if +they specify any name that doesn't have a password. -If @code{lower_case_table_names} is 1 @strong{MySQL} will convert all -table names to lower case on storage and lookup. Note that if you -change this option, you need to first convert your old table names to -lower case before starting @code{mysqld}. +@item +@strong{MySQL} user names can be up to 16 characters long; Unix user names +typically are limited to 8 characters. -@cindex variables, user -@cindex user variables -@cindex names, variables -@node Variables, Column types, Literals, Reference -@section User Variables +@item +@strong{MySQL} passwords have nothing to do with Unix passwords. There is no +necessary connection between the password you use to log in to a Unix machine +and the password you use to access a database on that machine. -@strong{MySQL} supports thread-specific variables with the -@code{@@variablename} syntax. A variable name may consist of -alphanumeric characters from the current character set and also -@samp{_}, @samp{$}, and @samp{.} . The default character set is -ISO-8859-1 Latin1; this may be changed with the -@code{--default-character-set} option to @code{mysqld}. @xref{Character -sets}. +@item +@strong{MySQL} encrypts passwords using a different algorithm than the +one used during the Unix login process. See the descriptions of the +@code{PASSWORD()} and @code{ENCRYPT()} functions in @ref{Miscellaneous +functions}. Note that even if the password is stored 'scrambled', and +knowing your 'scrambled' password is enough to be able to connect to +the @strong{MySQL} server! +@end itemize -Variables don't have to be initialized. They contain @code{NULL} by default -and can store an integer, real, or string value. All variables for -a thread are automatically freed when the thread exits. +@strong{MySQL} users and they privileges are normally created with the +@code{GRANT} command. @xref{GRANT}. -You can set a variable with the @code{SET} syntax: +When you login to a @strong{MySQL} server with a command line client you +should specify the password with @code{--password=your-password}. +@xref{Connecting}. @example -SET @@variable= @{ integer expression | real expression | string expression @} -[,@@variable= ...]. +mysql --user=monty --password=guess database_name @end example -You can also set a variable in an expression with the @code{@@variable:=expr} -syntax: +If you want the client to prompt for a password, you should use +@code{--password} without any argument @example -select @@t1:=(@@t2:=1)+@@t3:=4,@@t1,@@t2,@@t3; -+----------------------+------+------+------+ -| @@t1:=(@@t2:=1)+@@t3:=4 | @@t1 | @@t2 | @@t3 | -+----------------------+------+------+------+ -| 5 | 5 | 1 | 4 | -+----------------------+------+------+------+ +mysql --user=monty --password database_name @end example -(We had to use the @code{:=} syntax here, because @code{=} was reserved for -comparisons.) +or the short form: -User variables may be used where expressions are allowed. Note that -this does not currently include use in contexts where a number is explicitly -required, such as in the @code{LIMIT} clause of a @code{SELECT} statement, -or the @code{IGNORE number LINES} clause of a @code{LOAD DATA} statement. +@example +mysql -u monty -p database_name +@end example -@strong{NOTE:} In a @code{SELECT} statement, each expression is only -evaluated when it's sent to the client. This means that in the @code{HAVING}, -@code{GROUP BY}, or @code{ORDER BY} clause, you can't refer to an expression -that involves variables that are set in the @code{SELECT} part. For example, -the following statement will NOT work as expected: +Note that in the last example the password is @strong{NOT} 'database_name'. + +If you want to use the @code{-p} option to supply a password you should do like this: @example -SELECT (@@aa:=id) AS a, (@@aa+3) AS b FROM table_name HAVING b=5; +mysql -u monty -pguess database_name @end example -The reason is that @code{@@aa} will not contain the value of the current -row, but the value of @code{id} for the previous accepted row. +On some system the library call that @strong{MySQL} uses to prompt for a +password will automatically cut the password to 8 characters. Internally +@strong{MySQL} doesn't have any limit for the length of the password. -@cindex columns, types -@cindex types, columns -@node Column types, Functions, Variables, Reference -@section Column Types - -@strong{MySQL} supports a number of column types, which may be grouped into -three categories: numeric types, date and time types, and string (character) -types. This section first gives an overview of the types available and -summarizes the storage requirements for each column type, then provides a -more detailed description of the properties of the types in each category. -The overview is intentionally brief. The more detailed descriptions should -be consulted for additional information about particular column types, such -as the allowable formats in which you can specify values. - -The column types supported by @strong{MySQL} are listed below. -The following code letters are used in the descriptions: - -@cindex display size -@cindex sizes, display -@cindex digits -@cindex decimal point -@cindex brackets, square -@cindex square brackets -@table @code -@item M -Indicates the maximum display size. The maximum legal display size is 255. +@node Connecting, Password security, User names, Privilege system +@section Connecting to the MySQL Server +@cindex connecting, to the server +@cindex default hostname +@cindex hostname, default +@cindex server, connecting -@item D -Applies to floating-point types and indicates the number of digits -following the decimal point. The maximum possible value is 30, but -should be no greater than @code{M}-2. -@end table +@strong{MySQL} client programs generally require that you specify connection +parameters when you want to access a @strong{MySQL} server: the host you want +to connect to, your user name, and your password. For example, the +@code{mysql} client can be started like this (optional arguments are enclosed +between @samp{[} and @samp{]}): -Square brackets (@samp{[} and @samp{]}) indicate parts of type specifiers -that are optional. +@example +shell> mysql [-h host_name] [-u user_name] [-pyour_pass] +@end example -@tindex Types +Alternate forms of the @code{-h}, @code{-u}, and @code{-p} options are +@code{--host=host_name}, @code{--user=user_name}, and +@code{--password=your_pass}. Note that there is @emph{no space} between +@code{-p} or @code{--password=} and the password following it. -@c The @w{-number} stuff keeps a linebreak from occurring between -@c the - and number. +@strong{NOTE:} Specifying a password on the command line is not secure! +Any user on your system may then find out your password by typing a command +like: @code{ps auxww}. @xref{Option files}. -Note that if you specify @code{ZEROFILL} for a column, @strong{MySQL} will -automatically add the @code{UNSIGNED} attribute to the column. +@code{mysql} uses default values for connection parameters that are missing +from the command line: -@table @code -@tindex TINYINT -@item TINYINT[(M)] [UNSIGNED] [ZEROFILL] +@itemize @bullet +@item +The default hostname is @code{localhost}. -A very small integer. The signed range is @code{-128} to @code{127}. The -unsigned range is @code{0} to @code{255}. +@item +The default user name is your Unix login name. -@tindex SMALLINT -@item SMALLINT[(M)] [UNSIGNED] [ZEROFILL] +@item +No password is supplied if @code{-p} is missing. +@end itemize -A small integer. The signed range is @code{-32768} to @code{32767}. The -unsigned range is @code{0} to @code{65535}. +Thus, for a Unix user @code{joe}, the following commands are equivalent: -@tindex MEDIUMINT -@item MEDIUMINT[(M)] [UNSIGNED] [ZEROFILL] +@example +shell> mysql -h localhost -u joe +shell> mysql -h localhost +shell> mysql -u joe +shell> mysql +@end example -A medium-size integer. The signed range is @code{-8388608} to -@code{8388607}. The unsigned range is @code{0} to @code{16777215}. +Other @strong{MySQL} clients behave similarly. -@tindex INT -@item INT[(M)] [UNSIGNED] [ZEROFILL] +On Unix systems, you can specify different default values to be used when you +make a connection, so that you need not enter them on the command line each +time you invoke a client program. This can be done in a couple of ways: -A normal-size integer. The signed range is @code{-2147483648} to -@code{2147483647}. The unsigned range is @code{0} to @code{4294967295}. +@itemize @bullet +@item +@tindex .my.cnf file +You can specify connection parameters in the @code{[client]} section of the +@file{.my.cnf} configuration file in your home directory. The relevant +section of the file might look like this: -@tindex INTEGER -@item INTEGER[(M)] [UNSIGNED] [ZEROFILL] +@example +[client] +host=host_name +user=user_name +password=your_pass +@end example -This is a synonym for @code{INT}. +@xref{Option files}. -@tindex BIGINT -@item BIGINT[(M)] [UNSIGNED] [ZEROFILL] +@item +@tindex MYSQL_HOST environment variable +@tindex Environment variable, MYSQL_HOST +@tindex MYSQL_PWD environment variable +@tindex Environment variable, MYSQL_PWD +@tindex USER environment variable +@tindex Environment variable, USER +You can specify connection parameters using environment variables. The +host can be specified for @code{mysql} using @code{MYSQL_HOST}. The +@strong{MySQL} user name can be specified using @code{USER} (this is for +Windows only). The password can be specified using @code{MYSQL_PWD} +(but this is insecure; see the next section). @xref{Environment variables}. +@end itemize -A large integer. The signed range is @code{-9223372036854775808} to -@code{9223372036854775807}. The unsigned range is @code{0} to -@code{18446744073709551615}. +@node Password security, Privileges provided, Connecting, Privilege system +@section Keeping Your Password Secure -Some things you should be aware about @code{BIGINT} columns: +It is inadvisable to specify your password in a way that exposes it to +discovery by other users. The methods you can use to specify your password +when you run client programs are listed below, along with an assessment of +the risks of each method: @itemize @bullet @item -@cindex rounding errors -As all arithmetic is done using signed @code{BIGINT} or @code{DOUBLE} -values, so you shouldn't use unsigned big integers larger than -@code{9223372036854775807} (63 bits) except with bit functions! If you -do that, some of the last digits in the result may be wrong because of -rounding errors when converting the @code{BIGINT} to a @code{DOUBLE}. +Never give a normal user access to the @code{mysql.user} table. Knowing +the encrypted password for a user makes it possible to login as this +user. The passwords are only scrambled so that one shouldn't be able to +see the real password you used (if you happen to use a similar password +with your other applications). + @item -You can always store an exact integer value in a @code{BIGINT} column by -storing it as a string, as there is in this case there will be no -intermediate double representation. +Use a @code{-pyour_pass} or @code{--password=your_pass} option on the command +line. This is convenient but insecure, because your password becomes visible +to system status programs (such as @code{ps}) that may be invoked by other +users to display command lines. (@strong{MySQL} clients typically overwrite +the command-line argument with zeroes during their initialization sequence, +but there is still a brief interval during which the value is visible.) + @item -@samp{-}, @samp{+}, and @samp{*} will use @code{BIGINT} arithmetic when -both arguments are @code{INTEGER} values! This means that if you -multiply two big integers (or results from functions that return -integers) you may get unexpected results when the result is larger than -@code{9223372036854775807}. -@end itemize +Use a @code{-p} or @code{--password} option (with no @code{your_pass} value +specified). In this case, the client program solicits the password from +the terminal: +@findex -p option +@findex -password option -@cindex floating-point number -@tindex FLOAT -@tindex FLOAT(precision) -@item FLOAT(precision) [ZEROFILL] +@example +shell> mysql -u user_name -p +Enter password: ******** +@end example -A floating-point number. Cannot be unsigned. @code{precision} can be -@code{<=24} for a single-precision floating-point number and between 25 -and 53 for a double-precision floating-point number. These types are like -the @code{FLOAT} and @code{DOUBLE} types described immediately below. -@code{FLOAT(X)} has the same range as the corresponding @code{FLOAT} and -@code{DOUBLE} types, but the display size and number of decimals is undefined. +The @samp{*} characters represent your password. -In @strong{MySQL} Version 3.23, this is a true floating-point value. In -earlier @strong{MySQL} versions, @code{FLOAT(precision)} always has 2 decimals. +It is more secure to enter your password this way than to specify it on the +command line because it is not visible to other users. However, this method +of entering a password is suitable only for programs that you run +interactively. If you want to invoke a client from a script that runs +non-interactively, there is no opportunity to enter the password from the +terminal. On some systems, you may even find that the first line of your +script is read and interpreted (incorrectly) as your password! -Note that using @code{FLOAT} may give you some unexpected problems as -all calculation in @strong{MySQL} is done with double precision. -@xref{No matching rows}. +@item +@tindex .my.cnf file +Store your password in a configuration file. For example, you can list your +password in the @code{[client]} section of the @file{.my.cnf} file in your +home directory: -@cindex ODBC compatibility -@cindex compatibility, with ODBC -This syntax is provided for ODBC compatibility. +@example +[client] +password=your_pass +@end example -@tindex FLOAT -@tindex FLOAT(M,D) -@item FLOAT[(M,D)] [ZEROFILL] +If you store your password in @file{.my.cnf}, the file should not be group or +world readable or writable. Make sure the file's access mode is @code{400} +or @code{600}. -A small (single-precision) floating-point number. Cannot be unsigned. -Allowable values are @code{@w{-3.402823466E+38}} to -@code{@w{-1.175494351E-38}}, @code{0}, and @code{@w{1.175494351E-38}} to -@code{3.402823466E+38}. The M is the display width and D is the -number of decimals. @code{FLOAT} without an argument or with an argument of -<= 24 stands for a single-precision floating-point number. +@xref{Option files}. +@item +You can store your password in the @code{MYSQL_PWD} environment variable, but +this method must be considered extremely insecure and should not be used. +Some versions of @code{ps} include an option to display the environment of +running processes; your password will be in plain sight for all to see if +you set @code{MYSQL_PWD}. Even on systems without such a version of +@code{ps}, it is unwise to assume there is no other method to observe process +environments. @xref{Environment variables}. +@end itemize -@tindex DOUBLE -@tindex FLOAT(precision) -@item DOUBLE[(M,D)] [ZEROFILL] +All in all, the safest methods are to have the client program prompt for the +password or to specify the password in a properly protected @file{.my.cnf} +file. -A normal-size (double-precision) floating-point number. Cannot be -unsigned. Allowable values are @code{@w{-1.7976931348623157E+308}} to -@code{@w{-2.2250738585072014E-308}}, @code{0}, and -@code{2.2250738585072014E-308} to @code{1.7976931348623157E+308}. The M -is the display width and D is the number of decimals. @code{DOUBLE} -without an argument or @code{FLOAT(X)} where 25 <= X <= 53 stands for a -double-precision floating-point number. +@node Privileges provided, Privileges, Password security, Privilege system +@section Privileges Provided by MySQL +@cindex privilege information, location -@tindex DOUBLE PRECISION -@tindex REAL -@item DOUBLE PRECISION[(M,D)] [ZEROFILL] -@itemx REAL[(M,D)] [ZEROFILL] +Information about user privileges is stored in the @code{user}, @code{db}, +@code{host}, @code{tables_priv}, and @code{columns_priv} tables in the +@code{mysql} database (that is, in the database named @code{mysql}). The +@strong{MySQL} server reads the contents of these tables when it starts up +and under the circumstances indicated in @ref{Privilege changes}. -These are synonyms for @code{DOUBLE}. +The names used in this manual to refer to the privileges provided by +@strong{MySQL} are shown below, along with the table column name associated +with each privilege in the grant tables and the context in which the +privilege applies: -@tindex DECIMAL -@item DECIMAL[(M[,D])] [ZEROFILL] +@multitable @columnfractions .15 .25 .6 +@item @strong{Privilege} @tab @strong{Column} @tab @strong{Context} +@item @strong{select} @tab @code{Select_priv} @tab tables +@item @strong{insert} @tab @code{Insert_priv} @tab tables +@item @strong{update} @tab @code{Update_priv} @tab tables +@item @strong{delete} @tab @code{Delete_priv} @tab tables +@item @strong{index} @tab @code{Index_priv} @tab tables +@item @strong{alter} @tab @code{Alter_priv} @tab tables +@item @strong{create} @tab @code{Create_priv} @tab databases, tables, or indexes +@item @strong{drop} @tab @code{Drop_priv} @tab databases or tables +@item @strong{grant} @tab @code{Grant_priv} @tab databases or tables +@item @strong{references} @tab @code{References_priv} @tab databases or tables +@item @strong{reload} @tab @code{Reload_priv} @tab server administration +@item @strong{shutdown} @tab @code{Shutdown_priv} @tab server administration +@item @strong{process} @tab @code{Process_priv} @tab server administration +@item @strong{file} @tab @code{File_priv} @tab file access on server +@end multitable -An unpacked floating-point number. Cannot be unsigned. Behaves like a -@code{CHAR} column: ``unpacked'' means the number is stored as a string, -using one character for each digit of the value. The decimal point and, -for negative numbers, the @samp{-} sign, are not counted in M (but space -for these are reserved). If @code{D} is 0, values will have no decimal -point or fractional part. The maximum range of @code{DECIMAL} values is -the same as for @code{DOUBLE}, but the actual range for a given -@code{DECIMAL} column may be constrained by the choice of @code{M} and -@code{D}. +The @strong{select}, @strong{insert}, @strong{update}, and @strong{delete} +privileges allow you to perform operations on rows in existing tables in +a database. -If @code{D} is left out it's set to 0. If @code{M} is left out it's set to 10. +@code{SELECT} statements require the @strong{select} privilege only if they +actually retrieve rows from a table. You can execute certain @code{SELECT} +statements even without permission to access any of the databases on the +server. For example, you could use the @code{mysql} client as a simple +calculator: -Note that in @strong{MySQL} Version 3.22 the @code{M} argument had to -includes the space needed for the sign and the decimal point. +@example +mysql> SELECT 1+1; +mysql> SELECT PI()*2; +@end example -@tindex NUMERIC -@item NUMERIC(M,D) [ZEROFILL] +The @strong{index} privilege allows you to create or drop (remove) indexes. -This is a synonym for @code{DECIMAL}. +The @strong{alter} privilege allows you to use @code{ALTER TABLE}. -@tindex DATE -@item DATE +The @strong{create} and @strong{drop} privileges allow you to create new +databases and tables, or to drop (remove) existing databases and tables. -A date. The supported range is @code{'1000-01-01'} to @code{'9999-12-31'}. -@strong{MySQL} displays @code{DATE} values in @code{'YYYY-MM-DD'} format, but -allows you to assign values to @code{DATE} columns using either strings or -numbers. @xref{DATETIME}. +Note that if you grant the @strong{drop} privilege for the @code{mysql} +database to a user, that user can drop the database in which the +@strong{MySQL} access privileges are stored! -@tindex DATETIME -@item DATETIME +The @strong{grant} privilege allows you to give to other users those +privileges you yourself possess. -A date and time combination. The supported range is @code{'1000-01-01 -00:00:00'} to @code{'9999-12-31 23:59:59'}. @strong{MySQL} displays -@code{DATETIME} values in @code{'YYYY-MM-DD HH:MM:SS'} format, but allows you -to assign values to @code{DATETIME} columns using either strings or numbers. -@xref{DATETIME}. +The @strong{file} privilege gives you permission to read and write files on +the server using the @code{LOAD DATA INFILE} and @code{SELECT ... INTO +OUTFILE} statements. Any user to whom this privilege is granted can read or +write any file that the @strong{MySQL} server can read or write. -@tindex TIMESTAMP -@item TIMESTAMP[(M)] +The remaining privileges are used for administrative operations, which are +performed using the @code{mysqladmin} program. The table below shows which +@code{mysqladmin} commands each administrative privilege allows you to +execute: -A timestamp. The range is @code{'1970-01-01 00:00:00'} to sometime in the -year @code{2037}. @strong{MySQL} displays @code{TIMESTAMP} values in -@code{YYYYMMDDHHMMSS}, @code{YYMMDDHHMMSS}, @code{YYYYMMDD}, or @code{YYMMDD} -format, depending on whether @code{M} is @code{14} (or missing), @code{12}, -@code{8}, or @code{6}, but allows you to assign values to @code{TIMESTAMP} -columns using either strings or numbers. A @code{TIMESTAMP} column is useful -for recording the date and time of an @code{INSERT} or @code{UPDATE} -operation because it is automatically set to the date and time of the most -recent operation if you don't give it a value yourself. You can also set it -to the current date and time by assigning it a @code{NULL} value. @xref{Date -and time types}. +@multitable @columnfractions .15 .85 +@item @strong{Privilege} @tab @strong{Commands permitted to privilege holders} +@item @strong{reload} @tab @code{reload}, @code{refresh}, +@code{flush-privileges}, @code{flush-hosts}, @code{flush-logs}, and +@code{flush-tables} +@item @strong{shutdown} @tab @code{shutdown} +@item @strong{process} @tab @code{processlist}, @code{kill} +@end multitable -A @code{TIMESTAMP} is always stored in 4 bytes. The @code{M} argument only -affects how the @code{TIMESTAMP} column is displayed. +The @code{reload} command tells the server to re-read the grant tables. The +@code{refresh} command flushes all tables and opens and closes the log +files. @code{flush-privileges} is a synonym for @code{reload}. The other +@code{flush-*} commands perform functions similar to @code{refresh} but are +more limited in scope, and may be preferable in some instances. For example, +if you want to flush just the log files, @code{flush-logs} is a better choice +than @code{refresh}. -Note that @code{TIMESTAMP(X)} columns where X is 8 or 14 are reported to -be numbers while other @code{TIMESTAMP(X)} columns are reported to be -strings. This is just to ensure that one can reliably dump and restore -the table with these types! -@xref{DATETIME}. +The @code{shutdown} command shuts down the server. -@tindex TIME -@item TIME +The @code{processlist} command displays information about the threads +executing within the server. The @code{kill} command kills server threads. +You can always display or kill your own threads, but you need the +@strong{process} privilege to display or kill threads initiated by other +users. @xref{KILL}. -A time. The range is @code{'-838:59:59'} to @code{'838:59:59'}. -@strong{MySQL} displays @code{TIME} values in @code{'HH:MM:SS'} format, but -allows you to assign values to @code{TIME} columns using either strings or -numbers. @xref{TIME}. +It is a good idea in general to grant privileges only to those users who need +them, but you should exercise particular caution in granting certain +privileges: -@tindex YEAR -@item YEAR[(2|4)] +@itemize @bullet +@item +The @strong{grant} privilege allows users to give away their privileges to +other users. Two users with different privileges and with the @strong{grant} +privilege are able to combine privileges. -A year in 2- or 4-digit format (default is 4-digit). The allowable values -are @code{1901} to @code{2155}, @code{0000} in the 4-digit year format, -and 1970-2069 if you use the 2-digit format (70-69). @strong{MySQL} displays -@code{YEAR} values in @code{YYYY} format, but allows you to assign values to -@code{YEAR} columns using either strings or numbers. (The @code{YEAR} type is -new in @strong{MySQL} Version 3.22.). @xref{YEAR}. +@item +The @strong{alter} privilege may be used to subvert the privilege system +by renaming tables. -@tindex NATIONAL CHAR -@tindex NCHAR -@tindex CHAR -@tindex CHARACTER -@item [NATIONAL] CHAR(M) [BINARY] +@item +The @strong{file} privilege can be abused to read any world-readable file on +the server into a database table, the contents of which can then be +accessed using @code{SELECT}. This includes the contents of all databases +hosted by the server! -A fixed-length string that is always right-padded with spaces to the -specified length when stored. The range of @code{M} is 1 to 255 characters. -Trailing spaces are removed when the value is retrieved. @code{CHAR} values -are sorted and compared in case-insensitive fashion according to the -default character set unless the @code{BINARY} keyword is given. +@item +The @strong{shutdown} privilege can be abused to deny service to other +users entirely, by terminating the server. -@code{NATIONAL CHAR} (short form @code{NCHAR}) is the ANSI SQL way to -define that a CHAR column should use the default CHARACTER set. This is -the default in @strong{MySQL}. +@item +The @strong{process} privilege can be used to view the plain text of +currently executing queries, including queries that set or change passwords. -@code{CHAR} is a shorthand for @code{CHARACTER}. +@item +Privileges on the @code{mysql} database can be used to change passwords +and other access privilege information. (Passwords are stored +encrypted, so a malicious user cannot simply read them to know the plain +text password). If they can access the @code{mysql.user} password +column, they can use it to log into the @strong{MySQL} server +for the given user. (With sufficient privileges, the same user can +replace a password with a different one.) +@end itemize -@strong{MySQL} allows you to create a column of type -@code{CHAR(0)}. This is mainly useful when you have to be compliant with -some old applications that depend on the existence of a column but that do not -actually use the value. This is also quite nice when you need a -column that only can take 2 values: A @code{CHAR(0)}, that is not defined -as @code{NOT NULL}, will only occupy one bit and can only take 2 values: -@code{NULL} or @code{""}. @xref{CHAR}. +There are some things that you cannot do with the @strong{MySQL} +privilege system: -@tindex CHARACTER VARYING -@tindex CHAR VARYING -@tindex VARCHAR -@item [NATIONAL] VARCHAR(M) [BINARY] +@itemize @bullet +@item +You cannot explicitly specify that a given user should be denied access. +That is, you cannot explicitly match a user and then refuse the connection. -A variable-length string. @strong{NOTE:} Trailing spaces are removed when -the value is stored (this differs from the ANSI SQL specification). The range -of @code{M} is 1 to 255 characters. @code{VARCHAR} values are sorted and -compared in case-insensitive fashion unless the @code{BINARY} keyword is -given. @xref{Silent column changes}. +@item +You cannot specify that a user has privileges to create or drop tables +in a database but not to create or drop the database itself. +@end itemize -@code{VARCHAR} is a shorthand for @code{CHARACTER VARYING}. -@xref{CHAR}. +@node Privileges, Connection access, Privileges provided, Privilege system +@section How the Privilege System Works +@cindex privilege system, described -@tindex TINYBLOB -@tindex TINYTEXT -@item TINYBLOB -@itemx TINYTEXT +The @strong{MySQL} privilege system ensures that all users may do exactly the +things that they are supposed to be allowed to do. When you connect to a +@strong{MySQL} server, your identity is determined by @strong{the host from +which you connect} and @strong{the user name you specify}. The system grants +privileges according to your identity and @strong{what you want to do}. -A @code{BLOB} or @code{TEXT} column with a maximum length of 255 (2^8 - 1) -characters. @xref{Silent column changes}. @xref{BLOB}. +@strong{MySQL} considers both your hostname and user name in identifying you +because there is little reason to assume that a given user name belongs to +the same person everywhere on the Internet. For example, the user +@code{bill} who connects from @code{whitehouse.gov} need not be the same +person as the user @code{bill} who connects from @code{microsoft.com}. +@strong{MySQL} handles this by allowing you to distinguish users on different +hosts that happen to have the same name: you can grant @code{bill} one set +of privileges for connections from @code{whitehouse.gov}, and a different set +of privileges for connections from @code{microsoft.com}. -@tindex BLOB -@tindex TEXT -@item BLOB -@itemx TEXT +@strong{MySQL} access control involves two stages: -A @code{BLOB} or @code{TEXT} column with a maximum length of 65535 (2^16 - 1) -characters. @xref{Silent column changes}. @xref{BLOB}. +@itemize @bullet +@item +Stage 1: The server checks whether or not you are even allowed to connect. -@tindex MEDIUMBLOB -@tindex MEDIUMTEXT -@item MEDIUMBLOB -@itemx MEDIUMTEXT +@item +Stage 2: Assuming you can connect, the server checks each request you issue +to see whether or not you have sufficient privileges to perform it. For +example, if you try to select rows from a table in a database or drop a table +from the database, the server makes sure you have the @strong{select} +privilege for the table or the @strong{drop} privilege for the database. +@end itemize -A @code{BLOB} or @code{TEXT} column with a maximum length of 16777215 -(2^24 - 1) characters. @xref{Silent column changes}. @xref{BLOB}. +The server uses the @code{user}, @code{db}, and @code{host} tables in the +@code{mysql} database at both stages of access control. The fields in these +grant tables are shown below: -@tindex LONGBLOB -@tindex LONGTEXT -@item LONGBLOB -@itemx LONGTEXT - -A @code{BLOB} or @code{TEXT} column with a maximum length of 4294967295 -(2^32 - 1) characters. @xref{Silent column changes}. Note that because -the server/client protocol and MyISAM tables has currently a limit of -16M per communication packet / table row, you can't yet use this -the whole range of this type. @xref{BLOB}. - -@tindex ENUM -@item ENUM('value1','value2',...) - -An enumeration. A string object that can have only one value, chosen -from the list of values @code{'value1'}, @code{'value2'}, @code{...}, -@code{NULL} or the special @code{""} error value. An @code{ENUM} can -have a maximum of 65535 distinct values. @xref{ENUM}. - -@tindex SET -@item SET('value1','value2',...) - -A set. A string object that can have zero or more values, each of which must -be chosen from the list of values @code{'value1'}, @code{'value2'}, -@code{...} A @code{SET} can have a maximum of 64 members. @xref{SET}. -@end table +@multitable @columnfractions .2 .25 .25 .25 +@item @strong{Table name} @tab @code{user} @tab @code{db} @tab @code{host} -@menu -* Storage requirements:: Column type storage requirements -* Numeric types:: Numeric types -* Date and time types:: Date and time types -* String types:: String types -* Choosing types:: Choosing the right type for a column -* Indexes:: Column indexes -* Multiple-column indexes:: Multiple-column indexes -* Other-vendor column types:: Using column types from other database engines -@end menu +@item @strong{Scope fields} @tab @code{Host} @tab @code{Host} @tab @code{Host} +@item @tab @code{User} @tab @code{Db} @tab @code{Db} +@item @tab @code{Password} @tab @code{User} @tab -@cindex storage requirements, column type -@cindex columns, storage requirements +@item @strong{Privilege fields} @tab @code{Select_priv} @tab @code{Select_priv} @tab @code{Select_priv} +@item @tab @code{Insert_priv} @tab @code{Insert_priv} @tab @code{Insert_priv} +@item @tab @code{Update_priv} @tab @code{Update_priv} @tab @code{Update_priv} +@item @tab @code{Delete_priv} @tab @code{Delete_priv} @tab @code{Delete_priv} +@item @tab @code{Index_priv} @tab @code{Index_priv} @tab @code{Index_priv} +@item @tab @code{Alter_priv} @tab @code{Alter_priv} @tab @code{Alter_priv} +@item @tab @code{Create_priv} @tab @code{Create_priv} @tab @code{Create_priv} +@item @tab @code{Drop_priv} @tab @code{Drop_priv} @tab @code{Drop_priv} +@item @tab @code{Grant_priv} @tab @code{Grant_priv} @tab @code{Grant_priv} +@item @tab @code{References_priv} @tab @tab +@item @tab @code{Reload_priv} @tab @tab +@item @tab @code{Shutdown_priv} @tab @tab +@item @tab @code{Process_priv} @tab @tab +@item @tab @code{File_priv} @tab @tab +@end multitable -@node Storage requirements, Numeric types, Column types, Column types -@subsection Column Type Storage Requirements +For the second stage of access control (request verification), the server +may, if the request involves tables, additionally consult the +@code{tables_priv} and @code{columns_priv} tables. The fields in these +tables are shown below: -The storage requirements for each of the column types supported by -@strong{MySQL} are listed below by category. +@multitable @columnfractions .2 .25 .25 +@item @strong{Table name} @tab @code{tables_priv} @tab @code{columns_priv} -@cindex numeric types -@cindex types, numeric +@item @strong{Scope fields} @tab @code{Host} @tab @code{Host} +@item @tab @code{Db} @tab @code{Db} +@item @tab @code{User} @tab @code{User} +@item @tab @code{Table_name} @tab @code{Table_name} +@item @tab @tab @code{Column_name} -@subsubheading Storage requirements for numeric types +@item @strong{Privilege fields} @tab @code{Table_priv} @tab @code{Column_priv} +@item @tab @code{Column_priv} @tab -@multitable @columnfractions .5 .5 -@item @strong{Column type} @tab @strong{Storage required} -@item @code{TINYINT} @tab 1 byte -@item @code{SMALLINT} @tab 2 bytes -@item @code{MEDIUMINT} @tab 3 bytes -@item @code{INT} @tab 4 bytes -@item @code{INTEGER} @tab 4 bytes -@item @code{BIGINT} @tab 8 bytes -@item @code{FLOAT(X)} @tab 4 if X <= 24 or 8 if 25 <= X <= 53 -@item @code{FLOAT} @tab 4 bytes -@item @code{DOUBLE} @tab 8 bytes -@item @code{DOUBLE PRECISION} @tab 8 bytes -@item @code{REAL} @tab 8 bytes -@item @code{DECIMAL(M,D)} @tab @code{M+2} bytes if D > 0, @code{M+1} bytes if D = 0 (@code{D}+2, if @code{M < D}) -@item @code{NUMERIC(M,D)} @tab @code{M+2} bytes if D > 0, @code{M+1} bytes if D = 0 (@code{D}+2, if @code{M < D}) +@item @strong{Other fields} @tab @code{Timestamp} @tab @code{Timestamp} +@item @tab @code{Grantor} @tab @end multitable -@cindex date types -@cindex time types -@cindex types, date -@cindex types, time - -@subsubheading Storage requirements for date and time types - -@multitable @columnfractions .5 .5 -@item @strong{Column type} @tab @strong{Storage required} -@item @code{DATE} @tab 3 bytes -@item @code{DATETIME} @tab 8 bytes -@item @code{TIMESTAMP} @tab 4 bytes -@item @code{TIME} @tab 3 bytes -@item @code{YEAR} @tab 1 byte -@end multitable +Each grant table contains scope fields and privilege fields. -@subsubheading Storage requirements for string types +Scope fields determine the scope of each entry in the tables, that is, the +context in which the entry applies. For example, a @code{user} table entry +with @code{Host} and @code{User} values of @code{'thomas.loc.gov'} and +@code{'bob'} would be used for authenticating connections made to the server +by @code{bob} from the host @code{thomas.loc.gov}. Similarly, a @code{db} +table entry with @code{Host}, @code{User}, and @code{Db} fields of +@code{'thomas.loc.gov'}, @code{'bob'} and @code{'reports'} would be used when +@code{bob} connects from the host @code{thomas.loc.gov} to access the +@code{reports} database. The @code{tables_priv} and @code{columns_priv} +tables contain scope fields indicating tables or table/column combinations +to which each entry applies. -@multitable @columnfractions .5 .5 -@item @strong{Column type} @tab @strong{Storage required} -@item @code{CHAR(M)} @tab @code{M} bytes, @code{1 <= M <= 255} -@item @code{VARCHAR(M)} @tab @code{L}+1 bytes, where @code{L <= M} and -@code{1 <= M <= 255} -@item @code{TINYBLOB}, @code{TINYTEXT} @tab @code{L}+1 bytes, -where @code{L} < 2^8 -@item @code{BLOB}, @code{TEXT} @tab @code{L}+2 bytes, -where @code{L} < 2^16 -@item @code{MEDIUMBLOB}, @code{MEDIUMTEXT} @tab @code{L}+3 bytes, -where @code{L} < 2^24 -@item @code{LONGBLOB}, @code{LONGTEXT} @tab @code{L}+4 bytes, -where @code{L} < 2^32 -@item @code{ENUM('value1','value2',...)} @tab 1 or 2 bytes, depending on -the number of enumeration values (65535 values maximum) -@item @code{SET('value1','value2',...)} @tab 1, 2, 3, 4 or 8 bytes, depending -on the number of set members (64 members maximum) -@end multitable +@cindex case sensitivity, in access checking +For access-checking purposes, comparisons of @code{Host} values are +case insensitive. @code{User}, @code{Password}, @code{Db}, and +@code{Table_name} values are case sensitive. +@code{Column_name} values are case insensitive in @strong{MySQL} Version +3.22.12 or later. -@cindex BLOB, size -@cindex TEXT, size -@cindex VARCHAR, size -@code{VARCHAR} and the @code{BLOB} and @code{TEXT} types are variable-length -types, for which the storage requirements depend on the actual length of -column values (represented by @code{L} in the preceding table), rather than -on the type's maximum possible size. For example, a @code{VARCHAR(10)} -column can hold a string with a maximum length of 10 characters. The actual -storage required is the length of the string (@code{L}), plus 1 byte to -record the length of the string. For the string @code{'abcd'}, @code{L} is 4 -and the storage requirement is 5 bytes. +Privilege fields indicate the privileges granted by a table entry, that is, +what operations can be performed. The server combines the information in the +various grant tables to form a complete description of a user's privileges. +The rules used to do this are described in @ref{Request access}. -The @code{BLOB} and @code{TEXT} types require 1, 2, 3, or 4 bytes to record -the length of the column value, depending on the maximum possible length of -the type. @xref{BLOB}. +Scope fields are strings, declared as shown below; the default value for +each is the empty string: -If a table includes any variable-length column types, the record format will -also be variable-length. Note that when a table is created, @strong{MySQL} -may, under certain conditions, change a column from a variable-length type to a -fixed-length type, or vice-versa. @xref{Silent column changes}. +@multitable @columnfractions .15 .15 .7 +@item @strong{Field name} @tab @strong{Type} +@item @code{Host} @tab @code{CHAR(60)} +@item @code{User} @tab @code{CHAR(16)} +@item @code{Password} @tab @code{CHAR(16)} +@item @code{Db} @tab @code{CHAR(64)} @tab (@code{CHAR(60)} for the +@code{tables_priv} and @code{columns_priv} tables) +@item @code{Table_name} @tab @code{CHAR(60)} +@item @code{Column_name} @tab @code{CHAR(60)} +@end multitable -@cindex ENUM, size -The size of an @code{ENUM} object is determined by the number of -different enumeration values. One byte is used for enumerations with up -to 255 possible values. Two bytes are used for enumerations with up to -65535 values. @xref{ENUM}. +In the @code{user}, @code{db} and @code{host} tables, +all privilege fields are declared as @code{ENUM('N','Y')} --- each can have a +value of @code{'N'} or @code{'Y'}, and the default value is @code{'N'}. -@cindex SET, size -The size of a @code{SET} object is determined by the number of different -set members. If the set size is @code{N}, the object occupies @code{(N+7)/8} -bytes, rounded up to 1, 2, 3, 4, or 8 bytes. A @code{SET} can have a maximum -of 64 members. @xref{SET}. +In the @code{tables_priv} and @code{columns_priv} tables, the privilege +fields are declared as @code{SET} fields: -@node Numeric types, Date and time types, Storage requirements, Column types -@subsection Numeric Types +@multitable @columnfractions .2 .2 .6 +@item @strong{Table name} @tab @strong{Field name} @tab @strong{Possible set elements} +@item @code{tables_priv} @tab @code{Table_priv} @tab @code{'Select', 'Insert', +'Update', 'Delete', 'Create', 'Drop', 'Grant', 'References', 'Index', 'Alter'} +@item @code{tables_priv} @tab @code{Column_priv} @tab @code{'Select', 'Insert', +'Update', 'References'} +@item @code{columns_priv} @tab @code{Column_priv} @tab @code{'Select', 'Insert', +'Update', 'References'} +@end multitable -@strong{MySQL} supports all of the ANSI/ISO SQL92 numeric types. These -types include the exact numeric data types (@code{NUMERIC}, -@code{DECIMAL}, @code{INTEGER}, and @code{SMALLINT}), as well as the -approximate numeric data types (@code{FLOAT}, @code{REAL}, and -@code{DOUBLE PRECISION}). The keyword @code{INT} is a synonym for -@code{INTEGER}, and the keyword @code{DEC} is a synonym for -@code{DECIMAL}. +Briefly, the server uses the grant tables like this: -The @code{NUMERIC} and @code{DECIMAL} types are implemented as the same -type by @strong{MySQL}, as permitted by the SQL92 standard. They are -used for values for which it is important to preserve exact precision, -for example with monetary data. When declaring a column of one of these -types the precision and scale can be (and usually is) specified; for -example: +@itemize @bullet +@item +The @code{user} table scope fields determine whether to allow or reject +incoming connections. For allowed connections, any privileges granted in +the @code{user} table indicate the user's global (superuser) privileges. +These privileges apply to @strong{all} databases on the server. -@example - salary DECIMAL(9,2) -@end example +@item +The @code{db} and @code{host} tables are used together: -In this example, @code{9} (@code{precision}) represents the number of -significant decimal digits that will be stored for values, and -@code{2} (@code{scale}) represents the number of digits that will be -stored following the decimal point. In this case, therefore, the range -of values that can be stored in the @code{salary} column is from -@code{-9999999.99} to @code{9999999.99}. In ANSI/ISO SQL92, the syntax -@code{DECIMAL(p)} is equivalent to @code{DECIMAL(p,0)}. Similarly, the -syntax @code{DECIMAL} is equivalent to @code{DECIMAL(p,0)}, where the -implementation is allowed to decide the value of @code{p}. -@strong{MySQL} does not currently support either of these variant forms -of the @code{DECIMAL}/@code{NUMERIC} data types. This is not generally -a serious problem, as the principal benefits of these types derive from -the ability to control both precision and scale explicitly. +@itemize @minus +@item +The @code{db} table scope fields determine which users can access which +databases from which hosts. The privilege fields determine which operations +are allowed. -@code{DECIMAL} and @code{NUMERIC} values are stored as strings, rather -than as binary floating-point numbers, in order to preserve the decimal -precision of those values. One character is used for each digit of the -value, the decimal point (if @code{scale} > 0), and the @samp{-} sign -(for negative numbers). If @code{scale} is 0, @code{DECIMAL} and -@code{NUMERIC} values contain no decimal point or fractional part. +@item +The @code{host} table is used as an extension of the @code{db} table when you +want a given @code{db} table entry to apply to several hosts. For example, +if you want a user to be able to use a database from several hosts in +your network, leave the @code{Host} value empty in the user's @code{db} table +entry, then populate the @code{host} table with an entry for each of those +hosts. This mechanism is described more detail in @ref{Request access}. +@end itemize -The maximum range of @code{DECIMAL} and @code{NUMERIC} values is the -same as for @code{DOUBLE}, but the actual range for a given -@code{DECIMAL} or @code{NUMERIC} column can be constrained by the -@code{precision} or @code{scale} for a given column. When such a column -is assigned a value with more digits following the decimal point than -are allowed by the specified @code{scale}, the value is rounded to that -@code{scale}. When a @code{DECIMAL} or @code{NUMERIC} column is -assigned a value whose magnitude exceeds the range implied by the -specified (or defaulted) @code{precision} and @code{scale}, -@strong{MySQL} stores the value representing the corresponding end -point of that range. +@item +The @code{tables_priv} and @code{columns_priv} tables are similar to +the @code{db} table, but are more fine-grained: they apply at the +table and column levels rather than at the database level. +@end itemize -As an extension to the ANSI/ISO SQL92 standard, @strong{MySQL} also -supports the integral types @code{TINYINT}, @code{MEDIUMINT}, and -@code{BIGINT} as listed in the tables above. Another extension is -supported by @strong{MySQL} for optionally specifying the display width -of an integral value in parentheses following the base keyword for the -type (for example, @code{INT(4)}). This optional width specification is -used to left-pad the display of values whose width is less than the -width specified for the column, but does not constrain the range of -values that can be stored in the column, nor the number of digits that -will be displayed for values whose width exceeds that specified for the -column. When used in conjunction with the optional extension attribute -@code{ZEROFILL}, the default padding of spaces is replaced with zeroes. -For example, for a column declared as @code{INT(5) ZEROFILL}, a value -of @code{4} is retrieved as @code{00004}. Note that if you store larger -values than the display width in an integer column, you may experience -problems when @strong{MySQL} generates temporary tables for some -complicated joins, as in these cases @strong{MySQL} trusts that the -data did fit into the original column width. +Note that administrative privileges (@strong{reload}, @strong{shutdown}, +etc.) are specified only in the @code{user} table. This is because +administrative operations are operations on the server itself and are not +database-specific, so there is no reason to list such privileges in the +other grant tables. In fact, only the @code{user} table need +be consulted to determine whether or not you can perform an administrative +operation. -All integral types can have an optional (non-standard) attribute -@code{UNSIGNED}. Unsigned values can be used when you want to allow -only positive numbers in a column and you need a little bigger numeric -range for the column. +The @strong{file} privilege is specified only in the @code{user} table, too. +It is not an administrative privilege as such, but your ability to read or +write files on the server host is independent of the database you are +accessing. -The @code{FLOAT} type is used to represent approximate numeric data -types. The ANSI/ISO SQL92 standard allows an optional specification of -the precision (but not the range of the exponent) in bits following the -keyword @code{FLOAT} in parentheses. The @strong{MySQL} implementation -also supports this optional precision specification. When the keyword -@code{FLOAT} is used for a column type without a precision -specification, @strong{MySQL} uses four bytes to store the values. A -variant syntax is also supported, with two numbers given in parentheses -following the @code{FLOAT} keyword. With this option, the first number -continues to represent the storage requirements for the value in bytes, -and the second number specifies the number of digits to be stored and -displayed following the decimal point (as with @code{DECIMAL} and -@code{NUMERIC}). When @strong{MySQL} is asked to store a number for -such a column with more decimal digits following the decimal point than -specified for the column, the value is rounded to eliminate the extra -digits when the value is stored. +The @code{mysqld} server reads the contents of the grant tables once, when it +starts up. Changes to the grant tables take effect as indicated in +@ref{Privilege changes}. -The @code{REAL} and @code{DOUBLE PRECISION} types do not accept -precision specifications. As an extension to the ANSI/ISO SQL92 -standard, @strong{MySQL} recognizes @code{DOUBLE} as a synonym for the -@code{DOUBLE PRECISION} type. In contrast with the standard's -requirement that the precision for @code{REAL} be smaller than that used -for @code{DOUBLE PRECISION}, @strong{MySQL} implements both as 8-byte -double-precision floating-point values (when not running in ``ANSI mode''). -For maximum portability, code requiring storage of approximate numeric -data values should use @code{FLOAT} or @code{DOUBLE PRECISION} with no -specification of precision or number of decimal points. +When you modify the contents of the grant tables, it is a good idea to make +sure that your changes set up privileges the way you want. For help in +diagnosing problems, see @ref{Access denied}. For advice on security issues, +@pxref{Security}. -When asked to store a value in a numeric column that is outside the column -type's allowable range, @strong{MySQL} clips the value to the appropriate -endpoint of the range and stores the resulting value instead. +A useful +diagnostic tool is the @code{mysqlaccess} script, which Yves Carlier has +provided for the @strong{MySQL} distribution. Invoke @code{mysqlaccess} with +the @code{--help} option to find out how it works. +Note that @code{mysqlaccess} checks access using only the @code{user}, +@code{db} and @code{host} tables. It does not check table- or column-level +privileges. -For example, the range of an @code{INT} column is @code{-2147483648} to -@code{2147483647}. If you try to insert @code{-9999999999} into an -@code{INT} column, the value is clipped to the lower endpoint of the range, -and @code{-2147483648} is stored instead. Similarly, if you try to insert -@code{9999999999}, @code{2147483647} is stored instead. +@node Connection access, Request access, Privileges, Privilege system +@section Access Control, Stage 1: Connection Verification +@cindex access control +@cindex control access +@cindex connecting, verification +@cindex testing, connection to the server -If the @code{INT} column is @code{UNSIGNED}, the size of the column's -range is the same but its endpoints shift up to @code{0} and @code{4294967295}. -If you try to store @code{-9999999999} and @code{9999999999}, -the values stored in the column become @code{0} and @code{4294967296}. +When you attempt to connect to a @strong{MySQL} server, the server accepts or +rejects the connection based on your identity and whether or not you can +verify your identity by supplying the correct password. If not, the server +denies access to you completely. Otherwise, the server accepts the +connection, then enters Stage 2 and waits for requests. -Conversions that occur due to clipping are reported as ``warnings'' for -@code{ALTER TABLE}, @code{LOAD DATA INFILE}, @code{UPDATE}, and -multi-row @code{INSERT} statements. +Your identity is based on two pieces of information: -@cindex types, Date and Time -@cindex Date and Time types -@node Date and time types, String types, Numeric types, Column types -@subsection Date and Time Types +@itemize @bullet +@item +The host from which you connect -@menu -* Y2K issues:: Y2K issues and date types -* DATETIME:: The @code{DATETIME}, @code{DATE} and @code{TIMESTAMP} types -* TIME:: The @code{TIME} type -* YEAR:: The @code{YEAR} type -@end menu +@item +Your @strong{MySQL} user name +@end itemize -The date and time types are @code{DATETIME}, @code{DATE}, -@code{TIMESTAMP}, @code{TIME}, and @code{YEAR}. Each of these has a -range of legal values, as well as a ``zero'' value that is used when you -specify a really illegal value. Note that @strong{MySQL} allows you to store -certain 'not strictly' legal date values, for example @code{1999-11-31}. -The reason for this is that we think it's the responsibility of the -application to handle date checking, not the SQL servers. To make the -date checking 'fast', @strong{MySQL} only checks that the month is in -the range of 0-12 and the day is in the range of 0-31. The above ranges -are defined this way because @strong{MySQL} allows you to store, in a -@code{DATE} or @code{DATETIME} column, dates where the day or month-day -is zero. This is extremely useful for applications that need to store -a birth-date for which you don't know the exact date. In this case you -simply store the date like @code{1999-00-00} or @code{1999-01-00}. (You -cannot expect to get a correct value from functions like @code{DATE_SUB()} -or @code{DATE_ADD} for dates like these.) +Identity checking is performed using the three @code{user} table scope fields +(@code{Host}, @code{User}, and @code{Password}). The server accepts the +connection only if a @code{user} table entry matches your hostname and user +name, and you supply the correct password. -Here are some general considerations to keep in mind when working -with date and time types: +Values in the @code{user} table scope fields may be specified as follows: @itemize @bullet @item -@strong{MySQL} retrieves values for a given date or time type in a standard -format, but it attempts to interpret a variety of formats for values that -you supply (for example, when you specify a value to be assigned to or -compared to a date or time type). Nevertheless, only the formats described -in the following sections are supported. It is expected that you will supply -legal values, and unpredictable results may occur if you use values in other -formats. +A @code{Host} value may be a hostname or an IP number, or @code{'localhost'} +to indicate the local host. @item -Although @strong{MySQL} tries to interpret values in several formats, it -always expects the year part of date values to be leftmost. Dates must be -given in year-month-day order (for example, @code{'98-09-04'}), rather than -in the month-day-year or day-month-year orders commonly used elsewhere (for -example, @code{'09-04-98'}, @code{'04-09-98'}). +@cindex wildcards, in @code{mysql.user} table +You can use the wild-card characters @samp{%} and @samp{_} in the @code{Host} +field. @item -@strong{MySQL} automatically converts a date or time type value to a number -if the value is used in a numeric context, and vice versa. +A @code{Host} value of @code{'%'} matches any hostname. @item -When @strong{MySQL} encounters a value for a date or time type that is -out of range or otherwise illegal for the type (see the start of this -section), it converts the value to the ``zero'' value for that type. -(The exception is that out-of-range @code{TIME} values are clipped to -the appropriate endpoint of the @code{TIME} range.) The table below -shows the format of the ``zero'' value for each type: - -@multitable @columnfractions .3 .7 -@item @strong{Column type} @tab @strong{``Zero'' value} -@item @code{DATETIME} @tab @code{'0000-00-00 00:00:00'} -@item @code{DATE} @tab @code{'0000-00-00'} -@item @code{TIMESTAMP} @tab @code{00000000000000} (length depends on display size) -@item @code{TIME} @tab @code{'00:00:00'} -@item @code{YEAR} @tab @code{0000} -@end multitable +A blank @code{Host} value means that the privilege should be anded +with the entry in the @code{host} table that matches the given host name. +You can find more information about this in the next chapter. +@cindex netmask notation, in @code{mysql.user} table @item -The ``zero'' values are special, but you can store or refer to them -explicitly using the values shown in the table. You can also do this -using the values @code{'0'} or @code{0}, which are easier to write. +As of @strong{MySQL} Version 3.23, for @code{Host} values specified as +IP numbers, you can specify a netmask indicating how many address bits to +use for the network number. For example: -@item -``Zero'' date or time values used through @strong{MyODBC} are converted -automatically to @code{NULL} in @strong{MyODBC} Version 2.50.12 and above, -because ODBC can't handle such values. -@end itemize +@example +GRANT ALL PRIVILEGES on db.* to david@@'192.58.197.0/255.255.255.0'; +@end example -@cindex Year 2000 issues -@cindex date types, Y2K issues -@node Y2K issues, DATETIME, Date and time types, Date and time types -@subsubsection Y2K Issues and Date Types +This will allow everyone to connect from an IP where the following is true: -@strong{MySQL} itself is Y2K-safe (@pxref{Year 2000 compliance}), -but input values presented to @strong{MySQL} may not be. Any input -containing 2-digit year values is ambiguous, because the century is unknown. -Such values must be interpreted into 4-digit form because @strong{MySQL} stores -years internally using four digits. +@example +user_ip & netmask = host_ip. +@end example -For @code{DATETIME}, @code{DATE}, @code{TIMESTAMP}, and @code{YEAR} types, -@strong{MySQL} interprets dates with ambiguous year values using the -following rules: +In the above example all IP:s in the interval 192.58.197.0 - +192.58.197.255 can connect to the @strong{MySQL} server. -@itemize @bullet @item -Year values in the range @code{00-69} are converted to @code{2000-2069}. +@cindex anonymous user +Wild-card characters are not allowed in the @code{User} field, but you can +specify a blank value, which matches any name. If the @code{user} table +entry that matches an incoming connection has a blank user name, the user is +considered to be the anonymous user (the user with no name), rather than the +name that the client actually specified. This means that a blank user name +is used for all further access checking for the duration of the connection +(that is, during Stage 2). @item -Year values in the range @code{70-99} are converted to @code{1970-1999}. +The @code{Password} field can be blank. This does not mean that any password +matches, it means the user must connect without specifying a password. @end itemize -Remember that these rules provide only reasonable guesses as to what your -data mean. If the heuristics used by @strong{MySQL} don't produce the -correct values, you should provide unambiguous input containing 4-digit -year values. - -@code{ORDER BY} will sort 2-digit @code{YEAR/DATE/DATETIME} types properly. - -Note also that some functions like @code{MIN()} and @code{MAX()} will convert a -@code{TIMESTAMP/DATE} to a number. This means that a timestamp with a -2-digit year will not work properly with these functions. The fix in this -case is to convert the @code{TIMESTAMP/DATE} to 4-digit year format or -use something like @code{MIN(DATE_ADD(timestamp,INTERVAL 0 DAYS))}. +@findex PASSWORD() +Non-blank @code{Password} values represent encrypted passwords. +@strong{MySQL} does not store passwords in plaintext form for anyone to +see. Rather, the password supplied by a user who is attempting to +connect is encrypted (using the @code{PASSWORD()} function). The +encrypted password is then used when the client/server is checking if +the password is correct (This is done without the encrypted password +ever traveling over the connection.) Note that from @strong{MySQL}'s +point of view the encrypted password is the REAL password, so you should +not give anyone access to it! In particular, don't give normal users +read access to the tables in the @code{mysql} database! -@tindex DATETIME -@tindex DATE -@tindex TIMESTAMP -@node DATETIME, TIME, Y2K issues, Date and time types -@subsubsection The @code{DATETIME}, @code{DATE}, and @code{TIMESTAMP} Types +The examples below show how various combinations of @code{Host} and +@code{User} values in @code{user} table entries apply to incoming +connections: -The @code{DATETIME}, @code{DATE}, and @code{TIMESTAMP} types are related. -This section describes their characteristics, how they are similar, and how -they differ. +@multitable @columnfractions .25 .15 .60 +@item @code{Host} @strong{value} @tab @code{User} @strong{value} @tab @strong{Connections matched by entry} +@item @code{'thomas.loc.gov'} @tab @code{'fred'} @tab @code{fred}, connecting from @code{thomas.loc.gov} +@item @code{'thomas.loc.gov'} @tab @code{''} @tab Any user, connecting from @code{thomas.loc.gov} +@item @code{'%'} @tab @code{'fred'} @tab @code{fred}, connecting from any host +@item @code{'%'} @tab @code{''} @tab Any user, connecting from any host +@item @code{'%.loc.gov'} @tab @code{'fred'} @tab @code{fred}, connecting from any host in the @code{loc.gov} domain +@item @code{'x.y.%'} @tab @code{'fred'} @tab @code{fred}, connecting from @code{x.y.net}, @code{x.y.com},@code{x.y.edu}, etc. (this is probably not useful) +@item @code{'144.155.166.177'} @tab @code{'fred'} @tab @code{fred}, connecting from the host with IP address @code{144.155.166.177} +@item @code{'144.155.166.%'} @tab @code{'fred'} @tab @code{fred}, connecting from any host in the @code{144.155.166} class C subnet +@item @code{'144.155.166.0/255.255.255.0'} @tab @code{'fred'} @tab Same as previous example +@end multitable -The @code{DATETIME} type is used when you need values that contain both date -and time information. @strong{MySQL} retrieves and displays @code{DATETIME} -values in @code{'YYYY-MM-DD HH:MM:SS'} format. The supported range is -@code{'1000-01-01 00:00:00'} to @code{'9999-12-31 23:59:59'}. (``Supported'' -means that although earlier values might work, there is no guarantee that -they will.) +Because you can use IP wild-card values in the @code{Host} field (for example, +@code{'144.155.166.%'} to match every host on a subnet), there is the +possibility that someone might try to exploit this capability by naming a +host @code{144.155.166.somewhere.com}. To foil such attempts, @strong{MySQL} +disallows matching on hostnames that start with digits and a dot. Thus, if +you have a host named something like @code{1.2.foo.com}, its name will never +match the @code{Host} column of the grant tables. Only an IP number can +match an IP wild-card value. -The @code{DATE} type is used when you need only a date value, without a time -part. @strong{MySQL} retrieves and displays @code{DATE} values in -@code{'YYYY-MM-DD'} format. The supported range is @code{'1000-01-01'} to -@code{'9999-12-31'}. +An incoming connection may be matched by more than one entry in the +@code{user} table. For example, a connection from @code{thomas.loc.gov} by +@code{fred} would be matched by several of the entries just shown above. How +does the server choose which entry to use if more than one matches? The +server resolves this question by sorting the @code{user} table after reading +it at startup time, then looking through the entries in sorted order when a +user attempts to connect. The first matching entry is the one that is used. -The @code{TIMESTAMP} column type provides a type that you can use to -automatically mark @code{INSERT} or @code{UPDATE} operations with the current -date and time. If you have multiple @code{TIMESTAMP} columns, only the first -one is updated automatically. +@code{user} table sorting works as follows. Suppose the @code{user} table +looks like this: -Automatic updating of the first @code{TIMESTAMP} column occurs under any of -the following conditions: +@example ++-----------+----------+- +| Host | User | ... ++-----------+----------+- +| % | root | ... +| % | jeffrey | ... +| localhost | root | ... +| localhost | | ... ++-----------+----------+- +@end example -@itemize @bullet -@item -The column is not specified explicitly in an @code{INSERT} or -@code{LOAD DATA INFILE} statement. -@item -The column is not specified explicitly in an @code{UPDATE} statement and some -other column changes value. (Note that an @code{UPDATE} that sets a column -to the value it already has will not cause the @code{TIMESTAMP} column to be -updated, because if you set a column to its current value, @strong{MySQL} -ignores the update for efficiency.) -@item -You explicitly set the @code{TIMESTAMP} column to @code{NULL}. -@end itemize +When the server reads in the table, it orders the entries with the +most-specific @code{Host} values first (@code{'%'} in the @code{Host} column +means ``any host'' and is least specific). Entries with the same @code{Host} +value are ordered with the most-specific @code{User} values first (a blank +@code{User} value means ``any user'' and is least specific). The resulting +sorted @code{user} table looks like this: -@code{TIMESTAMP} columns other than the first may also be set to the current -date and time. Just set the column to @code{NULL} or to @code{NOW()}. +@example ++-----------+----------+- +| Host | User | ... ++-----------+----------+- +| localhost | root | ... +| localhost | | ... +| % | jeffrey | ... +| % | root | ... ++-----------+----------+- +@end example -You can set any @code{TIMESTAMP} column to a value different than the current -date and time by setting it explicitly to the desired value. This is true -even for the first @code{TIMESTAMP} column. You can use this property if, -for example, you want a @code{TIMESTAMP} to be set to the current date and -time when you create a row, but not to be changed whenever the row is updated -later: +@cindex grant tables, sorting +@cindex sorting, grant tables +@cindex @code{user} table, sorting +When a connection is attempted, the server looks through the sorted entries +and uses the first match found. For a connection from @code{localhost} by +@code{jeffrey}, the entries with @code{'localhost'} in the @code{Host} column +match first. Of those, the entry with the blank user name matches both the +connecting hostname and user name. (The @code{'%'/'jeffrey'} entry would +have matched, too, but it is not the first match in the table.) -@itemize @bullet -@item -Let @strong{MySQL} set the column when the row is created. -This will initialize it to the current date and time. +Here is another example. Suppose the @code{user} table looks like this: -@item -When you perform subsequent updates to other columns in the row, set -the @code{TIMESTAMP} column explicitly to its current value. -@end itemize +@example ++----------------+----------+- +| Host | User | ... ++----------------+----------+- +| % | jeffrey | ... +| thomas.loc.gov | | ... ++----------------+----------+- +@end example -On the other hand, you may find it just as easy to use a @code{DATETIME} -column that you initialize to @code{NOW()} when the row is created and -leave alone for subsequent updates. +The sorted table looks like this: -@code{TIMESTAMP} values may range from the beginning of 1970 to sometime in -the year 2037, with a resolution of one second. Values are displayed as -numbers. +@example ++----------------+----------+- +| Host | User | ... ++----------------+----------+- +| thomas.loc.gov | | ... +| % | jeffrey | ... ++----------------+----------+- +@end example -The format in which @strong{MySQL} retrieves and displays @code{TIMESTAMP} -values depends on the display size, as illustrated by the table below. The -`full' @code{TIMESTAMP} format is 14 digits, but @code{TIMESTAMP} columns may -be created with shorter display sizes: +A connection from @code{thomas.loc.gov} by @code{jeffrey} is matched by the +first entry, whereas a connection from @code{whitehouse.gov} by +@code{jeffrey} is matched by the second. -@multitable @columnfractions .3 .7 -@item @strong{Column type} @tab @strong{Display format} -@item @code{TIMESTAMP(14)} @tab @code{YYYYMMDDHHMMSS} -@item @code{TIMESTAMP(12)} @tab @code{YYMMDDHHMMSS} -@item @code{TIMESTAMP(10)} @tab @code{YYMMDDHHMM} -@item @code{TIMESTAMP(8)} @tab @code{YYYYMMDD} -@item @code{TIMESTAMP(6)} @tab @code{YYMMDD} -@item @code{TIMESTAMP(4)} @tab @code{YYMM} -@item @code{TIMESTAMP(2)} @tab @code{YY} -@end multitable +A common misconception is to think that for a given user name, all entries +that explicitly name that user will be used first when the server attempts to +find a match for the connection. This is simply not true. The previous +example illustrates this, where a connection from @code{thomas.loc.gov} by +@code{jeffrey} is first matched not by the entry containing @code{'jeffrey'} +as the @code{User} field value, but by the entry with no user name! -All @code{TIMESTAMP} columns have the same storage size, regardless of -display size. The most common display sizes are 6, 8, 12, and 14. You can -specify an arbitrary display size at table creation time, but values of 0 or -greater than 14 are coerced to 14. Odd-valued sizes in the range from 1 to -13 are coerced to the next higher even number. +If you have problems connecting to the server, print out the @code{user} +table and sort it by hand to see where the first match is being made. -You can specify @code{DATETIME}, @code{DATE}, and @code{TIMESTAMP} values using -any of a common set of formats: +@node Request access, Privilege changes, Connection access, Privilege system +@section Access Control, Stage 2: Request Verification -@itemize @bullet -@item -As a string in either @code{'YYYY-MM-DD HH:MM:SS'} or @code{'YY-MM-DD -HH:MM:SS'} format. A ``relaxed'' syntax is allowed---any punctuation -character may be used as the delimiter between date parts or time parts. -For example, @code{'98-12-31 11:30:45'}, @code{'98.12.31 11+30+45'}, -@code{'98/12/31 11*30*45'}, and @code{'98@@12@@31 11^30^45'} are -equivalent. +Once you establish a connection, the server enters Stage 2. For each request +that comes in on the connection, the server checks whether you have +sufficient privileges to perform it, based on the type of operation you wish +to perform. This is where the privilege fields in the grant tables come into +play. These privileges can come from any of the @code{user}, @code{db}, +@code{host}, @code{tables_priv}, or @code{columns_priv} tables. The grant +tables are manipulated with @code{GRANT} and @code{REVOKE} commands. +@xref{GRANT, , @code{GRANT}}. (You may find it helpful to refer to +@ref{Privileges}, which lists the fields present in each of the grant +tables.) -@item -As a string in either @code{'YYYY-MM-DD'} or @code{'YY-MM-DD'} format. -A ``relaxed'' syntax is allowed here, too. For example, @code{'98-12-31'}, -@code{'98.12.31'}, @code{'98/12/31'}, and @code{'98@@12@@31'} are -equivalent. +The @code{user} table grants privileges that are assigned to you on a global +basis and that apply no matter what the current database is. For example, if +the @code{user} table grants you the @strong{delete} privilege, you can +delete rows from any database on the server host! In other words, +@code{user} table privileges are superuser privileges. It is wise to grant +privileges in the @code{user} table only to superusers such as server or +database administrators. For other users, you should leave the privileges +in the @code{user} table set to @code{'N'} and grant privileges on a +database-specific basis only, using the @code{db} and @code{host} tables. + +@cindex Anonymous user +@cindex wild cards, in @code{mysql.db} table +@cindex wild cards, in @code{mysql.host} table +The @code{db} and @code{host} tables grant database-specific privileges. +Values in the scope fields may be specified as follows: +@itemize @bullet @item -As a string with no delimiters in either @code{'YYYYMMDDHHMMSS'} or -@code{'YYMMDDHHMMSS'} format, provided that the string makes sense as a -date. For example, @code{'19970523091528'} and @code{'970523091528'} are -interpreted as @code{'1997-05-23 09:15:28'}, but @code{'971122129015'} is -illegal (it has a nonsensical minute part) and becomes @code{'0000-00-00 -00:00:00'}. +The wild-card characters @samp{%} and @samp{_} can be used in the @code{Host} +and @code{Db} fields of either table. @item -As a string with no delimiters in either @code{'YYYYMMDD'} or @code{'YYMMDD'} -format, provided that the string makes sense as a date. For example, -@code{'19970523'} and @code{'970523'} are interpreted as -@code{'1997-05-23'}, but @code{'971332'} is illegal (it has nonsensical month -and day parts) and becomes @code{'0000-00-00'}. +A @code{'%'} @code{Host} value in the @code{db} table means ``any host.'' A +blank @code{Host} value in the @code{db} table means ``consult the +@code{host} table for further information.'' @item -As a number in either @code{YYYYMMDDHHMMSS} or @code{YYMMDDHHMMSS} -format, provided that the number makes sense as a date. For example, -@code{19830905132800} and @code{830905132800} are interpreted as -@code{'1983-09-05 13:28:00'}. +A @code{'%'} or blank @code{Host} value in the @code{host} table means ``any +host.'' @item -As a number in either @code{YYYYMMDD} or @code{YYMMDD} -format, provided that the number makes sense as a date. For example, -@code{19830905} and @code{830905} are interpreted as @code{'1983-09-05'}. +A @code{'%'} or blank @code{Db} value in either table means ``any database.'' @item -As the result of a function that returns a value that is acceptable -in a @code{DATETIME}, @code{DATE}, or @code{TIMESTAMP} context, such as -@code{NOW()} or @code{CURRENT_DATE}. +A blank @code{User} value in either table matches the anonymous user. @end itemize -Illegal @code{DATETIME}, @code{DATE}, or @code{TIMESTAMP} values are converted -to the ``zero'' value of the appropriate type (@code{'0000-00-00 00:00:00'}, -@code{'0000-00-00'}, or @code{00000000000000}). - -For values specified as strings that include date part delimiters, it is not -necessary to specify two digits for month or day values that are less than -@code{10}. @code{'1979-6-9'} is the same as @code{'1979-06-09'}. Similarly, -for values specified as strings that include time part delimiters, it is not -necessary to specify two digits for hour, month, or second values that are -less than @code{10}. @code{'1979-10-30 1:2:3'} is the same as -@code{'1979-10-30 01:02:03'}. - -Values specified as numbers should be 6, 8, 12, or 14 digits long. If the -number is 8 or 14 digits long, it is assumed to be in @code{YYYYMMDD} or -@code{YYYYMMDDHHMMSS} format and that the year is given by the first 4 -digits. If the number is 6 or 12 digits long, it is assumed to be in -@code{YYMMDD} or @code{YYMMDDHHMMSS} format and that the year is given by the -first 2 digits. Numbers that are not one of these lengths are interpreted -as though padded with leading zeros to the closest length. +@cindex grant tables, sorting +@cindex sorting, grant tables +@cindex @code{db} table, sorting +@cindex @code{host} table, sorting +The @code{db} and @code{host} tables are read in and sorted when the server +starts up (at the same time that it reads the @code{user} table). The +@code{db} table is sorted on the @code{Host}, @code{Db}, and @code{User} scope +fields, and the @code{host} table is sorted on the @code{Host} and @code{Db} +scope fields. As with the @code{user} table, sorting puts the most-specific +values first and least-specific values last, and when the server looks for +matching entries, it uses the first match that it finds. -@cindex non-delimited strings -@cindex strings, non-delimited -Values specified as non-delimited strings are interpreted using their length -as given. If the string is 8 or 14 characters long, the year is assumed to -be given by the first 4 characters. Otherwise the year is assumed to be -given by the first 2 characters. The string is interpreted from left to -right to find year, month, day, hour, minute, and second values, for as many -parts as are present in the string. This means you should not use strings -that have fewer than 6 characters. For example, if you specify @code{'9903'}, -thinking that will represent March, 1999, you will find that @strong{MySQL} -inserts a ``zero'' date into your table. This is because the year and month -values are @code{99} and @code{03}, but the day part is missing (zero), so -the value is not a legal date. -@code{TIMESTAMP} columns store legal values using the full precision with -which the value was specified, regardless of the display size. This has -several implications: +@cindex wild cards, in @code{mysql.tables_priv} table +@cindex wild cards, in @code{mysql.columns_priv} table +The @code{tables_priv} and @code{columns_priv} tables grant table- and +column-specific privileges. Values in the scope fields may be specified as +follows: @itemize @bullet @item -Always specify year, month, and day, even if your column types are -@code{TIMESTAMP(4)} or @code{TIMESTAMP(2)}. Otherwise, the value will not -be a legal date and @code{0} will be stored. - -@item -If you use @code{ALTER TABLE} to widen a narrow @code{TIMESTAMP} column, -information will be displayed that previously was ``hidden''. +The wild-card characters @samp{%} and @samp{_} +can be used in the @code{Host} field of either table. @item -Similarly, narrowing a @code{TIMESTAMP} column does not cause information to -be lost, except in the sense that less information is shown when the values -are displayed. +A @code{'%'} or blank @code{Host} value in either table means ``any host.'' @item -Although @code{TIMESTAMP} values are stored to full precision, the only -function that operates directly on the underlying stored value is -@code{UNIX_TIMESTAMP()}. Other functions operate on the formatted retrieved -value. This means you cannot use functions such as @code{HOUR()} or -@code{SECOND()} unless the relevant part of the @code{TIMESTAMP} value is -included in the formatted value. For example, the @code{HH} part of a -@code{TIMESTAMP} column is not displayed unless the display size is at least -10, so trying to use @code{HOUR()} on shorter @code{TIMESTAMP} values -produces a meaningless result. +The @code{Db}, @code{Table_name} and @code{Column_name} fields cannot contain +wild cards or be blank in either table. @end itemize -You can to some extent assign values of one date type to an object -of a different date type. However, there may be some alteration -of the value or loss of information: - -@itemize @bullet -@item -If you assign a @code{DATE} value to a @code{DATETIME} or @code{TIMESTAMP} -object, the time part of the resulting value is set to @code{'00:00:00'}, -because the @code{DATE} value contains no time information. +The @code{tables_priv} and @code{columns_priv} tables are sorted on +the @code{Host}, @code{Db}, and @code{User} fields. This is similar to +@code{db} table sorting, although the sorting is simpler because +only the @code{Host} field may contain wild cards. -@item -If you assign a @code{DATETIME} or @code{TIMESTAMP} value to a @code{DATE} -object, the time part of the resulting value is deleted, because the -@code{DATE} type stores no time information. +The request verification process is described below. (If you are familiar +with the access-checking source code, you will notice that the description +here differs slightly from the algorithm used in the code. The description +is equivalent to what the code actually does; it differs only to make the +explanation simpler.) -@item -Remember that although @code{DATETIME}, @code{DATE}, and @code{TIMESTAMP} -values all can be specified using the same set of formats, the types do not -all have the same range of values. For example, @code{TIMESTAMP} values -cannot be earlier than @code{1970} or later than @code{2037}. This means -that a date such as @code{'1968-01-01'}, while legal as a @code{DATETIME} or -@code{DATE} value, is not a valid @code{TIMESTAMP} value and will be -converted to @code{0} if assigned to such an object. -@end itemize +For administrative requests (@strong{shutdown}, @strong{reload}, etc.), the +server checks only the @code{user} table entry, because that is the only table +that specifies administrative privileges. Access is granted if the entry +allows the requested operation and denied otherwise. For example, if you +want to execute @code{mysqladmin shutdown} but your @code{user} table entry +doesn't grant the @strong{shutdown} privilege to you, access is denied +without even checking the @code{db} or @code{host} tables. (They +contain no @code{Shutdown_priv} column, so there is no need to do so.) -@cindex problems, date values -@cindex date values, problems -Be aware of certain pitfalls when specifying date values: +For database-related requests (@strong{insert}, @strong{update}, etc.), the +server first checks the user's global (superuser) privileges by looking in +the @code{user} table entry. If the entry allows the requested operation, +access is granted. If the global privileges in the @code{user} table are +insufficient, the server determines the user's database-specific privileges +by checking the @code{db} and @code{host} tables: -@itemize @bullet +@enumerate @item -The relaxed format allowed for values specified as strings can be deceiving. -For example, a value such as @code{'10:11:12'} might look like a time value -because of the @samp{:} delimiter, but if used in a date context will be -interpreted as the year @code{'2010-11-12'}. The value @code{'10:45:15'} -will be converted to @code{'0000-00-00'} because @code{'45'} is not a legal -month. +The server looks in the @code{db} table for a match on the @code{Host}, +@code{Db}, and @code{User} fields. The @code{Host} and @code{User} fields are +matched to the connecting user's hostname and @strong{MySQL} user name. The +@code{Db} field is matched to the database the user wants to access. If +there is no entry for the @code{Host} and @code{User}, access is denied. @item -Year values specified as two digits are ambiguous, because the century is -unknown. @strong{MySQL} interprets 2-digit year values using the following -rules: +If there is a matching @code{db} table entry and its @code{Host} field is +not blank, that entry defines the user's database-specific privileges. -@itemize @minus @item -Year values in the range @code{00-69} are converted to @code{2000-2069}. +If the matching @code{db} table entry's @code{Host} field is blank, it +signifies that the @code{host} table enumerates which hosts should be allowed +access to the database. In this case, a further lookup is done in the +@code{host} table to find a match on the @code{Host} and @code{Db} fields. +If no @code{host} table entry matches, access is denied. If there is a +match, the user's database-specific privileges are computed as the +intersection (@emph{not} the union!) of the privileges in the @code{db} and +@code{host} table entries, that is, the privileges that are @code{'Y'} in both +entries. (This way you can grant general privileges in the @code{db} table +entry and then selectively restrict them on a host-by-host basis using the +@code{host} table entries.) +@end enumerate -@item -Year values in the range @code{70-99} are converted to @code{1970-1999}. -@end itemize -@end itemize +After determining the database-specific privileges granted by the @code{db} +and @code{host} table entries, the server adds them to the global privileges +granted by the @code{user} table. If the result allows the requested +operation, access is granted. Otherwise, the server checks the user's +table and column privileges in the @code{tables_priv} and @code{columns_priv} +tables and adds those to the user's privileges. Access is allowed or denied +based on the result. -@tindex TIME -@node TIME, YEAR, DATETIME, Date and time types -@subsubsection The @code{TIME} Type +Expressed in boolean terms, the preceding description of how a user's +privileges are calculated may be summarized like this: -@strong{MySQL} retrieves and displays @code{TIME} values in @code{'HH:MM:SS'} -format (or @code{'HHH:MM:SS'} format for large hours values). @code{TIME} -values may range from @code{'-838:59:59'} to @code{'838:59:59'}. The reason -the hours part may be so large is that the @code{TIME} type may be used not -only to represent a time of day (which must be less than 24 hours), but also -elapsed time or a time interval between two events (which may be much greater -than 24 hours, or even negative). +@example +global privileges +OR (database privileges AND host privileges) +OR table privileges +OR column privileges +@end example -You can specify @code{TIME} values in a variety of formats: +It may not be apparent why, if the global @code{user} entry privileges are +initially found to be insufficient for the requested operation, the server +adds those privileges to the database-, table-, and column-specific privileges +later. The reason is that a request might require more than one type of +privilege. For example, if you execute an @code{INSERT ... SELECT} +statement, you need both @strong{insert} and @strong{select} privileges. +Your privileges might be such that the @code{user} table entry grants one +privilege and the @code{db} table entry grants the other. In this case, you +have the necessary privileges to perform the request, but the server cannot +tell that from either table by itself; the privileges granted by the entries +in both tables must be combined. -@itemize @bullet -@item -As a string in @code{'D HH:MM:SS.fraction'} format. (Note that -@strong{MySQL} doesn't yet store the fraction for the time column). One -can also use one of the following ``relaxed'' syntax: +@cindex @code{host} table +@cindex tables, @code{host} -@code{HH:MM:SS.fraction}, @code{HH:MM:SS}, @code{HH:MM}, @code{D HH:MM:SS}, -@code{D HH:MM}, @code{D HH} or @code{SS}. Here @code{D} is days between 0-33. +The @code{host} table can be used to maintain a list of secure servers. -@item -As a string with no delimiters in @code{'HHMMSS'} format, provided that -it makes sense as a time. For example, @code{'101112'} is understood as -@code{'10:11:12'}, but @code{'109712'} is illegal (it has a nonsensical -minute part) and becomes @code{'00:00:00'}. +At TcX, the @code{host} table contains a list of all machines on the local +network. These are granted all privileges. -@item -As a number in @code{HHMMSS} format, provided that it makes sense as a time. -For example, @code{101112} is understood as @code{'10:11:12'}. The following -alternative formats are also understood: @code{SS}, @code{MMSS},@code{HHMMSS}, -@code{HHMMSS.fraction}. Note that @strong{MySQL} doesn't yet store the -fraction part. +You can also use the @code{host} table to indicate hosts that are @emph{not} +secure. Suppose you have a machine @code{public.your.domain} that is located +in a public area that you do not consider secure. You can allow access to +all hosts on your network except that machine by using @code{host} table +entries +like this: -@item -As the result of a function that returns a value that is acceptable -in a @code{TIME} context, such as @code{CURRENT_TIME}. -@end itemize +@example ++--------------------+----+- +| Host | Db | ... ++--------------------+----+- +| public.your.domain | % | ... (all privileges set to 'N') +| %.your.domain | % | ... (all privileges set to 'Y') ++--------------------+----+- +@end example -For @code{TIME} values specified as strings that include a time part -delimiter, it is not necessary to specify two digits for hours, minutes, or -seconds values that are less than @code{10}. @code{'8:3:2'} is the same as -@code{'08:03:02'}. +@cindex privilege, changes +@cindex changes to privileges +@cindex tables, grant +@cindex grant tables -Be careful about assigning ``short'' @code{TIME} values to a @code{TIME} -column. Without semicolon, @strong{MySQL} interprets values using the -assumption that the rightmost digits represent seconds. (@strong{MySQL} -interprets @code{TIME} values as elapsed time rather than as time of -day.) For example, you might think of @code{'1112'} and @code{1112} as -meaning @code{'11:12:00'} (12 minutes after 11 o'clock), but -@strong{MySQL} interprets them as @code{'00:11:12'} (11 minutes, 12 seconds). -Similarly, @code{'12'} and @code{12} are interpreted as @code{'00:00:12'}. -@code{TIME} values with semicolon, instead, are always treated as -time of the day. That is @code{'11:12'} will mean @code{'11:12:00'}, -not @code{'00:11:12'}. - -Values that lie outside the @code{TIME} range -but are otherwise legal are clipped to the appropriate -endpoint of the range. For example, @code{'-850:00:00'} and -@code{'850:00:00'} are converted to @code{'-838:59:59'} and -@code{'838:59:59'}. +Naturally, you should always test your entries in the grant tables (for +example, using @code{mysqlaccess}) to make sure your access privileges are +actually set up the way you think they are. -Illegal @code{TIME} values are converted to @code{'00:00:00'}. Note that -because @code{'00:00:00'} is itself a legal @code{TIME} value, there is no way -to tell, from a value of @code{'00:00:00'} stored in a table, whether the -original value was specified as @code{'00:00:00'} or whether it was illegal. +@node Privilege changes, Default privileges, Request access, Privilege system +@section When Privilege Changes Take Effect -@tindex YEAR -@node YEAR, , TIME, Date and time types -@subsubsection The @code{YEAR} Type +When @code{mysqld} starts, all grant table contents are read into memory and +become effective at that point. -The @code{YEAR} type is a 1-byte type used for representing years. +Modifications to the grant tables that you perform using @code{GRANT}, +@code{REVOKE}, or @code{SET PASSWORD} are noticed by the server immediately. -@strong{MySQL} retrieves and displays @code{YEAR} values in @code{YYYY} -format. The range is @code{1901} to @code{2155}. +If you modify the grant tables manually (using @code{INSERT}, @code{UPDATE}, +etc.), you should execute a @code{FLUSH PRIVILEGES} statement or run +@code{mysqladmin flush-privileges} or @code{mysqladmin reload} to tell the +server to reload the grant tables. Otherwise your changes will have @emph{no +effect} until you restart the server. If you change the grant tables manually +but forget to reload the privileges, you will be wondering why your changes +don't seem to make any difference! -You can specify @code{YEAR} values in a variety of formats: +When the server notices that the grant tables have been changed, +existing client connections are affected as follows: @itemize @bullet @item -As a four-digit string in the range @code{'1901'} to @code{'2155'}. +Table and column privilege changes take effect with the client's next +request. @item -As a four-digit number in the range @code{1901} to @code{2155}. +Database privilege changes take effect at the next @code{USE db_name} +command. +@end itemize + +Global privilege changes and password changes take effect the next time the +client connects. + +@cindex privileges, default +@cindex default, privileges +@cindex root password +@cindex superuser +@cindex users, root +@cindex anonymous user +@cindex password, root user +@node Default privileges, Adding users, Privilege changes, Privilege system +@section Setting Up the Initial MySQL Privileges + +After installing @strong{MySQL}, you set up the initial access privileges by +running @code{scripts/mysql_install_db}. +@xref{Quick install}. +The @code{mysql_install_db} script starts up the @code{mysqld} +server, then initializes the grant tables to contain the following set +of privileges: +@itemize @bullet @item -As a two-digit string in the range @code{'00'} to @code{'99'}. Values in the -ranges @code{'00'} to @code{'69'} and @code{'70'} to @code{'99'} are -converted to @code{YEAR} values in the ranges @code{2000} to @code{2069} and -@code{1970} to @code{1999}. +The @strong{MySQL} @code{root} user is created as a superuser who can do +anything. Connections must be made from the local host. + +@strong{NOTE:} +The initial @code{root} password is empty, so anyone can connect as @code{root} +@emph{without a password} and be granted all privileges. @item -As a two-digit number in the range @code{1} to @code{99}. Values in the -ranges @code{1} to @code{69} and @code{70} to @code{99} are converted to -@code{YEAR} values in the ranges @code{2001} to @code{2069} and @code{1970} -to @code{1999}. Note that the range for two-digit numbers is slightly -different than the range for two-digit strings, because you cannot specify zero -directly as a number and have it be interpreted as @code{2000}. You -@emph{must} specify it as a string @code{'0'} or @code{'00'} or it will be -interpreted as @code{0000}. +@cindex anonymous user +An anonymous user is created that can do anything with databases that have a +name of @code{'test'} or starting with @code{'test_'}. Connections must be +made from the local host. This means any local user can connect without a +password and be treated as the anonymous user. @item -As the result of a function that returns a value that is acceptable -in a @code{YEAR} context, such as @code{NOW()}. +Other privileges are denied. For example, normal users can't use +@code{mysqladmin shutdown} or @code{mysqladmin processlist}. @end itemize -Illegal @code{YEAR} values are converted to @code{0000}. - -@cindex types, strings -@cindex string types -@node String types, Choosing types, Date and time types, Column types -@subsection String Types +@strong{NOTE:} The default privileges are different for Windows. +@xref{Windows running}. -@tindex CHAR -@tindex VARCHAR -@menu -* CHAR:: The @code{CHAR} and @code{VARCHAR} types -* BLOB:: The @code{BLOB} and @code{TEXT} types -* ENUM:: The @code{ENUM} type -* SET:: The @code{SET} type -@end menu +Because your installation is initially wide open, one of the first things you +should do is specify a password for the @strong{MySQL} +@code{root} user. You can do this as follows (note that you specify the +password using the @code{PASSWORD()} function): -The string types are @code{CHAR}, @code{VARCHAR}, @code{BLOB}, @code{TEXT}, -@code{ENUM}, and @code{SET}. This section describes how these types work, -their storage requirements, and how to use them in your queries. +@example +shell> mysql -u root mysql +mysql> UPDATE user SET Password=PASSWORD('new_password') + WHERE user='root'; +mysql> FLUSH PRIVILEGES; +@end example -@node CHAR, BLOB, String types, String types -@subsubsection The @code{CHAR} and @code{VARCHAR} Types +You can, in @strong{MySQL} Version 3.22 and above, use the @code{SET PASSWORD} +statement: -The @code{CHAR} and @code{VARCHAR} types are similar, but differ in the -way they are stored and retrieved. +@example +shell> mysql -u root mysql +mysql> SET PASSWORD FOR root=PASSWORD('new_password'); +@end example -The length of a @code{CHAR} column is fixed to the length that you declare -when you create the table. The length can be any value between 1 and 255. -(As of @strong{MySQL} Version 3.23, the length of @code{CHAR} may be 0 to 255.) -When @code{CHAR} values are stored, they are right-padded with spaces to the -specified length. When @code{CHAR} values are retrieved, trailing spaces are -removed. +Another way to set the password is by using the @code{mysqladmin} command: -Values in @code{VARCHAR} columns are variable-length strings. You can -declare a @code{VARCHAR} column to be any length between 1 and 255, just as -for @code{CHAR} columns. However, in contrast to @code{CHAR}, @code{VARCHAR} -values are stored using only as many characters as are needed, plus one byte -to record the length. Values are not padded; instead, trailing spaces are -removed when values are stored. (This space removal differs from the ANSI -SQL specification.) +@example +shell> mysqladmin -u root password new_password +@end example -If you assign a value to a @code{CHAR} or @code{VARCHAR} column that -exceeds the column's maximum length, the value is truncated to fit. +Only users with write/update access to the @code{mysql} database can change the +password for others users. All normal users (not anonymous ones) can only +change their own password with either of the above commands or with +@code{SET PASSWORD=PASSWORD('new password')}. -The table below illustrates the differences between the two types of columns -by showing the result of storing various string values into @code{CHAR(4)} -and @code{VARCHAR(4)} columns: +Note that if you update the password in the @code{user} table directly using +the first method, you must tell the server to re-read the grant tables (with +@code{FLUSH PRIVILEGES}), because the change will go unnoticed otherwise. -@c Need to use @(space) to make sure second column values retain spacing -@c in output for table below. +Once the @code{root} password has been set, thereafter you must supply that +password when you connect to the server as @code{root}. -@multitable @columnfractions .2 .15 .2 .2 .25 -@item @strong{Value} @tab @code{CHAR(4)} @tab @strong{Storage required} @tab @code{VARCHAR(4)} @tab @strong{Storage required} -@item @code{''} @tab @code{'@ @ @ @ '} @tab 4 bytes @tab @code{''} @tab 1 byte -@item @code{'ab'} @tab @code{'ab@ @ '} @tab 4 bytes @tab @code{'ab'} @tab 3 bytes -@item @code{'abcd'} @tab @code{'abcd'} @tab 4 bytes @tab @code{'abcd'} @tab 5 bytes -@item @code{'abcdefgh'} @tab @code{'abcd'} @tab 4 bytes @tab @code{'abcd'} @tab 5 bytes -@end multitable +You may wish to leave the @code{root} password blank so that you don't need +to specify it while you perform additional setup or testing. However, be sure +to set it before using your installation for any real production work. -The values retrieved from the @code{CHAR(4)} and @code{VARCHAR(4)} columns -will be the same in each case, because trailing spaces are removed from -@code{CHAR} columns upon retrieval. +See the @code{scripts/mysql_install_db} script to see how it sets up +the default privileges. You can use this as a basis to see how to +add other users. -Values in @code{CHAR} and @code{VARCHAR} columns are sorted and compared -in case-insensitive fashion, unless the @code{BINARY} attribute was -specified when the table was created. The @code{BINARY} attribute means -that column values are sorted and compared in case-sensitive fashion -according to the ASCII order of the machine where the @strong{MySQL} -server is running. @code{BINARY} doesn't affect how the column is stored -or retrieved. +If you want the initial privileges to be different than those just described +above, you can modify @code{mysql_install_db} before you run it. -The @code{BINARY} attribute is sticky. This means that if a column marked -@code{BINARY} is used in an expression, the whole expression is compared as a -@code{BINARY} value. +@cindex grant tables, re-creating +@cindex re-creating, grant tables +To re-create the grant tables completely, remove all the @file{.frm}, +@file{.MYI}, and @file{.MYD} files in the directory containing the +@code{mysql} database. (This is the directory named @file{mysql} under +the database directory, which is listed when you run @code{mysqld +--help}.) Then run the @code{mysql_install_db} script, possibly after +editing it first to have the privileges you want. -@strong{MySQL} may silently change the type of a @code{CHAR} or @code{VARCHAR} -column at table creation time. -@xref{Silent column changes}. +@strong{NOTE:} For @strong{MySQL} versions older than Version 3.22.10, +you should NOT delete the @file{.frm} files. If you accidentally do this, +you should copy them back from your @strong{MySQL} distribution before +running @code{mysql_install_db}. -@tindex BLOB -@tindex TEXT -@node BLOB, ENUM, CHAR, String types -@subsubsection The @code{BLOB} and @code{TEXT} Types +@cindex privileges, adding +@cindex adding, new user privileges +@cindex user privileges, adding +@findex GRANT statement +@findex statements, GRANT +@node Adding users, Passwords, Default privileges, Privilege system +@section Adding New Users to MySQL -A @code{BLOB} is a binary large object that can hold a variable amount of -data. The four @code{BLOB} types @code{TINYBLOB}, @code{BLOB}, -@code{MEDIUMBLOB}, and @code{LONGBLOB} differ only in the maximum length of -the values they can hold. -@xref{Storage requirements}. +You can add users two different ways: by using @code{GRANT} statements +or by manipulating the @strong{MySQL} grant tables directly. The +preferred method is to use @code{GRANT} statements, because they are +more concise and less error-prone. @xref{GRANT}. -The four @code{TEXT} types @code{TINYTEXT}, @code{TEXT}, @code{MEDIUMTEXT}, -and @code{LONGTEXT} correspond to the four @code{BLOB} types and have the -same maximum lengths and storage requirements. The only difference between -@code{BLOB} and @code{TEXT} types is that sorting and comparison is performed -in case-sensitive fashion for @code{BLOB} values and case-insensitive fashion -for @code{TEXT} values. In other words, a @code{TEXT} is a case-insensitive -@code{BLOB}. +There is also a lot of contributed programs like @code{phpmyadmin} that +can be used to create and administrate users. @xref{Contrib}. -If you assign a value to a @code{BLOB} or @code{TEXT} column that exceeds -the column type's maximum length, the value is truncated to fit. +The examples below show how to use the @code{mysql} client to set up new +users. These examples assume that privileges are set up according to the +defaults described in the previous section. This means that to make changes, +you must be on the same machine where @code{mysqld} is running, you must +connect as the @strong{MySQL} @code{root} user, and the @code{root} user must +have the @strong{insert} privilege for the @code{mysql} database and the +@strong{reload} administrative privilege. Also, if you have changed the +@code{root} user password, you must specify it for the @code{mysql} commands +below. -In most respects, you can regard a @code{TEXT} column as a @code{VARCHAR} -column that can be as big as you like. Similarly, you can regard a -@code{BLOB} column as a @code{VARCHAR BINARY} column. The differences are: +You can add new users by issuing @code{GRANT} statements: -@itemize @bullet -@item -You can have indexes on @code{BLOB} and @code{TEXT} columns with -@strong{MySQL} Version 3.23.2 and newer. Older versions of -@strong{MySQL} did not support this. +@example +shell> mysql --user=root mysql +mysql> GRANT ALL PRIVILEGES ON *.* TO monty@@localhost + IDENTIFIED BY 'some_pass' WITH GRANT OPTION; +mysql> GRANT ALL PRIVILEGES ON *.* TO monty@@"%" + IDENTIFIED BY 'some_pass' WITH GRANT OPTION; +mysql> GRANT RELOAD,PROCESS ON *.* TO admin@@localhost; +mysql> GRANT USAGE ON *.* TO dummy@@localhost; +@end example -@item -There is no trailing-space removal for @code{BLOB} and @code{TEXT} columns -when values are stored, as there is for @code{VARCHAR} columns. +These @code{GRANT} statements set up three new users: -@item -@cindex default values, @code{BLOB} and @code{TEXT} columns -@cindex @code{BLOB} columns, default values -@cindex @code{TEXT} columns, default values -@code{BLOB} and @code{TEXT} columns cannot have @code{DEFAULT} values. -@end itemize +@table @code +@item monty +A full superuser who can connect to the server from anywhere, but who must +use a password @code{'some_pass'} to do so. Note that we must issue +@code{GRANT} statements for both @code{monty@@localhost} and +@code{monty@@"%"}. If we don't add the entry with @code{localhost}, the +anonymous user entry for @code{localhost} that is created by +@code{mysql_install_db} will take precedence when we connect from the local +host, because it has a more specific @code{Host} field value and thus comes +earlier in the @code{user} table sort order. -@strong{MyODBC} defines @code{BLOB} values as @code{LONGVARBINARY} and -@code{TEXT} values as @code{LONGVARCHAR}. +@item admin +A user who can connect from @code{localhost} without a password and who is +granted the @strong{reload} and @strong{process} administrative privileges. +This allows the user to execute the @code{mysqladmin reload}, +@code{mysqladmin refresh}, and @code{mysqladmin flush-*} commands, as well as +@code{mysqladmin processlist} . No database-related privileges are granted. +(They can be granted later by issuing additional @code{GRANT} statements.) -Because @code{BLOB} and @code{TEXT} values may be extremely long, you -may run up against some constraints when using them: +@item dummy +A user who can connect without a password, but only from the local host. The +global privileges are all set to @code{'N'} --- the @code{USAGE} privilege +type allows you to create a user with no privileges. It is assumed that you +will grant database-specific privileges later. +@end table -@itemize @bullet -@item -If you want to use @code{GROUP BY} or @code{ORDER BY} on a @code{BLOB} or -@code{TEXT} column, you must convert the column value into a fixed-length -object. The standard way to do this is with the @code{SUBSTRING} -function. For example: +@findex INSERT statement, grant privileges +@findex statements, INSERT +You can also add the same user access information directly by issuing +@code{INSERT} statements and then telling the server to reload the grant +tables: @example -mysql> select comment from tbl_name,substring(comment,20) as substr - ORDER BY substr; +shell> mysql --user=root mysql +mysql> INSERT INTO user VALUES('localhost','monty',PASSWORD('some_pass'), + 'Y','Y','Y','Y','Y','Y','Y','Y','Y','Y','Y','Y','Y','Y'); +mysql> INSERT INTO user VALUES('%','monty',PASSWORD('some_pass'), + 'Y','Y','Y','Y','Y','Y','Y','Y','Y','Y','Y','Y','Y','Y'); +mysql> INSERT INTO user SET Host='localhost',User='admin', + Reload_priv='Y', Process_priv='Y'; +mysql> INSERT INTO user (Host,User,Password) + VALUES('localhost','dummy',''); +mysql> FLUSH PRIVILEGES; @end example -If you don't do this, only the first @code{max_sort_length} bytes of the -column are used when sorting. The default value of @code{max_sort_length} is -1024; this value can be changed using the @code{-O} option when starting the -@code{mysqld} server. You can group on an expression involving @code{BLOB} or -@code{TEXT} values by specifying the column position or by using an alias: - -@example -mysql> select id,substring(blob_col,1,100) from tbl_name - GROUP BY 2; -mysql> select id,substring(blob_col,1,100) as b from tbl_name - GROUP BY b; -@end example +Depending on your @strong{MySQL} version, you may have to use a different +number of @code{'Y'} values above (versions prior to Version 3.22.11 had fewer +privilege columns). For the @code{admin} user, the more readable extended +@code{INSERT} syntax that is available starting with Version 3.22.11 is used. -@item -The maximum size of a @code{BLOB} or @code{TEXT} object is determined by its -type, but the largest value you can actually transmit between the client and -server is determined by the amount of available memory and the size of the -communications buffers. You can change the message buffer size, but you must -do so on both the server and client ends. @xref{Server parameters}. -@end itemize +Note that to set up a superuser, you need only create a @code{user} table +entry with the privilege fields set to @code{'Y'}. No @code{db} or +@code{host} table entries are necessary. -Note that each @code{BLOB} or @code{TEXT} value is represented -internally by a separately allocated object. This is in contrast to all -other column types, for which storage is allocated once per column when -the table is opened. +The privilege columns in the @code{user} table were not set explicitly in the +last @code{INSERT} statement (for the @code{dummy} user), so those columns +are assigned the default value of @code{'N'}. This is the same thing that +@code{GRANT USAGE} does. -@tindex ENUM -@node ENUM, SET, BLOB, String types -@subsubsection The @code{ENUM} Type +The following example adds a user @code{custom} who can connect from hosts +@code{localhost}, @code{server.domain}, and @code{whitehouse.gov}. He wants +to access the @code{bankaccount} database only from @code{localhost}, +the @code{expenses} database only from @code{whitehouse.gov}, and +the @code{customer} database from all three hosts. He wants +to use the password @code{stupid} from all three hosts. -An @code{ENUM} is a string object whose value normally is chosen from a list -of allowed values that are enumerated explicitly in the column specification -at table creation time. +To set up this user's privileges using @code{GRANT} statements, run these +commands: -The value may also be the empty string (@code{""}) or @code{NULL} under -certain circumstances: +@example +shell> mysql --user=root mysql +mysql> GRANT SELECT,INSERT,UPDATE,DELETE,CREATE,DROP + ON bankaccount.* + TO custom@@localhost + IDENTIFIED BY 'stupid'; +mysql> GRANT SELECT,INSERT,UPDATE,DELETE,CREATE,DROP + ON expenses.* + TO custom@@whitehouse.gov + IDENTIFIED BY 'stupid'; +mysql> GRANT SELECT,INSERT,UPDATE,DELETE,CREATE,DROP + ON customer.* + TO custom@@'%' + IDENTIFIED BY 'stupid'; +@end example -@itemize @bullet -@item -If you insert an invalid value into an @code{ENUM} (that is, a string not -present in the list of allowed values), the empty string is inserted -instead as a special error value. +The reason that we do to grant statements for the user 'custom' is that +we want the give the user access to @strong{MySQL} both from the local +machine with Unix sockets and from the remote machine 'whitehouse.gov' +over TCP/IP. -@item -If an @code{ENUM} is declared @code{NULL}, @code{NULL} is also a legal value -for the column, and the default value is @code{NULL}. If an @code{ENUM} is -declared @code{NOT NULL}, the default value is the first element of the -list of allowed values. -@end itemize +To set up the user's privileges by modifying the grant tables directly, +run these commands (note the @code{FLUSH PRIVILEGES} at the end): -Each enumeration value has an index: +@example +shell> mysql --user=root mysql +mysql> INSERT INTO user (Host,User,Password) + VALUES('localhost','custom',PASSWORD('stupid')); +mysql> INSERT INTO user (Host,User,Password) + VALUES('server.domain','custom',PASSWORD('stupid')); +mysql> INSERT INTO user (Host,User,Password) + VALUES('whitehouse.gov','custom',PASSWORD('stupid')); +mysql> INSERT INTO db + (Host,Db,User,Select_priv,Insert_priv,Update_priv,Delete_priv, + Create_priv,Drop_priv) + VALUES + ('localhost','bankaccount','custom','Y','Y','Y','Y','Y','Y'); +mysql> INSERT INTO db + (Host,Db,User,Select_priv,Insert_priv,Update_priv,Delete_priv, + Create_priv,Drop_priv) + VALUES + ('whitehouse.gov','expenses','custom','Y','Y','Y','Y','Y','Y'); +mysql> INSERT INTO db + (Host,Db,User,Select_priv,Insert_priv,Update_priv,Delete_priv, + Create_priv,Drop_priv) + VALUES('%','customer','custom','Y','Y','Y','Y','Y','Y'); +mysql> FLUSH PRIVILEGES; +@end example -@itemize @bullet -@item -Values from the list of allowable elements in the column specification are -numbered beginning with 1. +The first three @code{INSERT} statements add @code{user} table entries that +allow user @code{custom} to connect from the various hosts with the given +password, but grant no permissions to him (all privileges are set to the +default value of @code{'N'}). The next three @code{INSERT} statements add +@code{db} table entries that grant privileges to @code{custom} for the +@code{bankaccount}, @code{expenses}, and @code{customer} databases, but only +when accessed from the proper hosts. As usual, when the grant tables are +modified directly, the server must be told to reload them (with +@code{FLUSH PRIVILEGES}) so that the privilege changes take effect. -@item -The index value of the empty string error value is 0. This means that you -can use the following @code{SELECT} statement to find rows into which invalid -@code{ENUM} values were assigned: +If you want to give a specific user access from any machine in a given +domain, you can issue a @code{GRANT} statement like the following: @example -mysql> SELECT * FROM tbl_name WHERE enum_col=0; +mysql> GRANT ... + ON *.* + TO myusername@@"%.mydomainname.com" + IDENTIFIED BY 'mypassword'; @end example -@item -The index of the @code{NULL} value is @code{NULL}. -@end itemize +To do the same thing by modifying the grant tables directly, do this: -For example, a column specified as @code{ENUM("one", "two", "three")} can -have any of the values shown below. The index of each value is also shown: +@example +mysql> INSERT INTO user VALUES ('%.mydomainname.com', 'myusername', + PASSWORD('mypassword'),...); +mysql> FLUSH PRIVILEGES; +@end example -@multitable @columnfractions .2 .8 -@item @strong{Value} @tab @strong{Index} -@item @code{NULL} @tab @code{NULL} -@item @code{""} @tab 0 -@item @code{"one"} @tab 1 -@item @code{"two"} @tab 2 -@item @code{"three"} @tab 3 -@end multitable +You can also use @code{xmysqladmin}, @code{mysql_webadmin}, and even +@code{xmysql} to insert, change, and update values in the grant tables. +You can find these utilities in the +@uref{http://www.mysql.com/Downloads/Contrib/,Contrib directory of the @strong{MySQL} +Website}. -An enumeration can have a maximum of 65535 elements. +@cindex passwords, setting +@findex PASSWORD() +@findex SET PASSWORD statement +@cindex setting, passwords +@node Passwords, Access denied, Adding users, Privilege system +@section Setting Up Passwords -Lettercase is irrelevant when you assign values to an @code{ENUM} column. -However, values retrieved from the column later have lettercase matching the -values that were used to specify the allowable values at table creation time. +In most cases you should use @code{GRANT} to set up your users/passwords, +so the following only applies for advanced users. @xref{GRANT, , @code{GRANT}}. -If you retrieve an @code{ENUM} in a numeric context, the column value's -index is returned. For example, you can retrieve numeric values from -an @code{ENUM} column like this: +The examples in the preceding sections illustrate an important principle: +when you store a non-empty password using @code{INSERT} or @code{UPDATE} +statements, you must use the @code{PASSWORD()} function to encrypt it. This +is because the @code{user} table stores passwords in encrypted form, not as +plaintext. If you forget that fact, you are likely to attempt to set +passwords like this: @example -mysql> SELECT enum_col+0 FROM tbl_name; +shell> mysql -u root mysql +mysql> INSERT INTO user (Host,User,Password) + VALUES('%','jeffrey','biscuit'); +mysql> FLUSH PRIVILEGES; @end example -If you store a number into an @code{ENUM}, the number is treated as an -index, and the value stored is the enumeration member with that index. -(However, this will not work with @code{LOAD DATA}, which treats all -input as strings.) - -@code{ENUM} values are sorted according to the order in which the enumeration -members were listed in the column specification. (In other words, -@code{ENUM} values are sorted according to their index numbers.) For -example, @code{"a"} sorts before @code{"b"} for @code{ENUM("a", "b")}, but -@code{"b"} sorts before @code{"a"} for @code{ENUM("b", "a")}. The empty -string sorts before non-empty strings, and @code{NULL} values sort before -all other enumeration values. - -If you want to get all possible values for an @code{ENUM} column, you should -use: @code{SHOW COLUMNS FROM table_name LIKE enum_column_name} and parse -the @code{ENUM} definition in the second column. - -@tindex SET -@node SET, , ENUM, String types -@subsubsection The @code{SET} Type +The result is that the plaintext value @code{'biscuit'} is stored as the +password in the @code{user} table. When the user @code{jeffrey} attempts to +connect to the server using this password, the @code{mysql} client encrypts +it with @code{PASSWORD()} and sends the result to the server. The server +compares the value in the @code{user} table (the encrypted value of +@code{'biscuit'}) to the encrypted password (which is @emph{not} +@code{'biscuit'}). The comparison fails and the server rejects the +connection: -A @code{SET} is a string object that can have zero or more values, each of -which must be chosen from a list of allowed values specified when the table -is created. @code{SET} column values that consist of multiple set members -are specified with members separated by commas (@samp{,}). A consequence of -this is that @code{SET} member values cannot themselves contain commas. +@example +shell> mysql -u jeffrey -pbiscuit test +Access denied +@end example -For example, a column specified as @code{SET("one", "two") NOT NULL} can have -any of these values: +Passwords must be encrypted when they are inserted in the @code{user} +table, so the @code{INSERT} statement should have been specified like this +instead: @example -"" -"one" -"two" -"one,two" +mysql> INSERT INTO user (Host,User,Password) + VALUES('%','jeffrey',PASSWORD('biscuit')); @end example +You must also use the @code{PASSWORD()} function when you use @code{SET +PASSWORD} statements: -A @code{SET} can have a maximum of 64 different members. +@example +mysql> SET PASSWORD FOR jeffrey@@"%" = PASSWORD('biscuit'); +@end example -@strong{MySQL} stores @code{SET} values numerically, with the low-order bit -of the stored value corresponding to the first set member. If you retrieve a -@code{SET} value in a numeric context, the value retrieved has bits set -corresponding to the set members that make up the column value. For example, -you can retrieve numeric values from a @code{SET} column like this: +If you set passwords using the @code{GRANT ... IDENTIFIED BY} statement +or the @code{mysqladmin password} command, the @code{PASSWORD()} function +is unnecessary. They both take care of encrypting the password for you, +so you would specify a password of @code{'biscuit'} like this: @example -mysql> SELECT set_col+0 FROM tbl_name; +mysql> GRANT USAGE ON *.* TO jeffrey@@"%" IDENTIFIED BY 'biscuit'; @end example -If a number is stored into a @code{SET} column, the bits that -are set in the binary representation of the number determine the -set members in the column value. Suppose a column is specified as -@code{SET("a","b","c","d")}. Then the members have the following bit -values: +or -@multitable @columnfractions .2 .2 .6 -@item @code{SET} @strong{member} @tab @strong{Decimal value} @tab @strong{Binary value} -@item @code{a} @tab @code{1} @tab @code{0001} -@item @code{b} @tab @code{2} @tab @code{0010} -@item @code{c} @tab @code{4} @tab @code{0100} -@item @code{d} @tab @code{8} @tab @code{1000} -@end multitable +@example +shell> mysqladmin -u jeffrey password biscuit +@end example -If you assign a value of @code{9} to this column, that is @code{1001} in -binary, so the first and fourth @code{SET} value members @code{"a"} and -@code{"d"} are selected and the resulting value is @code{"a,d"}. +@strong{NOTE:} @code{PASSWORD()} does not perform password encryption in the +same way that Unix passwords are encrypted. You should not assume that if +your Unix password and your @strong{MySQL} password are the same, that +@code{PASSWORD()} will result in the same encrypted value as is stored in the +Unix password file. @xref{User names}. -For a value containing more than one @code{SET} element, it does not matter -what order the elements are listed in when you insert the value. It also -does not matter how many times a given element is listed in the value. -When the value is retrieved later, each element in the value will appear -once, with elements listed according to the order in which they were -specified at table creation time. For example, if a column is specified as -@code{SET("a","b","c","d")}, then @code{"a,d"}, @code{"d,a"}, and -@code{"d,a,a,d,d"} will all appear as @code{"a,d"} when retrieved. +@node Access denied, , Passwords, Privilege system +@section Causes of @code{Access denied} Errors -@code{SET} values are sorted numerically. @code{NULL} values sort before -non-@code{NULL} @code{SET} values. +If you encounter @code{Access denied} errors when you try to connect to the +@strong{MySQL} server, the list below indicates some courses of +action you can take to correct the problem: -Normally, you perform a @code{SELECT} on a @code{SET} column using -the @code{LIKE} operator or the @code{FIND_IN_SET()} function: +@itemize @bullet +@item +After installing @strong{MySQL}, did you run the @code{mysql_install_db} +script to set up the initial grant table contents? If not, do so. +@xref{Default privileges}. Test the initial privileges by executing +this command: @example -mysql> SELECT * FROM tbl_name WHERE set_col LIKE '%value%'; -mysql> SELECT * FROM tbl_name WHERE FIND_IN_SET('value',set_col)>0; +shell> mysql -u root test @end example -But the following will also work: +The server should let you connect without error. You should also make sure +you have a file @file{user.MYD} in the @strong{MySQL} database directory. +Ordinarily, this is @file{PATH/var/mysql/user.MYD}, where @code{PATH} is the +pathname to the @strong{MySQL} installation root. + +@item +After a fresh installation, you should connect to the server and set up +your users and their access permissions: @example -mysql> SELECT * FROM tbl_name WHERE set_col = 'val1,val2'; -mysql> SELECT * FROM tbl_name WHERE set_col & 1; +shell> mysql -u root mysql @end example -The first of these statements looks for an exact match. The second looks -for values containing the first set member. - -If you want to get all possible values for a @code{SET} column, you should -use: @code{SHOW COLUMNS FROM table_name LIKE set_column_name} and parse -the @code{SET} definition in the second column. +The server should let you connect because the @strong{MySQL} @code{root} user +has no password initially. That is also a security risk, so setting the +@code{root} password is something you should do while you're setting up +your other @strong{MySQL} users. -@cindex types, columns -@cindex choosing types -@node Choosing types, Indexes, String types, Column types -@subsection Choosing the Right Type for a Column +If you try to connect as @code{root} and get this error: -For the most efficient use of storage, try to use the most precise type in -all cases. For example, if an integer column will be used for values in the -range between @code{1} and @code{99999}, @code{MEDIUMINT UNSIGNED} is the -best type. +@example +Access denied for user: '@@unknown' to database mysql +@end example -Accurate representation of monetary values is a common problem. In -@strong{MySQL}, you should use the @code{DECIMAL} type. This is stored as -a string, so no loss of accuracy should occur. If accuracy is not -too important, the @code{DOUBLE} type may also be good enough. +this means that you don't have an entry in the @code{user} table with a +@code{User} column value of @code{'root'} and that @code{mysqld} cannot +resolve the hostname for your client. In this case, you must restart the +server with the @code{--skip-grant-tables} option and edit your +@file{/etc/hosts} or @file{\windows\hosts} file to add an entry for your +host. -For high precision, you can always convert to a fixed-point type stored -in a @code{BIGINT}. This allows you to do all calculations with integers -and convert results back to floating-point values only when necessary. +@item +If you get an error like the following: -@cindex indexes, columns -@cindex columns, indexes -@cindex keys -@node Indexes, Multiple-column indexes, Choosing types, Column types -@subsection Column Indexes +@example +shell> mysqladmin -u root -pxxxx ver +Access denied for user: 'root@@localhost' (Using password: YES) +@end example -All @strong{MySQL} column types can be indexed. Use of indexes on the -relevant columns is the best way to improve the performance of @code{SELECT} -operations. +It means that you are using a wrong password. @xref{Passwords}. -The maximum number of keys and the maximum index length is defined per -table handler. @xref{Table types}. You can with all table handlers have -at least 16 keys and a total index length of at least 256 bytes. +If you have forgot the root password, you can restart @code{mysqld} with +@code{--skip-grant-tables} to change the password. You can find more +about this option later on in this manual section. -For @code{CHAR} and @code{VARCHAR} columns, you can index a prefix of a -column. This is much faster and requires less disk space than indexing the -whole column. The syntax to use in the @code{CREATE TABLE} statement to -index a column prefix looks like this: +If you get the above error even if you haven't specified a password, +this means that you a wrong password in some @code{my.ini} +file. @xref{Option files}. You can avoid using option files with the @code{--no-defaults} option, as follows: @example -KEY index_name (col_name(length)) +shell> mysqladmin --no-defaults -u root ver @end example -The example below creates an index for the first 10 characters of the -@code{name} column: +@item +@cindex @code{mysql_fix_privilege_tables} +If you updated an existing @strong{MySQL} installation from a version earlier +than Version 3.22.11 to Version 3.22.11 or later, did you run the +@code{mysql_fix_privilege_tables} script? If not, do so. The structure of +the grant tables changed with @strong{MySQL} Version 3.22.11 when the +@code{GRANT} statement became functional. -@example -mysql> CREATE TABLE test ( - name CHAR(200) NOT NULL, - KEY index_name (name(10))); -@end example +@item +If your privileges seem to have changed in the middle of a session, it may be +that a superuser has changed them. Reloading the grant tables affects new +client connections, but it also affects existing connections as indicated in +@ref{Privilege changes}. -For @code{BLOB} and @code{TEXT} columns, you must index a prefix of the -column. You cannot index the entire column. +@item +If you can't get your password to work, remember that you must use +the @code{PASSWORD()} function if you set the password with the +@code{INSERT}, @code{UPDATE}, or @code{SET PASSWORD} statements. The +@code{PASSWORD()} function is unnecessary if you specify the password using +the @code{GRANT ... INDENTIFIED BY} statement or the @code{mysqladmin +password} command. +@xref{Passwords}. -In @strong{MySQL} Version 3.23.23 or later, you can also create special -@strong{FULLTEXT} indexes. They are used for full-text search. Only the -@code{MyISAM} table type supports @code{FULLTEXT} indexes. They can be -created only from @code{VARCHAR} and @code{TEXT} columns. -Indexing always happens over the entire column and partial indexing is not -supported. See @ref{Fulltext Search} for details. +@item +@code{localhost} is a synonym for your local hostname, and is also the +default host to which clients try to connect if you specify no host +explicitly. However, connections to @code{localhost} do not work if you are +running on a system that uses MIT-pthreads (@code{localhost} connections are +made using Unix sockets, which are not supported by MIT-pthreads). To avoid +this problem on such systems, you should use the @code{--host} option to name +the server host explicitly. This will make a TCP/IP connection to the +@code{mysqld} server. In this case, you must have your real hostname in +@code{user} table entries on the server host. (This is true even if you are +running a client program on the same host as the server.) -@cindex multi-column indexes -@cindex indexes, multi-column -@cindex keys, multi-column -@node Multiple-column indexes, Other-vendor column types, Indexes, Column types -@subsection Multiple-column Indexes +@item +If you get an @code{Access denied} error when trying to connect to the +database with @code{mysql -u user_name db_name}, you may have a problem +with the @code{user} table. Check this by executing @code{mysql -u root +mysql} and issuing this SQL statement: -@strong{MySQL} can create indexes on multiple columns. An index may -consist of up to 15 columns. (On @code{CHAR} and @code{VARCHAR} columns you -can also use a prefix of the column as a part of an index). +@example +mysql> SELECT * FROM user; +@end example -A multiple-column index can be considered a sorted array containing values -that are created by concatenating the values of the indexed columns. +The result should include an entry with the @code{Host} and @code{User} +columns matching your computer's hostname and your @strong{MySQL} user name. -@strong{MySQL} uses multiple-column indexes in such a way that queries are -fast when you specify a known quantity for the first column of the index in a -@code{WHERE} clause, even if you don't specify values for the other columns. +@item +The @code{Access denied} error message will tell you who you are trying +to log in as, the host from which you are trying to connect, and whether +or not you were using a password. Normally, you should have one entry in +the @code{user} table that exactly matches the hostname and user name +that were given in the error message. For example if you get an error +message that contains @code{Using password: NO}, this means that you +tried to login without an password. -Suppose a table is created using the following specification: +@item +If you get the following error when you try to connect from a different host +than the one on which the @strong{MySQL} server is running, then there is no +row in the @code{user} table that matches that host: @example -mysql> CREATE TABLE test ( - id INT NOT NULL, - last_name CHAR(30) NOT NULL, - first_name CHAR(30) NOT NULL, - PRIMARY KEY (id), - INDEX name (last_name,first_name)); +Host ... is not allowed to connect to this MySQL server @end example -Then the index @code{name} is an index over @code{last_name} and -@code{first_name}. The index will be used for queries that specify -values in a known range for @code{last_name}, or for both @code{last_name} -and @code{first_name}. -Therefore, the @code{name} index will be used in the following queries: - -@example -mysql> SELECT * FROM test WHERE last_name="Widenius"; +You can fix this by using the command-line tool @code{mysql} (on the +server host!) to add a row to the @code{user}, @code{db}, or @code{host} +table for the user/hostname combination from which you are trying to +connect and then execute @code{mysqladmin flush-privileges}. If you are +not running @strong{MySQL} Version 3.22 and you don't know the IP number or +hostname of the machine from which you are connecting, you should put an +entry with @code{'%'} as the @code{Host} column value in the @code{user} +table and restart @code{mysqld} with the @code{--log} option on the +server machine. After trying to connect from the client machine, the +information in the @strong{MySQL} log will indicate how you really did +connect. (Then replace the @code{'%'} in the @code{user} table entry +with the actual hostname that shows up in the log. Otherwise, you'll +have a system that is insecure.) -mysql> SELECT * FROM test WHERE last_name="Widenius" - AND first_name="Michael"; +Another reason for this error on Linux is that you are using a binary +@strong{MySQL} version that is compiled with a different glibc version +than the one you are using. In this case you should either upgrade your +OS/glibc or download the source @strong{MySQL} version and compile this +yourself. A source RPM is normally trivial to compile and install, so +this isn't a big problem. -mysql> SELECT * FROM test WHERE last_name="Widenius" - AND (first_name="Michael" OR first_name="Monty"); +@item +If you get an error message where the hostname is not shown or where the +hostname is an IP, even if you try to connect with a hostname: -mysql> SELECT * FROM test WHERE last_name="Widenius" - AND first_name >="M" AND first_name < "N"; +@example +shell> mysqladmin -u root -pxxxx -h some-hostname ver +Access denied for user: 'root@' (Using password: YES) @end example -However, the @code{name} index will NOT be used in the following queries: +This means that @strong{MySQL} got some error when trying to resolve the +IP to a hostname. In this case you can execute @code{mysqladmin +flush-hosts} to reset the internal DNS cache. @xref{DNS}. -@example -mysql> SELECT * FROM test WHERE first_name="Michael"; +Some permanent solutions are: -mysql> SELECT * FROM test WHERE last_name="Widenius" - OR first_name="Michael"; -@end example +@itemize @minus +@item +Try to find out what is wrong with your DNS server and fix this. +@item +Specify IPs instead of hostnames in the @strong{MySQL} privilege tables. +@item +Start @code{mysqld} with @code{--skip-name-resolve}. +@item +Start @code{mysqld} with @code{--skip-host-cache}. +@item +Connect to @code{localhost} if you are running the server and the client +on the same machine. +@item +Put the client machine names in @code{/etc/hosts}. +@end itemize -For more information on the manner in which @strong{MySQL} uses indexes to -improve query performance, see @ref{MySQL indexes, , @strong{MySQL} -indexes}. +@item +If @code{mysql -u root test} works but @code{mysql -h your_hostname -u root +test} results in @code{Access denied}, then you may not have the correct name +for your host in the @code{user} table. A common problem here is that the +@code{Host} value in the user table entry specifies an unqualified hostname, +but your system's name resolution routines return a fully qualified domain +name (or vice-versa). For example, if you have an entry with host +@code{'tcx'} in the @code{user} table, but your DNS tells @strong{MySQL} that +your hostname is @code{'tcx.subnet.se'}, the entry will not work. Try adding +an entry to the @code{user} table that contains the IP number of your host as +the @code{Host} column value. (Alternatively, you could add an entry to the +@code{user} table with a @code{Host} value that contains a wild card---for +example, @code{'tcx.%'}. However, use of hostnames ending with @samp{%} is +@emph{insecure} and is @emph{not} recommended!) -@cindex types, portability -@cindex portability, types -@cindex columns, other types -@node Other-vendor column types, , Multiple-column indexes, Column types -@subsection Using Column Types from Other Database Engines +@item +If @code{mysql -u user_name test} works but @code{mysql -u user_name +other_db_name} doesn't work, you don't have an entry for @code{other_db_name} +listed in the @code{db} table. -To make it easier to use code written for SQL implementations from other -vendors, @strong{MySQL} maps column types as shown in the table below. These -mappings make it easier to move table definitions from other database engines -to @strong{MySQL}: +@item +If @code{mysql -u user_name db_name} works when executed on the server +machine, but @code{mysql -u host_name -u user_name db_name} doesn't work when +executed on another client machine, you don't have the client machine listed +in the @code{user} table or the @code{db} table. -@multitable @columnfractions .4 .6 -@item @strong{Other vendor type} @tab @strong{MySQL type} -@item @code{BINARY(NUM)} @tab @code{CHAR(NUM) BINARY} -@item @code{CHAR VARYING(NUM)} @tab @code{VARCHAR(NUM)} -@item @code{FLOAT4} @tab @code{FLOAT} -@item @code{FLOAT8} @tab @code{DOUBLE} -@item @code{INT1} @tab @code{TINYINT} -@item @code{INT2} @tab @code{SMALLINT} -@item @code{INT3} @tab @code{MEDIUMINT} -@item @code{INT4} @tab @code{INT} -@item @code{INT8} @tab @code{BIGINT} -@item @code{LONG VARBINARY} @tab @code{MEDIUMBLOB} -@item @code{LONG VARCHAR} @tab @code{MEDIUMTEXT} -@item @code{MIDDLEINT} @tab @code{MEDIUMINT} -@item @code{VARBINARY(NUM)} @tab @code{VARCHAR(NUM) BINARY} -@end multitable +@item +If you can't figure out why you get @code{Access denied}, remove from the +@code{user} table all entries that have @code{Host} values containing +wild cards (entries that contain @samp{%} or @samp{_}). A very common error +is to insert a new entry with @code{Host}=@code{'%'} and +@code{User}=@code{'some user'}, thinking that this will allow you to specify +@code{localhost} to connect from the same machine. The reason that this +doesn't work is that the default privileges include an entry with +@code{Host}=@code{'localhost'} and @code{User}=@code{''}. Because that entry +has a @code{Host} value @code{'localhost'} that is more specific than +@code{'%'}, it is used in preference to the new entry when connecting from +@code{localhost}! The correct procedure is to insert a second entry with +@code{Host}=@code{'localhost'} and @code{User}=@code{'some_user'}, or to +remove the entry with @code{Host}=@code{'localhost'} and +@code{User}=@code{''}. -Column type mapping occurs at table creation time. If you create a table -with types used by other vendors and then issue a @code{DESCRIBE tbl_name} -statement, @strong{MySQL} reports the table structure using the equivalent -@strong{MySQL} types. +@item +If you get the following error, you may have a problem with the @code{db} or +@code{host} table: -@cindex functions for @code{SELECT} and @code{WHERE} clauses -@node Functions, CREATE DATABASE, Column types, Reference -@section Functions for Use in @code{SELECT} and @code{WHERE} Clauses +@example +Access to database denied +@end example -A @code{select_expression} or @code{where_definition} in a SQL statement -can consist of any expression using the functions described below. +If the entry selected from the @code{db} table has an empty value in the +@code{Host} column, make sure there are one or more corresponding entries in +the @code{host} table specifying which hosts the @code{db} table entry +applies to. -An expression that contains @code{NULL} always produces a @code{NULL} value -unless otherwise indicated in the documentation for the operators and -functions involved in the expression. +If you get the error when using the SQL commands @code{SELECT ... +INTO OUTFILE} or @code{LOAD DATA INFILE}, your entry in the @code{user} table +probably doesn't have the @strong{file} privilege enabled. -@strong{NOTE:} There must be no whitespace between a function name and the -parenthesis following it. This helps the @strong{MySQL} parser distinguish -between function calls and references to tables or columns that happen to -have the same name as a function. Spaces around arguments are permitted, -though. +@item +@cindex configuration files +@cindex environment variables +@tindex .my.cnf file +Remember that client programs will use connection parameters specified +in configuration files or environment variables. @xref{Environment +variables}. If a client seems to be sending the wrong default +connection parameters when you don't specify them on the command line, +check your environment and the @file{.my.cnf} file in your home +directory. You might also check the system-wide @strong{MySQL} +configuration files, though it is far less likely that client connection +parameters will be specified there. @xref{Option files}. If you get +@code{Access denied} when you run a client without any options, make +sure you haven't specified an old password in any of your option files! +@xref{Option files}. -You can force @strong{MySQL} to accept spaces after the function name by -starting @code{mysqld} with @code{--ansi} or using the -@code{CLIENT_IGNORE_SPACE} to @code{mysql_connect()}, but in this case all -function names will become reserved words. @xref{ANSI mode}. +@item +If you make changes to the grant tables directly (using an @code{INSERT} or +@code{UPDATE} statement) and your changes seem to be ignored, remember +that you must issue a @code{FLUSH PRIVILEGES} statement or execute a +@code{mysqladmin flush-privileges} command to cause the server to re-read +the privilege tables. Otherwise your changes have no effect until the +next time the server is restarted. Remember that after you set the +@code{root} password with an @code{UPDATE} command, you won't need to +specify it until after you flush the privileges, because the server +won't know you've changed the password yet! -@need 2000 -For the sake of brevity, examples display the output from the @code{mysql} -program in abbreviated form. So this: +@item +If you have access problems with a Perl, PHP, Python, or ODBC program, try to +connect to the server with @code{mysql -u user_name db_name} or @code{mysql +-u user_name -pyour_pass db_name}. If you are able to connect using the +@code{mysql} client, there is a problem with your program and not with the +access privileges. (Note that there is no space between @code{-p} and the +password; you can also use the @code{--password=your_pass} syntax to specify +the password. If you use the @code{-p} option alone, @strong{MySQL} will +prompt you for the password.) -@example -mysql> select MOD(29,9); -1 rows in set (0.00 sec) +@item +For testing, start the @code{mysqld} daemon with the +@code{--skip-grant-tables} option. Then you can change the @strong{MySQL} +grant tables and use the @code{mysqlaccess} script to check whether or not +your modifications have the desired effect. When you are satisfied with your +changes, execute @code{mysqladmin flush-privileges} to tell the @code{mysqld} +server to start using the new grant tables. @strong{Note:} Reloading the +grant tables overrides the @code{--skip-grant-tables} option. This allows +you to tell the server to begin using the grant tables again without bringing +it down and restarting it. -+-----------+ -| mod(29,9) | -+-----------+ -| 2 | -+-----------+ -@end example +@item +If everything else fails, start the @code{mysqld} daemon with a debugging +option (for example, @code{--debug=d,general,query}). This will print host and +user information about attempted connections, as well as information about +each command issued. @xref{Making trace files}. -is displayed like this: +@item +If you have any other problems with the @strong{MySQL} grant tables and +feel you must post the problem to the mailing list, always provide a +dump of the @strong{MySQL} grant tables. You can dump the tables with +the @code{mysqldump mysql} command. As always, post your problem using +the @code{mysqlbug} script. @xref{Bug reports}. In some cases you may need +to restart @code{mysqld} with @code{--skip-grant-tables} to run +@code{mysqldump}. +@end itemize -@example -mysql> select MOD(29,9); - -> 2 -@end example +@node Reference, Table types, Privilege system, Top +@chapter MySQL Language Reference @menu -* Grouping functions:: Grouping functions -* Arithmetic functions:: Normal arithmetic operations -* Bit functions:: Bit functions -* Logical functions:: Logical operations -* Comparison functions:: Comparison operators -* String comparison functions:: String comparison functions -* Casts:: Cast operators -* Control flow functions:: Control flow functions -* Mathematical functions:: Mathematical functions -* String functions:: String functions -* Date and time functions:: Date and time functions -* Miscellaneous functions:: Miscellaneous functions -* Group by functions:: Functions for @code{GROUP BY} clause +* Literals:: Literals: How to write strings and numbers +* Variables:: User variables +* Column types:: Column types +* Functions:: Functions +* CREATE DATABASE:: @code{CREATE DATABASE} syntax +* DROP DATABASE:: @code{DROP DATABASE} syntax +* CREATE TABLE:: @code{CREATE TABLE} syntax +* ALTER TABLE:: @code{ALTER TABLE} syntax +* RENAME TABLE:: @code{RENAME TABLE} syntax +* DROP TABLE:: @code{DROP TABLE} syntax +* OPTIMIZE TABLE:: @code{OPTIMIZE TABLE} syntax +* CHECK TABLE:: @code{CHECK TABLE} syntax +* BACKUP TABLE:: @code{BACKUP TABLE} syntax +* RESTORE TABLE:: @code{RESTORE TABLE} syntax +* ANALYZE TABLE:: @code{ANALYZE TABLE} syntax +* REPAIR TABLE:: @code{REPAIR TABLE} syntax +* DELETE:: @code{DELETE} syntax +* TRUNCATE:: @code{TRUNCATE} syntax +* SELECT:: @code{SELECT} syntax +* JOIN:: @code{JOIN} syntax +* INSERT:: @code{INSERT} syntax +* REPLACE:: @code{REPLACE} syntax +* LOAD DATA:: @code{LOAD DATA INFILE} syntax +* UPDATE:: @code{UPDATE} syntax +* USE:: @code{USE} syntax +* FLUSH:: @code{FLUSH} syntax (clearing caches) +* KILL:: @code{KILL} syntax +* SHOW:: @code{SHOW} syntax (Get information about tables, columns, ...) +* EXPLAIN:: @code{EXPLAIN} syntax (Get information about a @code{SELECT}) +* DESCRIBE:: @code{DESCRIBE} syntax (Get information about names of columns) +* COMMIT:: @code{BEGIN/COMMIT/ROLLBACK} syntax +* LOCK TABLES:: @code{LOCK TABLES/UNLOCK TABLES} syntax +* SET OPTION:: @code{SET OPTION} syntax +* SET TRANSACTION:: @code{SET TRANSACTION} syntax +* GRANT:: @code{GRANT} and @code{REVOKE} syntax +* CREATE INDEX:: @code{CREATE INDEX} syntax +* DROP INDEX:: @code{DROP INDEX} syntax +* Comments:: Comment syntax +* CREATE FUNCTION:: @code{CREATE FUNCTION} syntax +* Reserved words:: Is @strong{MySQL} picky about reserved words? @end menu -@cindex functions, grouping -@cindex grouping, expressions -@node Grouping functions, Arithmetic functions, Functions, Functions -@subsection Grouping Functions -@table @code - -@findex () (parentheses) -@findex parentheses ( and ) -@item ( ... ) -Parentheses. Use these to force the order of evaluation in an expression: -@example -mysql> select 1+2*3; - -> 7 -mysql> select (1+2)*3; - -> 9 -@end example -@end table - -@node Arithmetic functions, Bit functions, Grouping functions, Functions -@subsection Normal Arithmetic Operations - -The usual arithmetic operators are available. Note that in the case of -@samp{-}, @samp{+}, and @samp{*}, the result is calculated with -@code{BIGINT} (64-bit) precision if both arguments are integers! - -@cindex operations, arithmetic -@cindex arithmetic expressions -@table @code -@findex + (addition) -@findex addition (+) -@item + -Addition: -@example -mysql> select 3+5; - -> 8 -@end example +@strong{MySQL} has a very complex, but intuitive and easy to learn SQL +interface. This chapter describes the various commands, types, and functions +you will need to know in order to use @strong{MySQL} efficiently and +effectively. This chapter also serves as a reference to all functionality +included in @strong{MySQL}. In order to use this chapter effectively, you +may find it useful to refer to the various indexes. -@findex - (subtraction) -@findex subtraction (-) -@item - -Subtraction: -@example -mysql> select 3-5; - -> -2 -@end example +@cindex strings, defined +@cindex strings, escaping characters +@cindex literals +@cindex escape characters +@cindex backslash, escape character +@node Literals, Variables, Reference, Reference +@section Literals: How to Write Strings and Numbers -@findex * (multiplication) -@findex multiplication (*) -@item * -Multiplication: -@example -mysql> select 3*5; - -> 15 -mysql> select 18014398509481984*18014398509481984.0; - -> 324518553658426726783156020576256.0 -mysql> select 18014398509481984*18014398509481984; - -> 0 -@end example +@menu +* String syntax:: Strings +* Number syntax:: Numbers +* Hexadecimal values:: Hexadecimal values +* NULL values:: @code{NULL} values +* Legal names:: Database, Table, Index, Column, and Alias Names +@end menu -The result of the last expression is incorrect because the result of the integer -multiplication exceeds the 64-bit range of @code{BIGINT} calculations. +This section describes the various ways to write strings and numbers in +@strong{MySQL}. It also covers the various nuances and ``gotchas'' that +you may run into when dealing with these basic types in @strong{MySQL}. -@findex / (division) -@findex division (/) -@item / -Division: -@example -mysql> select 3/5; - -> 0.60 -@end example +@node String syntax, Number syntax, Literals, Literals +@subsection Strings -Division by zero produces a @code{NULL} result: +A string is a sequence of characters, surrounded by either single quote +(@samp{'}) or double quote (@samp{"}) characters (only the single quote +if you run in ANSI mode). Examples: @example -mysql> select 102/(1-1); - -> NULL +'a string' +"another string" @end example -A division will be calculated with @code{BIGINT} arithmetic only if performed -in a context where its result is converted to an integer! -@end table - -@findex arithmetic functions -@findex bit functions -@findex functions, arithmetic -@findex functions, bit -@node Bit functions, Logical functions, Arithmetic functions, Functions -@subsection Bit Functions - -@strong{MySQL} uses @code{BIGINT} (64-bit) arithmetic for bit operations, so -these operators have a maximum range of 64 bits. +Within a string, certain sequences have special meaning. Each of these +sequences begins with a backslash (@samp{\}), known as the @emph{escape +character}. @strong{MySQL} recognizes the following escape sequences: +@c these aren't really functions, but that's probably the most reasonable index @table @code -@findex | (bitwise OR) -@findex OR, bitwise -@item | -Bitwise OR: -@example -mysql> select 29 | 15; - -> 31 -@end example - -@findex & (bitwise AND) -@findex AND, bitwise -@item & -Bitwise AND: -@example -mysql> select 29 & 15; - -> 13 -@end example +@findex \0 (ASCII 0) +@findex NUL +@item \0 +An ASCII 0 (@code{NUL}) character. -@findex << (left shift) -@item << -Shifts a longlong (@code{BIGINT}) number to the left: -@example -mysql> select 1 << 2; - -> 4 -@end example +@findex \' (single quote) +@findex single quote (\') +@item \' +A single quote (@samp{'}) character. -@findex >> (right shift) -@item >> -Shifts a longlong (@code{BIGINT}) number to the right: -@example -mysql> select 4 >> 2; - -> 1 -@end example +@findex \" (double quote) +@findex double quote (\") +@item \" +A double quote (@samp{"}) character. -@findex ~ -@item ~ -Invert all bits: -@example -mysql> select 5 & ~1; - -> 4 -@end example +@findex \b (backspace) +@findex backspace (\b) +@item \b +A backspace character. -@findex BIT_COUNT() -@item BIT_COUNT(N) -Returns the number of bits that are set in the argument @code{N}: -@example -mysql> select BIT_COUNT(29); - -> 4 -@end example -@end table +@findex \n (newline) +@findex newline (\n) +@item \n +A newline character. -@findex Logical functions -@findex Functions, logical -@node Logical functions, Comparison functions, Bit functions, Functions -@subsection Logical Operations +@findex \r (carriage return) +@findex return (\r) +@findex carriage return (\r) +@item \r +A carriage return character. -All logical functions return @code{1} (TRUE), @code{0} (FALSE) or -@code{NULL} (unknown, which is in most cases the same as FALSE): +@findex \t (tab) +@findex tab (\t) +@item \t +A tab character. -@table @code -@findex NOT, logical -@findex ! (logical NOT) -@item NOT -@itemx ! -Logical NOT. Returns @code{1} if the argument is @code{0}, otherwise returns -@code{0}. -Exception: @code{NOT NULL} returns @code{NULL}: -@example -mysql> select NOT 1; - -> 0 -mysql> select NOT NULL; - -> NULL -mysql> select ! (1+1); - -> 0 -mysql> select ! 1+1; - -> 1 -@end example -The last example returns @code{1} because the expression evaluates -the same way as @code{(!1)+1}. +@findex \z (Control-Z) ASCII(26) +@findex (Control-Z) \z +@item \z +ASCII(26) (Control-Z). This character can be encoded to allow you to +go around the problem that ASCII(26) stands for END-OF-FILE on Windows. +(ASCII(26) will cause problems if you try to use +@code{mysql database < filename}). -@findex OR, logical -@findex || (logical OR) -@item OR -@itemx || -Logical OR. Returns @code{1} if either argument is not @code{0} and not -@code{NULL}: -@example -mysql> select 1 || 0; - -> 1 -mysql> select 0 || 0; - -> 0 -mysql> select 1 || NULL; - -> 1 +@findex \\ (escape) +@findex escape (\\) +@item \\ +A backslash (@samp{\}) character. -@end example +@findex % (wild card character) +@findex Wild card character (%) +@item \% +A @samp{%} character. This is used to search for literal instances of +@samp{%} in contexts where @samp{%} would otherwise be interpreted +as a wild-card character. @xref{String comparison functions}. -@findex AND, logical -@findex && (logical AND) -@item AND -@itemx && -Logical AND. Returns @code{0} if either argument is @code{0} or @code{NULL}, -otherwise returns @code{1}: -@example -mysql> select 1 && NULL; - -> 0 -mysql> select 1 && 0; - -> 0 -@end example +@findex _ (wild card character) +@findex Wild card character (_) +@item \_ +A @samp{_} character. This is used to search for literal instances of +@samp{_} in contexts where @samp{_} would otherwise be interpreted +as a wild-card character. @xref{String comparison functions}. @end table -@cindex casts -@cindex type conversions -@findex comparison operators -@node Comparison functions, String comparison functions, Logical functions, Functions -@subsection Comparison Operators - -Comparison operations result in a value of @code{1} (TRUE), @code{0} (FALSE), -or @code{NULL}. These functions work for both numbers and strings. Strings -are automatically converted to numbers and numbers to strings as needed (as -in Perl). +Note that if you use @samp{\%} or @samp{\_} in some string contexts, these +will return the strings @samp{\%} and @samp{\_} and not @samp{%} and +@samp{_}. -@strong{MySQL} performs comparisons using the following -rules: +@cindex quotes, in strings +@noindent +There are several ways to include quotes within a string: @itemize @bullet @item -If one or both arguments are @code{NULL}, the result of the comparison -is @code{NULL}, except for the @code{<=>} operator. - -@item -If both arguments in a comparison operation are strings, they are compared as -strings. - -@item -If both arguments are integers, they are compared as integers. +A @samp{'} inside a string quoted with @samp{'} may be written as @samp{''}. @item -Hexadecimal values are treated as binary strings if not compared to a number. +A @samp{"} inside a string quoted with @samp{"} may be written as @samp{""}. @item -@cindex ODBC compatibility -@cindex compatibility, with ODBC -If one of the arguments is a @code{TIMESTAMP} or @code{DATETIME} column and -the other argument is a constant, the constant is converted -to a timestamp before the comparison is performed. This is done to be more -ODBC-friendly. +You can precede the quote character with an escape character (@samp{\}). @item -In all other cases, the arguments are compared as floating-point (real) -numbers. +A @samp{'} inside a string quoted with @samp{"} needs no special treatment +and need not be doubled or escaped. In the same way, @samp{"} inside a +string quoted with @samp{'} needs no special treatment. @end itemize -By default, string comparisons are done in case-independent fashion using the -current character set (ISO-8859-1 Latin1 by default, which also works -excellently for English). - -The examples below illustrate conversion of strings to numbers for comparison -operations: +The @code{SELECT} statements shown below demonstrate how quoting and +escaping work: @example -mysql> SELECT 1 > '6x'; - -> 0 -mysql> SELECT 7 > '6x'; - -> 1 -mysql> SELECT 0 > 'x6'; - -> 0 -mysql> SELECT 0 = 'x6'; - -> 1 +mysql> SELECT 'hello', '"hello"', '""hello""', 'hel''lo', '\'hello'; ++-------+---------+-----------+--------+--------+ +| hello | "hello" | ""hello"" | hel'lo | 'hello | ++-------+---------+-----------+--------+--------+ + +mysql> SELECT "hello", "'hello'", "''hello''", "hel""lo", "\"hello"; ++-------+---------+-----------+--------+--------+ +| hello | 'hello' | ''hello'' | hel"lo | "hello | ++-------+---------+-----------+--------+--------+ + +mysql> SELECT "This\nIs\nFour\nlines"; ++--------------------+ +| This +Is +Four +lines | ++--------------------+ @end example +@cindex quoting binary data + +If you want to insert binary data into a @code{BLOB} column, the following +characters must be represented by escape sequences: @table @code -@findex = (equal) -@findex equal (=) -@item = -Equal: -@example -mysql> select 1 = 0; - -> 0 -mysql> select '0' = 0; - -> 1 -mysql> select '0.0' = 0; - -> 1 -mysql> select '0.01' = 0; - -> 0 -mysql> select '.01' = 0.01; - -> 1 -@end example +@item NUL +ASCII 0. You should represent this by @samp{\0} (a backslash and an ASCII @samp{0} character). +@item \ +ASCII 92, backslash. Represent this by @samp{\\}. +@item ' +ASCII 39, single quote. Represent this by @samp{\'}. +@item " +ASCII 34, double quote. Represent this by @samp{\"}. +@end table -@findex <> (not equal) -@findex not equal (<>) -@findex != (not equal) -@findex not equal (!=) -@item <> -@itemx != -Not equal: -@example -mysql> select '.01' <> '0.01'; - -> 1 -mysql> select .01 <> '0.01'; - -> 0 -mysql> select 'zapp' <> 'zappp'; - -> 1 -@end example +@cindex quoting +@cindex @code{BLOB}, inserting binary data +@findex mysql_escape_string() +@findex DBI->quote +If you write C code, you can use the C API function +@code{mysql_escape_string()} to escape characters for the @code{INSERT} +statement. @xref{C API function overview}. In Perl, you can use the +@code{quote} method of the @code{DBI} package to convert special +characters to the proper escape sequences. @xref{Perl DBI Class, , Perl +@code{DBI} Class}. -@findex <= (less than or equal) -@findex less than or equal (<=) -@item <= -Less than or equal: -@example -mysql> select 0.1 <= 2; - -> 1 -@end example +You should use an escape function on any string that might contain any of the +special characters listed above! -@findex < (less than) -@findex less than (<) -@item < -Less than: -@example -mysql> select 2 <= 2; - -> 1 -@end example +@cindex numbers +@cindex valid numbers, examples +@cindex integers +@cindex floats +@cindex negative values +@node Number syntax, Hexadecimal values, String syntax, Literals +@subsection Numbers -@findex >= (greater than or equal) -@findex greater than or equal (>=) -@item >= -Greater than or equal: -@example -mysql> select 2 >= 2; - -> 1 -@end example +Integers are represented as a sequence of digits. Floats use @samp{.} as a +decimal separator. Either type of number may be preceded by @samp{-} to +indicate a negative value. + +Examples of valid integers: -@findex > (greater than) -@findex greater than (>) -@item > -Greater than: @example -mysql> select 2 > 2; - -> 0 +1221 +0 +-32 @end example -@cindex @code{NULL}, testing for null -@findex <=> (Equal to) -@item <=> -Null safe equal: +Examples of valid floating-point numbers: + @example -mysql> select 1 <=> 1, NULL <=> NULL, 1 <=> NULL; - -> 1 1 0 +294.42 +-32032.6809e+10 +148.00 @end example -@cindex @code{NULL}, testing for null -@findex IS NULL -@findex IS NOT NULL -@item IS NULL -@itemx IS NOT NULL -Test whether or not a value is or is not @code{NULL}: -@example -mysql> select 1 IS NULL, 0 IS NULL, NULL IS NULL; - -> 0 0 1 -mysql> select 1 IS NOT NULL, 0 IS NOT NULL, NULL IS NOT NULL; - -> 1 1 0 -@end example +An integer may be used in a floating-point context; it is interpreted +as the equivalent floating-point number. -@findex BETWEEN ... AND -@item expr BETWEEN min AND max -If @code{expr} is greater than or equal to @code{min} and @code{expr} is -less than or equal to @code{max}, @code{BETWEEN} returns @code{1}, -otherwise it returns @code{0}. This is equivalent to the expression -@code{(min <= expr AND expr <= max)} if all the arguments are of the -same type. The first argument (@code{expr}) determines how the -comparison is performed as follows: +@tindex hexadecimal values +@node Hexadecimal values, NULL values, Number syntax, Literals +@subsection Hexadecimal Values -@itemize @bullet -@item -If @code{expr} is a @code{TIMESTAMP}, @code{DATE}, or @code{DATETIME} -column, @code{MIN()} and @code{MAX()} are formatted to the same format if -they are constants. -@item -If @code{expr} is a case-insensitive string expression, a case-insensitive -string comparison is done. -@item -If @code{expr} is a case-sensitive string expression, a case-sensitive -string comparison is done. -@item -If @code{expr} is an integer expression, an integer comparison is done. -@item -Otherwise, a floating-point (real) comparison is done. -@end itemize +@strong{MySQL} supports hexadecimal values. In number context these act +like an integer (64-bit precision). In string context these act like a binary +string where each pair of hex digits is converted to a character: @example -mysql> select 1 BETWEEN 2 AND 3; - -> 0 -mysql> select 'b' BETWEEN 'a' AND 'c'; - -> 1 -mysql> select 2 BETWEEN 2 AND '3'; - -> 1 -mysql> select 2 BETWEEN 2 AND 'x-3'; - -> 0 +mysql> SELECT 0xa+0; + -> 10 +mysql> select 0x5061756c; + -> Paul @end example -@findex IN -@item expr IN (value,...) -Returns @code{1} if @code{expr} is any of the values in the @code{IN} list, -else returns @code{0}. If all values are constants, then all values are -evaluated according to the type of @code{expr} and sorted. The search for the -item is then done using a binary search. This means @code{IN} is very quick -if the @code{IN} value list consists entirely of constants. If @code{expr} -is a case-sensitive string expression, the string comparison is performed in -case-sensitive fashion: +Hexadecimal strings are often used by ODBC to give values for BLOB columns. -@example -mysql> select 2 IN (0,3,5,'wefwf'); - -> 0 -mysql> select 'wefwf' IN (0,3,5,'wefwf'); - -> 1 -@end example +@tindex NULL value +@node NULL values, Legal names, Hexadecimal values, Literals +@subsection @code{NULL} Values -@findex NOT IN -@item expr NOT IN (value,...) -Same as @code{NOT (expr IN (value,...))}. +The @code{NULL} value means ``no data'' and is different from values such +as @code{0} for numeric types or the empty string for string types. +@xref{Problems with NULL, , Problems with @code{NULL}}. -@findex ISNULL() -@item ISNULL(expr) -If @code{expr} is @code{NULL}, @code{ISNULL()} returns @code{1}, otherwise -it returns @code{0}: -@example -mysql> select ISNULL(1+1); - -> 0 -mysql> select ISNULL(1/0); - -> 1 -@end example +@code{NULL} may be represented by @code{\N} when using the text file import +or export formats (@code{LOAD DATA INFILE}, @code{SELECT ... INTO OUTFILE}). +@xref{LOAD DATA, , @code{LOAD DATA}}. -Note that a comparison of @code{NULL} values using @code{=} will always be -false! +@cindex names +@cindex legal names +@cindex databases, names +@cindex tables, names +@cindex indexes, names +@cindex columns, names +@cindex aliases, names +@node Legal names, , NULL values, Literals +@subsection Database, Table, Index, Column, and Alias Names -@cindex @code{NULL}, testing for null -@findex COALESCE() -@item COALESCE(list) -Returns first non-@code{NULL} element in list: +@menu +* Name case sensitivity:: Case sensitivity in names +@end menu -@example -mysql> select COALESCE(NULL,1); - -> 1 -mysql> select COALESCE(NULL,NULL,NULL); - -> NULL -@end example +Database, table, index, column, and alias names all follow the same rules in +@strong{MySQL}. -@findex INTERVAL() -@item INTERVAL(N,N1,N2,N3,...) -Returns @code{0} if @code{N} < @code{N1}, @code{1} if @code{N} < @code{N2} -and so on. All arguments are treated as integers. It is required that -@code{N1} < @code{N2} < @code{N3} < @code{...} < @code{Nn} for this function -to work correctly. This is because a binary search is used (very fast): +@tindex identifiers, quoting +@tindex quoting of identifiers +@tindex ` +@tindex " +Note that the rules changed starting with @strong{MySQL} Version 3.23.6 when we +introduced quoting of identifiers (database, table, and column names) +with @samp{`}. @samp{"} will also work to quote identifiers if you run +in ANSI mode. @xref{ANSI mode}. + +@multitable @columnfractions .15 .15 .70 +@item @strong{Identifier} @tab @strong{Max length} @tab @strong{Allowed characters} +@item Database @tab 64 @tab Any character that is allowed in a directory name except @samp{/} or @samp{.}. +@item Table @tab 64 @tab Any character that is allowed in a file name, except @samp{/} or @samp{.}. +@item Column @tab 64 @tab All characters. +@item Alias @tab 255 @tab All characters. +@end multitable + +Note that in addition to the above, you can't have ASCII(0) or ASCII(255) or +the quoting character in an identifier. + +Note that if the identifier is a restricted word or contains special characters +you must always quote it with @code{`} when you use it: @example -mysql> select INTERVAL(23, 1, 15, 17, 30, 44, 200); - -> 3 -mysql> select INTERVAL(10, 1, 10, 100, 1000); - -> 2 -mysql> select INTERVAL(22, 23, 30, 44, 200); - -> 0 +SELECT * from `select` where `select`.id > 100; @end example -@end table -@findex string comparison functions -@findex functions, string comparison -@node String comparison functions, Casts, Comparison functions, Functions -@subsection String Comparison Functions +In previous versions of @strong{MySQL}, the name rules are as follows: -@cindex case sensitivity, in string comparisons -@cindex string comparisons, case sensitivity -Normally, if any expression in a string comparison is case sensitive, the -comparison is performed in case-sensitive fashion. +@itemize @bullet +@item +A name may consist of alphanumeric characters from the current character set +and also @samp{_} and @samp{$}. The default character set is ISO-8859-1 +Latin1; this may be changed with the @code{--default-character-set} option +to @code{mysqld}. +@xref{Character sets}. -@table @code -@findex LIKE -@item expr LIKE pat [ESCAPE 'escape-char'] -Pattern matching using -SQL simple regular expression comparison. Returns @code{1} (TRUE) or @code{0} -(FALSE). With @code{LIKE} you can use the following two wild-card characters -in the pattern: +@item +A name may start with any character that is legal in a name. In particular, +a name may start with a number (this differs from many other database +systems!). However, a name cannot consist @emph{only} of numbers. -@multitable @columnfractions .1 .9 -@item @code{%} @tab Matches any number of characters, even zero characters -@item @code{_} @tab Matches exactly one character -@end multitable +@item +You cannot use the @samp{.} character in names because it is used to extend the +format by which you can refer to columns (see immediately below). +@end itemize -@example -mysql> select 'David!' LIKE 'David_'; - -> 1 -mysql> select 'David!' LIKE '%D%v%'; - -> 1 -@end example +It is recommended that you do not use names like @code{1e}, because +an expression like @code{1e+1} is ambiguous. It may be interpreted as the +expression @code{1e + 1} or as the number @code{1e+1}. -To test for literal instances of a wild-card character, precede the character -with the escape character. If you don't specify the @code{ESCAPE} character, -@samp{\} is assumed: +In @strong{MySQL} you can refer to a column using any of the following forms: -@multitable @columnfractions .1 .9 -@item @code{\%} @tab Matches one @code{%} character -@item @code{\_} @tab Matches one @code{_} character +@multitable @columnfractions .35 .65 +@item @strong{Column reference} @tab @strong{Meaning} +@item @code{col_name} @tab Column @code{col_name} +from whichever table used in the query contains a column of that name. +@item @code{tbl_name.col_name} @tab Column @code{col_name} from table +@code{tbl_name} of the current database. +@item @code{db_name.tbl_name.col_name} @tab Column @code{col_name} from table +@code{tbl_name} of the database @code{db_name}. This form is available in +@strong{MySQL} Version 3.22 or later. +@item +@code{`column_name`} @tab A column that is a keyword or contains special characters. @end multitable -@example -mysql> select 'David!' LIKE 'David\_'; - -> 0 -mysql> select 'David_' LIKE 'David\_'; - -> 1 -@end example +You need not specify a @code{tbl_name} or @code{db_name.tbl_name} prefix for +a column reference in a statement unless the reference would be ambiguous. +For example, suppose tables @code{t1} and @code{t2} each contain a column +@code{c}, and you retrieve @code{c} in a @code{SELECT} statement that uses +both @code{t1} and @code{t2}. In this case, @code{c} is ambiguous because it +is not unique among the tables used in the statement, so you must indicate +which table you mean by writing @code{t1.c} or @code{t2.c}. Similarly, if +you are retrieving from a table @code{t} in database @code{db1} and from a +table @code{t} in database @code{db2}, you must refer to columns in those +tables as @code{db1.t.col_name} and @code{db2.t.col_name}. -To specify a different escape character, use the @code{ESCAPE} clause: +@cindex ODBC compatibility +@cindex compatibility, with ODBC +The syntax @code{.tbl_name} means the table @code{tbl_name} in the current +database. This syntax is accepted for ODBC compatibility, because some ODBC +programs prefix table names with a @samp{.} character. -@example -mysql> select 'David_' LIKE 'David|_' ESCAPE '|'; - -> 1 -@end example +@cindex names, case-sensitivity +@cindex case-sensitivity, in names +@node Name case sensitivity, , Legal names, Legal names +@subsubsection Case Sensitivity in Names +@cindex database names, case sensitivity +@cindex table names, case sensitivity +@cindex column names, case sensitivity +@cindex alias names, case sensitivity -The following two statements illustrate that string comparisons are -case insensitive unless one of the operands is a binary string: +In @strong{MySQL}, databases and tables correspond to directories and files +within those directories. Consequently, the case sensitivity of the +underlying operating system determines the case sensitivity of database and +table names. This means database and table names are case sensitive in Unix +and case insensitive in Windows. @xref{Extensions to ANSI}. + +@strong{NOTE:} Although database and table names are case insensitive for +Windows, you should not refer to a given database or table using different +cases within the same query. The following query would not work because it +refers to a table both as @code{my_table} and as @code{MY_TABLE}: @example -mysql> select 'abc' LIKE 'ABC'; - -> 1 -mysql> SELECT 'abc' LIKE BINARY 'ABC'; - -> 0 +mysql> SELECT * FROM my_table WHERE MY_TABLE.col=1; @end example -@code{LIKE} is allowed on numeric expressions! (This is a @strong{MySQL} -extension to the ANSI SQL @code{LIKE}.) +Column names are case insensitive in all cases. + +Aliases on tables are case sensitive. The following query would not work +because it refers to the alias both as @code{a} and as @code{A}: @example -mysql> select 10 LIKE '1%'; - -> 1 +mysql> SELECT col_name FROM tbl_name AS a + WHERE a.col_name = 1 OR A.col_name = 2; @end example -Note: Because @strong{MySQL} uses the C escape syntax in strings (for example, -@samp{\n}), you must double any @samp{\} that you use in your @code{LIKE} -strings. For example, to search for @samp{\n}, specify it as @samp{\\n}. To -search for @samp{\}, specify it as @samp{\\\\} (the backslashes are stripped -once by the parser and another time when the pattern match is done, leaving -a single backslash to be matched). +Aliases on columns are case insensitive. -@findex NOT LIKE -@item expr NOT LIKE pat [ESCAPE 'escape-char'] -Same as @code{NOT (expr LIKE pat [ESCAPE 'escape-char'])}. +If you have a problem remembering the used cases for a table names, +adopt a consistent convention, such as always creating databases and +tables using lowercase names. -@cindex mSQL compatibility -@cindex compatibility, with mSQL -@findex REGEXP -@findex RLIKE -@item expr REGEXP pat -@itemx expr RLIKE pat -Performs a pattern match of a string expression @code{expr} against a pattern -@code{pat}. The pattern can be an extended regular expression. -@xref{Regexp}. Returns @code{1} if @code{expr} matches @code{pat}, otherwise -returns @code{0}. @code{RLIKE} is a synonym for @code{REGEXP}, provided for -@code{mSQL} compatibility. Note: Because @strong{MySQL} uses the C escape -syntax in strings (for example, @samp{\n}), you must double any @samp{\} that -you use in your @code{REGEXP} strings. As of @strong{MySQL} Version 3.23.4, -@code{REGEXP} is case insensitive for normal (not binary) strings: +One way to avoid this problem is to start @code{mysqld} with @code{-O +lower_case_table_names=1}. By default this option is 1 on Windows and 0 on +Unix. -@example -mysql> select 'Monty!' REGEXP 'm%y%%'; - -> 0 -mysql> select 'Monty!' REGEXP '.*'; - -> 1 -mysql> select 'new*\n*line' REGEXP 'new\\*.\\*line'; - -> 1 -mysql> select "a" REGEXP "A", "a" REGEXP BINARY "A"; - -> 1 0 -mysql> select "a" REGEXP "^[a-d]"; - -> 1 -@end example +If @code{lower_case_table_names} is 1 @strong{MySQL} will convert all +table names to lower case on storage and lookup. Note that if you +change this option, you need to first convert your old table names to +lower case before starting @code{mysqld}. -@item -@code{REGEXP} and @code{RLIKE} use the current character set (ISO-8859-1 -Latin1 by default) when deciding the type of a character. +@cindex variables, user +@cindex user variables +@cindex names, variables +@node Variables, Column types, Literals, Reference +@section User Variables -@findex NOT REGEXP -@item expr NOT REGEXP pat -@itemx expr NOT RLIKE pat -Same as @code{NOT (expr REGEXP pat)}. +@strong{MySQL} supports thread-specific variables with the +@code{@@variablename} syntax. A variable name may consist of +alphanumeric characters from the current character set and also +@samp{_}, @samp{$}, and @samp{.} . The default character set is +ISO-8859-1 Latin1; this may be changed with the +@code{--default-character-set} option to @code{mysqld}. @xref{Character +sets}. -@findex STRCMP() -@item STRCMP(expr1,expr2) -@code{STRCMP()} -returns @code{0} if the strings are the same, @code{-1} if the first -argument is smaller than the second according to the current sort order, -and @code{1} otherwise: +Variables don't have to be initialized. They contain @code{NULL} by default +and can store an integer, real, or string value. All variables for +a thread are automatically freed when the thread exits. + +You can set a variable with the @code{SET} syntax: @example -mysql> select STRCMP('text', 'text2'); - -> -1 -mysql> select STRCMP('text2', 'text'); - -> 1 -mysql> select STRCMP('text', 'text'); - -> 0 +SET @@variable= @{ integer expression | real expression | string expression @} +[,@@variable= ...]. @end example -@findex MATCH ... AGAINST() -@item MATCH (col1,col2,...) AGAINST (expr) -@code{MATCH ... AGAINST()} is used for full-text search and returns -relevance - similarity measure between the text in columns -@code{(col1,col2,...)} and the query @code{expr}. Relevance is a -positive floating-point number. Zero relevance means no similarity. -For @code{MATCH ... AGAINST()} to work, a @strong{FULLTEXT} index -must be created first. @xref{CREATE TABLE, , @code{CREATE TABLE}}. -@code{MATCH ... AGAINST()} is available in @strong{MySQL} Version -3.23.23 or later. For details and usage examples -@pxref{Fulltext Search}. -@end table - -@findex casts -@cindex cast operators -@cindex operators, cast -@node Casts, Control flow functions, String comparison functions, Functions -@subsection Cast Operators +You can also set a variable in an expression with the @code{@@variable:=expr} +syntax: -@table @code -@findex BINARY -@item @code{BINARY} -The @code{BINARY} operator casts the string following it to a binary string. -This is an easy way to force a column comparison to be case sensitive even -if the column isn't defined as @code{BINARY} or @code{BLOB}: @example -mysql> select "a" = "A"; - -> 1 -mysql> select BINARY "a" = "A"; - -> 0 +select @@t1:=(@@t2:=1)+@@t3:=4,@@t1,@@t2,@@t3; ++----------------------+------+------+------+ +| @@t1:=(@@t2:=1)+@@t3:=4 | @@t1 | @@t2 | @@t3 | ++----------------------+------+------+------+ +| 5 | 5 | 1 | 4 | ++----------------------+------+------+------+ @end example -@code{BINARY} was introduced in @strong{MySQL} Version 3.23.0. +(We had to use the @code{:=} syntax here, because @code{=} was reserved for +comparisons.) -Note that in some context @strong{MySQL} will not be able to use the -index efficiently when you cast an indexed column to @code{BINARY}. -@end table +User variables may be used where expressions are allowed. Note that +this does not currently include use in contexts where a number is explicitly +required, such as in the @code{LIMIT} clause of a @code{SELECT} statement, +or the @code{IGNORE number LINES} clause of a @code{LOAD DATA} statement. -If you want to compare a blob case-insensitively you can always convert -the blob to upper case before doing the comparison: +@strong{NOTE:} In a @code{SELECT} statement, each expression is only +evaluated when it's sent to the client. This means that in the @code{HAVING}, +@code{GROUP BY}, or @code{ORDER BY} clause, you can't refer to an expression +that involves variables that are set in the @code{SELECT} part. For example, +the following statement will NOT work as expected: @example -SELECT 'A' LIKE UPPER(blob_col) FROM table_name; +SELECT (@@aa:=id) AS a, (@@aa+3) AS b FROM table_name HAVING b=5; @end example -We plan to soon introduce casting between different character sets to -make string comparison even more flexible. +The reason is that @code{@@aa} will not contain the value of the current +row, but the value of @code{id} for the previous accepted row. -@findex control flow functions -@findex functions, control flow -@node Control flow functions, Mathematical functions, Casts, Functions -@subsection Control Flow Functions +@cindex columns, types +@cindex types, columns +@node Column types, Functions, Variables, Reference +@section Column Types -@table @code -@cindex @code{NULL}, testing for null -@findex IFNULL() -@item IFNULL(expr1,expr2) -If @code{expr1} is not @code{NULL}, @code{IFNULL()} returns @code{expr1}, -else it returns @code{expr2}. @code{IFNULL()} returns a numeric or string -value, depending on the context in which it is used: -@example -mysql> select IFNULL(1,0); - -> 1 -mysql> select IFNULL(NULL,10); - -> 10 -mysql> select IFNULL(1/0,10); - -> 10 -mysql> select IFNULL(1/0,'yes'); - -> 'yes' -@end example +@strong{MySQL} supports a number of column types, which may be grouped into +three categories: numeric types, date and time types, and string (character) +types. This section first gives an overview of the types available and +summarizes the storage requirements for each column type, then provides a +more detailed description of the properties of the types in each category. +The overview is intentionally brief. The more detailed descriptions should +be consulted for additional information about particular column types, such +as the allowable formats in which you can specify values. -@findex NULLIF() -@item NULLIF(expr1,expr2) -If @code{expr1 = expr2} is true, return @code{NULL} else return @code{expr1}. -This is the same as @code{CASE WHEN x = y THEN NULL ELSE x END}: -@example -mysql> select NULLIF(1,1); - -> NULL -mysql> select NULLIF(1,2); - -> 1 -@end example +The column types supported by @strong{MySQL} are listed below. +The following code letters are used in the descriptions: -Note that @code{expr1} is evaluated twice in @strong{MySQL} if the arguments -are equal. +@cindex display size +@cindex sizes, display +@cindex digits +@cindex decimal point +@cindex brackets, square +@cindex square brackets +@table @code +@item M +Indicates the maximum display size. The maximum legal display size is 255. -@findex IF() -@item IF(expr1,expr2,expr3) -If @code{expr1} is TRUE (@code{expr1 <> 0} and @code{expr1 <> NULL}) then -@code{IF()} returns @code{expr2}, else it returns @code{expr3}. -@code{IF()} returns a numeric or string value, depending on the context -in which it is used: +@item D +Applies to floating-point types and indicates the number of digits +following the decimal point. The maximum possible value is 30, but +should be no greater than @code{M}-2. +@end table -@example -mysql> select IF(1>2,2,3); - -> 3 -mysql> select IF(1<2,'yes','no'); - -> 'yes' -mysql> select IF(strcmp('test','test1'),'no','yes'); - -> 'no' -@end example +Square brackets (@samp{[} and @samp{]}) indicate parts of type specifiers +that are optional. -@code{expr1} is evaluated as an integer value, which means that if you are -testing floating-point or string values, you should do so using a comparison -operation: +@tindex Types -@example -mysql> select IF(0.1,1,0); - -> 0 -mysql> select IF(0.1<>0,1,0); - -> 1 -@end example +@c The @w{-number} stuff keeps a linebreak from occurring between +@c the - and number. -In the first case above, @code{IF(0.1)} returns @code{0} because @code{0.1} -is converted to an integer value, resulting in a test of @code{IF(0)}. This -may not be what you expect. In the second case, the comparison tests the -original floating-point value to see whether it is non-zero. The result -of the comparison is used as an integer. - -The default return type of @code{IF()} (which may matter when it is -stored into a temporary table) is calculated in @strong{MySQL} Version -3.23 as follows: +Note that if you specify @code{ZEROFILL} for a column, @strong{MySQL} will +automatically add the @code{UNSIGNED} attribute to the column. -@multitable @columnfractions .55 .45 -@item @strong{Expression} @tab @strong{Return value} -@item expr2 or expr3 returns string @tab string -@item expr2 or expr3 returns a floating-point value @tab floating-point -@item expr2 or expr3 returns an integer @tab integer -@end multitable +@table @code +@tindex TINYINT +@item TINYINT[(M)] [UNSIGNED] [ZEROFILL] -@findex CASE -@item CASE value WHEN [compare-value] THEN result [WHEN [compare-value] THEN result ...] [ELSE result] END -@item CASE WHEN [condition] THEN result [WHEN [condition] THEN result ...] [ELSE result] END +A very small integer. The signed range is @code{-128} to @code{127}. The +unsigned range is @code{0} to @code{255}. -The first version returns the @code{result} where -@code{value=compare-value}. The second version returns the result for -the first condition, which is true. If there was no matching result -value, then the result after @code{ELSE} is returned. If there is no -@code{ELSE} part then @code{NULL} is returned: +@tindex SMALLINT +@item SMALLINT[(M)] [UNSIGNED] [ZEROFILL] -@example -mysql> SELECT CASE 1 WHEN 1 THEN "one" WHEN 2 THEN "two" ELSE "more" END; - -> "one" -mysql> SELECT CASE WHEN 1>0 THEN "true" ELSE "false" END; - -> "true" -mysql> SELECT CASE BINARY "B" when "a" then 1 when "b" then 2 END; - -> NULL -@end example -@end table +A small integer. The signed range is @code{-32768} to @code{32767}. The +unsigned range is @code{0} to @code{65535}. -The type of the return value (@code{INTEGER}, @code{DOUBLE} or -@code{STRING}) is the same as the type of the first returned value (the -expression after the first @code{THEN}). +@tindex MEDIUMINT +@item MEDIUMINT[(M)] [UNSIGNED] [ZEROFILL] -@findex mathematical functions -@findex functions, mathematical -@node Mathematical functions, String functions, Control flow functions, Functions -@subsection Mathematical Functions -All mathematical functions return @code{NULL} in case of an error. +A medium-size integer. The signed range is @code{-8388608} to +@code{8388607}. The unsigned range is @code{0} to @code{16777215}. -@table @code -@findex - (unary minus) -@findex minus, unary (-) -@findex unary minus (-) -@item - -Unary minus. Changes the sign of the argument: -@example -mysql> select - 2; - -> -2 -@end example +@tindex INT +@item INT[(M)] [UNSIGNED] [ZEROFILL] -Note that if this operator is used with a @code{BIGINT}, the return value is a -@code{BIGINT}! This means that you should avoid using @code{-} on integers that -may have the value of @code{-2^63}! +A normal-size integer. The signed range is @code{-2147483648} to +@code{2147483647}. The unsigned range is @code{0} to @code{4294967295}. -@findex ABS() -@item ABS(X) -Returns the absolute value of @code{X}: -@example -mysql> select ABS(2); - -> 2 -mysql> select ABS(-32); - -> 32 -@end example +@tindex INTEGER +@item INTEGER[(M)] [UNSIGNED] [ZEROFILL] -This function is safe to use with @code{BIGINT} values. +This is a synonym for @code{INT}. -@findex SIGN() -@item SIGN(X) -Returns the sign of the argument as @code{-1}, @code{0}, or @code{1}, depending -on whether @code{X} is negative, zero, or positive: -@example -mysql> select SIGN(-32); - -> -1 -mysql> select SIGN(0); - -> 0 -mysql> select SIGN(234); - -> 1 -@end example +@tindex BIGINT +@item BIGINT[(M)] [UNSIGNED] [ZEROFILL] -@findex MOD() -@findex % (modulo) -@findex modulo (%) -@item MOD(N,M) -@itemx % -Modulo (like the @code{%} operator in C). -Returns the remainder of @code{N} divided by @code{M}: -@example -mysql> select MOD(234, 10); - -> 4 -mysql> select 253 % 7; - -> 1 -mysql> select MOD(29,9); - -> 2 -@end example +A large integer. The signed range is @code{-9223372036854775808} to +@code{9223372036854775807}. The unsigned range is @code{0} to +@code{18446744073709551615}. -This function is safe to use with @code{BIGINT} values. +Some things you should be aware about @code{BIGINT} columns: -@findex FLOOR() -@item FLOOR(X) -Returns the largest integer value not greater than @code{X}: -@example -mysql> select FLOOR(1.23); - -> 1 -mysql> select FLOOR(-1.23); - -> -2 -@end example +@itemize @bullet +@item +@cindex rounding errors +As all arithmetic is done using signed @code{BIGINT} or @code{DOUBLE} +values, so you shouldn't use unsigned big integers larger than +@code{9223372036854775807} (63 bits) except with bit functions! If you +do that, some of the last digits in the result may be wrong because of +rounding errors when converting the @code{BIGINT} to a @code{DOUBLE}. +@item +You can always store an exact integer value in a @code{BIGINT} column by +storing it as a string, as there is in this case there will be no +intermediate double representation. +@item +@samp{-}, @samp{+}, and @samp{*} will use @code{BIGINT} arithmetic when +both arguments are @code{INTEGER} values! This means that if you +multiply two big integers (or results from functions that return +integers) you may get unexpected results when the result is larger than +@code{9223372036854775807}. +@end itemize -Note that the return value is converted to a @code{BIGINT}! +@cindex floating-point number +@tindex FLOAT +@tindex FLOAT(precision) +@item FLOAT(precision) [ZEROFILL] -@findex CEILING() -@item CEILING(X) -Returns the smallest integer value not less than @code{X}: -@example -mysql> select CEILING(1.23); - -> 2 -mysql> select CEILING(-1.23); - -> -1 -@end example +A floating-point number. Cannot be unsigned. @code{precision} can be +@code{<=24} for a single-precision floating-point number and between 25 +and 53 for a double-precision floating-point number. These types are like +the @code{FLOAT} and @code{DOUBLE} types described immediately below. +@code{FLOAT(X)} has the same range as the corresponding @code{FLOAT} and +@code{DOUBLE} types, but the display size and number of decimals is undefined. -Note that the return value is converted to a @code{BIGINT}! +In @strong{MySQL} Version 3.23, this is a true floating-point value. In +earlier @strong{MySQL} versions, @code{FLOAT(precision)} always has 2 decimals. -@findex ROUND() -@item ROUND(X) -Returns the argument @code{X}, rounded to the nearest integer: -@example -mysql> select ROUND(-1.23); - -> -1 -mysql> select ROUND(-1.58); - -> -2 -mysql> select ROUND(1.58); - -> 2 -@end example +Note that using @code{FLOAT} may give you some unexpected problems as +all calculation in @strong{MySQL} is done with double precision. +@xref{No matching rows}. -Note that the behavior of @code{ROUND()} when the argument -is half way between two integers depends on the C library -implementation. Some round to the nearest even number, -always up, always down, or always towards zero. If you need -one kind of rounding, you should use a well-defined function -like @code{TRUNCATE()} or @code{FLOOR()} instead. +@cindex ODBC compatibility +@cindex compatibility, with ODBC +This syntax is provided for ODBC compatibility. -@findex ROUND() -@item ROUND(X,D) -Returns the argument @code{X}, rounded to a number with @code{D} decimals. -If @code{D} is @code{0}, the result will have no decimal point or fractional -part: +@tindex FLOAT +@tindex FLOAT(M,D) +@item FLOAT[(M,D)] [ZEROFILL] -@example -mysql> select ROUND(1.298, 1); - -> 1.3 -mysql> select ROUND(1.298, 0); - -> 1 -@end example +A small (single-precision) floating-point number. Cannot be unsigned. +Allowable values are @code{@w{-3.402823466E+38}} to +@code{@w{-1.175494351E-38}}, @code{0}, and @code{@w{1.175494351E-38}} to +@code{3.402823466E+38}. The M is the display width and D is the +number of decimals. @code{FLOAT} without an argument or with an argument of +<= 24 stands for a single-precision floating-point number. -@findex EXP() -@item EXP(X) -Returns the value of @code{e} (the base of natural logarithms) raised to -the power of @code{X}: -@example -mysql> select EXP(2); - -> 7.389056 -mysql> select EXP(-2); - -> 0.135335 -@end example -@findex LOG() -@item LOG(X) -Returns the natural logarithm of @code{X}: -@example -mysql> select LOG(2); - -> 0.693147 -mysql> select LOG(-2); - -> NULL -@end example -If you want the log of a number @code{X} to some arbitary base @code{B}, use -the formula @code{LOG(X)/LOG(B)}. +@tindex DOUBLE +@tindex FLOAT(precision) +@item DOUBLE[(M,D)] [ZEROFILL] -@findex LOG10() -@item LOG10(X) -Returns the base-10 logarithm of @code{X}: -@example -mysql> select LOG10(2); - -> 0.301030 -mysql> select LOG10(100); - -> 2.000000 -mysql> select LOG10(-100); - -> NULL -@end example +A normal-size (double-precision) floating-point number. Cannot be +unsigned. Allowable values are @code{@w{-1.7976931348623157E+308}} to +@code{@w{-2.2250738585072014E-308}}, @code{0}, and +@code{2.2250738585072014E-308} to @code{1.7976931348623157E+308}. The M +is the display width and D is the number of decimals. @code{DOUBLE} +without an argument or @code{FLOAT(X)} where 25 <= X <= 53 stands for a +double-precision floating-point number. -@findex POW() -@findex POWER() -@item POW(X,Y) -@itemx POWER(X,Y) -Returns the value of @code{X} raised to the power of @code{Y}: -@example -mysql> select POW(2,2); - -> 4.000000 -mysql> select POW(2,-2); - -> 0.250000 -@end example +@tindex DOUBLE PRECISION +@tindex REAL +@item DOUBLE PRECISION[(M,D)] [ZEROFILL] +@itemx REAL[(M,D)] [ZEROFILL] -@findex SQRT() -@item SQRT(X) -Returns the non-negative square root of @code{X}: -@example -mysql> select SQRT(4); - -> 2.000000 -mysql> select SQRT(20); - -> 4.472136 -@end example +These are synonyms for @code{DOUBLE}. -@findex PI() -@item PI() -Returns the value of PI. The default shown number of decimals is 5, but -@strong{MySQL} internally uses the full double precession for PI. -@example -mysql> select PI(); - -> 3.141593 -mysql> SELECT PI()+0.000000000000000000; - -> 3.141592653589793116 -@end example +@tindex DECIMAL +@item DECIMAL[(M[,D])] [ZEROFILL] -@findex COS() -@item COS(X) -Returns the cosine of @code{X}, where @code{X} is given in radians: -@example -mysql> select COS(PI()); - -> -1.000000 -@end example +An unpacked floating-point number. Cannot be unsigned. Behaves like a +@code{CHAR} column: ``unpacked'' means the number is stored as a string, +using one character for each digit of the value. The decimal point and, +for negative numbers, the @samp{-} sign, are not counted in M (but space +for these are reserved). If @code{D} is 0, values will have no decimal +point or fractional part. The maximum range of @code{DECIMAL} values is +the same as for @code{DOUBLE}, but the actual range for a given +@code{DECIMAL} column may be constrained by the choice of @code{M} and +@code{D}. -@findex SIN() -@item SIN(X) -Returns the sine of @code{X}, where @code{X} is given in radians: -@example -mysql> select SIN(PI()); - -> 0.000000 -@end example +If @code{D} is left out it's set to 0. If @code{M} is left out it's set to 10. -@findex TAN() -@item TAN(X) -Returns the tangent of @code{X}, where @code{X} is given in radians: -@example -mysql> select TAN(PI()+1); - -> 1.557408 -@end example +Note that in @strong{MySQL} Version 3.22 the @code{M} argument had to +includes the space needed for the sign and the decimal point. -@findex ACOS() -@item ACOS(X) -Returns the arc cosine of @code{X}, that is, the value whose cosine is -@code{X}. Returns @code{NULL} if @code{X} is not in the range @code{-1} to -@code{1}: -@example -mysql> select ACOS(1); - -> 0.000000 -mysql> select ACOS(1.0001); - -> NULL -mysql> select ACOS(0); - -> 1.570796 -@end example +@tindex NUMERIC +@item NUMERIC(M,D) [ZEROFILL] -@findex ASIN() -@item ASIN(X) -Returns the arc sine of @code{X}, that is, the value whose sine is -@code{X}. Returns @code{NULL} if @code{X} is not in the range @code{-1} to -@code{1}: -@example -mysql> select ASIN(0.2); - -> 0.201358 -mysql> select ASIN('foo'); - -> 0.000000 -@end example +This is a synonym for @code{DECIMAL}. -@findex ATAN() -@item ATAN(X) -Returns the arc tangent of @code{X}, that is, the value whose tangent is -@code{X}: -@example -mysql> select ATAN(2); - -> 1.107149 -mysql> select ATAN(-2); - -> -1.107149 -@end example +@tindex DATE +@item DATE -@findex ATAN2() -@item ATAN2(Y,X) -Returns the arc tangent of the two variables @code{X} and @code{Y}. It is -similar to calculating the arc tangent of @code{Y / X}, except that the -signs of both arguments are used to determine the quadrant of the -result: -@example -mysql> select ATAN(-2,2); - -> -0.785398 -mysql> select ATAN(PI(),0); - -> 1.570796 -@end example +A date. The supported range is @code{'1000-01-01'} to @code{'9999-12-31'}. +@strong{MySQL} displays @code{DATE} values in @code{'YYYY-MM-DD'} format, but +allows you to assign values to @code{DATE} columns using either strings or +numbers. @xref{DATETIME}. -@findex COT() -@item COT(X) -Returns the cotangent of @code{X}: -@example -mysql> select COT(12); - -> -1.57267341 -mysql> select COT(0); - -> NULL -@end example +@tindex DATETIME +@item DATETIME -@findex RAND() -@item RAND() -@itemx RAND(N) -Returns a random floating-point value in the range @code{0} to @code{1.0}. -If an integer argument @code{N} is specified, it is used as the seed value: -@example -mysql> select RAND(); - -> 0.5925 -mysql> select RAND(20); - -> 0.1811 -mysql> select RAND(20); - -> 0.1811 -mysql> select RAND(); - -> 0.2079 -mysql> select RAND(); - -> 0.7888 -@end example -You can't use a column with @code{RAND()} values in an @code{ORDER BY} -clause, because @code{ORDER BY} would evaluate the column multiple times. -In @strong{MySQL} Version 3.23, you can, however, do: -@code{SELECT * FROM table_name ORDER BY RAND()} +A date and time combination. The supported range is @code{'1000-01-01 +00:00:00'} to @code{'9999-12-31 23:59:59'}. @strong{MySQL} displays +@code{DATETIME} values in @code{'YYYY-MM-DD HH:MM:SS'} format, but allows you +to assign values to @code{DATETIME} columns using either strings or numbers. +@xref{DATETIME}. -This is useful to get a random sample of a set @code{SELECT * FROM -table1,table2 WHERE a=b AND c<d ORDER BY RAND() LIMIT 1000}. +@tindex TIMESTAMP +@item TIMESTAMP[(M)] -Note that a @code{RAND()} in a @code{WHERE} clause will be re-evaluated -every time the @code{WHERE} is executed. +A timestamp. The range is @code{'1970-01-01 00:00:00'} to sometime in the +year @code{2037}. @strong{MySQL} displays @code{TIMESTAMP} values in +@code{YYYYMMDDHHMMSS}, @code{YYMMDDHHMMSS}, @code{YYYYMMDD}, or @code{YYMMDD} +format, depending on whether @code{M} is @code{14} (or missing), @code{12}, +@code{8}, or @code{6}, but allows you to assign values to @code{TIMESTAMP} +columns using either strings or numbers. A @code{TIMESTAMP} column is useful +for recording the date and time of an @code{INSERT} or @code{UPDATE} +operation because it is automatically set to the date and time of the most +recent operation if you don't give it a value yourself. You can also set it +to the current date and time by assigning it a @code{NULL} value. @xref{Date +and time types}. -@findex LEAST() -@item LEAST(X,Y,...) -With two or more arguments, returns the smallest (minimum-valued) argument. -The arguments are compared using the following rules: +A @code{TIMESTAMP} is always stored in 4 bytes. The @code{M} argument only +affects how the @code{TIMESTAMP} column is displayed. -@itemize @bullet -@item -If the return value is used in an @code{INTEGER} context, or all arguments -are integer-valued, they are compared as integers. +Note that @code{TIMESTAMP(X)} columns where X is 8 or 14 are reported to +be numbers while other @code{TIMESTAMP(X)} columns are reported to be +strings. This is just to ensure that one can reliably dump and restore +the table with these types! +@xref{DATETIME}. -@item -If the return value is used in a @code{REAL} context, or all arguments are -real-valued, they are compared as reals. +@tindex TIME +@item TIME -@item -If any argument is a case-sensitive string, the arguments are compared -as case-sensitive strings. +A time. The range is @code{'-838:59:59'} to @code{'838:59:59'}. +@strong{MySQL} displays @code{TIME} values in @code{'HH:MM:SS'} format, but +allows you to assign values to @code{TIME} columns using either strings or +numbers. @xref{TIME}. -@item -In other cases, the arguments are compared as case-insensitive strings: -@end itemize +@tindex YEAR +@item YEAR[(2|4)] -@example -mysql> select LEAST(2,0); - -> 0 -mysql> select LEAST(34.0,3.0,5.0,767.0); - -> 3.0 -mysql> select LEAST("B","A","C"); - -> "A" -@end example -In @strong{MySQL} versions prior to Version 3.22.5, you can use @code{MIN()} -instead of @code{LEAST}. +A year in 2- or 4-digit format (default is 4-digit). The allowable values +are @code{1901} to @code{2155}, @code{0000} in the 4-digit year format, +and 1970-2069 if you use the 2-digit format (70-69). @strong{MySQL} displays +@code{YEAR} values in @code{YYYY} format, but allows you to assign values to +@code{YEAR} columns using either strings or numbers. (The @code{YEAR} type is +new in @strong{MySQL} Version 3.22.). @xref{YEAR}. -@findex GREATEST() -@item GREATEST(X,Y,...) -Returns the largest (maximum-valued) argument. -The arguments are compared using the same rules as for @code{LEAST}: -@example -mysql> select GREATEST(2,0); - -> 2 -mysql> select GREATEST(34.0,3.0,5.0,767.0); - -> 767.0 -mysql> select GREATEST("B","A","C"); - -> "C" -@end example -In @strong{MySQL} versions prior to Version 3.22.5, you can use @code{MAX()} -instead of @code{GREATEST}. +@tindex NATIONAL CHAR +@tindex NCHAR +@tindex CHAR +@tindex CHARACTER +@item [NATIONAL] CHAR(M) [BINARY] -@findex DEGREES() -@item DEGREES(X) -Returns the argument @code{X}, converted from radians to degrees: -@example -mysql> select DEGREES(PI()); - -> 180.000000 -@end example +A fixed-length string that is always right-padded with spaces to the +specified length when stored. The range of @code{M} is 1 to 255 characters. +Trailing spaces are removed when the value is retrieved. @code{CHAR} values +are sorted and compared in case-insensitive fashion according to the +default character set unless the @code{BINARY} keyword is given. -@findex RADIANS() -@item RADIANS(X) -Returns the argument @code{X}, converted from degrees to radians: -@example -mysql> select RADIANS(90); - -> 1.570796 -@end example +@code{NATIONAL CHAR} (short form @code{NCHAR}) is the ANSI SQL way to +define that a CHAR column should use the default CHARACTER set. This is +the default in @strong{MySQL}. -@findex TRUNCATE() -@item TRUNCATE(X,D) -Returns the number @code{X}, truncated to @code{D} decimals. If @code{D} -is @code{0}, the result will have no decimal point or fractional part: -@example -mysql> select TRUNCATE(1.223,1); - -> 1.2 -mysql> select TRUNCATE(1.999,1); - -> 1.9 -mysql> select TRUNCATE(1.999,0); - -> 1 -@end example +@code{CHAR} is a shorthand for @code{CHARACTER}. -Note that as decimal numbers are normally not stored as exact numbers in -computers, but as double values, you may be fooled by the following -result: +@strong{MySQL} allows you to create a column of type +@code{CHAR(0)}. This is mainly useful when you have to be compliant with +some old applications that depend on the existence of a column but that do not +actually use the value. This is also quite nice when you need a +column that only can take 2 values: A @code{CHAR(0)}, that is not defined +as @code{NOT NULL}, will only occupy one bit and can only take 2 values: +@code{NULL} or @code{""}. @xref{CHAR}. -@cindex rounding errors -@example -mysql> select TRUNCATE(10.28*100,0); - -> 1027 -@end example +@tindex CHARACTER VARYING +@tindex CHAR VARYING +@tindex VARCHAR +@item [NATIONAL] VARCHAR(M) [BINARY] -The above happens because 10.28 is actually stored as something like -10.2799999999999999. -@end table +A variable-length string. @strong{NOTE:} Trailing spaces are removed when +the value is stored (this differs from the ANSI SQL specification). The range +of @code{M} is 1 to 255 characters. @code{VARCHAR} values are sorted and +compared in case-insensitive fashion unless the @code{BINARY} keyword is +given. @xref{Silent column changes}. -@findex string functions -@findex functions, string -@node String functions, Date and time functions, Mathematical functions, Functions -@subsection String Functions +@code{VARCHAR} is a shorthand for @code{CHARACTER VARYING}. +@xref{CHAR}. -String-valued functions return @code{NULL} if the length of the result would -be greater than the @code{max_allowed_packet} server parameter. @xref{Server -parameters}. +@tindex TINYBLOB +@tindex TINYTEXT +@item TINYBLOB +@itemx TINYTEXT -For functions that operate on string positions, -the first position is numbered 1. +A @code{BLOB} or @code{TEXT} column with a maximum length of 255 (2^8 - 1) +characters. @xref{Silent column changes}. @xref{BLOB}. -@table @code -@findex ASCII() -@item ASCII(str) -Returns the ASCII code value of the leftmost character of the string -@code{str}. Returns @code{0} if @code{str} is the empty string. Returns -@code{NULL} if @code{str} is @code{NULL}: +@tindex BLOB +@tindex TEXT +@item BLOB +@itemx TEXT -@example -mysql> select ASCII('2'); - -> 50 -mysql> select ASCII(2); - -> 50 -mysql> select ASCII('dx'); - -> 100 -@end example +A @code{BLOB} or @code{TEXT} column with a maximum length of 65535 (2^16 - 1) +characters. @xref{Silent column changes}. @xref{BLOB}. -See also the @code{ORD()} function. +@tindex MEDIUMBLOB +@tindex MEDIUMTEXT +@item MEDIUMBLOB +@itemx MEDIUMTEXT -@findex ORD() -@item ORD(str) -If the leftmost character of the string str is a multi-byte character, -returns the code of multi-byte character by returning the ASCII code value -of the character in the format of: -@code{((first byte ASCII code)*256+(second byte ASCII code))[*256+third byte ASCII code...]}. -If the leftmost character is not a multi-byte character, returns the same -value as the like @code{ASCII()} function does: +A @code{BLOB} or @code{TEXT} column with a maximum length of 16777215 +(2^24 - 1) characters. @xref{Silent column changes}. @xref{BLOB}. -@example -mysql> select ORD('2'); - -> 50 -@end example +@tindex LONGBLOB +@tindex LONGTEXT +@item LONGBLOB +@itemx LONGTEXT -@findex CONV() -@item CONV(N,from_base,to_base) -Converts numbers between different number bases. Returns a string -representation of the number @code{N}, converted from base @code{from_base} -to base @code{to_base}. Returns @code{NULL} if any argument is @code{NULL}. -The argument @code{N} is interpreted as an integer, but may be specified as -an integer or a string. The minimum base is @code{2} and the maximum base is -@code{36}. If @code{to_base} is a negative number, @code{N} is regarded as a -signed number. Otherwise, @code{N} is treated as unsigned. @code{CONV} works -with 64-bit precision: +A @code{BLOB} or @code{TEXT} column with a maximum length of 4294967295 +(2^32 - 1) characters. @xref{Silent column changes}. Note that because +the server/client protocol and MyISAM tables has currently a limit of +16M per communication packet / table row, you can't yet use this +the whole range of this type. @xref{BLOB}. -@example -mysql> select CONV("a",16,2); - -> '1010' -mysql> select CONV("6E",18,8); - -> '172' -mysql> select CONV(-17,10,-18); - -> '-H' -mysql> select CONV(10+"10"+'10'+0xa,10,10); - -> '40' -@end example +@tindex ENUM +@item ENUM('value1','value2',...) -@findex BIN() -@item BIN(N) -Returns a string representation of the binary value of @code{N}, where -@code{N} is a longlong (@code{BIGINT}) number. This is equivalent to -@code{CONV(N,10,2)}. Returns @code{NULL} if @code{N} is @code{NULL}: +An enumeration. A string object that can have only one value, chosen +from the list of values @code{'value1'}, @code{'value2'}, @code{...}, +@code{NULL} or the special @code{""} error value. An @code{ENUM} can +have a maximum of 65535 distinct values. @xref{ENUM}. -@example -mysql> select BIN(12); - -> '1100' -@end example +@tindex SET +@item SET('value1','value2',...) -@findex OCT() -@item OCT(N) -Returns a string representation of the octal value of @code{N}, where -@code{N} is a longlong number. This is equivalent to @code{CONV(N,10,8)}. -Returns @code{NULL} if @code{N} is @code{NULL}: +A set. A string object that can have zero or more values, each of which must +be chosen from the list of values @code{'value1'}, @code{'value2'}, +@code{...} A @code{SET} can have a maximum of 64 members. @xref{SET}. +@end table -@example -mysql> select OCT(12); - -> '14' -@end example +@menu +* Storage requirements:: Column type storage requirements +* Numeric types:: Numeric types +* Date and time types:: Date and time types +* String types:: String types +* Choosing types:: Choosing the right type for a column +* Indexes:: Column indexes +* Multiple-column indexes:: Multiple-column indexes +* Other-vendor column types:: Using column types from other database engines +@end menu -@findex HEX() -@item HEX(N) -Returns a string representation of the hexadecimal value of @code{N}, where -@code{N} is a longlong (@code{BIGINT}) number. This is equivalent to -@code{CONV(N,10,16)}. Returns @code{NULL} if @code{N} is @code{NULL}: +@cindex storage requirements, column type +@cindex columns, storage requirements -@example -mysql> select HEX(255); - -> 'FF' -@end example +@node Storage requirements, Numeric types, Column types, Column types +@subsection Column Type Storage Requirements -@findex CHAR() -@item CHAR(N,...) -@code{CHAR()} interprets the arguments as integers and returns a string -consisting of the characters given by the ASCII code values of those -integers. @code{NULL} values are skipped: +The storage requirements for each of the column types supported by +@strong{MySQL} are listed below by category. -@example -mysql> select CHAR(77,121,83,81,'76'); - -> 'MySQL' -mysql> select CHAR(77,77.3,'77.3'); - -> 'MMM' -@end example +@cindex numeric types +@cindex types, numeric -@findex CONCAT() -@item CONCAT(str1,str2,...) -Returns the string that results from concatenating the arguments. Returns -@code{NULL} if any argument is @code{NULL}. May have more than 2 arguments. -A numeric argument is converted to the equivalent string form: +@subsubheading Storage requirements for numeric types -@example -mysql> select CONCAT('My', 'S', 'QL'); - -> 'MySQL' -mysql> select CONCAT('My', NULL, 'QL'); - -> NULL -mysql> select CONCAT(14.3); - -> '14.3' -@end example +@multitable @columnfractions .5 .5 +@item @strong{Column type} @tab @strong{Storage required} +@item @code{TINYINT} @tab 1 byte +@item @code{SMALLINT} @tab 2 bytes +@item @code{MEDIUMINT} @tab 3 bytes +@item @code{INT} @tab 4 bytes +@item @code{INTEGER} @tab 4 bytes +@item @code{BIGINT} @tab 8 bytes +@item @code{FLOAT(X)} @tab 4 if X <= 24 or 8 if 25 <= X <= 53 +@item @code{FLOAT} @tab 4 bytes +@item @code{DOUBLE} @tab 8 bytes +@item @code{DOUBLE PRECISION} @tab 8 bytes +@item @code{REAL} @tab 8 bytes +@item @code{DECIMAL(M,D)} @tab @code{M+2} bytes if D > 0, @code{M+1} bytes if D = 0 (@code{D}+2, if @code{M < D}) +@item @code{NUMERIC(M,D)} @tab @code{M+2} bytes if D > 0, @code{M+1} bytes if D = 0 (@code{D}+2, if @code{M < D}) +@end multitable -@findex CONCAT_WS() -@item CONCAT_WS(separator, str1, str2,...) +@cindex date types +@cindex time types +@cindex types, date +@cindex types, time -@code{CONCAT_WS()} stands for CONCAT With Separator and is a special form of -@code{CONCAT()}. The first argument is the separator for the rest of the -arguments. The separator can be a string as well as the rest of the -arguments. If the separator is @code{NULL}, the result will be @code{NULL}. -The function will skip any @code{NULL}s and empty strings, after the -separator argument. The separator will be added between the strings to be -concatenated: +@subsubheading Storage requirements for date and time types -@example -mysql> select CONCAT_WS(",","First name","Second name","Last Name"); - -> 'First name,Second name,Last Name' -mysql> select CONCAT_WS(",","First name",NULL,"Last Name"); - -> 'First name,Last Name' -@end example +@multitable @columnfractions .5 .5 +@item @strong{Column type} @tab @strong{Storage required} +@item @code{DATE} @tab 3 bytes +@item @code{DATETIME} @tab 8 bytes +@item @code{TIMESTAMP} @tab 4 bytes +@item @code{TIME} @tab 3 bytes +@item @code{YEAR} @tab 1 byte +@end multitable -@findex LENGTH() -@findex OCTET_LENGTH() -@findex CHAR_LENGTH() -@findex CHARACTER_LENGTH() -@item LENGTH(str) -@itemx OCTET_LENGTH(str) -@itemx CHAR_LENGTH(str) -@itemx CHARACTER_LENGTH(str) -Returns the length of the string @code{str}: +@subsubheading Storage requirements for string types -@example -mysql> select LENGTH('text'); - -> 4 -mysql> select OCTET_LENGTH('text'); - -> 4 -@end example +@multitable @columnfractions .5 .5 +@item @strong{Column type} @tab @strong{Storage required} +@item @code{CHAR(M)} @tab @code{M} bytes, @code{1 <= M <= 255} +@item @code{VARCHAR(M)} @tab @code{L}+1 bytes, where @code{L <= M} and +@code{1 <= M <= 255} +@item @code{TINYBLOB}, @code{TINYTEXT} @tab @code{L}+1 bytes, +where @code{L} < 2^8 +@item @code{BLOB}, @code{TEXT} @tab @code{L}+2 bytes, +where @code{L} < 2^16 +@item @code{MEDIUMBLOB}, @code{MEDIUMTEXT} @tab @code{L}+3 bytes, +where @code{L} < 2^24 +@item @code{LONGBLOB}, @code{LONGTEXT} @tab @code{L}+4 bytes, +where @code{L} < 2^32 +@item @code{ENUM('value1','value2',...)} @tab 1 or 2 bytes, depending on +the number of enumeration values (65535 values maximum) +@item @code{SET('value1','value2',...)} @tab 1, 2, 3, 4 or 8 bytes, depending +on the number of set members (64 members maximum) +@end multitable -Note that for @code{CHAR_LENGTH()}, multi-byte characters are only counted -once. +@cindex BLOB, size +@cindex TEXT, size +@cindex VARCHAR, size +@code{VARCHAR} and the @code{BLOB} and @code{TEXT} types are variable-length +types, for which the storage requirements depend on the actual length of +column values (represented by @code{L} in the preceding table), rather than +on the type's maximum possible size. For example, a @code{VARCHAR(10)} +column can hold a string with a maximum length of 10 characters. The actual +storage required is the length of the string (@code{L}), plus 1 byte to +record the length of the string. For the string @code{'abcd'}, @code{L} is 4 +and the storage requirement is 5 bytes. -@findex LOCATE() -@findex POSITION() -@item LOCATE(substr,str) -@itemx POSITION(substr IN str) -Returns the position of the first occurrence of substring @code{substr} -in string @code{str}. Returns @code{0} if @code{substr} is not in @code{str}: +The @code{BLOB} and @code{TEXT} types require 1, 2, 3, or 4 bytes to record +the length of the column value, depending on the maximum possible length of +the type. @xref{BLOB}. -@example -mysql> select LOCATE('bar', 'foobarbar'); - -> 4 -mysql> select LOCATE('xbar', 'foobar'); - -> 0 -@end example +If a table includes any variable-length column types, the record format will +also be variable-length. Note that when a table is created, @strong{MySQL} +may, under certain conditions, change a column from a variable-length type to a +fixed-length type, or vice-versa. @xref{Silent column changes}. -This function is multi-byte safe. +@cindex ENUM, size +The size of an @code{ENUM} object is determined by the number of +different enumeration values. One byte is used for enumerations with up +to 255 possible values. Two bytes are used for enumerations with up to +65535 values. @xref{ENUM}. -@findex LOCATE() -@item LOCATE(substr,str,pos) -Returns the position of the first occurrence of substring @code{substr} in -string @code{str}, starting at position @code{pos}. -Returns @code{0} if @code{substr} is not in @code{str}: +@cindex SET, size +The size of a @code{SET} object is determined by the number of different +set members. If the set size is @code{N}, the object occupies @code{(N+7)/8} +bytes, rounded up to 1, 2, 3, 4, or 8 bytes. A @code{SET} can have a maximum +of 64 members. @xref{SET}. -@example -mysql> select LOCATE('bar', 'foobarbar',5); - -> 7 -@end example +@node Numeric types, Date and time types, Storage requirements, Column types +@subsection Numeric Types -This function is multi-byte safe. +@strong{MySQL} supports all of the ANSI/ISO SQL92 numeric types. These +types include the exact numeric data types (@code{NUMERIC}, +@code{DECIMAL}, @code{INTEGER}, and @code{SMALLINT}), as well as the +approximate numeric data types (@code{FLOAT}, @code{REAL}, and +@code{DOUBLE PRECISION}). The keyword @code{INT} is a synonym for +@code{INTEGER}, and the keyword @code{DEC} is a synonym for +@code{DECIMAL}. -@findex INSTR() -@item INSTR(str,substr) -Returns the position of the first occurrence of substring @code{substr} in -string @code{str}. This is the same as the two-argument form of -@code{LOCATE()}, except that the arguments are swapped: +The @code{NUMERIC} and @code{DECIMAL} types are implemented as the same +type by @strong{MySQL}, as permitted by the SQL92 standard. They are +used for values for which it is important to preserve exact precision, +for example with monetary data. When declaring a column of one of these +types the precision and scale can be (and usually is) specified; for +example: @example -mysql> select INSTR('foobarbar', 'bar'); - -> 4 -mysql> select INSTR('xbar', 'foobar'); - -> 0 + salary DECIMAL(9,2) @end example -This function is multi-byte safe. - -@findex LPAD() -@item LPAD(str,len,padstr) -Returns the string @code{str}, left-padded with the string @code{padstr} -until @code{str} is @code{len} characters long. If @code{str} is longer -than @code{len'} then it will be shortened to @code{len} characters. +In this example, @code{9} (@code{precision}) represents the number of +significant decimal digits that will be stored for values, and +@code{2} (@code{scale}) represents the number of digits that will be +stored following the decimal point. In this case, therefore, the range +of values that can be stored in the @code{salary} column is from +@code{-9999999.99} to @code{9999999.99}. In ANSI/ISO SQL92, the syntax +@code{DECIMAL(p)} is equivalent to @code{DECIMAL(p,0)}. Similarly, the +syntax @code{DECIMAL} is equivalent to @code{DECIMAL(p,0)}, where the +implementation is allowed to decide the value of @code{p}. +@strong{MySQL} does not currently support either of these variant forms +of the @code{DECIMAL}/@code{NUMERIC} data types. This is not generally +a serious problem, as the principal benefits of these types derive from +the ability to control both precision and scale explicitly. -@example -mysql> select LPAD('hi',4,'??'); - -> '??hi' -@end example +@code{DECIMAL} and @code{NUMERIC} values are stored as strings, rather +than as binary floating-point numbers, in order to preserve the decimal +precision of those values. One character is used for each digit of the +value, the decimal point (if @code{scale} > 0), and the @samp{-} sign +(for negative numbers). If @code{scale} is 0, @code{DECIMAL} and +@code{NUMERIC} values contain no decimal point or fractional part. -@findex RPAD() -@item RPAD(str,len,padstr) -Returns the string @code{str}, right-padded with the string -@code{padstr} until @code{str} is @code{len} characters long. If -@code{str} is longer than @code{len'} then it will be shortened to -@code{len} characters. +The maximum range of @code{DECIMAL} and @code{NUMERIC} values is the +same as for @code{DOUBLE}, but the actual range for a given +@code{DECIMAL} or @code{NUMERIC} column can be constrained by the +@code{precision} or @code{scale} for a given column. When such a column +is assigned a value with more digits following the decimal point than +are allowed by the specified @code{scale}, the value is rounded to that +@code{scale}. When a @code{DECIMAL} or @code{NUMERIC} column is +assigned a value whose magnitude exceeds the range implied by the +specified (or defaulted) @code{precision} and @code{scale}, +@strong{MySQL} stores the value representing the corresponding end +point of that range. -@example -mysql> select RPAD('hi',5,'?'); - -> 'hi???' -@end example +As an extension to the ANSI/ISO SQL92 standard, @strong{MySQL} also +supports the integral types @code{TINYINT}, @code{MEDIUMINT}, and +@code{BIGINT} as listed in the tables above. Another extension is +supported by @strong{MySQL} for optionally specifying the display width +of an integral value in parentheses following the base keyword for the +type (for example, @code{INT(4)}). This optional width specification is +used to left-pad the display of values whose width is less than the +width specified for the column, but does not constrain the range of +values that can be stored in the column, nor the number of digits that +will be displayed for values whose width exceeds that specified for the +column. When used in conjunction with the optional extension attribute +@code{ZEROFILL}, the default padding of spaces is replaced with zeroes. +For example, for a column declared as @code{INT(5) ZEROFILL}, a value +of @code{4} is retrieved as @code{00004}. Note that if you store larger +values than the display width in an integer column, you may experience +problems when @strong{MySQL} generates temporary tables for some +complicated joins, as in these cases @strong{MySQL} trusts that the +data did fit into the original column width. -@findex LEFT() -@item LEFT(str,len) -Returns the leftmost @code{len} characters from the string @code{str}: +All integral types can have an optional (non-standard) attribute +@code{UNSIGNED}. Unsigned values can be used when you want to allow +only positive numbers in a column and you need a little bigger numeric +range for the column. -@example -mysql> select LEFT('foobarbar', 5); - -> 'fooba' -@end example +The @code{FLOAT} type is used to represent approximate numeric data +types. The ANSI/ISO SQL92 standard allows an optional specification of +the precision (but not the range of the exponent) in bits following the +keyword @code{FLOAT} in parentheses. The @strong{MySQL} implementation +also supports this optional precision specification. When the keyword +@code{FLOAT} is used for a column type without a precision +specification, @strong{MySQL} uses four bytes to store the values. A +variant syntax is also supported, with two numbers given in parentheses +following the @code{FLOAT} keyword. With this option, the first number +continues to represent the storage requirements for the value in bytes, +and the second number specifies the number of digits to be stored and +displayed following the decimal point (as with @code{DECIMAL} and +@code{NUMERIC}). When @strong{MySQL} is asked to store a number for +such a column with more decimal digits following the decimal point than +specified for the column, the value is rounded to eliminate the extra +digits when the value is stored. -This function is multi-byte safe. +The @code{REAL} and @code{DOUBLE PRECISION} types do not accept +precision specifications. As an extension to the ANSI/ISO SQL92 +standard, @strong{MySQL} recognizes @code{DOUBLE} as a synonym for the +@code{DOUBLE PRECISION} type. In contrast with the standard's +requirement that the precision for @code{REAL} be smaller than that used +for @code{DOUBLE PRECISION}, @strong{MySQL} implements both as 8-byte +double-precision floating-point values (when not running in ``ANSI mode''). +For maximum portability, code requiring storage of approximate numeric +data values should use @code{FLOAT} or @code{DOUBLE PRECISION} with no +specification of precision or number of decimal points. -@findex RIGHT() -@item RIGHT(str,len) -Returns the rightmost @code{len} characters from the string @code{str}: +When asked to store a value in a numeric column that is outside the column +type's allowable range, @strong{MySQL} clips the value to the appropriate +endpoint of the range and stores the resulting value instead. -@example -mysql> select RIGHT('foobarbar', 4); - -> 'rbar' -@end example +For example, the range of an @code{INT} column is @code{-2147483648} to +@code{2147483647}. If you try to insert @code{-9999999999} into an +@code{INT} column, the value is clipped to the lower endpoint of the range, +and @code{-2147483648} is stored instead. Similarly, if you try to insert +@code{9999999999}, @code{2147483647} is stored instead. -This function is multi-byte safe. +If the @code{INT} column is @code{UNSIGNED}, the size of the column's +range is the same but its endpoints shift up to @code{0} and @code{4294967295}. +If you try to store @code{-9999999999} and @code{9999999999}, +the values stored in the column become @code{0} and @code{4294967296}. -@findex SUBSTRING() -@findex MID() -@item SUBSTRING(str,pos,len) -@itemx SUBSTRING(str FROM pos FOR len) -@itemx MID(str,pos,len) -Returns a substring @code{len} characters long from string @code{str}, -starting at position @code{pos}. -The variant form that uses @code{FROM} is ANSI SQL92 syntax: +Conversions that occur due to clipping are reported as ``warnings'' for +@code{ALTER TABLE}, @code{LOAD DATA INFILE}, @code{UPDATE}, and +multi-row @code{INSERT} statements. -@example -mysql> select SUBSTRING('Quadratically',5,6); - -> 'ratica' -@end example - -This function is multi-byte safe. +@cindex types, Date and Time +@cindex Date and Time types +@node Date and time types, String types, Numeric types, Column types +@subsection Date and Time Types -@findex SUBSTRING() -@item SUBSTRING(str,pos) -@item SUBSTRING(str FROM pos) -Returns a substring from string @code{str} starting at position @code{pos}: +@menu +* Y2K issues:: Y2K issues and date types +* DATETIME:: The @code{DATETIME}, @code{DATE} and @code{TIMESTAMP} types +* TIME:: The @code{TIME} type +* YEAR:: The @code{YEAR} type +@end menu -@example -mysql> select SUBSTRING('Quadratically',5); - -> 'ratically' -mysql> select SUBSTRING('foobarbar' FROM 4); - -> 'barbar' -@end example +The date and time types are @code{DATETIME}, @code{DATE}, +@code{TIMESTAMP}, @code{TIME}, and @code{YEAR}. Each of these has a +range of legal values, as well as a ``zero'' value that is used when you +specify a really illegal value. Note that @strong{MySQL} allows you to store +certain 'not strictly' legal date values, for example @code{1999-11-31}. +The reason for this is that we think it's the responsibility of the +application to handle date checking, not the SQL servers. To make the +date checking 'fast', @strong{MySQL} only checks that the month is in +the range of 0-12 and the day is in the range of 0-31. The above ranges +are defined this way because @strong{MySQL} allows you to store, in a +@code{DATE} or @code{DATETIME} column, dates where the day or month-day +is zero. This is extremely useful for applications that need to store +a birth-date for which you don't know the exact date. In this case you +simply store the date like @code{1999-00-00} or @code{1999-01-00}. (You +cannot expect to get a correct value from functions like @code{DATE_SUB()} +or @code{DATE_ADD} for dates like these.) -This function is multi-byte safe. +Here are some general considerations to keep in mind when working +with date and time types: -@findex SUBSTRING_INDEX() -@item SUBSTRING_INDEX(str,delim,count) -Returns the substring from string @code{str} before @code{count} -occurrences of the delimiter @code{delim}. -If @code{count} is positive, everything to the left of the final delimiter -(counting from the left) is returned. -If @code{count} is negative, everything to the right of the final delimiter -(counting from the right) is returned: +@itemize @bullet +@item +@strong{MySQL} retrieves values for a given date or time type in a standard +format, but it attempts to interpret a variety of formats for values that +you supply (for example, when you specify a value to be assigned to or +compared to a date or time type). Nevertheless, only the formats described +in the following sections are supported. It is expected that you will supply +legal values, and unpredictable results may occur if you use values in other +formats. -@example -mysql> select SUBSTRING_INDEX('www.mysql.com', '.', 2); - -> 'www.mysql' -mysql> select SUBSTRING_INDEX('www.mysql.com', '.', -2); - -> 'mysql.com' -@end example +@item +Although @strong{MySQL} tries to interpret values in several formats, it +always expects the year part of date values to be leftmost. Dates must be +given in year-month-day order (for example, @code{'98-09-04'}), rather than +in the month-day-year or day-month-year orders commonly used elsewhere (for +example, @code{'09-04-98'}, @code{'04-09-98'}). -This function is multi-byte safe. +@item +@strong{MySQL} automatically converts a date or time type value to a number +if the value is used in a numeric context, and vice versa. -@findex LTRIM() -@item LTRIM(str) -Returns the string @code{str} with leading space characters removed: +@item +When @strong{MySQL} encounters a value for a date or time type that is +out of range or otherwise illegal for the type (see the start of this +section), it converts the value to the ``zero'' value for that type. +(The exception is that out-of-range @code{TIME} values are clipped to +the appropriate endpoint of the @code{TIME} range.) The table below +shows the format of the ``zero'' value for each type: -@example -mysql> select LTRIM(' barbar'); - -> 'barbar' -@end example +@multitable @columnfractions .3 .7 +@item @strong{Column type} @tab @strong{``Zero'' value} +@item @code{DATETIME} @tab @code{'0000-00-00 00:00:00'} +@item @code{DATE} @tab @code{'0000-00-00'} +@item @code{TIMESTAMP} @tab @code{00000000000000} (length depends on display size) +@item @code{TIME} @tab @code{'00:00:00'} +@item @code{YEAR} @tab @code{0000} +@end multitable -@findex RTRIM() -@item RTRIM(str) -Returns the string @code{str} with trailing space characters removed: +@item +The ``zero'' values are special, but you can store or refer to them +explicitly using the values shown in the table. You can also do this +using the values @code{'0'} or @code{0}, which are easier to write. -@example -mysql> select RTRIM('barbar '); - -> 'barbar' -@end example +@item +``Zero'' date or time values used through @strong{MyODBC} are converted +automatically to @code{NULL} in @strong{MyODBC} Version 2.50.12 and above, +because ODBC can't handle such values. +@end itemize -This function is multi-byte safe. +@cindex Year 2000 issues +@cindex date types, Y2K issues +@node Y2K issues, DATETIME, Date and time types, Date and time types +@subsubsection Y2K Issues and Date Types -@findex TRIM() -@item TRIM([[BOTH | LEADING | TRAILING] [remstr] FROM] str) -Returns the string @code{str} with all @code{remstr} prefixes and/or suffixes -removed. If none of the specifiers @code{BOTH}, @code{LEADING} or -@code{TRAILING} are given, @code{BOTH} is assumed. If @code{remstr} is not -specified, spaces are removed: +@strong{MySQL} itself is Y2K-safe (@pxref{Year 2000 compliance}), +but input values presented to @strong{MySQL} may not be. Any input +containing 2-digit year values is ambiguous, because the century is unknown. +Such values must be interpreted into 4-digit form because @strong{MySQL} stores +years internally using four digits. -@example -mysql> select TRIM(' bar '); - -> 'bar' -mysql> select TRIM(LEADING 'x' FROM 'xxxbarxxx'); - -> 'barxxx' -mysql> select TRIM(BOTH 'x' FROM 'xxxbarxxx'); - -> 'bar' -mysql> select TRIM(TRAILING 'xyz' FROM 'barxxyz'); - -> 'barx' -@end example +For @code{DATETIME}, @code{DATE}, @code{TIMESTAMP}, and @code{YEAR} types, +@strong{MySQL} interprets dates with ambiguous year values using the +following rules: -This function is multi-byte safe. +@itemize @bullet +@item +Year values in the range @code{00-69} are converted to @code{2000-2069}. -@findex SOUNDEX() -@item SOUNDEX(str) -Returns a soundex string from @code{str}. Two strings that sound almost the -same should have identical soundex strings. A standard soundex string -is 4 characters long, but the @code{SOUNDEX()} function returns an -arbitrarily long string. You can use @code{SUBSTRING()} on the result to get -a standard soundex string. All non-alphanumeric characters are ignored -in the given string. All international alpha characters outside the A-Z range -are treated as vowels: +@item +Year values in the range @code{70-99} are converted to @code{1970-1999}. +@end itemize -@example -mysql> select SOUNDEX('Hello'); - -> 'H400' -mysql> select SOUNDEX('Quadratically'); - -> 'Q36324' -@end example +Remember that these rules provide only reasonable guesses as to what your +data mean. If the heuristics used by @strong{MySQL} don't produce the +correct values, you should provide unambiguous input containing 4-digit +year values. -@findex SPACE() -@item SPACE(N) -Returns a string consisting of @code{N} space characters: +@code{ORDER BY} will sort 2-digit @code{YEAR/DATE/DATETIME} types properly. -@example -mysql> select SPACE(6); - -> ' ' -@end example +Note also that some functions like @code{MIN()} and @code{MAX()} will convert a +@code{TIMESTAMP/DATE} to a number. This means that a timestamp with a +2-digit year will not work properly with these functions. The fix in this +case is to convert the @code{TIMESTAMP/DATE} to 4-digit year format or +use something like @code{MIN(DATE_ADD(timestamp,INTERVAL 0 DAYS))}. -@findex REPLACE() -@item REPLACE(str,from_str,to_str) -Returns the string @code{str} with all all occurrences of the string -@code{from_str} replaced by the string @code{to_str}: +@tindex DATETIME +@tindex DATE +@tindex TIMESTAMP +@node DATETIME, TIME, Y2K issues, Date and time types +@subsubsection The @code{DATETIME}, @code{DATE}, and @code{TIMESTAMP} Types -@example -mysql> select REPLACE('www.mysql.com', 'w', 'Ww'); - -> 'WwWwWw.mysql.com' -@end example +The @code{DATETIME}, @code{DATE}, and @code{TIMESTAMP} types are related. +This section describes their characteristics, how they are similar, and how +they differ. -This function is multi-byte safe. +The @code{DATETIME} type is used when you need values that contain both date +and time information. @strong{MySQL} retrieves and displays @code{DATETIME} +values in @code{'YYYY-MM-DD HH:MM:SS'} format. The supported range is +@code{'1000-01-01 00:00:00'} to @code{'9999-12-31 23:59:59'}. (``Supported'' +means that although earlier values might work, there is no guarantee that +they will.) -@findex REPEAT() -@item REPEAT(str,count) -Returns a string consisting of the string @code{str} repeated @code{count} -times. If @code{count <= 0}, returns an empty string. Returns @code{NULL} if -@code{str} or @code{count} are @code{NULL}: +The @code{DATE} type is used when you need only a date value, without a time +part. @strong{MySQL} retrieves and displays @code{DATE} values in +@code{'YYYY-MM-DD'} format. The supported range is @code{'1000-01-01'} to +@code{'9999-12-31'}. -@example -mysql> select REPEAT('MySQL', 3); - -> 'MySQLMySQLMySQL' -@end example +The @code{TIMESTAMP} column type provides a type that you can use to +automatically mark @code{INSERT} or @code{UPDATE} operations with the current +date and time. If you have multiple @code{TIMESTAMP} columns, only the first +one is updated automatically. -@findex REVERSE() -@item REVERSE(str) -Returns the string @code{str} with the order of the characters reversed: +Automatic updating of the first @code{TIMESTAMP} column occurs under any of +the following conditions: -@example -mysql> select REVERSE('abc'); - -> 'cba' -@end example +@itemize @bullet +@item +The column is not specified explicitly in an @code{INSERT} or +@code{LOAD DATA INFILE} statement. +@item +The column is not specified explicitly in an @code{UPDATE} statement and some +other column changes value. (Note that an @code{UPDATE} that sets a column +to the value it already has will not cause the @code{TIMESTAMP} column to be +updated, because if you set a column to its current value, @strong{MySQL} +ignores the update for efficiency.) +@item +You explicitly set the @code{TIMESTAMP} column to @code{NULL}. +@end itemize -This function is multi-byte safe. +@code{TIMESTAMP} columns other than the first may also be set to the current +date and time. Just set the column to @code{NULL} or to @code{NOW()}. -@findex INSERT() -@item INSERT(str,pos,len,newstr) -Returns the string @code{str}, with the substring beginning at position -@code{pos} and @code{len} characters long replaced by the string -@code{newstr}: +You can set any @code{TIMESTAMP} column to a value different than the current +date and time by setting it explicitly to the desired value. This is true +even for the first @code{TIMESTAMP} column. You can use this property if, +for example, you want a @code{TIMESTAMP} to be set to the current date and +time when you create a row, but not to be changed whenever the row is updated +later: -@example -mysql> select INSERT('Quadratic', 3, 4, 'What'); - -> 'QuWhattic' -@end example +@itemize @bullet +@item +Let @strong{MySQL} set the column when the row is created. +This will initialize it to the current date and time. -This function is multi-byte safe. +@item +When you perform subsequent updates to other columns in the row, set +the @code{TIMESTAMP} column explicitly to its current value. +@end itemize -@findex ELT() -@item ELT(N,str1,str2,str3,...) -Returns @code{str1} if @code{N} = @code{1}, @code{str2} if @code{N} = -@code{2}, and so on. Returns @code{NULL} if @code{N} is less than @code{1} -or greater than the number of arguments. @code{ELT()} is the complement of -@code{FIELD()}: +On the other hand, you may find it just as easy to use a @code{DATETIME} +column that you initialize to @code{NOW()} when the row is created and +leave alone for subsequent updates. -@example -mysql> select ELT(1, 'ej', 'Heja', 'hej', 'foo'); - -> 'ej' -mysql> select ELT(4, 'ej', 'Heja', 'hej', 'foo'); - -> 'foo' -@end example +@code{TIMESTAMP} values may range from the beginning of 1970 to sometime in +the year 2037, with a resolution of one second. Values are displayed as +numbers. -@findex FIELD() -@item FIELD(str,str1,str2,str3,...) -Returns the index of @code{str} in the @code{str1}, @code{str2}, -@code{str3}, @code{...} list. -Returns @code{0} if @code{str} is not found. -@code{FIELD()} is the complement of @code{ELT()}: +The format in which @strong{MySQL} retrieves and displays @code{TIMESTAMP} +values depends on the display size, as illustrated by the table below. The +`full' @code{TIMESTAMP} format is 14 digits, but @code{TIMESTAMP} columns may +be created with shorter display sizes: -@example -mysql> select FIELD('ej', 'Hej', 'ej', 'Heja', 'hej', 'foo'); - -> 2 -mysql> select FIELD('fo', 'Hej', 'ej', 'Heja', 'hej', 'foo'); - -> 0 -@end example +@multitable @columnfractions .3 .7 +@item @strong{Column type} @tab @strong{Display format} +@item @code{TIMESTAMP(14)} @tab @code{YYYYMMDDHHMMSS} +@item @code{TIMESTAMP(12)} @tab @code{YYMMDDHHMMSS} +@item @code{TIMESTAMP(10)} @tab @code{YYMMDDHHMM} +@item @code{TIMESTAMP(8)} @tab @code{YYYYMMDD} +@item @code{TIMESTAMP(6)} @tab @code{YYMMDD} +@item @code{TIMESTAMP(4)} @tab @code{YYMM} +@item @code{TIMESTAMP(2)} @tab @code{YY} +@end multitable -@findex FIND_IN_SET() -@item FIND_IN_SET(str,strlist) -Returns a value @code{1} to @code{N} if the string @code{str} is in the list -@code{strlist} consisting of @code{N} substrings. A string list is a string -composed of substrings separated by @samp{,} characters. If the first -argument is a constant string and the second is a column of type @code{SET}, -the @code{FIND_IN_SET()} function is optimized to use bit arithmetic! -Returns @code{0} if @code{str} is not in @code{strlist} or if @code{strlist} -is the empty string. Returns @code{NULL} if either argument is @code{NULL}. -This function will not work properly if the first argument contains a -@samp{,}: +All @code{TIMESTAMP} columns have the same storage size, regardless of +display size. The most common display sizes are 6, 8, 12, and 14. You can +specify an arbitrary display size at table creation time, but values of 0 or +greater than 14 are coerced to 14. Odd-valued sizes in the range from 1 to +13 are coerced to the next higher even number. -@example -mysql> SELECT FIND_IN_SET('b','a,b,c,d'); - -> 2 -@end example +You can specify @code{DATETIME}, @code{DATE}, and @code{TIMESTAMP} values using +any of a common set of formats: -@findex MAKE_SET() -@item MAKE_SET(bits,str1,str2,...) -Returns a set (a string containing substrings separated by @samp{,} -characters) consisting of the strings that have the corresponding bit in -@code{bits} set. @code{str1} corresponds to bit 0, @code{str2} to bit 1, -etc. @code{NULL} strings in @code{str1}, @code{str2}, @code{...} -are not appended to the result: +@itemize @bullet +@item +As a string in either @code{'YYYY-MM-DD HH:MM:SS'} or @code{'YY-MM-DD +HH:MM:SS'} format. A ``relaxed'' syntax is allowed---any punctuation +character may be used as the delimiter between date parts or time parts. +For example, @code{'98-12-31 11:30:45'}, @code{'98.12.31 11+30+45'}, +@code{'98/12/31 11*30*45'}, and @code{'98@@12@@31 11^30^45'} are +equivalent. -@example -mysql> SELECT MAKE_SET(1,'a','b','c'); - -> 'a' -mysql> SELECT MAKE_SET(1 | 4,'hello','nice','world'); - -> 'hello,world' -mysql> SELECT MAKE_SET(0,'a','b','c'); - -> '' -@end example +@item +As a string in either @code{'YYYY-MM-DD'} or @code{'YY-MM-DD'} format. +A ``relaxed'' syntax is allowed here, too. For example, @code{'98-12-31'}, +@code{'98.12.31'}, @code{'98/12/31'}, and @code{'98@@12@@31'} are +equivalent. -@findex EXPORT_SET() -@item EXPORT_SET(bits,on,off,[separator,[number_of_bits]]) -Returns a string where for every bit set in 'bit', you get an 'on' string -and for every reset bit you get an 'off' string. Each string is separated -with 'separator' (default ',') and only 'number_of_bits' (default 64) of -'bits' is used: +@item +As a string with no delimiters in either @code{'YYYYMMDDHHMMSS'} or +@code{'YYMMDDHHMMSS'} format, provided that the string makes sense as a +date. For example, @code{'19970523091528'} and @code{'970523091528'} are +interpreted as @code{'1997-05-23 09:15:28'}, but @code{'971122129015'} is +illegal (it has a nonsensical minute part) and becomes @code{'0000-00-00 +00:00:00'}. -@example -mysql> select EXPORT_SET(5,'Y','N',',',4) - -> Y,N,Y,N -@end example +@item +As a string with no delimiters in either @code{'YYYYMMDD'} or @code{'YYMMDD'} +format, provided that the string makes sense as a date. For example, +@code{'19970523'} and @code{'970523'} are interpreted as +@code{'1997-05-23'}, but @code{'971332'} is illegal (it has nonsensical month +and day parts) and becomes @code{'0000-00-00'}. -@findex LCASE() -@findex LOWER() -@item LCASE(str) -@itemx LOWER(str) -Returns the string @code{str} with all characters changed to lowercase -according to the current character set mapping (the default is ISO-8859-1 -Latin1): +@item +As a number in either @code{YYYYMMDDHHMMSS} or @code{YYMMDDHHMMSS} +format, provided that the number makes sense as a date. For example, +@code{19830905132800} and @code{830905132800} are interpreted as +@code{'1983-09-05 13:28:00'}. -@example -mysql> select LCASE('QUADRATICALLY'); - -> 'quadratically' -@end example +@item +As a number in either @code{YYYYMMDD} or @code{YYMMDD} +format, provided that the number makes sense as a date. For example, +@code{19830905} and @code{830905} are interpreted as @code{'1983-09-05'}. -This function is multi-byte safe. +@item +As the result of a function that returns a value that is acceptable +in a @code{DATETIME}, @code{DATE}, or @code{TIMESTAMP} context, such as +@code{NOW()} or @code{CURRENT_DATE}. +@end itemize -@findex UCASE() -@findex UPPER() -@item UCASE(str) -@itemx UPPER(str) -Returns the string @code{str} with all characters changed to uppercase -according to the current character set mapping (the default is ISO-8859-1 -Latin1): +Illegal @code{DATETIME}, @code{DATE}, or @code{TIMESTAMP} values are converted +to the ``zero'' value of the appropriate type (@code{'0000-00-00 00:00:00'}, +@code{'0000-00-00'}, or @code{00000000000000}). -@example -mysql> select UCASE('Hej'); - -> 'HEJ' -@end example +For values specified as strings that include date part delimiters, it is not +necessary to specify two digits for month or day values that are less than +@code{10}. @code{'1979-6-9'} is the same as @code{'1979-06-09'}. Similarly, +for values specified as strings that include time part delimiters, it is not +necessary to specify two digits for hour, month, or second values that are +less than @code{10}. @code{'1979-10-30 1:2:3'} is the same as +@code{'1979-10-30 01:02:03'}. -This function is multi-byte safe. +Values specified as numbers should be 6, 8, 12, or 14 digits long. If the +number is 8 or 14 digits long, it is assumed to be in @code{YYYYMMDD} or +@code{YYYYMMDDHHMMSS} format and that the year is given by the first 4 +digits. If the number is 6 or 12 digits long, it is assumed to be in +@code{YYMMDD} or @code{YYMMDDHHMMSS} format and that the year is given by the +first 2 digits. Numbers that are not one of these lengths are interpreted +as though padded with leading zeros to the closest length. -@findex FILE -@findex LOAD_FILE() -@item LOAD_FILE(file_name) -Reads the file and returns the file contents as a string. The file -must be on the server, you must specify the full pathname to the -file, and you must have the @strong{file} privilege. The file must -be readable by all and be smaller than @code{max_allowed_packet}. +@cindex non-delimited strings +@cindex strings, non-delimited +Values specified as non-delimited strings are interpreted using their length +as given. If the string is 8 or 14 characters long, the year is assumed to +be given by the first 4 characters. Otherwise the year is assumed to be +given by the first 2 characters. The string is interpreted from left to +right to find year, month, day, hour, minute, and second values, for as many +parts as are present in the string. This means you should not use strings +that have fewer than 6 characters. For example, if you specify @code{'9903'}, +thinking that will represent March, 1999, you will find that @strong{MySQL} +inserts a ``zero'' date into your table. This is because the year and month +values are @code{99} and @code{03}, but the day part is missing (zero), so +the value is not a legal date. -If the file doesn't exist or can't be read due to one of the above reasons, -the function returns @code{NULL}: +@code{TIMESTAMP} columns store legal values using the full precision with +which the value was specified, regardless of the display size. This has +several implications: -@example -mysql> UPDATE table_name - SET blob_column=LOAD_FILE("/tmp/picture") - WHERE id=1; -@end example -@end table +@itemize @bullet +@item +Always specify year, month, and day, even if your column types are +@code{TIMESTAMP(4)} or @code{TIMESTAMP(2)}. Otherwise, the value will not +be a legal date and @code{0} will be stored. -If you are not using @strong{MySQL} Version 3.23, you have to do the reading -of the file inside your application and create an @code{INSERT} statement -to update the database with the file information. One way to do this, if -you are using the @strong{MySQL}++ library, can be found at -@uref{http://www.mysql.com/documentation/mysql++/mysql++-examples.html}. +@item +If you use @code{ALTER TABLE} to widen a narrow @code{TIMESTAMP} column, +information will be displayed that previously was ``hidden''. -@strong{MySQL} automatically converts numbers to strings as necessary, and -vice-versa: +@item +Similarly, narrowing a @code{TIMESTAMP} column does not cause information to +be lost, except in the sense that less information is shown when the values +are displayed. -@example -mysql> SELECT 1+"1"; - -> 2 -mysql> SELECT CONCAT(2,' test'); - -> '2 test' -@end example +@item +Although @code{TIMESTAMP} values are stored to full precision, the only +function that operates directly on the underlying stored value is +@code{UNIX_TIMESTAMP()}. Other functions operate on the formatted retrieved +value. This means you cannot use functions such as @code{HOUR()} or +@code{SECOND()} unless the relevant part of the @code{TIMESTAMP} value is +included in the formatted value. For example, the @code{HH} part of a +@code{TIMESTAMP} column is not displayed unless the display size is at least +10, so trying to use @code{HOUR()} on shorter @code{TIMESTAMP} values +produces a meaningless result. +@end itemize -If you want to convert a number to a string explicitly, pass it as the -argument to @code{CONCAT()}. +You can to some extent assign values of one date type to an object +of a different date type. However, there may be some alteration +of the value or loss of information: -If a string function is given a binary string as an argument, the resulting -string is also a binary string. A number converted to a string is treated as -a binary string. This only affects comparisons. +@itemize @bullet +@item +If you assign a @code{DATE} value to a @code{DATETIME} or @code{TIMESTAMP} +object, the time part of the resulting value is set to @code{'00:00:00'}, +because the @code{DATE} value contains no time information. -@findex date and time functions -@findex functions, date and time -@node Date and time functions, Miscellaneous functions, String functions, Functions -@subsection Date and Time Functions +@item +If you assign a @code{DATETIME} or @code{TIMESTAMP} value to a @code{DATE} +object, the time part of the resulting value is deleted, because the +@code{DATE} type stores no time information. -See @ref{Date and time types} for a description of the range of values -each type has and the valid formats in which date and time values may be -specified. +@item +Remember that although @code{DATETIME}, @code{DATE}, and @code{TIMESTAMP} +values all can be specified using the same set of formats, the types do not +all have the same range of values. For example, @code{TIMESTAMP} values +cannot be earlier than @code{1970} or later than @code{2037}. This means +that a date such as @code{'1968-01-01'}, while legal as a @code{DATETIME} or +@code{DATE} value, is not a valid @code{TIMESTAMP} value and will be +converted to @code{0} if assigned to such an object. +@end itemize -Here is an example that uses date functions. The query below selects -all records with a @code{date_col} value from within the last 30 days: +@cindex problems, date values +@cindex date values, problems +Be aware of certain pitfalls when specifying date values: -@example -mysql> SELECT something FROM table - WHERE TO_DAYS(NOW()) - TO_DAYS(date_col) <= 30; -@end example +@itemize @bullet +@item +The relaxed format allowed for values specified as strings can be deceiving. +For example, a value such as @code{'10:11:12'} might look like a time value +because of the @samp{:} delimiter, but if used in a date context will be +interpreted as the year @code{'2010-11-12'}. The value @code{'10:45:15'} +will be converted to @code{'0000-00-00'} because @code{'45'} is not a legal +month. -@table @code -@findex DAYOFWEEK() -@item DAYOFWEEK(date) -Returns the weekday index +@item +Year values specified as two digits are ambiguous, because the century is +unknown. @strong{MySQL} interprets 2-digit year values using the following +rules: -for @code{date} (@code{1} = Sunday, @code{2} = Monday, ... @code{7} = -Saturday). These index values correspond to the ODBC standard: +@itemize @minus +@item +Year values in the range @code{00-69} are converted to @code{2000-2069}. -@example -mysql> select DAYOFWEEK('1998-02-03'); - -> 3 -@end example +@item +Year values in the range @code{70-99} are converted to @code{1970-1999}. +@end itemize +@end itemize -@findex WEEKDAY() -@item WEEKDAY(date) -Returns the weekday index for -@code{date} (@code{0} = Monday, @code{1} = Tuesday, ... @code{6} = Sunday): +@tindex TIME +@node TIME, YEAR, DATETIME, Date and time types +@subsubsection The @code{TIME} Type -@example -mysql> select WEEKDAY('1997-10-04 22:23:00'); - -> 5 -mysql> select WEEKDAY('1997-11-05'); - -> 2 -@end example +@strong{MySQL} retrieves and displays @code{TIME} values in @code{'HH:MM:SS'} +format (or @code{'HHH:MM:SS'} format for large hours values). @code{TIME} +values may range from @code{'-838:59:59'} to @code{'838:59:59'}. The reason +the hours part may be so large is that the @code{TIME} type may be used not +only to represent a time of day (which must be less than 24 hours), but also +elapsed time or a time interval between two events (which may be much greater +than 24 hours, or even negative). -@findex DAYOFMONTH() -@item DAYOFMONTH(date) -Returns the day of the month for @code{date}, in the range @code{1} to -@code{31}: +You can specify @code{TIME} values in a variety of formats: -@example -mysql> select DAYOFMONTH('1998-02-03'); - -> 3 -@end example +@itemize @bullet +@item +As a string in @code{'D HH:MM:SS.fraction'} format. (Note that +@strong{MySQL} doesn't yet store the fraction for the time column). One +can also use one of the following ``relaxed'' syntax: -@findex DAYOFYEAR() -@item DAYOFYEAR(date) -Returns the day of the year for @code{date}, in the range @code{1} to -@code{366}: +@code{HH:MM:SS.fraction}, @code{HH:MM:SS}, @code{HH:MM}, @code{D HH:MM:SS}, +@code{D HH:MM}, @code{D HH} or @code{SS}. Here @code{D} is days between 0-33. -@example -mysql> select DAYOFYEAR('1998-02-03'); - -> 34 -@end example +@item +As a string with no delimiters in @code{'HHMMSS'} format, provided that +it makes sense as a time. For example, @code{'101112'} is understood as +@code{'10:11:12'}, but @code{'109712'} is illegal (it has a nonsensical +minute part) and becomes @code{'00:00:00'}. -@findex MONTH() -@item MONTH(date) -Returns the month for @code{date}, in the range @code{1} to @code{12}: +@item +As a number in @code{HHMMSS} format, provided that it makes sense as a time. +For example, @code{101112} is understood as @code{'10:11:12'}. The following +alternative formats are also understood: @code{SS}, @code{MMSS},@code{HHMMSS}, +@code{HHMMSS.fraction}. Note that @strong{MySQL} doesn't yet store the +fraction part. -@example -mysql> select MONTH('1998-02-03'); - -> 2 -@end example +@item +As the result of a function that returns a value that is acceptable +in a @code{TIME} context, such as @code{CURRENT_TIME}. +@end itemize -@findex DAYNAME() -@item DAYNAME(date) -Returns the name of the weekday for @code{date}: +For @code{TIME} values specified as strings that include a time part +delimiter, it is not necessary to specify two digits for hours, minutes, or +seconds values that are less than @code{10}. @code{'8:3:2'} is the same as +@code{'08:03:02'}. -@example -mysql> select DAYNAME("1998-02-05"); - -> 'Thursday' -@end example +Be careful about assigning ``short'' @code{TIME} values to a @code{TIME} +column. Without semicolon, @strong{MySQL} interprets values using the +assumption that the rightmost digits represent seconds. (@strong{MySQL} +interprets @code{TIME} values as elapsed time rather than as time of +day.) For example, you might think of @code{'1112'} and @code{1112} as +meaning @code{'11:12:00'} (12 minutes after 11 o'clock), but +@strong{MySQL} interprets them as @code{'00:11:12'} (11 minutes, 12 seconds). +Similarly, @code{'12'} and @code{12} are interpreted as @code{'00:00:12'}. +@code{TIME} values with semicolon, instead, are always treated as +time of the day. That is @code{'11:12'} will mean @code{'11:12:00'}, +not @code{'00:11:12'}. -@findex MONTHNAME() -@item MONTHNAME(date) -Returns the name of the month for @code{date}: +Values that lie outside the @code{TIME} range +but are otherwise legal are clipped to the appropriate +endpoint of the range. For example, @code{'-850:00:00'} and +@code{'850:00:00'} are converted to @code{'-838:59:59'} and +@code{'838:59:59'}. -@example -mysql> select MONTHNAME("1998-02-05"); - -> 'February' -@end example +Illegal @code{TIME} values are converted to @code{'00:00:00'}. Note that +because @code{'00:00:00'} is itself a legal @code{TIME} value, there is no way +to tell, from a value of @code{'00:00:00'} stored in a table, whether the +original value was specified as @code{'00:00:00'} or whether it was illegal. -@findex QUARTER() -@item QUARTER(date) -Returns the quarter of the year for @code{date}, in the range @code{1} -to @code{4}: +@tindex YEAR +@node YEAR, , TIME, Date and time types +@subsubsection The @code{YEAR} Type -@example -mysql> select QUARTER('98-04-01'); - -> 2 -@end example +The @code{YEAR} type is a 1-byte type used for representing years. -@findex WEEK() -@item WEEK(date) -@itemx WEEK(date,first) -With a single argument, returns the week for @code{date}, in the range -@code{0} to @code{53} (yes, there may be the beginnings of a week 53), -for locations where Sunday is the first day of the week. The -two-argument form of @code{WEEK()} allows you to specify whether the -week starts on Sunday or Monday. The week starts on Sunday if the -second argument is @code{0}, on Monday if the second argument is -@code{1}: +@strong{MySQL} retrieves and displays @code{YEAR} values in @code{YYYY} +format. The range is @code{1901} to @code{2155}. -@example -mysql> select WEEK('1998-02-20'); - -> 7 -mysql> select WEEK('1998-02-20',0); - -> 7 -mysql> select WEEK('1998-02-20',1); - -> 8 -mysql> select WEEK('1998-12-31',1); - -> 53 -@end example +You can specify @code{YEAR} values in a variety of formats: -@findex YEAR() -@item YEAR(date) -Returns the year for @code{date}, in the range @code{1000} to @code{9999}: +@itemize @bullet +@item +As a four-digit string in the range @code{'1901'} to @code{'2155'}. -@example -mysql> select YEAR('98-02-03'); - -> 1998 -@end example +@item +As a four-digit number in the range @code{1901} to @code{2155}. -@item YEARWEEK(date) -@itemx YEARWEEK(date,first) -Returns year and week for a date. The second arguments works exactly -like the second argument to @code{WEEK()}. Note that the year may be -different from the year in the date argument for the first and the last -week of the year: +@item +As a two-digit string in the range @code{'00'} to @code{'99'}. Values in the +ranges @code{'00'} to @code{'69'} and @code{'70'} to @code{'99'} are +converted to @code{YEAR} values in the ranges @code{2000} to @code{2069} and +@code{1970} to @code{1999}. -@example -mysql> select YEARWEEK('1987-01-01'); - -> 198653 -@end example +@item +As a two-digit number in the range @code{1} to @code{99}. Values in the +ranges @code{1} to @code{69} and @code{70} to @code{99} are converted to +@code{YEAR} values in the ranges @code{2001} to @code{2069} and @code{1970} +to @code{1999}. Note that the range for two-digit numbers is slightly +different than the range for two-digit strings, because you cannot specify zero +directly as a number and have it be interpreted as @code{2000}. You +@emph{must} specify it as a string @code{'0'} or @code{'00'} or it will be +interpreted as @code{0000}. -@findex HOUR() -@item HOUR(time) -Returns the hour for @code{time}, in the range @code{0} to @code{23}: +@item +As the result of a function that returns a value that is acceptable +in a @code{YEAR} context, such as @code{NOW()}. +@end itemize -@example -mysql> select HOUR('10:05:03'); - -> 10 -@end example +Illegal @code{YEAR} values are converted to @code{0000}. -@findex MINUTE() -@item MINUTE(time) -Returns the minute for @code{time}, in the range @code{0} to @code{59}: +@cindex types, strings +@cindex string types +@node String types, Choosing types, Date and time types, Column types +@subsection String Types -@example -mysql> select MINUTE('98-02-03 10:05:03'); - -> 5 -@end example +@tindex CHAR +@tindex VARCHAR +@menu +* CHAR:: The @code{CHAR} and @code{VARCHAR} types +* BLOB:: The @code{BLOB} and @code{TEXT} types +* ENUM:: The @code{ENUM} type +* SET:: The @code{SET} type +@end menu -@findex SECOND() -@item SECOND(time) -Returns the second for @code{time}, in the range @code{0} to @code{59}: +The string types are @code{CHAR}, @code{VARCHAR}, @code{BLOB}, @code{TEXT}, +@code{ENUM}, and @code{SET}. This section describes how these types work, +their storage requirements, and how to use them in your queries. -@example -mysql> select SECOND('10:05:03'); - -> 3 -@end example +@node CHAR, BLOB, String types, String types +@subsubsection The @code{CHAR} and @code{VARCHAR} Types -@findex PERIOD_ADD() -@item PERIOD_ADD(P,N) -Adds @code{N} months to period @code{P} (in the format @code{YYMM} or -@code{YYYYMM}). Returns a value in the format @code{YYYYMM}. +The @code{CHAR} and @code{VARCHAR} types are similar, but differ in the +way they are stored and retrieved. -Note that the period argument @code{P} is @emph{not} a date value: +The length of a @code{CHAR} column is fixed to the length that you declare +when you create the table. The length can be any value between 1 and 255. +(As of @strong{MySQL} Version 3.23, the length of @code{CHAR} may be 0 to 255.) +When @code{CHAR} values are stored, they are right-padded with spaces to the +specified length. When @code{CHAR} values are retrieved, trailing spaces are +removed. -@example -mysql> select PERIOD_ADD(9801,2); - -> 199803 -@end example +Values in @code{VARCHAR} columns are variable-length strings. You can +declare a @code{VARCHAR} column to be any length between 1 and 255, just as +for @code{CHAR} columns. However, in contrast to @code{CHAR}, @code{VARCHAR} +values are stored using only as many characters as are needed, plus one byte +to record the length. Values are not padded; instead, trailing spaces are +removed when values are stored. (This space removal differs from the ANSI +SQL specification.) -@findex PERIOD_DIFF() -@item PERIOD_DIFF(P1,P2) -Returns the number of months between periods @code{P1} and @code{P2}. -@code{P1} and @code{P2} should be in the format @code{YYMM} or @code{YYYYMM}. +If you assign a value to a @code{CHAR} or @code{VARCHAR} column that +exceeds the column's maximum length, the value is truncated to fit. -Note that the period arguments @code{P1} and @code{P2} are @emph{not} -date values: +The table below illustrates the differences between the two types of columns +by showing the result of storing various string values into @code{CHAR(4)} +and @code{VARCHAR(4)} columns: -@example -mysql> select PERIOD_DIFF(9802,199703); - -> 11 -@end example +@c Need to use @(space) to make sure second column values retain spacing +@c in output for table below. -@findex DATE_ADD() -@findex DATE_SUB() -@findex ADDDATE() -@findex SUBDATE() -@findex EXTRACT(type FROM date) -@item DATE_ADD(date,INTERVAL expr type) -@itemx DATE_SUB(date,INTERVAL expr type) -@itemx ADDDATE(date,INTERVAL expr type) -@itemx SUBDATE(date,INTERVAL expr type) +@multitable @columnfractions .2 .15 .2 .2 .25 +@item @strong{Value} @tab @code{CHAR(4)} @tab @strong{Storage required} @tab @code{VARCHAR(4)} @tab @strong{Storage required} +@item @code{''} @tab @code{'@ @ @ @ '} @tab 4 bytes @tab @code{''} @tab 1 byte +@item @code{'ab'} @tab @code{'ab@ @ '} @tab 4 bytes @tab @code{'ab'} @tab 3 bytes +@item @code{'abcd'} @tab @code{'abcd'} @tab 4 bytes @tab @code{'abcd'} @tab 5 bytes +@item @code{'abcdefgh'} @tab @code{'abcd'} @tab 4 bytes @tab @code{'abcd'} @tab 5 bytes +@end multitable -These functions perform date arithmetic. They are new for @strong{MySQL} -Version 3.22. @code{ADDDATE()} and @code{SUBDATE()} are synonyms for -@code{DATE_ADD()} and @code{DATE_SUB()}. +The values retrieved from the @code{CHAR(4)} and @code{VARCHAR(4)} columns +will be the same in each case, because trailing spaces are removed from +@code{CHAR} columns upon retrieval. -In @strong{MySQL} Version 3.23, you can use @code{+} and @code{-} instead of -@code{DATE_ADD()} and @code{DATE_SUB()} if the expression on the right side is -a date or datetime column. (See example) +Values in @code{CHAR} and @code{VARCHAR} columns are sorted and compared +in case-insensitive fashion, unless the @code{BINARY} attribute was +specified when the table was created. The @code{BINARY} attribute means +that column values are sorted and compared in case-sensitive fashion +according to the ASCII order of the machine where the @strong{MySQL} +server is running. @code{BINARY} doesn't affect how the column is stored +or retrieved. -@code{date} is a @code{DATETIME} or @code{DATE} value specifying the starting -date. @code{expr} is an expression specifying the interval value to be added -or subtracted from the starting date. @code{expr} is a string; it may start -with a @samp{-} for negative intervals. @code{type} is a keyword indicating -how the expression should be interpreted. +The @code{BINARY} attribute is sticky. This means that if a column marked +@code{BINARY} is used in an expression, the whole expression is compared as a +@code{BINARY} value. -The @code{EXTRACT(type FROM date)} function returns the 'type' -interval from the date. +@strong{MySQL} may silently change the type of a @code{CHAR} or @code{VARCHAR} +column at table creation time. +@xref{Silent column changes}. -The following table shows how the @code{type} and @code{expr} arguments -are related: +@tindex BLOB +@tindex TEXT +@node BLOB, ENUM, CHAR, String types +@subsubsection The @code{BLOB} and @code{TEXT} Types -@multitable @columnfractions .5 .5 -@item @code{type} @strong{value} @tab @strong{Expected} @code{expr} @strong{format} -@item @code{SECOND} @tab @code{SECONDS} -@item @code{MINUTE} @tab @code{MINUTES} -@item @code{HOUR} @tab @code{HOURS} -@item @code{DAY} @tab @code{DAYS} -@item @code{MONTH} @tab @code{MONTHS} -@item @code{YEAR} @tab @code{YEARS} -@item @code{MINUTE_SECOND} @tab @code{"MINUTES:SECONDS"} -@item @code{HOUR_MINUTE} @tab @code{"HOURS:MINUTES"} -@item @code{DAY_HOUR} @tab @code{"DAYS HOURS"} -@item @code{YEAR_MONTH} @tab @code{"YEARS-MONTHS"} -@item @code{HOUR_SECOND} @tab @code{"HOURS:MINUTES:SECONDS"} -@item @code{DAY_MINUTE} @tab @code{"DAYS HOURS:MINUTES"} -@item @code{DAY_SECOND} @tab @code{"DAYS HOURS:MINUTES:SECONDS"} -@end multitable +A @code{BLOB} is a binary large object that can hold a variable amount of +data. The four @code{BLOB} types @code{TINYBLOB}, @code{BLOB}, +@code{MEDIUMBLOB}, and @code{LONGBLOB} differ only in the maximum length of +the values they can hold. +@xref{Storage requirements}. -@strong{MySQL} allows any punctuation delimiter in the @code{expr} format. -Those shown in the table are the suggested delimiters. If the @code{date} -argument is a @code{DATE} value and your calculations involve only -@code{YEAR}, @code{MONTH}, and @code{DAY} parts (that is, no time parts), the -result is a @code{DATE} value. Otherwise the result is a @code{DATETIME} -value: +The four @code{TEXT} types @code{TINYTEXT}, @code{TEXT}, @code{MEDIUMTEXT}, +and @code{LONGTEXT} correspond to the four @code{BLOB} types and have the +same maximum lengths and storage requirements. The only difference between +@code{BLOB} and @code{TEXT} types is that sorting and comparison is performed +in case-sensitive fashion for @code{BLOB} values and case-insensitive fashion +for @code{TEXT} values. In other words, a @code{TEXT} is a case-insensitive +@code{BLOB}. -@example -mysql> SELECT "1997-12-31 23:59:59" + INTERVAL 1 SECOND; - -> 1998-01-01 00:00:00 -mysql> SELECT INTERVAL 1 DAY + "1997-12-31"; - -> 1998-01-01 -mysql> SELECT "1998-01-01" - INTERVAL 1 SECOND; - -> 1997-12-31 23:59:59 -mysql> SELECT DATE_ADD("1997-12-31 23:59:59", - INTERVAL 1 SECOND); - -> 1998-01-01 00:00:00 -mysql> SELECT DATE_ADD("1997-12-31 23:59:59", - INTERVAL 1 DAY); - -> 1998-01-01 23:59:59 -mysql> SELECT DATE_ADD("1997-12-31 23:59:59", - INTERVAL "1:1" MINUTE_SECOND); - -> 1998-01-01 00:01:00 -mysql> SELECT DATE_SUB("1998-01-01 00:00:00", - INTERVAL "1 1:1:1" DAY_SECOND); - -> 1997-12-30 22:58:59 -mysql> SELECT DATE_ADD("1998-01-01 00:00:00", - INTERVAL "-1 10" DAY_HOUR); - -> 1997-12-30 14:00:00 -mysql> SELECT DATE_SUB("1998-01-02", INTERVAL 31 DAY); - -> 1997-12-02 -mysql> SELECT EXTRACT(YEAR FROM "1999-07-02"); - -> 1999 -mysql> SELECT EXTRACT(YEAR_MONTH FROM "1999-07-02 01:02:03"); - -> 199907 -mysql> SELECT EXTRACT(DAY_MINUTE FROM "1999-07-02 01:02:03"); - -> 20102 -@end example +If you assign a value to a @code{BLOB} or @code{TEXT} column that exceeds +the column type's maximum length, the value is truncated to fit. -If you specify an interval value that is too short (does not include all the -interval parts that would be expected from the @code{type} keyword), -@strong{MySQL} assumes you have left out the leftmost parts of the interval -value. For example, if you specify a @code{type} of @code{DAY_SECOND}, the -value of @code{expr} is expected to have days, hours, minutes, and seconds -parts. If you specify a value like @code{"1:10"}, @strong{MySQL} assumes -that the days and hours parts are missing and the value represents minutes -and seconds. In other words, @code{"1:10" DAY_SECOND} is interpreted in such -a way that it is equivalent to @code{"1:10" MINUTE_SECOND}. This is -analogous to the way that @strong{MySQL} interprets @code{TIME} values -as representing elapsed time rather than as time of day. +In most respects, you can regard a @code{TEXT} column as a @code{VARCHAR} +column that can be as big as you like. Similarly, you can regard a +@code{BLOB} column as a @code{VARCHAR BINARY} column. The differences are: -Note that if you add or subtract a date value against something that -contains a time part, the date value will be automatically converted to a -datetime value: +@itemize @bullet +@item +You can have indexes on @code{BLOB} and @code{TEXT} columns with +@strong{MySQL} Version 3.23.2 and newer. Older versions of +@strong{MySQL} did not support this. -@example -mysql> select date_add("1999-01-01", interval 1 day); - -> 1999-01-02 -mysql> select date_add("1999-01-01", interval 1 hour); - -> 1999-01-01 01:00:00 -@end example +@item +There is no trailing-space removal for @code{BLOB} and @code{TEXT} columns +when values are stored, as there is for @code{VARCHAR} columns. -If you use really incorrect dates, the result is @code{NULL}. If you add -@code{MONTH}, @code{YEAR_MONTH}, or @code{YEAR} and the resulting date -has a day that is larger than the maximum day for the new month, the day is -adjusted to the maximum days in the new month: +@item +@cindex default values, @code{BLOB} and @code{TEXT} columns +@cindex @code{BLOB} columns, default values +@cindex @code{TEXT} columns, default values +@code{BLOB} and @code{TEXT} columns cannot have @code{DEFAULT} values. +@end itemize -@example -mysql> select DATE_ADD('1998-01-30', Interval 1 month); - -> 1998-02-28 -@end example +@strong{MyODBC} defines @code{BLOB} values as @code{LONGVARBINARY} and +@code{TEXT} values as @code{LONGVARCHAR}. -Note from the preceding example that the word @code{INTERVAL} and the -@code{type} keyword are not case sensitive. +Because @code{BLOB} and @code{TEXT} values may be extremely long, you +may run up against some constraints when using them: -@findex TO_DAYS() -@item TO_DAYS(date) -Given a date @code{date}, returns a daynumber (the number of days since year -0): +@itemize @bullet +@item +If you want to use @code{GROUP BY} or @code{ORDER BY} on a @code{BLOB} or +@code{TEXT} column, you must convert the column value into a fixed-length +object. The standard way to do this is with the @code{SUBSTRING} +function. For example: @example -mysql> select TO_DAYS(950501); - -> 728779 -mysql> select TO_DAYS('1997-10-07'); - -> 729669 +mysql> select comment from tbl_name,substring(comment,20) as substr + ORDER BY substr; @end example -@code{TO_DAYS()} is not intended for use with values that precede the advent -of the Gregorian calendar (1582), because it doesn't take into account the -days that were lost when the calendar was changed. - -@findex FROM_DAYS() -@item FROM_DAYS(N) -Given a daynumber @code{N}, returns a @code{DATE} value: +If you don't do this, only the first @code{max_sort_length} bytes of the +column are used when sorting. The default value of @code{max_sort_length} is +1024; this value can be changed using the @code{-O} option when starting the +@code{mysqld} server. You can group on an expression involving @code{BLOB} or +@code{TEXT} values by specifying the column position or by using an alias: @example -mysql> select FROM_DAYS(729669); - -> '1997-10-07' +mysql> select id,substring(blob_col,1,100) from tbl_name + GROUP BY 2; +mysql> select id,substring(blob_col,1,100) as b from tbl_name + GROUP BY b; @end example -@code{FROM_DAYS()} is not intended for use with values that precede the -advent of the Gregorian calendar (1582), because it doesn't take into account -the days that were lost when the calendar was changed. +@item +The maximum size of a @code{BLOB} or @code{TEXT} object is determined by its +type, but the largest value you can actually transmit between the client and +server is determined by the amount of available memory and the size of the +communications buffers. You can change the message buffer size, but you must +do so on both the server and client ends. @xref{Server parameters}. +@end itemize -@findex DATE_FORMAT() -@item DATE_FORMAT(date,format) -Formats the @code{date} value according to the @code{format} string. The -following specifiers may be used in the @code{format} string: -@multitable @columnfractions .1 .6 -@item @code{%M} @tab Month name (@code{January}..@code{December}) -@item @code{%W} @tab Weekday name (@code{Sunday}..@code{Saturday}) -@item @code{%D} @tab Day of the month with English suffix (@code{1st}, @code{2nd}, @code{3rd}, etc.) -@item @code{%Y} @tab Year, numeric, 4 digits -@item @code{%y} @tab Year, numeric, 2 digits -@item @code{%X} @tab Year for the week where Sunday is the first day of the week, numeric, 4 digits, used with '%V' -@item @code{%x} @tab Year for the week, where Monday is the first day of the week, numeric, 4 digits, used with '%v' -@item @code{%a} @tab Abbreviated weekday name (@code{Sun}..@code{Sat}) -@item @code{%d} @tab Day of the month, numeric (@code{00}..@code{31}) -@item @code{%e} @tab Day of the month, numeric (@code{0}..@code{31}) -@item @code{%m} @tab Month, numeric (@code{01}..@code{12}) -@item @code{%c} @tab Month, numeric (@code{1}..@code{12}) -@item @code{%b} @tab Abbreviated month name (@code{Jan}..@code{Dec}) -@item @code{%j} @tab Day of year (@code{001}..@code{366}) -@item @code{%H} @tab Hour (@code{00}..@code{23}) -@item @code{%k} @tab Hour (@code{0}..@code{23}) -@item @code{%h} @tab Hour (@code{01}..@code{12}) -@item @code{%I} @tab Hour (@code{01}..@code{12}) -@item @code{%l} @tab Hour (@code{1}..@code{12}) -@item @code{%i} @tab Minutes, numeric (@code{00}..@code{59}) -@item @code{%r} @tab Time, 12-hour (@code{hh:mm:ss [AP]M}) -@item @code{%T} @tab Time, 24-hour (@code{hh:mm:ss}) -@item @code{%S} @tab Seconds (@code{00}..@code{59}) -@item @code{%s} @tab Seconds (@code{00}..@code{59}) -@item @code{%p} @tab @code{AM} or @code{PM} -@item @code{%w} @tab Day of the week (@code{0}=Sunday..@code{6}=Saturday) -@item @code{%U} @tab Week (@code{0}..@code{53}), where Sunday is the first day of the week -@item @code{%u} @tab Week (@code{0}..@code{53}), where Monday is the first day of the week -@item @code{%V} @tab Week (@code{1}..@code{53}), where Sunday is the first day of the week. Used with '%X' -@item @code{%v} @tab Week (@code{1}..@code{53}), where Monday is the first day of the week. Used with '%x' -@item @code{%%} @tab A literal @samp{%}. -@end multitable +Note that each @code{BLOB} or @code{TEXT} value is represented +internally by a separately allocated object. This is in contrast to all +other column types, for which storage is allocated once per column when +the table is opened. -All other characters are just copied to the result without interpretation: +@tindex ENUM +@node ENUM, SET, BLOB, String types +@subsubsection The @code{ENUM} Type -@example -mysql> select DATE_FORMAT('1997-10-04 22:23:00', '%W %M %Y'); - -> 'Saturday October 1997' -mysql> select DATE_FORMAT('1997-10-04 22:23:00', '%H:%i:%s'); - -> '22:23:00' -mysql> select DATE_FORMAT('1997-10-04 22:23:00', - '%D %y %a %d %m %b %j'); - -> '4th 97 Sat 04 10 Oct 277' -mysql> select DATE_FORMAT('1997-10-04 22:23:00', - '%H %k %I %r %T %S %w'); - -> '22 22 10 10:23:00 PM 22:23:00 00 6' -mysql> select DATE_FORMAT('1999-01-01', '%X %V'); - -> '1998 52' -@end example +An @code{ENUM} is a string object whose value normally is chosen from a list +of allowed values that are enumerated explicitly in the column specification +at table creation time. -As of @strong{MySQL} Version 3.23, the @samp{%} character is required before -format specifier characters. In earlier versions of @strong{MySQL}, -@samp{%} was optional. +The value may also be the empty string (@code{""}) or @code{NULL} under +certain circumstances: -@findex TIME_FORMAT() -@item TIME_FORMAT(time,format) -This is used like the @code{DATE_FORMAT()} function above, but the -@code{format} string may contain only those format specifiers that handle -hours, minutes, and seconds. Other specifiers produce a @code{NULL} value or -@code{0}. +@itemize @bullet +@item +If you insert an invalid value into an @code{ENUM} (that is, a string not +present in the list of allowed values), the empty string is inserted +instead as a special error value. -@findex CURDATE() -@findex CURRENT_DATE -@item CURDATE() -@itemx CURRENT_DATE -Returns today's date as a value in @code{'YYYY-MM-DD'} or @code{YYYYMMDD} -format, depending on whether the function is used in a string or numeric -context: +@item +If an @code{ENUM} is declared @code{NULL}, @code{NULL} is also a legal value +for the column, and the default value is @code{NULL}. If an @code{ENUM} is +declared @code{NOT NULL}, the default value is the first element of the +list of allowed values. +@end itemize -@example -mysql> select CURDATE(); - -> '1997-12-15' -mysql> select CURDATE() + 0; - -> 19971215 -@end example +Each enumeration value has an index: -@findex CURTIME() -@findex CURRENT_TIME -@item CURTIME() -@itemx CURRENT_TIME -Returns the current time as a value in @code{'HH:MM:SS'} or @code{HHMMSS} -format, depending on whether the function is used in a string or numeric -context: +@itemize @bullet +@item +Values from the list of allowable elements in the column specification are +numbered beginning with 1. + +@item +The index value of the empty string error value is 0. This means that you +can use the following @code{SELECT} statement to find rows into which invalid +@code{ENUM} values were assigned: @example -mysql> select CURTIME(); - -> '23:50:26' -mysql> select CURTIME() + 0; - -> 235026 +mysql> SELECT * FROM tbl_name WHERE enum_col=0; @end example -@findex NOW() -@findex SYSDATE() -@findex CURRENT_TIMESTAMP -@item NOW() -@itemx SYSDATE() -@itemx CURRENT_TIMESTAMP -Returns the current date and time as a value in @code{'YYYY-MM-DD HH:MM:SS'} -or @code{YYYYMMDDHHMMSS} format, depending on whether the function is used in -a string or numeric context: +@item +The index of the @code{NULL} value is @code{NULL}. +@end itemize -@example -mysql> select NOW(); - -> '1997-12-15 23:50:26' -mysql> select NOW() + 0; - -> 19971215235026 -@end example +For example, a column specified as @code{ENUM("one", "two", "three")} can +have any of the values shown below. The index of each value is also shown: -@findex UNIX_TIMESTAMP() -@item UNIX_TIMESTAMP() -@itemx UNIX_TIMESTAMP(date) -If called with no argument, returns a Unix timestamp (seconds since -@code{'1970-01-01 00:00:00'} GMT). If @code{UNIX_TIMESTAMP()} is called with -a @code{date} argument, it returns the value of the argument as seconds since -@code{'1970-01-01 00:00:00'} GMT. @code{date} may be a @code{DATE} string, -a @code{DATETIME} string, a @code{TIMESTAMP}, or a number in the format -@code{YYMMDD} or @code{YYYYMMDD} in local time: +@multitable @columnfractions .2 .8 +@item @strong{Value} @tab @strong{Index} +@item @code{NULL} @tab @code{NULL} +@item @code{""} @tab 0 +@item @code{"one"} @tab 1 +@item @code{"two"} @tab 2 +@item @code{"three"} @tab 3 +@end multitable -@example -mysql> select UNIX_TIMESTAMP(); - -> 882226357 -mysql> select UNIX_TIMESTAMP('1997-10-04 22:23:00'); - -> 875996580 -@end example +An enumeration can have a maximum of 65535 elements. -When @code{UNIX_TIMESTAMP} is used on a @code{TIMESTAMP} column, the function -will receive the value directly, with no implicit -``string-to-unix-timestamp'' conversion. -If you give @code{UNIX_TIMESTAMP()} a wrong or out-of-range date, it will -return 0. +Lettercase is irrelevant when you assign values to an @code{ENUM} column. +However, values retrieved from the column later have lettercase matching the +values that were used to specify the allowable values at table creation time. -@findex FROM_UNIXTIME() -@item FROM_UNIXTIME(unix_timestamp) -Returns a representation of the @code{unix_timestamp} argument as a value in -@code{'YYYY-MM-DD HH:MM:SS'} or @code{YYYYMMDDHHMMSS} format, depending on -whether the function is used in a string or numeric context: +If you retrieve an @code{ENUM} in a numeric context, the column value's +index is returned. For example, you can retrieve numeric values from +an @code{ENUM} column like this: @example -mysql> select FROM_UNIXTIME(875996580); - -> '1997-10-04 22:23:00' -mysql> select FROM_UNIXTIME(875996580) + 0; - -> 19971004222300 +mysql> SELECT enum_col+0 FROM tbl_name; @end example -@findex FROM_UNIXTIME() -@item FROM_UNIXTIME(unix_timestamp,format) -Returns a string representation of the Unix timestamp, formatted according to -the @code{format} string. @code{format} may contain the same specifiers as -those listed in the entry for the @code{DATE_FORMAT()} function: +If you store a number into an @code{ENUM}, the number is treated as an +index, and the value stored is the enumeration member with that index. +(However, this will not work with @code{LOAD DATA}, which treats all +input as strings.) -@example -mysql> select FROM_UNIXTIME(UNIX_TIMESTAMP(), - '%Y %D %M %h:%i:%s %x'); - -> '1997 23rd December 03:43:30 x' -@end example +@code{ENUM} values are sorted according to the order in which the enumeration +members were listed in the column specification. (In other words, +@code{ENUM} values are sorted according to their index numbers.) For +example, @code{"a"} sorts before @code{"b"} for @code{ENUM("a", "b")}, but +@code{"b"} sorts before @code{"a"} for @code{ENUM("b", "a")}. The empty +string sorts before non-empty strings, and @code{NULL} values sort before +all other enumeration values. -@findex SEC_TO_TIME() -@item SEC_TO_TIME(seconds) -Returns the @code{seconds} argument, converted to hours, minutes, and seconds, -as a value in @code{'HH:MM:SS'} or @code{HHMMSS} format, depending on whether -the function is used in a string or numeric context: +If you want to get all possible values for an @code{ENUM} column, you should +use: @code{SHOW COLUMNS FROM table_name LIKE enum_column_name} and parse +the @code{ENUM} definition in the second column. -@example -mysql> select SEC_TO_TIME(2378); - -> '00:39:38' -mysql> select SEC_TO_TIME(2378) + 0; - -> 3938 -@end example +@tindex SET +@node SET, , ENUM, String types +@subsubsection The @code{SET} Type -@findex TIME_TO_SEC() -@item TIME_TO_SEC(time) -Returns the @code{time} argument, converted to seconds: +A @code{SET} is a string object that can have zero or more values, each of +which must be chosen from a list of allowed values specified when the table +is created. @code{SET} column values that consist of multiple set members +are specified with members separated by commas (@samp{,}). A consequence of +this is that @code{SET} member values cannot themselves contain commas. + +For example, a column specified as @code{SET("one", "two") NOT NULL} can have +any of these values: @example -mysql> select TIME_TO_SEC('22:23:00'); - -> 80580 -mysql> select TIME_TO_SEC('00:39:38'); - -> 2378 +"" +"one" +"two" +"one,two" @end example -@end table -@findex miscellaneous functions -@findex functions, miscellaneous -@node Miscellaneous functions, Group by functions, Date and time functions, Functions -@subsection Miscellaneous Functions -@table @code -@findex DATABASE() -@item DATABASE() -Returns the current database name: +A @code{SET} can have a maximum of 64 different members. + +@strong{MySQL} stores @code{SET} values numerically, with the low-order bit +of the stored value corresponding to the first set member. If you retrieve a +@code{SET} value in a numeric context, the value retrieved has bits set +corresponding to the set members that make up the column value. For example, +you can retrieve numeric values from a @code{SET} column like this: @example -mysql> select DATABASE(); - -> 'test' +mysql> SELECT set_col+0 FROM tbl_name; @end example -If there is no current database, @code{DATABASE()} returns the empty string. +If a number is stored into a @code{SET} column, the bits that +are set in the binary representation of the number determine the +set members in the column value. Suppose a column is specified as +@code{SET("a","b","c","d")}. Then the members have the following bit +values: -@findex USER() -@findex SYSTEM_USER() -@findex SESSION_USER() -@item USER() -@itemx SYSTEM_USER() -@itemx SESSION_USER() -Returns the current @strong{MySQL} user name: +@multitable @columnfractions .2 .2 .6 +@item @code{SET} @strong{member} @tab @strong{Decimal value} @tab @strong{Binary value} +@item @code{a} @tab @code{1} @tab @code{0001} +@item @code{b} @tab @code{2} @tab @code{0010} +@item @code{c} @tab @code{4} @tab @code{0100} +@item @code{d} @tab @code{8} @tab @code{1000} +@end multitable -@example -mysql> select USER(); - -> 'davida@@localhost' -@end example +If you assign a value of @code{9} to this column, that is @code{1001} in +binary, so the first and fourth @code{SET} value members @code{"a"} and +@code{"d"} are selected and the resulting value is @code{"a,d"}. -In @strong{MySQL} Version 3.22.11 or later, this includes the client hostname -as well as the user name. You can extract just the user name part like this -(which works whether or not the value includes a hostname part): +For a value containing more than one @code{SET} element, it does not matter +what order the elements are listed in when you insert the value. It also +does not matter how many times a given element is listed in the value. +When the value is retrieved later, each element in the value will appear +once, with elements listed according to the order in which they were +specified at table creation time. For example, if a column is specified as +@code{SET("a","b","c","d")}, then @code{"a,d"}, @code{"d,a"}, and +@code{"d,a,a,d,d"} will all appear as @code{"a,d"} when retrieved. + +@code{SET} values are sorted numerically. @code{NULL} values sort before +non-@code{NULL} @code{SET} values. + +Normally, you perform a @code{SELECT} on a @code{SET} column using +the @code{LIKE} operator or the @code{FIND_IN_SET()} function: @example -mysql> select substring_index(USER(),"@@",1); - -> 'davida' +mysql> SELECT * FROM tbl_name WHERE set_col LIKE '%value%'; +mysql> SELECT * FROM tbl_name WHERE FIND_IN_SET('value',set_col)>0; @end example -@findex PASSWORD() -@item PASSWORD(str) -Calculates a password string from the plaintext password @code{str}. This is -the function that is used for encrypting @strong{MySQL} passwords for storage -in the @code{Password} column of the @code{user} grant table: +But the following will also work: @example -mysql> select PASSWORD('badpwd'); - -> '7f84554057dd964b' +mysql> SELECT * FROM tbl_name WHERE set_col = 'val1,val2'; +mysql> SELECT * FROM tbl_name WHERE set_col & 1; @end example -@cindex password encryption, reversibility of -@code{PASSWORD()} encryption is non-reversible. +The first of these statements looks for an exact match. The second looks +for values containing the first set member. -@code{PASSWORD()} does not perform password encryption in the same way that -Unix passwords are encrypted. You should not assume that if your Unix -password and your @strong{MySQL} password are the same, @code{PASSWORD()} -will result in the same encrypted value as is stored in the Unix password -file. See @code{ENCRYPT()}. +If you want to get all possible values for a @code{SET} column, you should +use: @code{SHOW COLUMNS FROM table_name LIKE set_column_name} and parse +the @code{SET} definition in the second column. -@findex ENCRYPT() -@item ENCRYPT(str[,salt]) -Encrypt @code{str} using the Unix @code{crypt()} system call. The -@code{salt} argument should be a string with two characters. -(As of @strong{MySQL} Version 3.22.16, @code{salt} may be longer than two characters.): +@cindex types, columns +@cindex choosing types +@node Choosing types, Indexes, String types, Column types +@subsection Choosing the Right Type for a Column -@example -mysql> select ENCRYPT("hello"); - -> 'VxuFAJXVARROc' -@end example +For the most efficient use of storage, try to use the most precise type in +all cases. For example, if an integer column will be used for values in the +range between @code{1} and @code{99999}, @code{MEDIUMINT UNSIGNED} is the +best type. -If @code{crypt()} is not available on your system, @code{ENCRYPT()} always -returns @code{NULL}. +Accurate representation of monetary values is a common problem. In +@strong{MySQL}, you should use the @code{DECIMAL} type. This is stored as +a string, so no loss of accuracy should occur. If accuracy is not +too important, the @code{DOUBLE} type may also be good enough. -@code{ENCRYPT()} ignores all but the first 8 characters of @code{str}, at -least on some systems. This will be determined by the behavior of the -underlying @code{crypt()} system call. +For high precision, you can always convert to a fixed-point type stored +in a @code{BIGINT}. This allows you to do all calculations with integers +and convert results back to floating-point values only when necessary. -@findex ENCODE() -@item ENCODE(str,pass_str) -Encrypt @code{str} using @code{pass_str} as the password. -To decrypt the result, use @code{DECODE()}. +@cindex indexes, columns +@cindex columns, indexes +@cindex keys +@node Indexes, Multiple-column indexes, Choosing types, Column types +@subsection Column Indexes -The results is a binary string of the same length as @code{string}. -If you want to save it in a column, use a @code{BLOB} column type. +All @strong{MySQL} column types can be indexed. Use of indexes on the +relevant columns is the best way to improve the performance of @code{SELECT} +operations. -@findex DECODE() -@item DECODE(crypt_str,pass_str) -Descrypts the encrypted string @code{crypt_str} using @code{pass_str} as the -password. @code{crypt_str} should be a string returned from -@code{ENCODE()}. +The maximum number of keys and the maximum index length is defined per +table handler. @xref{Table types}. You can with all table handlers have +at least 16 keys and a total index length of at least 256 bytes. -@findex MD5() -@item MD5(string) -Calculates a MD5 checksum for the string. Value is returned as a 32 long -hex number that may, for example, be used as a hash key: +For @code{CHAR} and @code{VARCHAR} columns, you can index a prefix of a +column. This is much faster and requires less disk space than indexing the +whole column. The syntax to use in the @code{CREATE TABLE} statement to +index a column prefix looks like this: @example -mysql> select MD5("testing"); - -> 'ae2b1fca515949e5d54fb22b8ed95575' +KEY index_name (col_name(length)) @end example -This is an "RSA Data Security, Inc. MD5 Message-Digest Algorithm". - -@findex LAST_INSERT_ID([expr]) -@item LAST_INSERT_ID([expr]) -Returns the last automatically generated value that was inserted into an -@code{AUTO_INCREMENT} column. -@xref{mysql_insert_id,, @code{mysql_insert_id()}}. +The example below creates an index for the first 10 characters of the +@code{name} column: @example -mysql> select LAST_INSERT_ID(); - -> 195 +mysql> CREATE TABLE test ( + name CHAR(200) NOT NULL, + KEY index_name (name(10))); @end example -The last ID that was generated is maintained in the server on a -per-connection basis. It will not be changed by another client. It will not -even be changed if you update another @code{AUTO_INCREMENT} column with a -non-magic value (that is, a value that is not @code{NULL} and not @code{0}). +For @code{BLOB} and @code{TEXT} columns, you must index a prefix of the +column. You cannot index the entire column. -If you insert many rows at the same time with an insert statement, -@code{LAST_INSERT_ID()} returns the value for the first inserted row. -The reason for this is so that you it makes it possible to easily reproduce -the same @code{INSERT} statement against some other server. +In @strong{MySQL} Version 3.23.23 or later, you can also create special +@strong{FULLTEXT} indexes. They are used for full-text search. Only the +@code{MyISAM} table type supports @code{FULLTEXT} indexes. They can be +created only from @code{VARCHAR} and @code{TEXT} columns. +Indexing always happens over the entire column and partial indexing is not +supported. See @ref{Fulltext Search} for details. -@cindex sequence emulation -If @code{expr} is given as an argument to @code{LAST_INSERT_ID()} in an -@code{UPDATE} clause, then the value of the argument is returned as a -@code{LAST_INSERT_ID()} value. This can be used to simulate sequences. +@cindex multi-column indexes +@cindex indexes, multi-column +@cindex keys, multi-column +@node Multiple-column indexes, Other-vendor column types, Indexes, Column types +@subsection Multiple-column Indexes -First create the table: +@strong{MySQL} can create indexes on multiple columns. An index may +consist of up to 15 columns. (On @code{CHAR} and @code{VARCHAR} columns you +can also use a prefix of the column as a part of an index). -@example -mysql> create table sequence (id int not null); -mysql> insert into sequence values (0); -@end example +A multiple-column index can be considered a sorted array containing values +that are created by concatenating the values of the indexed columns. -Then the table can be used to generate sequence numbers like this: +@strong{MySQL} uses multiple-column indexes in such a way that queries are +fast when you specify a known quantity for the first column of the index in a +@code{WHERE} clause, even if you don't specify values for the other columns. + +Suppose a table is created using the following specification: @example -mysql> update sequence set id=LAST_INSERT_ID(id+1); +mysql> CREATE TABLE test ( + id INT NOT NULL, + last_name CHAR(30) NOT NULL, + first_name CHAR(30) NOT NULL, + PRIMARY KEY (id), + INDEX name (last_name,first_name)); @end example -You can generate sequences without calling @code{LAST_INSERT_ID()}, but the -utility of using the function this way is that the ID value is maintained in -the server as the last automatically generated value. You can retrieve the -new ID as you would read any normal @code{AUTO_INCREMENT} value in -@strong{MySQL}. For example, @code{LAST_INSERT_ID()} (without an argument) -will return the new ID. The C API function @code{mysql_insert_id()} -can also be used to get the value. - -@findex FORMAT() -@item FORMAT(X,D) -Formats the number @code{X} to a format like @code{'#,###,###.##'}, rounded -to @code{D} decimals. If @code{D} is @code{0}, the result will have no -decimal point or fractional part: +Then the index @code{name} is an index over @code{last_name} and +@code{first_name}. The index will be used for queries that specify +values in a known range for @code{last_name}, or for both @code{last_name} +and @code{first_name}. +Therefore, the @code{name} index will be used in the following queries: @example -mysql> select FORMAT(12332.123456, 4); - -> '12,332.1235' -mysql> select FORMAT(12332.1,4); - -> '12,332.1000' -mysql> select FORMAT(12332.2,0); - -> '12,332' -@end example +mysql> SELECT * FROM test WHERE last_name="Widenius"; -@findex VERSION() -@item VERSION() -Returns a string indicating the @strong{MySQL} server version: +mysql> SELECT * FROM test WHERE last_name="Widenius" + AND first_name="Michael"; -@example -mysql> select VERSION(); - -> '3.23.13-log' -@end example +mysql> SELECT * FROM test WHERE last_name="Widenius" + AND (first_name="Michael" OR first_name="Monty"); -Note that if your version ends with @code{-log} this means that logging is -enabled. +mysql> SELECT * FROM test WHERE last_name="Widenius" + AND first_name >="M" AND first_name < "N"; +@end example -@findex CONNECTION_ID() -@item CONNECTION_ID() -Returns the connection id (@code{thread_id}) for the connection. -Every connection has its own unique id: +However, the @code{name} index will NOT be used in the following queries: @example -mysql> select CONNECTION_ID(); - -> 1 +mysql> SELECT * FROM test WHERE first_name="Michael"; + +mysql> SELECT * FROM test WHERE last_name="Widenius" + OR first_name="Michael"; @end example -@cindex timeout -@findex GET_LOCK() -@item GET_LOCK(str,timeout) -Tries to obtain a lock with a name given by the string @code{str}, with a -timeout of @code{timeout} seconds. Returns @code{1} if the lock was obtained -successfully, @code{0} if the attempt timed out, or @code{NULL} if an error -occurred (such as running out of memory or the thread was killed with -@code{mysqladmin kill}). A lock is released when you execute -@code{RELEASE_LOCK()}, execute a new @code{GET_LOCK()}, or the thread -terminates. This function can be used to implement application locks or to -simulate record locks. It blocks requests by other clients for locks with -the same name; clients that agree on a given lock string name can use the -string to perform cooperative advisory locking: +For more information on the manner in which @strong{MySQL} uses indexes to +improve query performance, see @ref{MySQL indexes, , @strong{MySQL} +indexes}. -@example -mysql> select GET_LOCK("lock1",10); - -> 1 -mysql> select GET_LOCK("lock2",10); - -> 1 -mysql> select RELEASE_LOCK("lock2"); - -> 1 -mysql> select RELEASE_LOCK("lock1"); - -> NULL -@end example +@cindex types, portability +@cindex portability, types +@cindex columns, other types +@node Other-vendor column types, , Multiple-column indexes, Column types +@subsection Using Column Types from Other Database Engines -Note that the second @code{RELEASE_LOCK()} call returns @code{NULL} because -the lock @code{"lock1"} was automatically released by the second -@code{GET_LOCK()} call. +To make it easier to use code written for SQL implementations from other +vendors, @strong{MySQL} maps column types as shown in the table below. These +mappings make it easier to move table definitions from other database engines +to @strong{MySQL}: -@findex RELEASE_LOCK() -@item RELEASE_LOCK(str) -Releases the lock named by the string @code{str} that was obtained with -@code{GET_LOCK()}. Returns @code{1} if the lock was released, @code{0} if the -lock wasn't locked by this thread (in which case the lock is not released), -and @code{NULL} if the named lock didn't exist. The lock will not exist if -it was never obtained by a call to @code{GET_LOCK()} or if it already has -been released. +@multitable @columnfractions .4 .6 +@item @strong{Other vendor type} @tab @strong{MySQL type} +@item @code{BINARY(NUM)} @tab @code{CHAR(NUM) BINARY} +@item @code{CHAR VARYING(NUM)} @tab @code{VARCHAR(NUM)} +@item @code{FLOAT4} @tab @code{FLOAT} +@item @code{FLOAT8} @tab @code{DOUBLE} +@item @code{INT1} @tab @code{TINYINT} +@item @code{INT2} @tab @code{SMALLINT} +@item @code{INT3} @tab @code{MEDIUMINT} +@item @code{INT4} @tab @code{INT} +@item @code{INT8} @tab @code{BIGINT} +@item @code{LONG VARBINARY} @tab @code{MEDIUMBLOB} +@item @code{LONG VARCHAR} @tab @code{MEDIUMTEXT} +@item @code{MIDDLEINT} @tab @code{MEDIUMINT} +@item @code{VARBINARY(NUM)} @tab @code{VARCHAR(NUM) BINARY} +@end multitable -@findex BENCHMARK() -@item BENCHMARK(count,expr) -The @code{BENCHMARK()} function executes the expression @code{expr} -repeatedly @code{count} times. It may be used to time how fast @strong{MySQL} -processes the expression. The result value is always @code{0}. The intended -use is in the @code{mysql} client, which reports query execution times: +Column type mapping occurs at table creation time. If you create a table +with types used by other vendors and then issue a @code{DESCRIBE tbl_name} +statement, @strong{MySQL} reports the table structure using the equivalent +@strong{MySQL} types. -@example -mysql> select BENCHMARK(1000000,encode("hello","goodbye")); -+----------------------------------------------+ -| BENCHMARK(1000000,encode("hello","goodbye")) | -+----------------------------------------------+ -| 0 | -+----------------------------------------------+ -1 row in set (4.74 sec) -@end example +@cindex functions for @code{SELECT} and @code{WHERE} clauses +@node Functions, CREATE DATABASE, Column types, Reference +@section Functions for Use in @code{SELECT} and @code{WHERE} Clauses -The time reported is elapsed time on the client end, not CPU time on the -server end. It may be advisable to execute @code{BENCHMARK()} several -times, and interpret the result with regard to how heavily loaded the -server machine is. +A @code{select_expression} or @code{where_definition} in a SQL statement +can consist of any expression using the functions described below. -@findex INET_NTOA() -@item INET_NTOA(expr) -Returns the network address (4 or 8 byte) for the numeric expression: +An expression that contains @code{NULL} always produces a @code{NULL} value +unless otherwise indicated in the documentation for the operators and +functions involved in the expression. -@example -mysql> select INET_NTOA(3520061480); - -> "209.207.224.40" -@end example +@strong{NOTE:} There must be no whitespace between a function name and the +parenthesis following it. This helps the @strong{MySQL} parser distinguish +between function calls and references to tables or columns that happen to +have the same name as a function. Spaces around arguments are permitted, +though. -@findex INET_ATON() -@item INET_ATON(expr) -Returns an integer that represents the numeric value for a network address. -Addresses may be 4 or 8 byte addresses: +You can force @strong{MySQL} to accept spaces after the function name by +starting @code{mysqld} with @code{--ansi} or using the +@code{CLIENT_IGNORE_SPACE} to @code{mysql_connect()}, but in this case all +function names will become reserved words. @xref{ANSI mode}. + +@need 2000 +For the sake of brevity, examples display the output from the @code{mysql} +program in abbreviated form. So this: @example -mysql> select INET_ATON("209.207.224.40"); - -> 3520061480 -@end example +mysql> select MOD(29,9); +1 rows in set (0.00 sec) -The generated number is always in network byte order; For example the -above number is calculated as @code{209*255^3 + 207*255^2 + 224*255 +40}. ++-----------+ +| mod(29,9) | ++-----------+ +| 2 | ++-----------+ +@end example -@findex MASTER_POS_WAIT() -@item MASTER_POS_WAIT(log_name, log_pos) -Blocks until the slave reaches the specified position in the master log during -replication. If master information is not initialized, returns NULL. If the -slave is not running, will block and wait until it is started and goes to or -past -the specified position. If the slave is already past the specified position, -returns immediately. The return value is the number of log events it had to -wait to get to the specified position, or NULL in case of error. Useful for -control of master-slave synchronization, but was originally written to -facilitate replication testing. -@end table +is displayed like this: -@findex GROUP BY functions -@findex functions, GROUP BY -@node Group by functions, , Miscellaneous functions, Functions -@subsection Functions for Use with @code{GROUP BY} Clauses +@example +mysql> select MOD(29,9); + -> 2 +@end example -If you use a group function in a statement containing no @code{GROUP BY} -clause, it is equivalent to grouping on all rows. +@menu +* Grouping functions:: Grouping functions +* Arithmetic functions:: Normal arithmetic operations +* Bit functions:: Bit functions +* Logical functions:: Logical operations +* Comparison functions:: Comparison operators +* String comparison functions:: String comparison functions +* Casts:: Cast operators +* Control flow functions:: Control flow functions +* Mathematical functions:: Mathematical functions +* String functions:: String functions +* Date and time functions:: Date and time functions +* Miscellaneous functions:: Miscellaneous functions +* Group by functions:: Functions for @code{GROUP BY} clause +@end menu +@cindex functions, grouping +@cindex grouping, expressions +@node Grouping functions, Arithmetic functions, Functions, Functions +@subsection Grouping Functions @table @code -@findex COUNT() -@item COUNT(expr) -Returns a count of the number of non-@code{NULL} values in the rows -retrieved by a @code{SELECT} statement: +@findex () (parentheses) +@findex parentheses ( and ) +@item ( ... ) +Parentheses. Use these to force the order of evaluation in an expression: @example -mysql> select student.student_name,COUNT(*) - from student,course - where student.student_id=course.student_id - GROUP BY student_name; - +mysql> select 1+2*3; + -> 7 +mysql> select (1+2)*3; + -> 9 @end example +@end table -@code{COUNT(*)} is somewhat different in that it returns a count of -the number of rows retrieved, whether or not they contain @code{NULL} -values. +@node Arithmetic functions, Bit functions, Grouping functions, Functions +@subsection Normal Arithmetic Operations -@code{COUNT(*)} is optimized to -return very quickly if the @code{SELECT} retrieves from one table, no -other columns are retrieved, and there is no @code{WHERE} clause. -For example: +The usual arithmetic operators are available. Note that in the case of +@samp{-}, @samp{+}, and @samp{*}, the result is calculated with +@code{BIGINT} (64-bit) precision if both arguments are integers! +@cindex operations, arithmetic +@cindex arithmetic expressions +@table @code +@findex + (addition) +@findex addition (+) +@item + +Addition: @example -mysql> select COUNT(*) from student; +mysql> select 3+5; + -> 8 @end example -@findex COUNT(DISTINCT) -@findex DISTINCT -@item COUNT(DISTINCT expr,[expr...]) -Returns a count of the number of different non-@code{NULL} values: - +@findex - (subtraction) +@findex subtraction (-) +@item - +Subtraction: @example -mysql> select COUNT(DISTINCT results) from student; +mysql> select 3-5; + -> -2 @end example -In @strong{MySQL} you can get the number of distinct expression -combinations that don't contain NULL by giving a list of expressions. -In ANSI SQL you would have to do a concatenation of all expressions -inside @code{CODE(DISTINCT ..)}. - -@findex AVG() -@item AVG(expr) -Returns the average value of @code{expr}: - +@findex * (multiplication) +@findex multiplication (*) +@item * +Multiplication: @example -mysql> select student_name, AVG(test_score) - from student - GROUP BY student_name; +mysql> select 3*5; + -> 15 +mysql> select 18014398509481984*18014398509481984.0; + -> 324518553658426726783156020576256.0 +mysql> select 18014398509481984*18014398509481984; + -> 0 @end example -@findex MIN() -@findex MAX() -@item MIN(expr) -@itemx MAX(expr) -Returns the minimum or maximum value of @code{expr}. @code{MIN()} and -@code{MAX()} may take a string argument; in such cases they return the -minimum or maximum string value. @xref{MySQL indexes}. +The result of the last expression is incorrect because the result of the integer +multiplication exceeds the 64-bit range of @code{BIGINT} calculations. +@findex / (division) +@findex division (/) +@item / +Division: @example -mysql> select student_name, MIN(test_score), MAX(test_score) - from student - GROUP BY student_name; +mysql> select 3/5; + -> 0.60 @end example -@findex SUM() -@item SUM(expr) -Returns the sum of @code{expr}. Note that if the return set has no rows, -it returns NULL! - -@findex STD() -@findex STDDEV() -@cindex Oracle compatibility -@cindex compatibility, with Oracle -@item STD(expr) -@itemx STDDEV(expr) -Returns the standard deviation of @code{expr}. This is an extension to -ANSI SQL. The @code{STDDEV()} form of this function is provided for Oracle -compatibility. - -@findex BIT_OR() -@item BIT_OR(expr) -Returns the bitwise @code{OR} of all bits in @code{expr}. The calculation is -performed with 64-bit (@code{BIGINT}) precision. - -@findex BIT_AND() -@item BIT_AND(expr) -Returns the bitwise @code{AND} of all bits in @code{expr}. The calculation is -performed with 64-bit (@code{BIGINT}) precision. -@end table - -@cindex @code{GROUP BY}, extensions to ANSI SQL -@strong{MySQL} has extended the use of @code{GROUP BY}. You can use columns or -calculations in the @code{SELECT} expressions that don't appear in -the @code{GROUP BY} part. This stands for @emph{any possible value for this -group}. You can use this to get better performance by avoiding sorting and -grouping on unnecessary items. For example, you don't need to group on -@code{customer.name} in the following query: +Division by zero produces a @code{NULL} result: @example -mysql> select order.custid,customer.name,max(payments) - from order,customer - where order.custid = customer.custid - GROUP BY order.custid; +mysql> select 102/(1-1); + -> NULL @end example -In ANSI SQL, you would have to add @code{customer.name} to the @code{GROUP -BY} clause. In @strong{MySQL}, the name is redundant if you don't run in -ANSI mode. +A division will be calculated with @code{BIGINT} arithmetic only if performed +in a context where its result is converted to an integer! +@end table -@strong{Don't use this feature} if the columns you omit from the -@code{GROUP BY} part aren't unique in the group! You will get -unpredictable results. +@findex arithmetic functions +@findex bit functions +@findex functions, arithmetic +@findex functions, bit +@node Bit functions, Logical functions, Arithmetic functions, Functions +@subsection Bit Functions -In some cases, you can use @code{MIN()} and @code{MAX()} to obtain a specific -column value even if it isn't unique. The following gives the value of -@code{column} from the row containing the smallest value in the @code{sort} -column: +@strong{MySQL} uses @code{BIGINT} (64-bit) arithmetic for bit operations, so +these operators have a maximum range of 64 bits. +@table @code +@findex | (bitwise OR) +@findex OR, bitwise +@item | +Bitwise OR: @example -substr(MIN(concat(rpad(sort,6,' '),column)),7) +mysql> select 29 | 15; + -> 31 @end example -@xref{example-Maximum-column-group-row}. - -@cindex @code{ORDER BY}, aliases in -@cindex aliases, in @code{ORDER BY} clauses -@cindex @code{GROUP BY}, aliases in -@cindex aliases, in @code{GROUP BY} clauses -@cindex expression aliases -@cindex aliases, for expressions -Note that if you are using @strong{MySQL} Version 3.22 (or earlier) or if -you are trying to follow ANSI SQL, you can't use expressions in @code{GROUP -BY} or @code{ORDER BY} clauses. You can work around this limitation by -using an alias for the expression: - +@findex & (bitwise AND) +@findex AND, bitwise +@item & +Bitwise AND: @example -mysql> SELECT id,FLOOR(value/100) AS val FROM tbl_name - GROUP BY id,val ORDER BY val; +mysql> select 29 & 15; + -> 13 @end example -In @strong{MySQL} Version 3.23 you can do: +@findex << (left shift) +@item << +Shifts a longlong (@code{BIGINT}) number to the left: +@example +mysql> select 1 << 2; + -> 4 +@end example +@findex >> (right shift) +@item >> +Shifts a longlong (@code{BIGINT}) number to the right: @example -mysql> SELECT id,FLOOR(value/100) FROM tbl_name ORDER BY RAND(); +mysql> select 4 >> 2; + -> 1 @end example -@findex CREATE DATABASE -@node CREATE DATABASE, DROP DATABASE, Functions, Reference -@section @code{CREATE DATABASE} Syntax +@findex ~ +@item ~ +Invert all bits: +@example +mysql> select 5 & ~1; + -> 4 +@end example +@findex BIT_COUNT() +@item BIT_COUNT(N) +Returns the number of bits that are set in the argument @code{N}: @example -CREATE DATABASE [IF NOT EXISTS] db_name +mysql> select BIT_COUNT(29); + -> 4 @end example +@end table -@code{CREATE DATABASE} creates a database with the given name. Rules for -allowable database names are given in @ref{Legal names}. An error occurs if -the database already exists and you didn't specify @code{IF NOT EXISTS}. +@findex Logical functions +@findex Functions, logical +@node Logical functions, Comparison functions, Bit functions, Functions +@subsection Logical Operations -Databases in @strong{MySQL} are implemented as directories containing files -that correspond to tables in the database. Because there are no tables in a -database when it is initially created, the @code{CREATE DATABASE} statement -only creates a directory under the @strong{MySQL} data directory. +All logical functions return @code{1} (TRUE), @code{0} (FALSE) or +@code{NULL} (unknown, which is in most cases the same as FALSE): -@cindex @code{mysqladmin} -You can also create databases with @code{mysqladmin}. -@xref{Programs}. +@table @code +@findex NOT, logical +@findex ! (logical NOT) +@item NOT +@itemx ! +Logical NOT. Returns @code{1} if the argument is @code{0}, otherwise returns +@code{0}. +Exception: @code{NOT NULL} returns @code{NULL}: +@example +mysql> select NOT 1; + -> 0 +mysql> select NOT NULL; + -> NULL +mysql> select ! (1+1); + -> 0 +mysql> select ! 1+1; + -> 1 +@end example +The last example returns @code{1} because the expression evaluates +the same way as @code{(!1)+1}. -@findex DROP DATABASE -@node DROP DATABASE, CREATE TABLE, CREATE DATABASE, Reference -@section @code{DROP DATABASE} Syntax +@findex OR, logical +@findex || (logical OR) +@item OR +@itemx || +Logical OR. Returns @code{1} if either argument is not @code{0} and not +@code{NULL}: +@example +mysql> select 1 || 0; + -> 1 +mysql> select 0 || 0; + -> 0 +mysql> select 1 || NULL; + -> 1 + +@end example +@findex AND, logical +@findex && (logical AND) +@item AND +@itemx && +Logical AND. Returns @code{0} if either argument is @code{0} or @code{NULL}, +otherwise returns @code{1}: @example -DROP DATABASE [IF EXISTS] db_name +mysql> select 1 && NULL; + -> 0 +mysql> select 1 && 0; + -> 0 @end example +@end table -@code{DROP DATABASE} drops all tables in the database and deletes the -database. If you do a @code{DROP DATABASE} on a symbolic linked -database, both the link and the original database is deleted. @strong{Be -VERY careful with this command!} +@cindex casts +@cindex type conversions +@findex comparison operators +@node Comparison functions, String comparison functions, Logical functions, Functions +@subsection Comparison Operators -@code{DROP DATABASE} returns the number of files that were removed from -the database directory. Normally, this is three times the number of -tables, because normally each table corresponds to a @file{.MYD} file, a -@file{.MYI} file, and a @file{.frm} file. +Comparison operations result in a value of @code{1} (TRUE), @code{0} (FALSE), +or @code{NULL}. These functions work for both numbers and strings. Strings +are automatically converted to numbers and numbers to strings as needed (as +in Perl). -The @code{DROP DATABASE} command removes from the given database -directory all files with the following extensions: +@strong{MySQL} performs comparisons using the following +rules: -@multitable @columnfractions .25 .25 .25 .25 -@item .BAK @tab .DAT @tab .HSH @tab .ISD -@item .ISM @tab .ISM @tab .MRG @tab .MYD -@item .MYI @tab .db @tab .frm -@end multitable +@itemize @bullet +@item +If one or both arguments are @code{NULL}, the result of the comparison +is @code{NULL}, except for the @code{<=>} operator. -All subdirectories that consists of 2 digits (@code{RAID} directories) -are also removed. +@item +If both arguments in a comparison operation are strings, they are compared as +strings. -In @strong{MySQL} Version 3.22 or later, you can use the keywords -@code{IF EXISTS} to prevent an error from occurring if the database doesn't -exist. - -@cindex @code{mysqladmin} -You can also drop databases with @code{mysqladmin}. @xref{Programs}. - -@findex CREATE TABLE -@node CREATE TABLE, ALTER TABLE, DROP DATABASE, Reference -@section @code{CREATE TABLE} Syntax - -@menu -* Silent column changes:: Silent column changes -@end menu +@item +If both arguments are integers, they are compared as integers. -@example -CREATE [TEMPORARY] TABLE [IF NOT EXISTS] tbl_name [(create_definition,...)] -[table_options] [select_statement] +@item +Hexadecimal values are treated as binary strings if not compared to a number. -create_definition: - col_name type [NOT NULL | NULL] [DEFAULT default_value] [AUTO_INCREMENT] - [PRIMARY KEY] [reference_definition] - or PRIMARY KEY (index_col_name,...) - or KEY [index_name] (index_col_name,...) - or INDEX [index_name] (index_col_name,...) - or UNIQUE [INDEX] [index_name] (index_col_name,...) - or FULLTEXT [INDEX] [index_name] (index_col_name,...) - or [CONSTRAINT symbol] FOREIGN KEY index_name (index_col_name,...) - [reference_definition] - or CHECK (expr) +@item +@cindex ODBC compatibility +@cindex compatibility, with ODBC +If one of the arguments is a @code{TIMESTAMP} or @code{DATETIME} column and +the other argument is a constant, the constant is converted +to a timestamp before the comparison is performed. This is done to be more +ODBC-friendly. -type: - TINYINT[(length)] [UNSIGNED] [ZEROFILL] - or SMALLINT[(length)] [UNSIGNED] [ZEROFILL] - or MEDIUMINT[(length)] [UNSIGNED] [ZEROFILL] - or INT[(length)] [UNSIGNED] [ZEROFILL] - or INTEGER[(length)] [UNSIGNED] [ZEROFILL] - or BIGINT[(length)] [UNSIGNED] [ZEROFILL] - or REAL[(length,decimals)] [UNSIGNED] [ZEROFILL] - or DOUBLE[(length,decimals)] [UNSIGNED] [ZEROFILL] - or FLOAT[(length,decimals)] [UNSIGNED] [ZEROFILL] - or DECIMAL(length,decimals) [UNSIGNED] [ZEROFILL] - or NUMERIC(length,decimals) [UNSIGNED] [ZEROFILL] - or CHAR(length) [BINARY] - or VARCHAR(length) [BINARY] - or DATE - or TIME - or TIMESTAMP - or DATETIME - or TINYBLOB - or BLOB - or MEDIUMBLOB - or LONGBLOB - or TINYTEXT - or TEXT - or MEDIUMTEXT - or LONGTEXT - or ENUM(value1,value2,value3,...) - or SET(value1,value2,value3,...) +@item +In all other cases, the arguments are compared as floating-point (real) +numbers. +@end itemize -index_col_name: - col_name [(length)] +By default, string comparisons are done in case-independent fashion using the +current character set (ISO-8859-1 Latin1 by default, which also works +excellently for English). -reference_definition: - REFERENCES tbl_name [(index_col_name,...)] - [MATCH FULL | MATCH PARTIAL] - [ON DELETE reference_option] - [ON UPDATE reference_option] +The examples below illustrate conversion of strings to numbers for comparison +operations: -reference_option: - RESTRICT | CASCADE | SET NULL | NO ACTION | SET DEFAULT +@example +mysql> SELECT 1 > '6x'; + -> 0 +mysql> SELECT 7 > '6x'; + -> 1 +mysql> SELECT 0 > 'x6'; + -> 0 +mysql> SELECT 0 = 'x6'; + -> 1 +@end example -table_options: - TYPE = @{BDB | HEAP | ISAM | InnoDB | MERGE | MYISAM @} -or AUTO_INCREMENT = # -or AVG_ROW_LENGTH = # -or CHECKSUM = @{0 | 1@} -or COMMENT = "string" -or MAX_ROWS = # -or MIN_ROWS = # -or PACK_KEYS = @{0 | 1@} -or PASSWORD = "string" -or DELAY_KEY_WRITE = @{0 | 1@} -or ROW_FORMAT= @{ default | dynamic | fixed | compressed @} -or RAID_TYPE= @{1 | STRIPED | RAID0 @} RAID_CHUNKS=# RAID_CHUNKSIZE=# -or UNION = (table_name,[table_name...]) -or DATA DIRECTORY="directory" -or INDEX DIRECTORY="directory" +@table @code +@findex = (equal) +@findex equal (=) +@item = +Equal: +@example +mysql> select 1 = 0; + -> 0 +mysql> select '0' = 0; + -> 1 +mysql> select '0.0' = 0; + -> 1 +mysql> select '0.01' = 0; + -> 0 +mysql> select '.01' = 0.01; + -> 1 +@end example -select_statement: - [IGNORE | REPLACE] SELECT ... (Some legal select statement) +@findex <> (not equal) +@findex not equal (<>) +@findex != (not equal) +@findex not equal (!=) +@item <> +@itemx != +Not equal: +@example +mysql> select '.01' <> '0.01'; + -> 1 +mysql> select .01 <> '0.01'; + -> 0 +mysql> select 'zapp' <> 'zappp'; + -> 1 @end example -@code{CREATE TABLE} -creates a table with the given name in the current database. Rules for -allowable table names are given in @ref{Legal names}. An error occurs if -there is no current database or if the table already exists. +@findex <= (less than or equal) +@findex less than or equal (<=) +@item <= +Less than or equal: +@example +mysql> select 0.1 <= 2; + -> 1 +@end example -In @strong{MySQL} Version 3.22 or later, the table name can be specified as -@code{db_name.tbl_name}. This works whether or not there is a current -database. +@findex < (less than) +@findex less than (<) +@item < +Less than: +@example +mysql> select 2 <= 2; + -> 1 +@end example -In @strong{MySQL} Version 3.23, you can use the @code{TEMPORARY} keyword when -you create a table. A temporary table will automatically be deleted if a -connection dies and the name is per connection. This means that two different -connections can both use the same temporary table name without conflicting -with each other or with an existing table of the same name. (The existing table -is hidden until the temporary table is deleted). +@findex >= (greater than or equal) +@findex greater than or equal (>=) +@item >= +Greater than or equal: +@example +mysql> select 2 >= 2; + -> 1 +@end example -In @strong{MySQL} Version 3.23 or later, you can use the keywords -@code{IF NOT EXISTS} so that an error does not occur if the table already -exists. Note that there is no verification that the table structures are -identical. +@findex > (greater than) +@findex greater than (>) +@item > +Greater than: +@example +mysql> select 2 > 2; + -> 0 +@end example -Each table @code{tbl_name} is represented by some files in the database -directory. In the case of MyISAM-type tables you will get: +@cindex @code{NULL}, testing for null +@findex <=> (Equal to) +@item <=> +Null safe equal: +@example +mysql> select 1 <=> 1, NULL <=> NULL, 1 <=> NULL; + -> 1 1 0 +@end example -@multitable @columnfractions .2 .8 -@item @strong{File} @tab @strong{Purpose} -@item @code{tbl_name.frm} @tab Table definition (form) file -@item @code{tbl_name.MYD} @tab Data file -@item @code{tbl_name.MYI} @tab Index file -@end multitable +@cindex @code{NULL}, testing for null +@findex IS NULL +@findex IS NOT NULL +@item IS NULL +@itemx IS NOT NULL +Test whether or not a value is or is not @code{NULL}: +@example +mysql> select 1 IS NULL, 0 IS NULL, NULL IS NULL; + -> 0 0 1 +mysql> select 1 IS NOT NULL, 0 IS NOT NULL, NULL IS NOT NULL; + -> 1 1 0 +@end example -For more information on the properties of the various column types, see -@ref{Column types}: +@findex BETWEEN ... AND +@item expr BETWEEN min AND max +If @code{expr} is greater than or equal to @code{min} and @code{expr} is +less than or equal to @code{max}, @code{BETWEEN} returns @code{1}, +otherwise it returns @code{0}. This is equivalent to the expression +@code{(min <= expr AND expr <= max)} if all the arguments are of the +same type. The first argument (@code{expr}) determines how the +comparison is performed as follows: @itemize @bullet @item -If neither @code{NULL} nor @code{NOT NULL} is specified, the column -is treated as though @code{NULL} had been specified. - +If @code{expr} is a @code{TIMESTAMP}, @code{DATE}, or @code{DATETIME} +column, @code{MIN()} and @code{MAX()} are formatted to the same format if +they are constants. @item -An integer column may have the additional attribute @code{AUTO_INCREMENT}. -When you insert a value of @code{NULL} (recommended) or @code{0} into an -@code{AUTO_INCREMENT} column, the column is set to @code{value+1}, where -@code{value} is the largest value for the column currently in the table. -@code{AUTO_INCREMENT} sequences begin with @code{1}. -@xref{mysql_insert_id, , @code{mysql_insert_id()}}. - -If you delete the row containing the maximum value for an -@code{AUTO_INCREMENT} column, the value will be reused with an -@code{ISAM}, or @code{BDB} table but not with a -@code{MyISAM} or @code{InnoDB} table. If you delete all rows in the table -with @code{DELETE FROM table_name} (without a @code{WHERE}) in -@code{AUTOCOMMIT} mode, the sequence starts over for all table types. - -@strong{NOTE:} There can be only one @code{AUTO_INCREMENT} column per -table, and it must be indexed. @strong{MySQL} Version 3.23 will also only -work properly if the auto_increment column only has positive -values. Inserting a negative number is regarded as inserting a very large -positive number. This is done to avoid precision problems when -numbers 'wrap' over from positive to negative and also to ensure that one -doesn't accidentally get an auto_increment column that contains 0. - -@cindex ODBC compatibility -@cindex compatibility, with ODBC -To make @strong{MySQL} compatible with some ODBC applications, you can find -the last inserted row with the following query: +If @code{expr} is a case-insensitive string expression, a case-insensitive +string comparison is done. +@item +If @code{expr} is a case-sensitive string expression, a case-sensitive +string comparison is done. +@item +If @code{expr} is an integer expression, an integer comparison is done. +@item +Otherwise, a floating-point (real) comparison is done. +@end itemize @example -SELECT * FROM tbl_name WHERE auto_col IS NULL +mysql> select 1 BETWEEN 2 AND 3; + -> 0 +mysql> select 'b' BETWEEN 'a' AND 'c'; + -> 1 +mysql> select 2 BETWEEN 2 AND '3'; + -> 1 +mysql> select 2 BETWEEN 2 AND 'x-3'; + -> 0 @end example -@item -@code{NULL} values are handled differently for @code{TIMESTAMP} columns than -for other column types. You cannot store a literal @code{NULL} in a -@code{TIMESTAMP} column; setting the column to @code{NULL} sets it to the -current date and time. Because @code{TIMESTAMP} columns behave this way, the -@code{NULL} and @code{NOT NULL} attributes do not apply in the normal way and -are ignored if you specify them. - -On the other hand, to make it easier for @strong{MySQL} clients to use -@code{TIMESTAMP} columns, the server reports that such columns may be -assigned @code{NULL} values (which is true), even though @code{TIMESTAMP} -never actually will contain a @code{NULL} value. You can see this when you -use @code{DESCRIBE tbl_name} to get a description of your table. - -Note that setting a @code{TIMESTAMP} column to @code{0} is not the same -as setting it to @code{NULL}, because @code{0} is a valid @code{TIMESTAMP} -value. +@findex IN +@item expr IN (value,...) +Returns @code{1} if @code{expr} is any of the values in the @code{IN} list, +else returns @code{0}. If all values are constants, then all values are +evaluated according to the type of @code{expr} and sorted. The search for the +item is then done using a binary search. This means @code{IN} is very quick +if the @code{IN} value list consists entirely of constants. If @code{expr} +is a case-sensitive string expression, the string comparison is performed in +case-sensitive fashion: -@item -If no @code{DEFAULT} value is specified for a column, @strong{MySQL} -automatically assigns one. +@example +mysql> select 2 IN (0,3,5,'wefwf'); + -> 0 +mysql> select 'wefwf' IN (0,3,5,'wefwf'); + -> 1 +@end example -If the column may take @code{NULL} as a value, the default value is -@code{NULL}. +@findex NOT IN +@item expr NOT IN (value,...) +Same as @code{NOT (expr IN (value,...))}. -If the column is declared as @code{NOT NULL}, the default value depends on -the column type: +@findex ISNULL() +@item ISNULL(expr) +If @code{expr} is @code{NULL}, @code{ISNULL()} returns @code{1}, otherwise +it returns @code{0}: +@example +mysql> select ISNULL(1+1); + -> 0 +mysql> select ISNULL(1/0); + -> 1 +@end example -@itemize @minus -@item -For numeric types other than those declared with the @code{AUTO_INCREMENT} -attribute, the default is @code{0}. For an @code{AUTO_INCREMENT} column, the -default value is the next value in the sequence. +Note that a comparison of @code{NULL} values using @code{=} will always be +false! -@item -For date and time types other than @code{TIMESTAMP}, the default is the -appropriate zero value for the type. For the first @code{TIMESTAMP} -column in a table, the default value is the current date and time. -@xref{Date and time types}. +@cindex @code{NULL}, testing for null +@findex COALESCE() +@item COALESCE(list) +Returns first non-@code{NULL} element in list: -@item -For string types other than @code{ENUM}, the default value is the empty string. -For @code{ENUM}, the default is the first enumeration value. -@end itemize +@example +mysql> select COALESCE(NULL,1); + -> 1 +mysql> select COALESCE(NULL,NULL,NULL); + -> NULL +@end example -Default values must be constants. This means, for example, that you cannot -set the default for a date column to be the value of a function such as -@code{NOW()} or @code{CURRENT_DATE}. +@findex INTERVAL() +@item INTERVAL(N,N1,N2,N3,...) +Returns @code{0} if @code{N} < @code{N1}, @code{1} if @code{N} < @code{N2} +and so on. All arguments are treated as integers. It is required that +@code{N1} < @code{N2} < @code{N3} < @code{...} < @code{Nn} for this function +to work correctly. This is because a binary search is used (very fast): -@item -@code{KEY} is a synonym for @code{INDEX}. +@example +mysql> select INTERVAL(23, 1, 15, 17, 30, 44, 200); + -> 3 +mysql> select INTERVAL(10, 1, 10, 100, 1000); + -> 2 +mysql> select INTERVAL(22, 23, 30, 44, 200); + -> 0 +@end example +@end table -@item -In @strong{MySQL}, a @code{UNIQUE} key can have only distinct values. An -error occurs if you try to add a new row with a key that matches an existing -row. +@findex string comparison functions +@findex functions, string comparison +@node String comparison functions, Casts, Comparison functions, Functions +@subsection String Comparison Functions -@item -@tindex PRIMARY KEY -A @code{PRIMARY KEY} is a unique @code{KEY} with the extra constraint -that all key columns must be defined as @code{NOT NULL}. In @strong{MySQL} -the key is named @code{PRIMARY}. A table can have only one @code{PRIMARY KEY}. -If you don't have a @code{PRIMARY KEY} and some applications ask for the -@code{PRIMARY KEY} in your tables, @strong{MySQL} will return the first -@code{UNIQUE} key, which doesn't have any @code{NULL} columns, as the -@code{PRIMARY KEY}. +@cindex case sensitivity, in string comparisons +@cindex string comparisons, case sensitivity +Normally, if any expression in a string comparison is case sensitive, the +comparison is performed in case-sensitive fashion. -@item -A @code{PRIMARY KEY} can be a multiple-column index. However, you cannot -create a multiple-column index using the @code{PRIMARY KEY} key attibute in a -column specification. Doing so will mark only that single column as primary. -You must use the @code{PRIMARY KEY(index_col_name, ...)} syntax. +@table @code +@findex LIKE +@item expr LIKE pat [ESCAPE 'escape-char'] +Pattern matching using +SQL simple regular expression comparison. Returns @code{1} (TRUE) or @code{0} +(FALSE). With @code{LIKE} you can use the following two wild-card characters +in the pattern: -@item -If the @code{PRIMARY} or @code{UNIQUE} key consists of only one column and this -is of type integer, you can also refer to it as @code{_rowid} -(new in Version 3.23.11). +@multitable @columnfractions .1 .9 +@item @code{%} @tab Matches any number of characters, even zero characters +@item @code{_} @tab Matches exactly one character +@end multitable -@item -If you don't assign a name to an index, the index will be assigned the same -name as the first @code{index_col_name}, with an optional suffix (@code{_2}, -@code{_3}, @code{...}) to make it unique. You can see index names for a -table using @code{SHOW INDEX FROM tbl_name}. -@xref{SHOW, , @code{SHOW}}. +@example +mysql> select 'David!' LIKE 'David_'; + -> 1 +mysql> select 'David!' LIKE '%D%v%'; + -> 1 +@end example -@item -@cindex @code{NULL} values, and indexes -@cindex indexes, and @code{NULL} values -Only the @code{MyISAM} table type supports indexes on columns that can have -@code{NULL} values. In other cases you must declare such columns -@code{NOT NULL} or an error results. +To test for literal instances of a wild-card character, precede the character +with the escape character. If you don't specify the @code{ESCAPE} character, +@samp{\} is assumed: -@item -With @code{col_name(length)} syntax, you can specify an index that -uses only a part of a @code{CHAR} or @code{VARCHAR} column. This can -make the index file much smaller. -@xref{Indexes}. +@multitable @columnfractions .1 .9 +@item @code{\%} @tab Matches one @code{%} character +@item @code{\_} @tab Matches one @code{_} character +@end multitable -@item -@cindex @code{BLOB} columns, indexing -@cindex indexes, and @code{BLOB} columns -@cindex @code{TEXT} columns, indexing -@cindex indexes, and @code{TEXT} columns -Only the @code{MyISAM} table type supports indexing on @code{BLOB} and -@code{TEXT} columns. When putting an index on a @code{BLOB} or @code{TEXT} -column you MUST always specify the length of the index: @example -CREATE TABLE test (blob_col BLOB, index(blob_col(10))); +mysql> select 'David!' LIKE 'David\_'; + -> 0 +mysql> select 'David_' LIKE 'David\_'; + -> 1 @end example -@item -When you use @code{ORDER BY} or @code{GROUP BY} with a @code{TEXT} or -@code{BLOB} column, only the first @code{max_sort_length} bytes are used. -@xref{BLOB, , @code{BLOB}}. - -@item -In @strong{MySQL} Version 3.23.23 or later, you can also create special -@strong{FULLTEXT} indexes. They are used for full-text search. Only the -@code{MyISAM} table type supports @code{FULLTEXT} indexes. They can be created -only from @code{VARCHAR} and @code{TEXT} columns. -Indexing always happens over the entire column, partial indexing is not -supported. See @ref{Fulltext Search} for details of operation. - -@item -The @code{FOREIGN KEY}, @code{CHECK}, and @code{REFERENCES} clauses don't -actually do anything. The syntax for them is provided only for compatibility, -to make it easier to port code from other SQL servers and to run applications -that create tables with references. -@xref{Missing functions}. +To specify a different escape character, use the @code{ESCAPE} clause: -@item -Each @code{NULL} column takes one bit extra, rounded up to the nearest byte. +@example +mysql> select 'David_' LIKE 'David|_' ESCAPE '|'; + -> 1 +@end example -@item -The maximum record length in bytes can be calculated as follows: +The following two statements illustrate that string comparisons are +case insensitive unless one of the operands is a binary string: @example -row length = 1 - + (sum of column lengths) - + (number of NULL columns + 7)/8 - + (number of variable-length columns) +mysql> select 'abc' LIKE 'ABC'; + -> 1 +mysql> SELECT 'abc' LIKE BINARY 'ABC'; + -> 0 @end example -@item -The @code{table_options} and @code{SELECT} options are only -implemented in @strong{MySQL} Version 3.23 and above. +@code{LIKE} is allowed on numeric expressions! (This is a @strong{MySQL} +extension to the ANSI SQL @code{LIKE}.) -The different table types are: +@example +mysql> select 10 LIKE '1%'; + -> 1 +@end example -@multitable @columnfractions .20 .80 -@item BDB or Berkeley_db @tab Transaction-safe tables with page locking. @xref{BDB}. -@item HEAP @tab The data for this table is only stored in memory. @xref{HEAP}. -@item ISAM @tab The original table handler. @xref{ISAM}. -@item InnoDB @tab Transaction-safe tables with row locking. @xref{InnoDB}. -@item MERGE @tab A collection of MyISAM tables used as one table. @xref{MERGE}. -@item MyISAM @tab The new binary portable table handler that is replacing ISAM. @xref{MyISAM}. -@end multitable -@xref{Table types}. - -If a table type is specified, and that particular type is not available, -@strong{MySQL} will choose the closest table type to the one that you have -specified. For example, if @code{TYPE=BDB} is specified, and that distribution -of @strong{MySQL} does not support @code{BDB} tables, the table will be created -as @code{MyISAM} instead. - -The other table options are used to optimize the behavior of the -table. In most cases, you don't have to specify any of them. -The options work for all table types, if not otherwise indicated: - -@multitable @columnfractions .20 .80 -@item @code{AUTO_INCREMENT} @tab The next auto_increment value you want to set for your table (MyISAM). -@item @code{AVG_ROW_LENGTH} @tab An approximation of the average row length for your table. You only need to set this for large tables with variable size records. -@item @code{CHECKSUM} @tab Set this to 1 if you want @strong{MySQL} to maintain a checksum for all rows (makes the table a little slower to update but makes it easier to find corrupted tables) (MyISAM). -@item @code{COMMENT} @tab A 60-character comment for your table. -@item @code{MAX_ROWS} @tab Max number of rows you plan to store in the table. -@item @code{MIN_ROWS} @tab Minimum number of rows you plan to store in the table. -@item @code{PACK_KEYS} @tab Set this to 1 if you want to have a smaller index. This usually makes updates slower and reads faster (MyISAM, ISAM). -@item @code{PASSWORD} @tab Encrypt the @code{.frm} file with a password. This option doesn't do anything in the standard @strong{MySQL} version. -@item @code{DELAY_KEY_WRITE} @tab Set this to 1 if want to delay key table updates until the table is closed (MyISAM). -@item @code{ROW_FORMAT} @tab Defines how the rows should be stored. Currently you can only use the DYNAMIC and STATIC options for MyISAM tables. -@end multitable +Note: Because @strong{MySQL} uses the C escape syntax in strings (for example, +@samp{\n}), you must double any @samp{\} that you use in your @code{LIKE} +strings. For example, to search for @samp{\n}, specify it as @samp{\\n}. To +search for @samp{\}, specify it as @samp{\\\\} (the backslashes are stripped +once by the parser and another time when the pattern match is done, leaving +a single backslash to be matched). -When you use a @code{MyISAM} table, @strong{MySQL} uses the product of -@code{max_rows * avg_row_length} to decide how big the resulting table -will be. If you don't specify any of the above options, the maximum size -for a table will be 4G (or 2G if your operating systems only supports 2G -tables). The reason for this is just to keep down the pointer sizes -to make the index smaller and faster if you don't really need big files. +@findex NOT LIKE +@item expr NOT LIKE pat [ESCAPE 'escape-char'] +Same as @code{NOT (expr LIKE pat [ESCAPE 'escape-char'])}. -If you don't use @code{PACK_KEYS}, the default is to only pack strings, -not numbers. If you use @code{PACK_KEYS=1}, numbers will be packed as well. +@cindex mSQL compatibility +@cindex compatibility, with mSQL +@findex REGEXP +@findex RLIKE +@item expr REGEXP pat +@itemx expr RLIKE pat +Performs a pattern match of a string expression @code{expr} against a pattern +@code{pat}. The pattern can be an extended regular expression. +@xref{Regexp}. Returns @code{1} if @code{expr} matches @code{pat}, otherwise +returns @code{0}. @code{RLIKE} is a synonym for @code{REGEXP}, provided for +@code{mSQL} compatibility. Note: Because @strong{MySQL} uses the C escape +syntax in strings (for example, @samp{\n}), you must double any @samp{\} that +you use in your @code{REGEXP} strings. As of @strong{MySQL} Version 3.23.4, +@code{REGEXP} is case insensitive for normal (not binary) strings: -When packing binary number keys, @strong{MySQL} will use prefix compression. -This means that you will only get a big benefit of this if you have -many numbers that are the same. Prefix compression means that every -key needs one extra byte to indicate how many bytes of the previous key are -the same for the next key (note that the pointer to the row is stored -in high-byte-first-order directly after the key, to improve -compression.) This means that if you have many equal keys on two rows -in a row, all following 'same' keys will usually only take 2 bytes -(including the pointer to the row). Compare this to the ordinary case -where the following keys will take storage_size_for_key + -pointer_size (usually 4). On the other hand, if all keys are -totally different, you will lose 1 byte per key, if the key isn't a -key that can have @code{NULL} values (In this case the packed key length will -be stored in the same byte that is used to mark if a key is @code{NULL}.) +@example +mysql> select 'Monty!' REGEXP 'm%y%%'; + -> 0 +mysql> select 'Monty!' REGEXP '.*'; + -> 1 +mysql> select 'new*\n*line' REGEXP 'new\\*.\\*line'; + -> 1 +mysql> select "a" REGEXP "A", "a" REGEXP BINARY "A"; + -> 1 0 +mysql> select "a" REGEXP "^[a-d]"; + -> 1 +@end example @item -If you specify a @code{SELECT} after the @code{CREATE} statement, -@strong{MySQL} will create new fields for all elements in the -@code{SELECT}. For example: +@code{REGEXP} and @code{RLIKE} use the current character set (ISO-8859-1 +Latin1 by default) when deciding the type of a character. -@example -mysql> CREATE TABLE test (a int not null auto_increment, - primary key (a), key(b)) - TYPE=MyISAM SELECT b,c from test2; -@end example +@findex NOT REGEXP +@item expr NOT REGEXP pat +@itemx expr NOT RLIKE pat +Same as @code{NOT (expr REGEXP pat)}. -This will create a @code{MyISAM} table with three columns, a, b, and c. -Notice that the columns from the @code{SELECT} statement are appended to -the right side of the table, not overlapped onto it. Take the following -example: +@findex STRCMP() +@item STRCMP(expr1,expr2) +@code{STRCMP()} +returns @code{0} if the strings are the same, @code{-1} if the first +argument is smaller than the second according to the current sort order, +and @code{1} otherwise: @example -mysql> select * from foo; -+---+ -| n | -+---+ -| 1 | -+---+ - -mysql> create table bar (m int) select n from foo; -Query OK, 1 row affected (0.02 sec) -Records: 1 Duplicates: 0 Warnings: 0 - -mysql> select * from bar; -+------+---+ -| m | n | -+------+---+ -| NULL | 1 | -+------+---+ -1 row in set (0.00 sec) +mysql> select STRCMP('text', 'text2'); + -> -1 +mysql> select STRCMP('text2', 'text'); + -> 1 +mysql> select STRCMP('text', 'text'); + -> 0 @end example -For each row in table @code{foo}, a row is inserted in @code{bar} with -the values from @code{foo} and default values for the new columns. - -If any errors occur while copying the data to the table, it will -automatically be deleted. - -To ensure that the update log/binary log can be used to re-create the -original tables, @strong{MySQL} will not allow concurrent inserts during -@code{CREATE TABLE .... SELECT}. -@item -The @code{RAID_TYPE} option will help you to break the 2G/4G limit for -the MyISAM data file (not the index file) on -operating systems that don't support big files. You can get also more speed -from the I/O bottleneck by putting @code{RAID} directories on different -physical disks. @code{RAID_TYPE} will work on any OS, as long as you have -configured @strong{MySQL} with @code{--with-raid}. For now the only allowed -@code{RAID_TYPE} is @code{STRIPED} (@code{1} and @code{RAID0} are aliases -for this). +@findex MATCH ... AGAINST() +@item MATCH (col1,col2,...) AGAINST (expr) +@code{MATCH ... AGAINST()} is used for full-text search and returns +relevance - similarity measure between the text in columns +@code{(col1,col2,...)} and the query @code{expr}. Relevance is a +positive floating-point number. Zero relevance means no similarity. +For @code{MATCH ... AGAINST()} to work, a @strong{FULLTEXT} index +must be created first. @xref{CREATE TABLE, , @code{CREATE TABLE}}. +@code{MATCH ... AGAINST()} is available in @strong{MySQL} Version +3.23.23 or later. For details and usage examples +@pxref{Fulltext Search}. +@end table -If you specify @code{RAID_TYPE=STRIPED} for a @code{MyISAM} table, -@code{MyISAM} will create @code{RAID_CHUNKS} subdirectories named 00, -01, 02 in the database directory. In each of these directories -@code{MyISAM} will create a @code{table_name.MYD}. When writing data -to the data file, the @code{RAID} handler will map the first -@code{RAID_CHUNKSIZE} *1024 bytes to the first file, the next -@code{RAID_CHUNKSIZE} *1024 bytes to the next file and so on. -@item -@code{UNION} is used when you want to use a collection of identical -tables as one. This only works with MERGE tables. @xref{MERGE}. +@findex casts +@cindex cast operators +@cindex operators, cast +@node Casts, Control flow functions, String comparison functions, Functions +@subsection Cast Operators -For the moment you need to have @code{SELECT}, @code{UPDATE}, and -@code{DELETE} privileges on the tables you map to a @code{MERGE} table. -All mapped tables must be in the same database as the @code{MERGE} table. -@item -In the created table the @code{PRIMARY} key will be placed first, followed -by all @code{UNIQUE} keys and then the normal keys. This helps the -@strong{MySQL} optimizer to prioritize which key to use and also more quickly -detect duplicated @code{UNIQUE} keys. +@table @code +@findex BINARY +@item @code{BINARY} +The @code{BINARY} operator casts the string following it to a binary string. +This is an easy way to force a column comparison to be case sensitive even +if the column isn't defined as @code{BINARY} or @code{BLOB}: +@example +mysql> select "a" = "A"; + -> 1 +mysql> select BINARY "a" = "A"; + -> 0 +@end example -@item -By using @code{DATA DIRECTORY="directory"} or @code{INDEX -DIRECTORY="directory"} you can specify where the table handler should -put it's table and index files. This only works for @code{MyISAM} tables -in @code{MySQL} 4.0, when you are not using the @code{--skip-symlink} -option. @xref{Symbolic links to tables}. +@code{BINARY} was introduced in @strong{MySQL} Version 3.23.0. -@end itemize +Note that in some context @strong{MySQL} will not be able to use the +index efficiently when you cast an indexed column to @code{BINARY}. +@end table -@cindex silent column changes -@node Silent column changes, , CREATE TABLE, CREATE TABLE -@subsection Silent Column Specification Changes +If you want to compare a blob case-insensitively you can always convert +the blob to upper case before doing the comparison: -In some cases, @strong{MySQL} silently changes a column specification from -that given in a @code{CREATE TABLE} statement. (This may also occur with -@code{ALTER TABLE}.): +@example +SELECT 'A' LIKE UPPER(blob_col) FROM table_name; +@end example -@itemize @bullet -@item -@code{VARCHAR} columns with a length less than four are changed to -@code{CHAR}. +We plan to soon introduce casting between different character sets to +make string comparison even more flexible. -@item -If any column in a table has a variable length, the entire row is -variable-length as a result. Therefore, if a table contains any -variable-length columns (@code{VARCHAR}, @code{TEXT}, or @code{BLOB}), -all @code{CHAR} columns longer than three characters are changed to -@code{VARCHAR} columns. This doesn't affect how you use the columns in -any way; in @strong{MySQL}, @code{VARCHAR} is just a different way to -store characters. @strong{MySQL} performs this conversion because it -saves space and makes table operations faster. @xref{Table types}. +@findex control flow functions +@findex functions, control flow +@node Control flow functions, Mathematical functions, Casts, Functions +@subsection Control Flow Functions -@item -@code{TIMESTAMP} display sizes must be even and in the range from 2 to 14. -If you specify a display size of 0 or greater than 14, the size is coerced -to 14. Odd-valued sizes in the range from 1 to 13 are coerced -to the next higher even number. +@table @code +@cindex @code{NULL}, testing for null +@findex IFNULL() +@item IFNULL(expr1,expr2) +If @code{expr1} is not @code{NULL}, @code{IFNULL()} returns @code{expr1}, +else it returns @code{expr2}. @code{IFNULL()} returns a numeric or string +value, depending on the context in which it is used: +@example +mysql> select IFNULL(1,0); + -> 1 +mysql> select IFNULL(NULL,10); + -> 10 +mysql> select IFNULL(1/0,10); + -> 10 +mysql> select IFNULL(1/0,'yes'); + -> 'yes' +@end example -@item -You cannot store a literal @code{NULL} in a @code{TIMESTAMP} column; setting -it to @code{NULL} sets it to the current date and time. Because -@code{TIMESTAMP} columns behave this way, the @code{NULL} and @code{NOT NULL} -attributes do not apply in the normal way and are ignored if you specify -them. @code{DESCRIBE tbl_name} always reports that a @code{TIMESTAMP} -column may be assigned @code{NULL} values. +@findex NULLIF() +@item NULLIF(expr1,expr2) +If @code{expr1 = expr2} is true, return @code{NULL} else return @code{expr1}. +This is the same as @code{CASE WHEN x = y THEN NULL ELSE x END}: +@example +mysql> select NULLIF(1,1); + -> NULL +mysql> select NULLIF(1,2); + -> 1 +@end example -@item -@strong{MySQL} maps certain column types used by other SQL database vendors -to @strong{MySQL} types. @xref{Other-vendor column types}. -@end itemize +Note that @code{expr1} is evaluated twice in @strong{MySQL} if the arguments +are equal. -If you want to see whether or not @strong{MySQL} used a column type other -than the one you specified, issue a @code{DESCRIBE tbl_name} statement after -creating or altering your table. +@findex IF() +@item IF(expr1,expr2,expr3) +If @code{expr1} is TRUE (@code{expr1 <> 0} and @code{expr1 <> NULL}) then +@code{IF()} returns @code{expr2}, else it returns @code{expr3}. +@code{IF()} returns a numeric or string value, depending on the context +in which it is used: -@cindex @code{myisampack} -Certain other column type changes may occur if you compress a table -using @code{myisampack}. @xref{Compressed format}. +@example +mysql> select IF(1>2,2,3); + -> 3 +mysql> select IF(1<2,'yes','no'); + -> 'yes' +mysql> select IF(strcmp('test','test1'),'no','yes'); + -> 'no' +@end example -@findex ALTER TABLE -@node ALTER TABLE, RENAME TABLE, CREATE TABLE, Reference -@section @code{ALTER TABLE} Syntax +@code{expr1} is evaluated as an integer value, which means that if you are +testing floating-point or string values, you should do so using a comparison +operation: @example -ALTER [IGNORE] TABLE tbl_name alter_spec [, alter_spec ...] - -alter_specification: - ADD [COLUMN] create_definition [FIRST | AFTER column_name ] - or ADD [COLUMN] (create_definition, create_definition,...) - or ADD INDEX [index_name] (index_col_name,...) - or ADD PRIMARY KEY (index_col_name,...) - or ADD UNIQUE [index_name] (index_col_name,...) - or ADD FULLTEXT [index_name] (index_col_name,...) - or ADD [CONSTRAINT symbol] FOREIGN KEY index_name (index_col_name,...) - [reference_definition] - or ALTER [COLUMN] col_name @{SET DEFAULT literal | DROP DEFAULT@} - or CHANGE [COLUMN] old_col_name create_definition - or MODIFY [COLUMN] create_definition - or DROP [COLUMN] col_name - or DROP PRIMARY KEY - or DROP INDEX index_name - or RENAME [TO] new_tbl_name - or ORDER BY col - or table_options +mysql> select IF(0.1,1,0); + -> 0 +mysql> select IF(0.1<>0,1,0); + -> 1 @end example -@code{ALTER TABLE} allows you to change the structure of an existing table. -For example, you can add or delete columns, create or destroy indexes, change -the type of existing columns, or rename columns or the table itself. You can -also change the comment for the table and type of the table. -@xref{CREATE TABLE, , @code{CREATE TABLE}}. +In the first case above, @code{IF(0.1)} returns @code{0} because @code{0.1} +is converted to an integer value, resulting in a test of @code{IF(0)}. This +may not be what you expect. In the second case, the comparison tests the +original floating-point value to see whether it is non-zero. The result +of the comparison is used as an integer. -If you use @code{ALTER TABLE} to change a column specification but -@code{DESCRIBE tbl_name} indicates that your column was not changed, it is -possible that @strong{MySQL} ignored your modification for one of the reasons -described in @ref{Silent column changes}. For example, if you try to change -a @code{VARCHAR} column to @code{CHAR}, @strong{MySQL} will still use -@code{VARCHAR} if the table contains other variable-length columns. +The default return type of @code{IF()} (which may matter when it is +stored into a temporary table) is calculated in @strong{MySQL} Version +3.23 as follows: -@code{ALTER TABLE} works by making a temporary copy of the original table. -The alteration is performed on the copy, then the original table is -deleted and the new one is renamed. This is done in such a way that -all updates are automatically redirected to the new table without -any failed updates. While @code{ALTER TABLE} is executing, the original -table is readable by other clients. Updates and writes to the table -are stalled until the new table is ready. +@multitable @columnfractions .55 .45 +@item @strong{Expression} @tab @strong{Return value} +@item expr2 or expr3 returns string @tab string +@item expr2 or expr3 returns a floating-point value @tab floating-point +@item expr2 or expr3 returns an integer @tab integer +@end multitable -Note that if you use any other option to @code{ALTER TABLE} than -@code{RENAME}, @strong{MySQL} will always create a temporary table, even -if the data wouldn't strictly need to be copied (like when you change the -name of a column). We plan to fix this in the future, but as one doesn't -normally do @code{ALTER TABLE} that often this isn't that high on our TODO. +@findex CASE +@item CASE value WHEN [compare-value] THEN result [WHEN [compare-value] THEN result ...] [ELSE result] END +@item CASE WHEN [condition] THEN result [WHEN [condition] THEN result ...] [ELSE result] END -@itemize @bullet -@item -To use @code{ALTER TABLE}, you need @strong{ALTER}, @strong{INSERT}, -and @strong{CREATE} privileges on the table. +The first version returns the @code{result} where +@code{value=compare-value}. The second version returns the result for +the first condition, which is true. If there was no matching result +value, then the result after @code{ELSE} is returned. If there is no +@code{ELSE} part then @code{NULL} is returned: -@item -@code{IGNORE} is a @strong{MySQL} extension to ANSI SQL92. -It controls how @code{ALTER TABLE} works if there are duplicates on -unique keys in the new table. -If @code{IGNORE} isn't specified, the copy is aborted and rolled back. -If @code{IGNORE} is specified, then for rows with duplicates on a unique -key, only the first row is used; the others are deleted. +@example +mysql> SELECT CASE 1 WHEN 1 THEN "one" WHEN 2 THEN "two" ELSE "more" END; + -> "one" +mysql> SELECT CASE WHEN 1>0 THEN "true" ELSE "false" END; + -> "true" +mysql> SELECT CASE BINARY "B" when "a" then 1 when "b" then 2 END; + -> NULL +@end example +@end table -@item -You can issue multiple @code{ADD}, @code{ALTER}, @code{DROP}, and -@code{CHANGE} clauses in a single @code{ALTER TABLE} statement. This is a -@strong{MySQL} extension to ANSI SQL92, which allows only one of each clause -per @code{ALTER TABLE} statement. +The type of the return value (@code{INTEGER}, @code{DOUBLE} or +@code{STRING}) is the same as the type of the first returned value (the +expression after the first @code{THEN}). -@item -@code{CHANGE col_name}, @code{DROP col_name}, and @code{DROP -INDEX} are @strong{MySQL} extensions to ANSI SQL92. +@findex mathematical functions +@findex functions, mathematical +@node Mathematical functions, String functions, Control flow functions, Functions +@subsection Mathematical Functions +All mathematical functions return @code{NULL} in case of an error. -@item -@code{MODIFY} is an Oracle extension to @code{ALTER TABLE}. +@table @code +@findex - (unary minus) +@findex minus, unary (-) +@findex unary minus (-) +@item - +Unary minus. Changes the sign of the argument: +@example +mysql> select - 2; + -> -2 +@end example -@item -The optional word @code{COLUMN} is a pure noise word and can be omitted. +Note that if this operator is used with a @code{BIGINT}, the return value is a +@code{BIGINT}! This means that you should avoid using @code{-} on integers that +may have the value of @code{-2^63}! -@item -If you use @code{ALTER TABLE tbl_name RENAME TO new_name} without any other -options, @strong{MySQL} simply renames the files that correspond to the table -@code{tbl_name}. There is no need to create the temporary table. -@xref{RENAME TABLE,, @code{RENAME TABLE}}. +@findex ABS() +@item ABS(X) +Returns the absolute value of @code{X}: +@example +mysql> select ABS(2); + -> 2 +mysql> select ABS(-32); + -> 32 +@end example -@item -@code{create_definition} clauses use the same syntax for @code{ADD} and -@code{CHANGE} as for @code{CREATE TABLE}. Note that this syntax includes -the column name, not just the column type. -@xref{CREATE TABLE, , @code{CREATE TABLE}}. +This function is safe to use with @code{BIGINT} values. -@item -You can rename a column using a @code{CHANGE old_col_name create_definition} -clause. To do so, specify the old and new column names and the type that -the column currently has. For example, to rename an @code{INTEGER} column -from @code{a} to @code{b}, you can do this: +@findex SIGN() +@item SIGN(X) +Returns the sign of the argument as @code{-1}, @code{0}, or @code{1}, depending +on whether @code{X} is negative, zero, or positive: +@example +mysql> select SIGN(-32); + -> -1 +mysql> select SIGN(0); + -> 0 +mysql> select SIGN(234); + -> 1 +@end example +@findex MOD() +@findex % (modulo) +@findex modulo (%) +@item MOD(N,M) +@itemx % +Modulo (like the @code{%} operator in C). +Returns the remainder of @code{N} divided by @code{M}: @example -mysql> ALTER TABLE t1 CHANGE a b INTEGER; +mysql> select MOD(234, 10); + -> 4 +mysql> select 253 % 7; + -> 1 +mysql> select MOD(29,9); + -> 2 @end example -If you want to change a column's type but not the name, @code{CHANGE} -syntax still requires two column names even if they are the same. For -example: +This function is safe to use with @code{BIGINT} values. +@findex FLOOR() +@item FLOOR(X) +Returns the largest integer value not greater than @code{X}: @example -mysql> ALTER TABLE t1 CHANGE b b BIGINT NOT NULL; +mysql> select FLOOR(1.23); + -> 1 +mysql> select FLOOR(-1.23); + -> -2 @end example -However, as of @strong{MySQL} Version 3.22.16a, you can also use @code{MODIFY} -to change a column's type without renaming it: +Note that the return value is converted to a @code{BIGINT}! +@findex CEILING() +@item CEILING(X) +Returns the smallest integer value not less than @code{X}: @example -mysql> ALTER TABLE t1 MODIFY b BIGINT NOT NULL; +mysql> select CEILING(1.23); + -> 2 +mysql> select CEILING(-1.23); + -> -1 @end example -@item -If you use @code{CHANGE} or @code{MODIFY} to shorten a column for which -an index exists on part of the column (for instance, if you have an index -on the first 10 characters of a @code{VARCHAR} column), you cannot make -the column shorter than the number of characters that are indexed. - -@item -When you change a column type using @code{CHANGE} or @code{MODIFY}, -@strong{MySQL} tries to convert data to the new type as well as possible. - -@item -In @strong{MySQL} Version 3.22 or later, you can use @code{FIRST} or -@code{ADD ... AFTER col_name} to add a column at a specific position within -a table row. The default is to add the column last. - -@findex ALTER COLUMN -@item -@code{ALTER COLUMN} specifies a new default value for a column -or removes the old default value. -If the old default is removed and the column can be @code{NULL}, the new -default is @code{NULL}. If the column cannot be @code{NULL}, @strong{MySQL} -assigns a default value, as described in -@ref{CREATE TABLE, , @code{CREATE TABLE}}. - -@findex DROP INDEX -@item -@code{DROP INDEX} removes an index. This is a @strong{MySQL} extension to -ANSI SQL92. @xref{DROP INDEX}. - -@item -If columns are dropped from a table, the columns are also removed from any -index of which they are a part. If all columns that make up an index are -dropped, the index is dropped as well. - -@item -If a table contains only one column, the column cannot be dropped. -If what you intend is to remove the table, use @code{DROP TABLE} instead. - -@findex DROP PRIMARY KEY -@item -@code{DROP PRIMARY KEY} drops the primary index. If no such -index exists, it drops the first @code{UNIQUE} index in the table. -(@strong{MySQL} marks the first @code{UNIQUE} key as the @code{PRIMARY KEY} -if no @code{PRIMARY KEY} was specified explicitly.) +Note that the return value is converted to a @code{BIGINT}! -@findex ORDER BY -@item -@code{ORDER BY} allows you to create the new table with the rows in a -specific order. Note that the table will not remain in this order after -inserts and deletes. In some cases, it may make sorting easier for -@strong{MySQL} if the table is in order by the column that you wish to -order it by later. This option is mainly useful when you know that you -are mostly going to query the rows in a certain order; By using this -option after big changes to the table, you may be able to get higher -performance. +@findex ROUND() +@item ROUND(X) +Returns the argument @code{X}, rounded to the nearest integer: +@example +mysql> select ROUND(-1.23); + -> -1 +mysql> select ROUND(-1.58); + -> -2 +mysql> select ROUND(1.58); + -> 2 +@end example -@findex ALTER TABLE -@item -If you use @code{ALTER TABLE} on a @code{MyISAM} table, all non-unique -indexes are created in a separate batch (like in @code{REPAIR}). -This should make @code{ALTER TABLE} much faster when you have many indexes. +Note that the behavior of @code{ROUND()} when the argument +is half way between two integers depends on the C library +implementation. Some round to the nearest even number, +always up, always down, or always towards zero. If you need +one kind of rounding, you should use a well-defined function +like @code{TRUNCATE()} or @code{FLOOR()} instead. -@item -@findex mysql_info() -With the C API function @code{mysql_info()}, you can find out how many -records were copied, and (when @code{IGNORE} is used) how many records were -deleted due to duplication of unique key values. +@findex ROUND() +@item ROUND(X,D) +Returns the argument @code{X}, rounded to a number with @code{D} decimals. +If @code{D} is @code{0}, the result will have no decimal point or fractional +part: -@item -@cindex foreign keys -@cindex references -The @code{FOREIGN KEY}, @code{CHECK}, and @code{REFERENCES} clauses don't -actually do anything. The syntax for them is provided only for compatibility, -to make it easier to port code from other SQL servers and to run applications -that create tables with references. -@xref{Missing functions}. -@end itemize +@example +mysql> select ROUND(1.298, 1); + -> 1.3 +mysql> select ROUND(1.298, 0); + -> 1 +@end example -Here is an example that shows some of the uses of @code{ALTER TABLE}. We -begin with a table @code{t1} that is created as shown below: +@findex EXP() +@item EXP(X) +Returns the value of @code{e} (the base of natural logarithms) raised to +the power of @code{X}: +@example +mysql> select EXP(2); + -> 7.389056 +mysql> select EXP(-2); + -> 0.135335 +@end example +@findex LOG() +@item LOG(X) +Returns the natural logarithm of @code{X}: +@example +mysql> select LOG(2); + -> 0.693147 +mysql> select LOG(-2); + -> NULL +@end example +If you want the log of a number @code{X} to some arbitary base @code{B}, use +the formula @code{LOG(X)/LOG(B)}. +@findex LOG10() +@item LOG10(X) +Returns the base-10 logarithm of @code{X}: @example -mysql> CREATE TABLE t1 (a INTEGER,b CHAR(10)); +mysql> select LOG10(2); + -> 0.301030 +mysql> select LOG10(100); + -> 2.000000 +mysql> select LOG10(-100); + -> NULL @end example -To rename the table from @code{t1} to @code{t2}: +@findex POW() +@findex POWER() +@item POW(X,Y) +@itemx POWER(X,Y) +Returns the value of @code{X} raised to the power of @code{Y}: +@example +mysql> select POW(2,2); + -> 4.000000 +mysql> select POW(2,-2); + -> 0.250000 +@end example +@findex SQRT() +@item SQRT(X) +Returns the non-negative square root of @code{X}: @example -mysql> ALTER TABLE t1 RENAME t2; +mysql> select SQRT(4); + -> 2.000000 +mysql> select SQRT(20); + -> 4.472136 @end example -To change column @code{a} from @code{INTEGER} to @code{TINYINT NOT NULL} -(leaving the name the same), and to change column @code{b} from -@code{CHAR(10)} to @code{CHAR(20)} as well as renaming it from @code{b} to -@code{c}: +@findex PI() +@item PI() +Returns the value of PI. The default shown number of decimals is 5, but +@strong{MySQL} internally uses the full double precession for PI. +@example +mysql> select PI(); + -> 3.141593 +mysql> SELECT PI()+0.000000000000000000; + -> 3.141592653589793116 +@end example +@findex COS() +@item COS(X) +Returns the cosine of @code{X}, where @code{X} is given in radians: @example -mysql> ALTER TABLE t2 MODIFY a TINYINT NOT NULL, CHANGE b c CHAR(20); +mysql> select COS(PI()); + -> -1.000000 @end example -To add a new @code{TIMESTAMP} column named @code{d}: +@findex SIN() +@item SIN(X) +Returns the sine of @code{X}, where @code{X} is given in radians: +@example +mysql> select SIN(PI()); + -> 0.000000 +@end example +@findex TAN() +@item TAN(X) +Returns the tangent of @code{X}, where @code{X} is given in radians: @example -mysql> ALTER TABLE t2 ADD d TIMESTAMP; +mysql> select TAN(PI()+1); + -> 1.557408 @end example -To add an index on column @code{d}, and make column @code{a} the primary key: +@findex ACOS() +@item ACOS(X) +Returns the arc cosine of @code{X}, that is, the value whose cosine is +@code{X}. Returns @code{NULL} if @code{X} is not in the range @code{-1} to +@code{1}: +@example +mysql> select ACOS(1); + -> 0.000000 +mysql> select ACOS(1.0001); + -> NULL +mysql> select ACOS(0); + -> 1.570796 +@end example +@findex ASIN() +@item ASIN(X) +Returns the arc sine of @code{X}, that is, the value whose sine is +@code{X}. Returns @code{NULL} if @code{X} is not in the range @code{-1} to +@code{1}: @example -mysql> ALTER TABLE t2 ADD INDEX (d), ADD PRIMARY KEY (a); +mysql> select ASIN(0.2); + -> 0.201358 +mysql> select ASIN('foo'); + -> 0.000000 @end example -To remove column @code{c}: +@findex ATAN() +@item ATAN(X) +Returns the arc tangent of @code{X}, that is, the value whose tangent is +@code{X}: +@example +mysql> select ATAN(2); + -> 1.107149 +mysql> select ATAN(-2); + -> -1.107149 +@end example +@findex ATAN2() +@item ATAN2(Y,X) +Returns the arc tangent of the two variables @code{X} and @code{Y}. It is +similar to calculating the arc tangent of @code{Y / X}, except that the +signs of both arguments are used to determine the quadrant of the +result: @example -mysql> ALTER TABLE t2 DROP COLUMN c; +mysql> select ATAN(-2,2); + -> -0.785398 +mysql> select ATAN(PI(),0); + -> 1.570796 @end example -To add a new @code{AUTO_INCREMENT} integer column named @code{c}: +@findex COT() +@item COT(X) +Returns the cotangent of @code{X}: +@example +mysql> select COT(12); + -> -1.57267341 +mysql> select COT(0); + -> NULL +@end example +@findex RAND() +@item RAND() +@itemx RAND(N) +Returns a random floating-point value in the range @code{0} to @code{1.0}. +If an integer argument @code{N} is specified, it is used as the seed value: @example -mysql> ALTER TABLE t2 ADD c INT UNSIGNED NOT NULL AUTO_INCREMENT, - ADD INDEX (c); +mysql> select RAND(); + -> 0.5925 +mysql> select RAND(20); + -> 0.1811 +mysql> select RAND(20); + -> 0.1811 +mysql> select RAND(); + -> 0.2079 +mysql> select RAND(); + -> 0.7888 @end example +You can't use a column with @code{RAND()} values in an @code{ORDER BY} +clause, because @code{ORDER BY} would evaluate the column multiple times. +In @strong{MySQL} Version 3.23, you can, however, do: +@code{SELECT * FROM table_name ORDER BY RAND()} -Note that we indexed @code{c}, because @code{AUTO_INCREMENT} columns must be -indexed, and also that we declare @code{c} as @code{NOT NULL}, because -indexed columns cannot be @code{NULL}. +This is useful to get a random sample of a set @code{SELECT * FROM +table1,table2 WHERE a=b AND c<d ORDER BY RAND() LIMIT 1000}. -When you add an @code{AUTO_INCREMENT} column, column values are filled in -with sequence numbers for you automatically. You can set the first -sequence number by executing @code{SET INSERT_ID=#} before -@code{ALTER TABLE} or using the @code{AUTO_INCREMENT = #} table option. -@xref{SET OPTION}. +Note that a @code{RAND()} in a @code{WHERE} clause will be re-evaluated +every time the @code{WHERE} is executed. -With MyISAM tables, if you don't change the @code{AUTO_INCREMENT} -column, the sequence number will not be affected. If you drop an -@code{AUTO_INCREMENT} column and then add another @code{AUTO_INCREMENT} -column, the numbers will start from 1 again. +@findex LEAST() +@item LEAST(X,Y,...) +With two or more arguments, returns the smallest (minimum-valued) argument. +The arguments are compared using the following rules: -@xref{ALTER TABLE problems}. +@itemize @bullet +@item +If the return value is used in an @code{INTEGER} context, or all arguments +are integer-valued, they are compared as integers. -@findex RENAME TABLE -@node RENAME TABLE, DROP TABLE, ALTER TABLE, Reference -@section @code{RENAME TABLE} Syntax +@item +If the return value is used in a @code{REAL} context, or all arguments are +real-valued, they are compared as reals. -@example -RENAME TABLE tbl_name TO new_table_name[, tbl_name2 TO new_table_name2,...] -@end example +@item +If any argument is a case-sensitive string, the arguments are compared +as case-sensitive strings. -The rename is done atomically, which means that no other thread can -access any of the tables while the rename is running. This makes it -possible to replace a table with an empty one: +@item +In other cases, the arguments are compared as case-insensitive strings: +@end itemize @example -CREATE TABLE new_table (...); -RENAME TABLE old_table TO backup_table, new_table TO old_table; +mysql> select LEAST(2,0); + -> 0 +mysql> select LEAST(34.0,3.0,5.0,767.0); + -> 3.0 +mysql> select LEAST("B","A","C"); + -> "A" @end example +In @strong{MySQL} versions prior to Version 3.22.5, you can use @code{MIN()} +instead of @code{LEAST}. -The rename is done from left to right, which means that if you want to -swap two tables names, you have to: - +@findex GREATEST() +@item GREATEST(X,Y,...) +Returns the largest (maximum-valued) argument. +The arguments are compared using the same rules as for @code{LEAST}: @example -RENAME TABLE old_table TO backup_table, - new_table TO old_table, - backup_table TO new_table; +mysql> select GREATEST(2,0); + -> 2 +mysql> select GREATEST(34.0,3.0,5.0,767.0); + -> 767.0 +mysql> select GREATEST("B","A","C"); + -> "C" @end example +In @strong{MySQL} versions prior to Version 3.22.5, you can use @code{MAX()} +instead of @code{GREATEST}. -As long as two databases are on the same disk you can also rename -from one database to another: - +@findex DEGREES() +@item DEGREES(X) +Returns the argument @code{X}, converted from radians to degrees: @example -RENAME TABLE current_database.table_name TO other_database.table_name; +mysql> select DEGREES(PI()); + -> 180.000000 @end example -When you execute @code{RENAME}, you can't have any locked tables or -active transactions. You must also have the @code{ALTER} and @code{DROP} -privilege on the original table and @code{CREATE} and @code{INSERT} -privilege on the new table. +@findex RADIANS() +@item RADIANS(X) +Returns the argument @code{X}, converted from degrees to radians: +@example +mysql> select RADIANS(90); + -> 1.570796 +@end example -If @strong{MySQL} encounters any errors in a multiple table rename, it -will do a reverse rename for all renamed tables to get everything back -to the original state. +@findex TRUNCATE() +@item TRUNCATE(X,D) +Returns the number @code{X}, truncated to @code{D} decimals. If @code{D} +is @code{0}, the result will have no decimal point or fractional part: +@example +mysql> select TRUNCATE(1.223,1); + -> 1.2 +mysql> select TRUNCATE(1.999,1); + -> 1.9 +mysql> select TRUNCATE(1.999,0); + -> 1 +@end example -@findex DROP TABLE -@node DROP TABLE, OPTIMIZE TABLE, RENAME TABLE, Reference -@section @code{DROP TABLE} Syntax +Note that as decimal numbers are normally not stored as exact numbers in +computers, but as double values, you may be fooled by the following +result: +@cindex rounding errors @example -DROP TABLE [IF EXISTS] tbl_name [, tbl_name,...] [RESTRICT | CASCADE] +mysql> select TRUNCATE(10.28*100,0); + -> 1027 @end example -@code{DROP TABLE} removes one or more tables. All table data and the table -definition are @emph{removed}, so @strong{be careful} with this command! +The above happens because 10.28 is actually stored as something like +10.2799999999999999. +@end table -In @strong{MySQL} Version 3.22 or later, you can use the keywords -@code{IF EXISTS} to prevent an error from occurring for tables that don't -exist. +@findex string functions +@findex functions, string +@node String functions, Date and time functions, Mathematical functions, Functions +@subsection String Functions -@code{RESTRICT} and @code{CASCADE} are allowed to make porting easier. -For the moment they don't do anything. +String-valued functions return @code{NULL} if the length of the result would +be greater than the @code{max_allowed_packet} server parameter. @xref{Server +parameters}. -@strong{NOTE}: @code{DROP TABLE} is not transaction-safe and will -automatically commit any active transactions. +For functions that operate on string positions, +the first position is numbered 1. -@cindex tables, defragment -@cindex tables, fragmentation -@findex OPTIMIZE TABLE -@node OPTIMIZE TABLE, CHECK TABLE, DROP TABLE, Reference -@section @code{OPTIMIZE TABLE} Syntax +@table @code +@findex ASCII() +@item ASCII(str) +Returns the ASCII code value of the leftmost character of the string +@code{str}. Returns @code{0} if @code{str} is the empty string. Returns +@code{NULL} if @code{str} is @code{NULL}: @example -OPTIMIZE TABLE tbl_name[,tbl_name]... +mysql> select ASCII('2'); + -> 50 +mysql> select ASCII(2); + -> 50 +mysql> select ASCII('dx'); + -> 100 @end example -@code{OPTIMIZE TABLE} should be used if you have deleted a large part of a -table or if you have made many changes to a table with variable-length rows -(tables that have @code{VARCHAR}, @code{BLOB}, or @code{TEXT} columns). -Deleted records are maintained in a linked list and subsequent @code{INSERT} -operations reuse old record positions. You can use @code{OPTIMIZE TABLE} to -reclaim the unused space and to defragment the data file. - -For the moment @code{OPTIMIZE TABLE} only works on @strong{MyISAM} and -@code{BDB} tables. For @code{BDB} tables, @code{OPTIMIZE TABLE} is -currently mapped to @code{ANALYZE TABLE}. @xref{ANALYZE TABLE}. +See also the @code{ORD()} function. -You can get optimize table to work on other table types by starting -@code{mysqld} with @code{--skip-new} or @code{--safe-mode}, but in this -case @code{OPTIMIZE TABLE} is just mapped to @code{ALTER TABLE}. +@findex ORD() +@item ORD(str) +If the leftmost character of the string str is a multi-byte character, +returns the code of multi-byte character by returning the ASCII code value +of the character in the format of: +@code{((first byte ASCII code)*256+(second byte ASCII code))[*256+third byte ASCII code...]}. +If the leftmost character is not a multi-byte character, returns the same +value as the like @code{ASCII()} function does: -@code{OPTIMIZE TABLE} works the following way: -@itemize @bullet -@item -If the table has deleted or split rows, repair the table. -@item -If the index pages are not sorted, sort them. -@item -If the statistics are not up to date (and the repair couldn't be done -by sorting the index), update them. -@end itemize +@example +mysql> select ORD('2'); + -> 50 +@end example -@code{OPTIMIZE TABLE} for @code{MyISAM} tables is equvialent of running -@code{myisamchk --quick --check-changed-tables --sort-index --analyze} -on the table. +@findex CONV() +@item CONV(N,from_base,to_base) +Converts numbers between different number bases. Returns a string +representation of the number @code{N}, converted from base @code{from_base} +to base @code{to_base}. Returns @code{NULL} if any argument is @code{NULL}. +The argument @code{N} is interpreted as an integer, but may be specified as +an integer or a string. The minimum base is @code{2} and the maximum base is +@code{36}. If @code{to_base} is a negative number, @code{N} is regarded as a +signed number. Otherwise, @code{N} is treated as unsigned. @code{CONV} works +with 64-bit precision: -Note that the table is locked during the time @code{OPTIMIZE TABLE} is -running! +@example +mysql> select CONV("a",16,2); + -> '1010' +mysql> select CONV("6E",18,8); + -> '172' +mysql> select CONV(-17,10,-18); + -> '-H' +mysql> select CONV(10+"10"+'10'+0xa,10,10); + -> '40' +@end example -@findex CHECK TABLE -@node CHECK TABLE, BACKUP TABLE, OPTIMIZE TABLE, Reference -@section @code{CHECK TABLE} Syntax +@findex BIN() +@item BIN(N) +Returns a string representation of the binary value of @code{N}, where +@code{N} is a longlong (@code{BIGINT}) number. This is equivalent to +@code{CONV(N,10,2)}. Returns @code{NULL} if @code{N} is @code{NULL}: @example -CHECK TABLE tbl_name[,tbl_name...] [option [option...]] - -option = QUICK | FAST | MEDIUM | EXTEND | CHANGED +mysql> select BIN(12); + -> '1100' @end example -@code{CHECK TABLE} only works on @code{MyISAM} tables. On -@code{MyISAM} tables it's the same thing as running @code{myisamchk -m -table_name} on the table. +@findex OCT() +@item OCT(N) +Returns a string representation of the octal value of @code{N}, where +@code{N} is a longlong number. This is equivalent to @code{CONV(N,10,8)}. +Returns @code{NULL} if @code{N} is @code{NULL}: -If you don't specify any option @code{MEDIUM} is used. +@example +mysql> select OCT(12); + -> '14' +@end example -Checks the table(s) for errors. For @code{MyISAM} tables the key statistics -is updated. The command returns a table with the following columns: +@findex HEX() +@item HEX(N) +Returns a string representation of the hexadecimal value of @code{N}, where +@code{N} is a longlong (@code{BIGINT}) number. This is equivalent to +@code{CONV(N,10,16)}. Returns @code{NULL} if @code{N} is @code{NULL}: -@multitable @columnfractions .35 .65 -@item @strong{Column} @tab @strong{Value} -@item Table @tab Table name. -@item Op @tab Always ``check''. -@item Msg_type @tab One of @code{status}, @code{error}, @code{info}, or @code{warning}. -@item Msg_text @tab The message. -@end multitable - -Note that you can get many rows of information for each checked -table. The last row will be of @code{Msg_type status} and should -normally be @code{OK}. If you don't get @code{OK}, or @code{Not -checked} you should normally run a repair of the table. @xref{Table -maintenance}. @code{Not checked} means that the table the given @code{TYPE} -told @strong{MySQL} that there wasn't any need to check the table. - -The different check types stand for the following: +@example +mysql> select HEX(255); + -> 'FF' +@end example -@multitable @columnfractions .20 .80 -@item @strong{Type} @tab @strong{Meaning} -@item @code{QUICK} @tab Don't scan the rows to check for wrong links. -@item @code{FAST} @tab Only check tables which haven't been closed properly. -@item @code{CHANGED} @tab Only check tables which have been changed since last check or haven't been closed properly. -@item @code{MEDIUM} @tab Scan rows to verify that deleted links are ok. This also calculates a key checksum for the rows and verifies this with a calcualted checksum for the keys. -@item @code{EXTENDED} @tab Do a full key lookup for all keys for each row. This ensures that the table is 100 % consistent, but will take a long time! -@end multitable +@findex CHAR() +@item CHAR(N,...) +@code{CHAR()} interprets the arguments as integers and returns a string +consisting of the characters given by the ASCII code values of those +integers. @code{NULL} values are skipped: -For dynamic sized @code{MyISAM} tables a started check will always -do a @code{MEDIUM} check. For static size rows we skip the row scan -for @code{QUICK} and @code{FAST} as the rows are very seldom corrupted. +@example +mysql> select CHAR(77,121,83,81,'76'); + -> 'MySQL' +mysql> select CHAR(77,77.3,'77.3'); + -> 'MMM' +@end example -You can combine check options as in: +@findex CONCAT() +@item CONCAT(str1,str2,...) +Returns the string that results from concatenating the arguments. Returns +@code{NULL} if any argument is @code{NULL}. May have more than 2 arguments. +A numeric argument is converted to the equivalent string form: @example -CHECK TABLE test_table FAST QUICK; +mysql> select CONCAT('My', 'S', 'QL'); + -> 'MySQL' +mysql> select CONCAT('My', NULL, 'QL'); + -> NULL +mysql> select CONCAT(14.3); + -> '14.3' @end example -Which only would do a quick check on the table if it wasn't closed properly. +@findex CONCAT_WS() +@item CONCAT_WS(separator, str1, str2,...) -@strong{NOTE:} that in some case @code{CHECK TABLE} will change the -table! This happens if the table is marked as 'corrupted' or 'not -closed properly' but @code{CHECK TABLE} didn't find any problems in the -table. In this case @code{CHECK TABLE} will mark the table as ok. +@code{CONCAT_WS()} stands for CONCAT With Separator and is a special form of +@code{CONCAT()}. The first argument is the separator for the rest of the +arguments. The separator can be a string as well as the rest of the +arguments. If the separator is @code{NULL}, the result will be @code{NULL}. +The function will skip any @code{NULL}s and empty strings, after the +separator argument. The separator will be added between the strings to be +concatenated: -If a table is corrupted, then it's most likely that the problem is in -the indexes and not in the data part. All of the above check types -checks the indexes throughly and should thus find most errors. +@example +mysql> select CONCAT_WS(",","First name","Second name","Last Name"); + -> 'First name,Second name,Last Name' +mysql> select CONCAT_WS(",","First name",NULL,"Last Name"); + -> 'First name,Last Name' +@end example -If you just want to check a table that you assume is ok, you should use -no check options or the @code{QUICK} option. The later should be used -when you are in a hurry and can take the very small risk that -@code{QUICK} didn't find an error in the data file (In most cases -@strong{MySQL} should find, under normal usage, any error in the data -file. If this happens then the table will be marked as 'corrupted', -in which case the table can't be used until it's repaired). +@findex LENGTH() +@findex OCTET_LENGTH() +@findex CHAR_LENGTH() +@findex CHARACTER_LENGTH() +@item LENGTH(str) +@itemx OCTET_LENGTH(str) +@itemx CHAR_LENGTH(str) +@itemx CHARACTER_LENGTH(str) +Returns the length of the string @code{str}: -@code{FAST} and @code{CHANGED} are mostly intended to be used from a -script (for example to be executed from cron) if you want to check your -table from time to time. In most cases you @code{FAST} is to be prefered -over @code{CHANGED}. (The only case when it isn't is when you suspect a -bug you have found a bug in the @code{MyISAM} code.). +@example +mysql> select LENGTH('text'); + -> 4 +mysql> select OCTET_LENGTH('text'); + -> 4 +@end example -@code{EXTENDED} is only to be used after you have run a normal check but -still get strange errors from a table when @strong{MySQL} tries to -update a row or find a row by key (this is VERY unlikely to happen if a -normal check has succeeded!). +Note that for @code{CHAR_LENGTH()}, multi-byte characters are only counted +once. -Some things reported by check table, can't be corrected automatically: +@findex LOCATE() +@findex POSITION() +@item LOCATE(substr,str) +@itemx POSITION(substr IN str) +Returns the position of the first occurrence of substring @code{substr} +in string @code{str}. Returns @code{0} if @code{substr} is not in @code{str}: -@itemize @bullet -@item -@code{Found row where the auto_increment column has the value 0}. +@example +mysql> select LOCATE('bar', 'foobarbar'); + -> 4 +mysql> select LOCATE('xbar', 'foobar'); + -> 0 +@end example -This means that you have in the table a row where the -@code{auto_increment} index column contains the value 0. -(It's possible to create a row where the auto_increment column is 0 by -explicitely setting the column to 0 with an @code{UPDATE} statement) +This function is multi-byte safe. -This isn't an error in itself, but could cause trouble if you decide to -dump the table and restore it or do an @code{ALTER TABLE} on the -table. In this case the auto_increment column will change value, -according to the rules of auto_increment columns, which could cause -problems like a duplicate key error. +@findex LOCATE() +@item LOCATE(substr,str,pos) +Returns the position of the first occurrence of substring @code{substr} in +string @code{str}, starting at position @code{pos}. +Returns @code{0} if @code{substr} is not in @code{str}: -To get rid of the warning, just execute an @code{UPDATE} statement -to set the column to some other value than 0. -@end itemize +@example +mysql> select LOCATE('bar', 'foobarbar',5); + -> 7 +@end example +This function is multi-byte safe. -@findex BACKUP TABLE -@node BACKUP TABLE, RESTORE TABLE, CHECK TABLE, Reference -@section @code{BACKUP TABLE} Syntax +@findex INSTR() +@item INSTR(str,substr) +Returns the position of the first occurrence of substring @code{substr} in +string @code{str}. This is the same as the two-argument form of +@code{LOCATE()}, except that the arguments are swapped: @example -BACKUP TABLE tbl_name[,tbl_name...] TO '/path/to/backup/directory' +mysql> select INSTR('foobarbar', 'bar'); + -> 4 +mysql> select INSTR('xbar', 'foobar'); + -> 0 @end example -Make a copy of all the table files to the backup directory that are the -minimum needed to restore it. Currenlty only works for @code{MyISAM} -tables. For @code{MyISAM} table, copies @code{.frm} (definition) and -@code{.MYD} (data) files. The index file can be rebuilt from those two. - -Before using this command, please see @xref{Backup}. +This function is multi-byte safe. -During the backup, read lock will be held for each table, one at time, -as they are being backed up. If you want to backup several tables as -a snapshot, you must first issue @code{LOCK TABLES} obtaining a read -lock for each table in the group. +@findex LPAD() +@item LPAD(str,len,padstr) +Returns the string @code{str}, left-padded with the string @code{padstr} +until @code{str} is @code{len} characters long. If @code{str} is longer +than @code{len'} then it will be shortened to @code{len} characters. -The command returns a table with the following columns: +@example +mysql> select LPAD('hi',4,'??'); + -> '??hi' +@end example -@multitable @columnfractions .35 .65 -@item @strong{Column} @tab @strong{Value} -@item Table @tab Table name -@item Op @tab Always ``backup'' -@item Msg_type @tab One of @code{status}, @code{error}, @code{info} or @code{warning}. -@item Msg_text @tab The message. -@end multitable +@findex RPAD() +@item RPAD(str,len,padstr) +Returns the string @code{str}, right-padded with the string +@code{padstr} until @code{str} is @code{len} characters long. If +@code{str} is longer than @code{len'} then it will be shortened to +@code{len} characters. -Note that @code{BACKUP TABLE} is only available in @strong{MySQL} -version 3.23.25 and later. +@example +mysql> select RPAD('hi',5,'?'); + -> 'hi???' +@end example -@findex RESTORE TABLE -@node RESTORE TABLE, ANALYZE TABLE, BACKUP TABLE, Reference -@section @code{RESTORE TABLE} Syntax +@findex LEFT() +@item LEFT(str,len) +Returns the leftmost @code{len} characters from the string @code{str}: @example -RESTORE TABLE tbl_name[,tbl_name...] FROM '/path/to/backup/directory' +mysql> select LEFT('foobarbar', 5); + -> 'fooba' @end example -Restores the table(s) from the backup that was made with -@code{BACKUP TABLE}. Existing tables will not be overwritten - if you -try to restore over an existing table, you will get an error. Restore -will take longer than BACKUP due to the need to rebuilt the index. The -more keys you have, the longer it is going to take. Just as -@code{BACKUP TABLE}, currently only works of @code{MyISAM} tables. - +This function is multi-byte safe. -The command returns a table with the following columns: +@findex RIGHT() +@item RIGHT(str,len) +Returns the rightmost @code{len} characters from the string @code{str}: -@multitable @columnfractions .35 .65 -@item @strong{Column} @tab @strong{Value} -@item Table @tab Table name -@item Op @tab Always ``restore'' -@item Msg_type @tab One of @code{status}, @code{error}, @code{info} or @code{warning}. -@item Msg_text @tab The message. -@end multitable +@example +mysql> select RIGHT('foobarbar', 4); + -> 'rbar' +@end example +This function is multi-byte safe. -@findex ANALYZE TABLE -@node ANALYZE TABLE, REPAIR TABLE, RESTORE TABLE, Reference -@section @code{ANALYZE TABLE} Syntax +@findex SUBSTRING() +@findex MID() +@item SUBSTRING(str,pos,len) +@itemx SUBSTRING(str FROM pos FOR len) +@itemx MID(str,pos,len) +Returns a substring @code{len} characters long from string @code{str}, +starting at position @code{pos}. +The variant form that uses @code{FROM} is ANSI SQL92 syntax: @example -ANALYZE TABLE tbl_name[,tbl_name...] +mysql> select SUBSTRING('Quadratically',5,6); + -> 'ratica' @end example -Analyze and store the key distribution for the table. During the -analyze the table is locked with a read lock. This works on -@code{MyISAM} and @code{BDB} tables. +This function is multi-byte safe. -This is equivalent to running @code{myisamchk -a} on the table. +@findex SUBSTRING() +@item SUBSTRING(str,pos) +@item SUBSTRING(str FROM pos) +Returns a substring from string @code{str} starting at position @code{pos}: -@strong{MySQL} uses the stored key distribution to decide in which order -tables should be joined when one does a join on something else than a -constant. +@example +mysql> select SUBSTRING('Quadratically',5); + -> 'ratically' +mysql> select SUBSTRING('foobarbar' FROM 4); + -> 'barbar' +@end example -The command returns a table with the following columns: +This function is multi-byte safe. -@multitable @columnfractions .35 .65 -@item @strong{Column} @tab @strong{Value} -@item Table @tab Table name -@item Op @tab Always ``analyze'' -@item Msg_type @tab One of @code{status}, @code{error}, @code{info} or @code{warning}. -@item Msg_text @tab The message. -@end multitable +@findex SUBSTRING_INDEX() +@item SUBSTRING_INDEX(str,delim,count) +Returns the substring from string @code{str} before @code{count} +occurrences of the delimiter @code{delim}. +If @code{count} is positive, everything to the left of the final delimiter +(counting from the left) is returned. +If @code{count} is negative, everything to the right of the final delimiter +(counting from the right) is returned: -You can check the stored key distribution with the @code{SHOW INDEX} command. -@xref{SHOW DATABASE INFO}. +@example +mysql> select SUBSTRING_INDEX('www.mysql.com', '.', 2); + -> 'www.mysql' +mysql> select SUBSTRING_INDEX('www.mysql.com', '.', -2); + -> 'mysql.com' +@end example -If the table hasn't changed since the last @code{ANALYZE TABLE} command, -the table will not be analyzed again. +This function is multi-byte safe. -@findex REPAIR TABLE -@node REPAIR TABLE, DELETE, ANALYZE TABLE, Reference -@section @code{REPAIR TABLE} Syntax +@findex LTRIM() +@item LTRIM(str) +Returns the string @code{str} with leading space characters removed: @example -REPAIR TABLE tbl_name[,tbl_name...] [QUICK] [EXTENDED] +mysql> select LTRIM(' barbar'); + -> 'barbar' @end example -@code{REPAIR TABLE} only works on @code{MyISAM} tables and is the same -as running @code{myisamchk -r table_name} on the table. - -Normally you should never have to run this command, but if disaster strikes -you are very likely to get back all your data from a MyISAM table with -@code{REPAIR TABLE}. If your tables get corrupted a lot you should -try to find the reason for this! @xref{Crashing}. @xref{MyISAM table problems}. +@findex RTRIM() +@item RTRIM(str) +Returns the string @code{str} with trailing space characters removed: -@code{REPAIR TABLE} repairs a possible corrupted table. The command returns a -table with the following columns: +@example +mysql> select RTRIM('barbar '); + -> 'barbar' +@end example -@multitable @columnfractions .35 .65 -@item @strong{Column} @tab @strong{Value} -@item Table @tab Table name -@item Op @tab Always ``repair'' -@item Msg_type @tab One of @code{status}, @code{error}, @code{info} or @code{warning}. -@item Msg_text @tab The message. -@end multitable +This function is multi-byte safe. -Note that you can get many rows of information for each repaired -table. The last one row will be of @code{Msg_type status} and should -normally be @code{OK}. If you don't get @code{OK}, you should try -repairing the table with @code{myisamchk -o}, as @code{REPAIR TABLE} -does not yet implement all the options of @code{myisamchk}. In the near -future, we will make it more flexible. +@findex TRIM() +@item TRIM([[BOTH | LEADING | TRAILING] [remstr] FROM] str) +Returns the string @code{str} with all @code{remstr} prefixes and/or suffixes +removed. If none of the specifiers @code{BOTH}, @code{LEADING} or +@code{TRAILING} are given, @code{BOTH} is assumed. If @code{remstr} is not +specified, spaces are removed: -If @code{QUICK} is given then @strong{MySQL} will try to do a -@code{REPAIR} of only the index tree. +@example +mysql> select TRIM(' bar '); + -> 'bar' +mysql> select TRIM(LEADING 'x' FROM 'xxxbarxxx'); + -> 'barxxx' +mysql> select TRIM(BOTH 'x' FROM 'xxxbarxxx'); + -> 'bar' +mysql> select TRIM(TRAILING 'xyz' FROM 'barxxyz'); + -> 'barx' +@end example -If you use @code{EXTENDED} then @strong{MySQL} will create the index row -by row instead of creating one index at a time with sorting; This may be -better than sorting on fixed-length keys if you have long @code{char()} -keys that compress very good. +This function is multi-byte safe. -@findex DELETE -@node DELETE, TRUNCATE, REPAIR TABLE, Reference -@section @code{DELETE} Syntax +@findex SOUNDEX() +@item SOUNDEX(str) +Returns a soundex string from @code{str}. Two strings that sound almost the +same should have identical soundex strings. A standard soundex string +is 4 characters long, but the @code{SOUNDEX()} function returns an +arbitrarily long string. You can use @code{SUBSTRING()} on the result to get +a standard soundex string. All non-alphanumeric characters are ignored +in the given string. All international alpha characters outside the A-Z range +are treated as vowels: @example -DELETE [LOW_PRIORITY] FROM tbl_name - [WHERE where_definition] - [LIMIT rows] +mysql> select SOUNDEX('Hello'); + -> 'H400' +mysql> select SOUNDEX('Quadratically'); + -> 'Q36324' @end example -@code{DELETE} deletes rows from @code{tbl_name} that satisfy the condition -given by @code{where_definition}, and returns the number of records deleted. +@findex SPACE() +@item SPACE(N) +Returns a string consisting of @code{N} space characters: -@c If an @code{ORDER BY} clause is used, the rows will be deleted in that order. -@c This is really only useful in conjunction with @code{LIMIT}. For example: +@example +mysql> select SPACE(6); + -> ' ' +@end example -@c @example -@c DELETE FROM somelog -@c WHERE user = 'jcole' -@c ORDER BY timestamp -@c LIMIT 1 -@c @end example +@findex REPLACE() +@item REPLACE(str,from_str,to_str) +Returns the string @code{str} with all all occurrences of the string +@code{from_str} replaced by the string @code{to_str}: -@c This will delete the oldest entry (by @code{timestamp}) where the row matches -@c the @code{WHERE} clause. +@example +mysql> select REPLACE('www.mysql.com', 'w', 'Ww'); + -> 'WwWwWw.mysql.com' +@end example -If you issue a @code{DELETE} with no @code{WHERE} clause, all rows are -deleted. If you do this in @code{AUTOCOMMIT} mode, this works as -@code{TRUNCATE}. @xref{TRUNCATE}. One problem with this is that -@code{DELETE} will return zero as the number of affected records, but -this will be fixed in 4.0. +This function is multi-byte safe. -If you really want to know how many records are deleted when you are deleting -all rows, and are willing to suffer a speed penalty, you can use a -@code{DELETE} statement of this form: +@findex REPEAT() +@item REPEAT(str,count) +Returns a string consisting of the string @code{str} repeated @code{count} +times. If @code{count <= 0}, returns an empty string. Returns @code{NULL} if +@code{str} or @code{count} are @code{NULL}: @example -mysql> DELETE FROM tbl_name WHERE 1>0; +mysql> select REPEAT('MySQL', 3); + -> 'MySQLMySQLMySQL' @end example -Note that this is MUCH slower than @code{DELETE FROM tbl_name} with no -@code{WHERE} clause, because it deletes rows one at a time. - -If you specify the keyword @code{LOW_PRIORITY}, execution of the -@code{DELETE} is delayed until no other clients are reading from the table. +@findex REVERSE() +@item REVERSE(str) +Returns the string @code{str} with the order of the characters reversed: -Deleted records are maintained in a linked list and subsequent @code{INSERT} -operations reuse old record positions. To reclaim unused space and reduce -file sizes, use the @code{OPTIMIZE TABLE} statement or the @code{myisamchk} -utility to reorganize tables. @code{OPTIMIZE TABLE} is easier, but -@code{myisamchk} is faster. -See @ref{OPTIMIZE TABLE, , @code{OPTIMIZE TABLE}} and @ref{Optimization}. +@example +mysql> select REVERSE('abc'); + -> 'cba' +@end example -The @strong{MySQL}-specific @code{LIMIT rows} option to @code{DELETE} tells -the server the maximum number of rows to be deleted before control is -returned to the client. This can be used to ensure that a specific -@code{DELETE} command doesn't take too much time. You can simply repeat -the @code{DELETE} command until the number of affected rows is less than -the @code{LIMIT} value. +This function is multi-byte safe. -@findex TRUNCATE -@node TRUNCATE, SELECT, DELETE, Reference -@section @code{TRUNCATE} Syntax +@findex INSERT() +@item INSERT(str,pos,len,newstr) +Returns the string @code{str}, with the substring beginning at position +@code{pos} and @code{len} characters long replaced by the string +@code{newstr}: @example -TRUNCATE TABLE table_name +mysql> select INSERT('Quadratic', 3, 4, 'What'); + -> 'QuWhattic' @end example -Is in 3.23 and the same thing as @code{DELETE FROM table_name}. @xref{DELETE}. -The differences are: - -@itemize @bullet -@item -Implemented as a drop and re-create of the table, which makes this -much faster when deleting many rows. -@item -Not transaction-safe; @code{TRUNCATE TABLE} will automatically end the current -transaction as if @code{COMMIT} would have been called. -@item -Doesn't return the number of deleted rows. -@item -As long as the table definition file @file{table_name.frm} is -valid, the table can be re-created this way, even if the data or index -files have become corrupted. -@end itemize - -@code{TRUNCATE} is an Oracle SQL extension. +This function is multi-byte safe. -@findex SELECT -@node SELECT, JOIN, TRUNCATE, Reference -@section @code{SELECT} Syntax +@findex ELT() +@item ELT(N,str1,str2,str3,...) +Returns @code{str1} if @code{N} = @code{1}, @code{str2} if @code{N} = +@code{2}, and so on. Returns @code{NULL} if @code{N} is less than @code{1} +or greater than the number of arguments. @code{ELT()} is the complement of +@code{FIELD()}: -@c help SELECT @example -SELECT [STRAIGHT_JOIN] [SQL_SMALL_RESULT] [SQL_BIG_RESULT] [SQL_BUFFER_RESULT] - [HIGH_PRIORITY] - [DISTINCT | DISTINCTROW | ALL] - select_expression,... - [INTO @{OUTFILE | DUMPFILE@} 'file_name' export_options] - [FROM table_references - [WHERE where_definition] - [GROUP BY @{unsigned_integer | col_name | formula@} [ASC | DESC], ...] - [HAVING where_definition] - [ORDER BY @{unsigned_integer | col_name | formula@} [ASC | DESC] ,...] - [LIMIT [offset,] rows] - [PROCEDURE procedure_name] - [FOR UPDATE | LOCK IN SHARE MODE]] +mysql> select ELT(1, 'ej', 'Heja', 'hej', 'foo'); + -> 'ej' +mysql> select ELT(4, 'ej', 'Heja', 'hej', 'foo'); + -> 'foo' @end example -@c help end - -@code{SELECT} is used to retrieve rows selected from one or more tables. -@code{select_expression} indicates the columns you want to retrieve. -@code{SELECT} may also be used to retrieve rows computed without reference to -any table. For example: -@example -mysql> SELECT 1 + 1; - -> 2 -@end example - -All keywords used must be given in exactly the order shown above. For example, -a @code{HAVING} clause must come after any @code{GROUP BY} clause and before -any @code{ORDER BY} clause. - -@itemize @bullet - -@item -@cindex aliases, on expressions -@cindex expression aliases -A @code{SELECT} expression may be given an alias using @code{AS}. The alias -is used as the expression's column name and can be used with -@code{ORDER BY} or @code{HAVING} clauses. For example: +@findex FIELD() +@item FIELD(str,str1,str2,str3,...) +Returns the index of @code{str} in the @code{str1}, @code{str2}, +@code{str3}, @code{...} list. +Returns @code{0} if @code{str} is not found. +@code{FIELD()} is the complement of @code{ELT()}: @example -mysql> select concat(last_name,', ',first_name) AS full_name - from mytable ORDER BY full_name; +mysql> select FIELD('ej', 'Hej', 'ej', 'Heja', 'hej', 'foo'); + -> 2 +mysql> select FIELD('fo', 'Hej', 'ej', 'Heja', 'hej', 'foo'); + -> 0 @end example -@item -The @code{FROM table_references} clause indicates the tables from which to -retrieve rows. If you name more than one table, you are performing a -join. For information on join syntax, see @ref{JOIN, , @code{JOIN}}. - -@item -You can refer to a column as @code{col_name}, @code{tbl_name.col_name}, or -@code{db_name.tbl_name.col_name}. You need not specify a @code{tbl_name} or -@code{db_name.tbl_name} prefix for a column reference in a @code{SELECT} -statement unless the reference would be ambiguous. See @ref{Legal names}, -for examples of ambiguity that require the more explicit column reference -forms. - -@item -@cindex aliases, for tables -@cindex table aliases -A table reference may be aliased using @code{tbl_name [AS] alias_name}: +@findex FIND_IN_SET() +@item FIND_IN_SET(str,strlist) +Returns a value @code{1} to @code{N} if the string @code{str} is in the list +@code{strlist} consisting of @code{N} substrings. A string list is a string +composed of substrings separated by @samp{,} characters. If the first +argument is a constant string and the second is a column of type @code{SET}, +the @code{FIND_IN_SET()} function is optimized to use bit arithmetic! +Returns @code{0} if @code{str} is not in @code{strlist} or if @code{strlist} +is the empty string. Returns @code{NULL} if either argument is @code{NULL}. +This function will not work properly if the first argument contains a +@samp{,}: @example -mysql> select t1.name, t2.salary from employee AS t1, info AS t2 - where t1.name = t2.name; -mysql> select t1.name, t2.salary from employee t1, info t2 - where t1.name = t2.name; +mysql> SELECT FIND_IN_SET('b','a,b,c,d'); + -> 2 @end example -@item -Columns selected for output may be referred to in @code{ORDER BY} and -@code{GROUP BY} clauses using column names, column aliases, or column -positions. Column positions begin with 1: +@findex MAKE_SET() +@item MAKE_SET(bits,str1,str2,...) +Returns a set (a string containing substrings separated by @samp{,} +characters) consisting of the strings that have the corresponding bit in +@code{bits} set. @code{str1} corresponds to bit 0, @code{str2} to bit 1, +etc. @code{NULL} strings in @code{str1}, @code{str2}, @code{...} +are not appended to the result: @example -mysql> select college, region, seed from tournament - ORDER BY region, seed; -mysql> select college, region AS r, seed AS s from tournament - ORDER BY r, s; -mysql> select college, region, seed from tournament - ORDER BY 2, 3; +mysql> SELECT MAKE_SET(1,'a','b','c'); + -> 'a' +mysql> SELECT MAKE_SET(1 | 4,'hello','nice','world'); + -> 'hello,world' +mysql> SELECT MAKE_SET(0,'a','b','c'); + -> '' @end example -To sort in reverse order, add the @code{DESC} (descending) keyword to the -name of the column in the @code{ORDER BY} clause that you are sorting by. -The default is ascending order; this may be specified explicitly using -the @code{ASC} keyword. - -@item -You can in the @code{WHERE} clause use any of the functions that -@strong{MySQL} support. @xref{Functions}. - -@item -The @code{HAVING} clause can refer to any column or alias named in the -@code{select_expression}. It is applied last, just before items are sent to -the client, with no optimization. Don't use @code{HAVING} for items that -should be in the @code{WHERE} clause. For example, do not write this: +@findex EXPORT_SET() +@item EXPORT_SET(bits,on,off,[separator,[number_of_bits]]) +Returns a string where for every bit set in 'bit', you get an 'on' string +and for every reset bit you get an 'off' string. Each string is separated +with 'separator' (default ',') and only 'number_of_bits' (default 64) of +'bits' is used: @example -mysql> select col_name from tbl_name HAVING col_name > 0; +mysql> select EXPORT_SET(5,'Y','N',',',4) + -> Y,N,Y,N @end example -Write this instead: +@findex LCASE() +@findex LOWER() +@item LCASE(str) +@itemx LOWER(str) +Returns the string @code{str} with all characters changed to lowercase +according to the current character set mapping (the default is ISO-8859-1 +Latin1): @example -mysql> select col_name from tbl_name WHERE col_name > 0; +mysql> select LCASE('QUADRATICALLY'); + -> 'quadratically' @end example -In @strong{MySQL} Version 3.22.5 or later, you can also write queries like this: - -@example -mysql> select user,max(salary) from users - group by user HAVING max(salary)>10; -@end example +This function is multi-byte safe. -In older @strong{MySQL} versions, you can write this instead: +@findex UCASE() +@findex UPPER() +@item UCASE(str) +@itemx UPPER(str) +Returns the string @code{str} with all characters changed to uppercase +according to the current character set mapping (the default is ISO-8859-1 +Latin1): @example -mysql> select user,max(salary) AS sum from users - group by user HAVING sum>10; +mysql> select UCASE('Hej'); + -> 'HEJ' @end example -@item -@code{SQL_SMALL_RESULT}, @code{SQL_BIG_RESULT}, @code{SQL_BUFFER_RESULT}, -@code{STRAIGHT_JOIN}, and @code{HIGH_PRIORITY} are @strong{MySQL} extensions -to ANSI SQL92. - -@item -@code{HIGH_PRIORITY} will give the @code{SELECT} higher priority than -a statement that updates a table. You should only use this for queries -that are very fast and must be done at once. A @code{SELECT HIGH_PRIORITY} -query will run if the table is locked for read even if there is an update -statement that is waiting for the table to be free. +This function is multi-byte safe. -@item -@code{SQL_BIG_RESULT} can be used with @code{GROUP BY} or @code{DISTINCT} -to tell the optimizer that the result set will have many rows. In this case, -@strong{MySQL} will directly use disk-based temporary tables if needed. -@strong{MySQL} will also, in this case, prefer sorting to doing a -temporary table with a key on the @code{GROUP BY} elements. +@findex FILE +@findex LOAD_FILE() +@item LOAD_FILE(file_name) +Reads the file and returns the file contents as a string. The file +must be on the server, you must specify the full pathname to the +file, and you must have the @strong{file} privilege. The file must +be readable by all and be smaller than @code{max_allowed_packet}. -@item -@cindex @code{GROUP BY}, extensions to ANSI SQL -If you use @code{GROUP BY}, the output rows will be sorted according to the -@code{GROUP BY} as if you would have had an @code{ORDER BY} over all the fields -in the @code{GROUP BY}. @strong{MySQL} has extended the @code{GROUP BY} so that -you can also specify @code{ASC} and @code{DESC} to @code{GROUP BY}: +If the file doesn't exist or can't be read due to one of the above reasons, +the function returns @code{NULL}: @example -SELECT a,COUNT(b) FROM test_table GROUP BY a DESC +mysql> UPDATE table_name + SET blob_column=LOAD_FILE("/tmp/picture") + WHERE id=1; @end example +@end table -@item -@strong{MySQL} has extended the use of @code{GROUP BY} to allow you to -select fields which are not mentioned in the @code{GROUP BY} clause. -If you are not getting the results you expect from your query, please -read the @code{GROUP BY} description. -@xref{Group by functions}. +If you are not using @strong{MySQL} Version 3.23, you have to do the reading +of the file inside your application and create an @code{INSERT} statement +to update the database with the file information. One way to do this, if +you are using the @strong{MySQL}++ library, can be found at +@uref{http://www.mysql.com/documentation/mysql++/mysql++-examples.html}. -@item -@code{SQL_BUFFER_RESULT} will force the result to be put into a temporary -table. This will help @strong{MySQL} free the table locks early and will help -in cases where it takes a long time to send the result set to the client. +@strong{MySQL} automatically converts numbers to strings as necessary, and +vice-versa: -@item -@code{SQL_SMALL_RESULT}, a @strong{MySQL}-specific option, can be used -with @code{GROUP BY} or @code{DISTINCT} to tell the optimizer that the -result set will be small. In this case, @strong{MySQL} will use fast -temporary tables to store the resulting table instead of using sorting. In -@strong{MySQL} Version 3.23 this shouldn't normally be needed. +@example +mysql> SELECT 1+"1"; + -> 2 +mysql> SELECT CONCAT(2,' test'); + -> '2 test' +@end example -@item -@code{STRAIGHT_JOIN} forces the optimizer to join the tables in the order in -which they are listed in the @code{FROM} clause. You can use this to speed up -a query if the optimizer joins the tables in non-optimal order. -@xref{EXPLAIN, , @code{EXPLAIN}}. +If you want to convert a number to a string explicitly, pass it as the +argument to @code{CONCAT()}. -@item -The @code{LIMIT} clause can be used to constrain the number of rows returned -by the @code{SELECT} statement. @code{LIMIT} takes one or two numeric -arguments. +If a string function is given a binary string as an argument, the resulting +string is also a binary string. A number converted to a string is treated as +a binary string. This only affects comparisons. -If two arguments are given, the first specifies the offset of the first row to -return, the second specifies the maximum number of rows to return. -The offset of the initial row is 0 (not 1): +@findex date and time functions +@findex functions, date and time +@node Date and time functions, Miscellaneous functions, String functions, Functions +@subsection Date and Time Functions -@example -mysql> select * from table LIMIT 5,10; # Retrieve rows 6-15 -@end example +See @ref{Date and time types} for a description of the range of values +each type has and the valid formats in which date and time values may be +specified. -If one argument is given, it indicates the maximum number of rows to return: +Here is an example that uses date functions. The query below selects +all records with a @code{date_col} value from within the last 30 days: @example -mysql> select * from table LIMIT 5; # Retrieve first 5 rows +mysql> SELECT something FROM table + WHERE TO_DAYS(NOW()) - TO_DAYS(date_col) <= 30; @end example -In other words, @code{LIMIT n} is equivalent to @code{LIMIT 0,n}. +@table @code +@findex DAYOFWEEK() +@item DAYOFWEEK(date) +Returns the weekday index -@item -@tindex /etc/passwd -The @code{SELECT ... INTO OUTFILE 'file_name'} form of @code{SELECT} writes -the selected rows to a file. The file is created on the server host and -cannot already exist (among other things, this prevents database tables and -files such as @file{/etc/passwd} from being destroyed). You must have the -@strong{file} privilege on the server host to use this form of @code{SELECT}. +for @code{date} (@code{1} = Sunday, @code{2} = Monday, ... @code{7} = +Saturday). These index values correspond to the ODBC standard: -@code{SELECT ... INTO OUTFILE} is mainly intended to let you very -quickly dump a table on the server machine. If you want to create the -resulting file on some other host than the server host you can't use -@code{SELECT ... INTO OUTFILE}. In this case you should instead use some -client program like @code{mysqldump --tab} or @code{mysql -e "SELECT -..." > outfile} to generate the file. +@example +mysql> select DAYOFWEEK('1998-02-03'); + -> 3 +@end example -@code{SELECT ... INTO OUTFILE} is the complement of @code{LOAD DATA -INFILE}; the syntax for the @code{export_options} part of the statement -consists of the same @code{FIELDS} and @code{LINES} clauses that are used -with the @code{LOAD DATA INFILE} statement. -@xref{LOAD DATA, , @code{LOAD DATA}}. +@findex WEEKDAY() +@item WEEKDAY(date) +Returns the weekday index for +@code{date} (@code{0} = Monday, @code{1} = Tuesday, ... @code{6} = Sunday): -In the resulting text file, only the following characters are escaped by -the @code{ESCAPED BY} character: +@example +mysql> select WEEKDAY('1997-10-04 22:23:00'); + -> 5 +mysql> select WEEKDAY('1997-11-05'); + -> 2 +@end example -@itemize @bullet -@item The @code{ESCAPED BY} character -@item The first character in @code{FIELDS TERMINATED BY} -@item The first character in @code{LINES TERMINATED BY} -@end itemize +@findex DAYOFMONTH() +@item DAYOFMONTH(date) +Returns the day of the month for @code{date}, in the range @code{1} to +@code{31}: -Additionally, @code{ASCII 0} is converted to @code{ESCAPED BY} followed by 0 -(@code{ASCII 48}). +@example +mysql> select DAYOFMONTH('1998-02-03'); + -> 3 +@end example -The reason for the above is that you MUST escape any @code{FIELDS -TERMINATED BY}, @code{ESCAPED BY}, or @code{LINES TERMINATED BY} -characters to reliably be able to read the file back. @code{ASCII 0} is -escaped to make it easier to view with some pagers. +@findex DAYOFYEAR() +@item DAYOFYEAR(date) +Returns the day of the year for @code{date}, in the range @code{1} to +@code{366}: -As the resulting file doesn't have to conform to the SQL syntax, nothing -else need be escaped. +@example +mysql> select DAYOFYEAR('1998-02-03'); + -> 34 +@end example -Here follows an example of getting a file in the format used by many -old programs. +@findex MONTH() +@item MONTH(date) +Returns the month for @code{date}, in the range @code{1} to @code{12}: @example -SELECT a,b,a+b INTO OUTFILE "/tmp/result.text" -FIELDS TERMINATED BY ',' OPTIONALLY ENCLOSED BY '"' -LINES TERMINATED BY "\n" -FROM test_table; +mysql> select MONTH('1998-02-03'); + -> 2 @end example -@item -@findex DUMPFILE -If you use @code{INTO DUMPFILE} instead of @code{INTO OUTFILE}, @strong{MySQL} -will only write one row into the file, without any column or line -terminations and without any escaping. This is useful if you want to -store a blob in a file. +@findex DAYNAME() +@item DAYNAME(date) +Returns the name of the weekday for @code{date}: -@item -Note that any file created by @code{INTO OUTFILE} and @code{INTO -DUMPFILE} is going to be readable for all users! The reason is that the -@strong{MySQL} server can't create a file that is owned by anyone else -than the user it's running as (you should never run @code{mysqld} as root), -the file has to be word readable so that you can retrieve the rows. +@example +mysql> select DAYNAME("1998-02-05"); + -> 'Thursday' +@end example -@item -If you are using @code{FOR UPDATE} on a table handler with page/row locks, -the examined rows will be write locked. -@end itemize +@findex MONTHNAME() +@item MONTHNAME(date) +Returns the name of the month for @code{date}: -@findex JOIN -@findex INNER JOIN -@findex CROSS JOIN -@findex LEFT JOIN -@findex LEFT OUTER JOIN -@findex NATURAL LEFT JOIN -@findex NATURAL LEFT OUTER JOIN -@findex RIGHT JOIN -@findex RIGHT OUTER JOIN -@findex NATURAL RIGHT JOIN -@findex NATURAL RIGHT OUTER JOIN -@findex STRAIGHT_JOIN -@node JOIN, INSERT, SELECT, Reference -@section @code{JOIN} Syntax +@example +mysql> select MONTHNAME("1998-02-05"); + -> 'February' +@end example -@strong{MySQL} supports the following @code{JOIN} syntaxes for use in -@code{SELECT} statements: +@findex QUARTER() +@item QUARTER(date) +Returns the quarter of the year for @code{date}, in the range @code{1} +to @code{4}: @example -table_reference, table_reference -table_reference [CROSS] JOIN table_reference -table_reference INNER JOIN table_reference join_condition -table_reference STRAIGHT_JOIN table_reference -table_reference LEFT [OUTER] JOIN table_reference join_condition -table_reference LEFT [OUTER] JOIN table_reference -table_reference NATURAL [LEFT [OUTER]] JOIN table_reference -@{ oj table_reference LEFT OUTER JOIN table_reference ON conditional_expr @} -table_reference RIGHT [OUTER] JOIN table_reference join_condition -table_reference RIGHT [OUTER] JOIN table_reference -table_reference NATURAL [RIGHT [OUTER]] JOIN table_reference +mysql> select QUARTER('98-04-01'); + -> 2 @end example -Where @code{table_reference} is defined as: +@findex WEEK() +@item WEEK(date) +@itemx WEEK(date,first) +With a single argument, returns the week for @code{date}, in the range +@code{0} to @code{53} (yes, there may be the beginnings of a week 53), +for locations where Sunday is the first day of the week. The +two-argument form of @code{WEEK()} allows you to specify whether the +week starts on Sunday or Monday. The week starts on Sunday if the +second argument is @code{0}, on Monday if the second argument is +@code{1}: + @example -table_name [[AS] alias] [USE INDEX (key_list)] [IGNORE INDEX (key_list)] +mysql> select WEEK('1998-02-20'); + -> 7 +mysql> select WEEK('1998-02-20',0); + -> 7 +mysql> select WEEK('1998-02-20',1); + -> 8 +mysql> select WEEK('1998-12-31',1); + -> 53 @end example -and @code{join_condition} is defined as: +@findex YEAR() +@item YEAR(date) +Returns the year for @code{date}, in the range @code{1000} to @code{9999}: @example -ON conditional_expr | -USING (column_list) +mysql> select YEAR('98-02-03'); + -> 1998 @end example -Note that in versions before Version 3.23.16, the @code{INNER JOIN} didn't take -a join condition! +@item YEARWEEK(date) +@itemx YEARWEEK(date,first) +Returns year and week for a date. The second arguments works exactly +like the second argument to @code{WEEK()}. Note that the year may be +different from the year in the date argument for the first and the last +week of the year: -@cindex ODBC compatibility -@cindex compatibility, with ODBC -The last @code{LEFT OUTER JOIN} syntax shown above exists only for -compatibility with ODBC: +@example +mysql> select YEARWEEK('1987-01-01'); + -> 198653 +@end example -@itemize @bullet -@item -A table reference may be aliased using @code{tbl_name AS alias_name} or -@code{tbl_name alias_name}: +@findex HOUR() +@item HOUR(time) +Returns the hour for @code{time}, in the range @code{0} to @code{23}: @example -mysql> select t1.name, t2.salary from employee AS t1, info AS t2 - where t1.name = t2.name; +mysql> select HOUR('10:05:03'); + -> 10 @end example -@item -@code{INNER JOIN} and @code{,} (comma) are semantically equivalent. -Both do a full join between the tables used. Normally, you specify how -the tables should be linked in the @code{WHERE} condition. +@findex MINUTE() +@item MINUTE(time) +Returns the minute for @code{time}, in the range @code{0} to @code{59}: -@item -The @code{ON} conditional is any conditional of the form that may be used in -a @code{WHERE} clause. +@example +mysql> select MINUTE('98-02-03 10:05:03'); + -> 5 +@end example -@item -If there is no matching record for the right table in the @code{ON} or -@code{USING} part in a @code{LEFT JOIN}, a row with all columns set to -@code{NULL} is used for the right table. You can use this fact to find -records in a table that have no counterpart in another table: +@findex SECOND() +@item SECOND(time) +Returns the second for @code{time}, in the range @code{0} to @code{59}: @example -mysql> select table1.* from table1 - LEFT JOIN table2 ON table1.id=table2.id - where table2.id is NULL; +mysql> select SECOND('10:05:03'); + -> 3 @end example -This example finds all rows in @code{table1} with an @code{id} value that is -not present in @code{table2} (that is, all rows in @code{table1} with no -corresponding row in @code{table2}). This assumes that @code{table2.id} is -declared @code{NOT NULL}, of course. @xref{LEFT JOIN optimization}. +@findex PERIOD_ADD() +@item PERIOD_ADD(P,N) +Adds @code{N} months to period @code{P} (in the format @code{YYMM} or +@code{YYYYMM}). Returns a value in the format @code{YYYYMM}. -@item -The @code{USING} @code{(column_list)} clause names a list of columns that must -exist in both tables. A @code{USING} clause such as: +Note that the period argument @code{P} is @emph{not} a date value: @example -A LEFT JOIN B USING (C1,C2,C3,...) +mysql> select PERIOD_ADD(9801,2); + -> 199803 @end example -is defined to be semantically identical to an @code{ON} expression like -this: +@findex PERIOD_DIFF() +@item PERIOD_DIFF(P1,P2) +Returns the number of months between periods @code{P1} and @code{P2}. +@code{P1} and @code{P2} should be in the format @code{YYMM} or @code{YYYYMM}. + +Note that the period arguments @code{P1} and @code{P2} are @emph{not} +date values: @example -A.C1=B.C1 AND A.C2=B.C2 AND A.C3=B.C3,... +mysql> select PERIOD_DIFF(9802,199703); + -> 11 @end example -@item -The @code{NATURAL [LEFT] JOIN} of two tables is defined to be -semantically equivalent to an @code{INNER JOIN} or a @code{LEFT JOIN} -with a @code{USING} clause that names all columns that exist in both -tables. +@findex DATE_ADD() +@findex DATE_SUB() +@findex ADDDATE() +@findex SUBDATE() +@findex EXTRACT(type FROM date) +@item DATE_ADD(date,INTERVAL expr type) +@itemx DATE_SUB(date,INTERVAL expr type) +@itemx ADDDATE(date,INTERVAL expr type) +@itemx SUBDATE(date,INTERVAL expr type) -@item -@code{RIGHT JOIN} works analogously as @code{LEFT JOIN}. To keep code -portable across databases, it's recommended to use @code{LEFT JOIN} -instead of @code{RIGHT JOIN}. +These functions perform date arithmetic. They are new for @strong{MySQL} +Version 3.22. @code{ADDDATE()} and @code{SUBDATE()} are synonyms for +@code{DATE_ADD()} and @code{DATE_SUB()}. -@item -@code{STRAIGHT_JOIN} is identical to @code{JOIN}, except that the left table -is always read before the right table. This can be used for those (few) -cases where the join optimizer puts the tables in the wrong order. +In @strong{MySQL} Version 3.23, you can use @code{+} and @code{-} instead of +@code{DATE_ADD()} and @code{DATE_SUB()} if the expression on the right side is +a date or datetime column. (See example) -@item -As of @strong{MySQL} Version 3.23.12, you can give hints about which -index @strong{MySQL} should use when retrieving information from a -table. This is useful if @code{EXPLAIN} shows that @strong{MySQL} is -using the wrong index. By specifying @code{USE INDEX (key_list)}, you -can tell @strong{MySQL} to use only one of the specified indexes to -find rows in the table. The alternative syntax @code{IGNORE INDEX -(key_list)} can be used to tell @strong{MySQL} to not use some -particular index. -@end itemize +@code{date} is a @code{DATETIME} or @code{DATE} value specifying the starting +date. @code{expr} is an expression specifying the interval value to be added +or subtracted from the starting date. @code{expr} is a string; it may start +with a @samp{-} for negative intervals. @code{type} is a keyword indicating +how the expression should be interpreted. -Some examples: +The @code{EXTRACT(type FROM date)} function returns the 'type' +interval from the date. + +The following table shows how the @code{type} and @code{expr} arguments +are related: + +@multitable @columnfractions .5 .5 +@item @code{type} @strong{value} @tab @strong{Expected} @code{expr} @strong{format} +@item @code{SECOND} @tab @code{SECONDS} +@item @code{MINUTE} @tab @code{MINUTES} +@item @code{HOUR} @tab @code{HOURS} +@item @code{DAY} @tab @code{DAYS} +@item @code{MONTH} @tab @code{MONTHS} +@item @code{YEAR} @tab @code{YEARS} +@item @code{MINUTE_SECOND} @tab @code{"MINUTES:SECONDS"} +@item @code{HOUR_MINUTE} @tab @code{"HOURS:MINUTES"} +@item @code{DAY_HOUR} @tab @code{"DAYS HOURS"} +@item @code{YEAR_MONTH} @tab @code{"YEARS-MONTHS"} +@item @code{HOUR_SECOND} @tab @code{"HOURS:MINUTES:SECONDS"} +@item @code{DAY_MINUTE} @tab @code{"DAYS HOURS:MINUTES"} +@item @code{DAY_SECOND} @tab @code{"DAYS HOURS:MINUTES:SECONDS"} +@end multitable + +@strong{MySQL} allows any punctuation delimiter in the @code{expr} format. +Those shown in the table are the suggested delimiters. If the @code{date} +argument is a @code{DATE} value and your calculations involve only +@code{YEAR}, @code{MONTH}, and @code{DAY} parts (that is, no time parts), the +result is a @code{DATE} value. Otherwise the result is a @code{DATETIME} +value: @example -mysql> select * from table1,table2 where table1.id=table2.id; -mysql> select * from table1 LEFT JOIN table2 ON table1.id=table2.id; -mysql> select * from table1 LEFT JOIN table2 USING (id); -mysql> select * from table1 LEFT JOIN table2 ON table1.id=table2.id - LEFT JOIN table3 ON table2.id=table3.id; -mysql> select * from table1 USE INDEX (key1,key2) WHERE key1=1 and key2=2 AND - key3=3; -mysql> select * from table1 IGNORE INDEX (key3) WHERE key1=1 and key2=2 AND - key3=3; +mysql> SELECT "1997-12-31 23:59:59" + INTERVAL 1 SECOND; + -> 1998-01-01 00:00:00 +mysql> SELECT INTERVAL 1 DAY + "1997-12-31"; + -> 1998-01-01 +mysql> SELECT "1998-01-01" - INTERVAL 1 SECOND; + -> 1997-12-31 23:59:59 +mysql> SELECT DATE_ADD("1997-12-31 23:59:59", + INTERVAL 1 SECOND); + -> 1998-01-01 00:00:00 +mysql> SELECT DATE_ADD("1997-12-31 23:59:59", + INTERVAL 1 DAY); + -> 1998-01-01 23:59:59 +mysql> SELECT DATE_ADD("1997-12-31 23:59:59", + INTERVAL "1:1" MINUTE_SECOND); + -> 1998-01-01 00:01:00 +mysql> SELECT DATE_SUB("1998-01-01 00:00:00", + INTERVAL "1 1:1:1" DAY_SECOND); + -> 1997-12-30 22:58:59 +mysql> SELECT DATE_ADD("1998-01-01 00:00:00", + INTERVAL "-1 10" DAY_HOUR); + -> 1997-12-30 14:00:00 +mysql> SELECT DATE_SUB("1998-01-02", INTERVAL 31 DAY); + -> 1997-12-02 +mysql> SELECT EXTRACT(YEAR FROM "1999-07-02"); + -> 1999 +mysql> SELECT EXTRACT(YEAR_MONTH FROM "1999-07-02 01:02:03"); + -> 199907 +mysql> SELECT EXTRACT(DAY_MINUTE FROM "1999-07-02 01:02:03"); + -> 20102 @end example -@xref{LEFT JOIN optimization, , @code{LEFT JOIN} optimization}. +If you specify an interval value that is too short (does not include all the +interval parts that would be expected from the @code{type} keyword), +@strong{MySQL} assumes you have left out the leftmost parts of the interval +value. For example, if you specify a @code{type} of @code{DAY_SECOND}, the +value of @code{expr} is expected to have days, hours, minutes, and seconds +parts. If you specify a value like @code{"1:10"}, @strong{MySQL} assumes +that the days and hours parts are missing and the value represents minutes +and seconds. In other words, @code{"1:10" DAY_SECOND} is interpreted in such +a way that it is equivalent to @code{"1:10" MINUTE_SECOND}. This is +analogous to the way that @strong{MySQL} interprets @code{TIME} values +as representing elapsed time rather than as time of day. -@findex INSERT -@node INSERT, REPLACE, JOIN, Reference -@section @code{INSERT} Syntax +Note that if you add or subtract a date value against something that +contains a time part, the date value will be automatically converted to a +datetime value: @example - INSERT [LOW_PRIORITY | DELAYED] [IGNORE] - [INTO] tbl_name [(col_name,...)] - VALUES (expression,...),(...),... -or INSERT [LOW_PRIORITY | DELAYED] [IGNORE] - [INTO] tbl_name [(col_name,...)] - SELECT ... -or INSERT [LOW_PRIORITY | DELAYED] [IGNORE] - [INTO] tbl_name - SET col_name=expression, col_name=expression, ... +mysql> select date_add("1999-01-01", interval 1 day); + -> 1999-01-02 +mysql> select date_add("1999-01-01", interval 1 hour); + -> 1999-01-01 01:00:00 +@end example + +If you use really incorrect dates, the result is @code{NULL}. If you add +@code{MONTH}, @code{YEAR_MONTH}, or @code{YEAR} and the resulting date +has a day that is larger than the maximum day for the new month, the day is +adjusted to the maximum days in the new month: +@example +mysql> select DATE_ADD('1998-01-30', Interval 1 month); + -> 1998-02-28 @end example -@code{INSERT} inserts new rows into an existing table. The @code{INSERT -... VALUES} form of the statement inserts rows based on explicitly -specified values. The @code{INSERT ... SELECT} form inserts rows -selected from another table or tables. The @code{INSERT ... VALUES} -form with multiple value lists is supported in @strong{MySQL} Version -3.22.5 or later. The @code{col_name=expression} syntax is supported in -@strong{MySQL} Version 3.22.10 or later. +Note from the preceding example that the word @code{INTERVAL} and the +@code{type} keyword are not case sensitive. -@code{tbl_name} is the table into which rows should be inserted. The column -name list or the @code{SET} clause indicates which columns the statement -specifies values for: +@findex TO_DAYS() +@item TO_DAYS(date) +Given a date @code{date}, returns a daynumber (the number of days since year +0): -@itemize @bullet -@item -If you specify no column list for @code{INSERT ... VALUES} or @code{INSERT -... SELECT}, values for all columns must be provided in the -@code{VALUES()} list or by the @code{SELECT}. If you don't know the order of -the columns in the table, use @code{DESCRIBE tbl_name} to find out. +@example +mysql> select TO_DAYS(950501); + -> 728779 +mysql> select TO_DAYS('1997-10-07'); + -> 729669 +@end example -@item -Any column not explicitly given a value is set to its default value. For -example, if you specify a column list that doesn't name all the columns in -the table, unnamed columns are set to their default values. Default value -assignment is described in @ref{CREATE TABLE, , @code{CREATE TABLE}}. +@code{TO_DAYS()} is not intended for use with values that precede the advent +of the Gregorian calendar (1582), because it doesn't take into account the +days that were lost when the calendar was changed. -@item -An @code{expression} may refer to any column that was set earlier in a value -list. For example, you can say this: +@findex FROM_DAYS() +@item FROM_DAYS(N) +Given a daynumber @code{N}, returns a @code{DATE} value: @example -mysql> INSERT INTO tbl_name (col1,col2) VALUES(15,col1*2); +mysql> select FROM_DAYS(729669); + -> '1997-10-07' @end example -But not this: +@code{FROM_DAYS()} is not intended for use with values that precede the +advent of the Gregorian calendar (1582), because it doesn't take into account +the days that were lost when the calendar was changed. + +@findex DATE_FORMAT() +@item DATE_FORMAT(date,format) +Formats the @code{date} value according to the @code{format} string. The +following specifiers may be used in the @code{format} string: +@multitable @columnfractions .1 .6 +@item @code{%M} @tab Month name (@code{January}..@code{December}) +@item @code{%W} @tab Weekday name (@code{Sunday}..@code{Saturday}) +@item @code{%D} @tab Day of the month with English suffix (@code{1st}, @code{2nd}, @code{3rd}, etc.) +@item @code{%Y} @tab Year, numeric, 4 digits +@item @code{%y} @tab Year, numeric, 2 digits +@item @code{%X} @tab Year for the week where Sunday is the first day of the week, numeric, 4 digits, used with '%V' +@item @code{%x} @tab Year for the week, where Monday is the first day of the week, numeric, 4 digits, used with '%v' +@item @code{%a} @tab Abbreviated weekday name (@code{Sun}..@code{Sat}) +@item @code{%d} @tab Day of the month, numeric (@code{00}..@code{31}) +@item @code{%e} @tab Day of the month, numeric (@code{0}..@code{31}) +@item @code{%m} @tab Month, numeric (@code{01}..@code{12}) +@item @code{%c} @tab Month, numeric (@code{1}..@code{12}) +@item @code{%b} @tab Abbreviated month name (@code{Jan}..@code{Dec}) +@item @code{%j} @tab Day of year (@code{001}..@code{366}) +@item @code{%H} @tab Hour (@code{00}..@code{23}) +@item @code{%k} @tab Hour (@code{0}..@code{23}) +@item @code{%h} @tab Hour (@code{01}..@code{12}) +@item @code{%I} @tab Hour (@code{01}..@code{12}) +@item @code{%l} @tab Hour (@code{1}..@code{12}) +@item @code{%i} @tab Minutes, numeric (@code{00}..@code{59}) +@item @code{%r} @tab Time, 12-hour (@code{hh:mm:ss [AP]M}) +@item @code{%T} @tab Time, 24-hour (@code{hh:mm:ss}) +@item @code{%S} @tab Seconds (@code{00}..@code{59}) +@item @code{%s} @tab Seconds (@code{00}..@code{59}) +@item @code{%p} @tab @code{AM} or @code{PM} +@item @code{%w} @tab Day of the week (@code{0}=Sunday..@code{6}=Saturday) +@item @code{%U} @tab Week (@code{0}..@code{53}), where Sunday is the first day of the week +@item @code{%u} @tab Week (@code{0}..@code{53}), where Monday is the first day of the week +@item @code{%V} @tab Week (@code{1}..@code{53}), where Sunday is the first day of the week. Used with '%X' +@item @code{%v} @tab Week (@code{1}..@code{53}), where Monday is the first day of the week. Used with '%x' +@item @code{%%} @tab A literal @samp{%}. +@end multitable + +All other characters are just copied to the result without interpretation: @example -mysql> INSERT INTO tbl_name (col1,col2) VALUES(col2*2,15); +mysql> select DATE_FORMAT('1997-10-04 22:23:00', '%W %M %Y'); + -> 'Saturday October 1997' +mysql> select DATE_FORMAT('1997-10-04 22:23:00', '%H:%i:%s'); + -> '22:23:00' +mysql> select DATE_FORMAT('1997-10-04 22:23:00', + '%D %y %a %d %m %b %j'); + -> '4th 97 Sat 04 10 Oct 277' +mysql> select DATE_FORMAT('1997-10-04 22:23:00', + '%H %k %I %r %T %S %w'); + -> '22 22 10 10:23:00 PM 22:23:00 00 6' +mysql> select DATE_FORMAT('1999-01-01', '%X %V'); + -> '1998 52' @end example -@item -If you specify the keyword @code{LOW_PRIORITY}, execution of the -@code{INSERT} is delayed until no other clients are reading from the -table. In this case the client has to wait until the insert statement -is completed, which may take a long time if the table is in heavy -use. This is in contrast to @code{INSERT DELAYED}, which lets the client -continue at once. @xref{INSERT DELAYED}. Note that @code{LOW_PRIORITY} -should normally not be used with @code{MyISAM} tables as this disables -concurrent inserts. @xref{MyISAM}. +As of @strong{MySQL} Version 3.23, the @samp{%} character is required before +format specifier characters. In earlier versions of @strong{MySQL}, +@samp{%} was optional. -@item -If you specify the keyword @code{IGNORE} in an @code{INSERT} with many value -rows, any rows that duplicate an existing @code{PRIMARY} or @code{UNIQUE} -key in the table are ignored and are not inserted. If you do not specify -@code{IGNORE}, the insert is aborted if there is any row that duplicates an -existing key value. You can determine with the C API function -@code{mysql_info()} how many rows were inserted into the table. +@findex TIME_FORMAT() +@item TIME_FORMAT(time,format) +This is used like the @code{DATE_FORMAT()} function above, but the +@code{format} string may contain only those format specifiers that handle +hours, minutes, and seconds. Other specifiers produce a @code{NULL} value or +@code{0}. -@item -If @strong{MySQL} was configured using the @code{DONT_USE_DEFAULT_FIELDS} -option, @code{INSERT} statements generate an error unless you explicitly -specify values for all columns that require a non-@code{NULL} value. -@xref{configure options, , @code{configure} options}. +@findex CURDATE() +@findex CURRENT_DATE +@item CURDATE() +@itemx CURRENT_DATE +Returns today's date as a value in @code{'YYYY-MM-DD'} or @code{YYYYMMDD} +format, depending on whether the function is used in a string or numeric +context: -@item -You can find the value used for an @code{AUTO_INCREMENT} column -with the @code{mysql_insert_id} function. -@xref{mysql_insert_id, , @code{mysql_insert_id()}}. -@end itemize +@example +mysql> select CURDATE(); + -> '1997-12-15' +mysql> select CURDATE() + 0; + -> 19971215 +@end example -@findex mysql_info() -If you use @code{INSERT ... SELECT} or an @code{INSERT ... VALUES} -statement with multiple value lists, you can use the C API function -@code{mysql_info()} to get information about the query. The format of the -information string is shown below: +@findex CURTIME() +@findex CURRENT_TIME +@item CURTIME() +@itemx CURRENT_TIME +Returns the current time as a value in @code{'HH:MM:SS'} or @code{HHMMSS} +format, depending on whether the function is used in a string or numeric +context: @example -Records: 100 Duplicates: 0 Warnings: 0 +mysql> select CURTIME(); + -> '23:50:26' +mysql> select CURTIME() + 0; + -> 235026 @end example -@code{Duplicates} indicates the number of rows that couldn't be inserted -because they would duplicate some existing unique index value. -@code{Warnings} indicates the number of attempts to insert column values that -were problematic in some way. Warnings can occur under any of the following -conditions: +@findex NOW() +@findex SYSDATE() +@findex CURRENT_TIMESTAMP +@item NOW() +@itemx SYSDATE() +@itemx CURRENT_TIMESTAMP +Returns the current date and time as a value in @code{'YYYY-MM-DD HH:MM:SS'} +or @code{YYYYMMDDHHMMSS} format, depending on whether the function is used in +a string or numeric context: -@itemize @bullet -@item -Inserting @code{NULL} into a column that has been declared @code{NOT NULL}. -The column is set to its default value. +@example +mysql> select NOW(); + -> '1997-12-15 23:50:26' +mysql> select NOW() + 0; + -> 19971215235026 +@end example -@item -Setting a numeric column to a value that lies outside the column's range. -The value is clipped to the appropriate endpoint of the range. +@findex UNIX_TIMESTAMP() +@item UNIX_TIMESTAMP() +@itemx UNIX_TIMESTAMP(date) +If called with no argument, returns a Unix timestamp (seconds since +@code{'1970-01-01 00:00:00'} GMT). If @code{UNIX_TIMESTAMP()} is called with +a @code{date} argument, it returns the value of the argument as seconds since +@code{'1970-01-01 00:00:00'} GMT. @code{date} may be a @code{DATE} string, +a @code{DATETIME} string, a @code{TIMESTAMP}, or a number in the format +@code{YYMMDD} or @code{YYYYMMDD} in local time: -@item -Setting a numeric column to a value such as @code{'10.34 a'}. The trailing -garbage is stripped and the remaining numeric part is inserted. If the value -doesn't make sense as a number at all, the column is set to @code{0}. +@example +mysql> select UNIX_TIMESTAMP(); + -> 882226357 +mysql> select UNIX_TIMESTAMP('1997-10-04 22:23:00'); + -> 875996580 +@end example -@item -Inserting a string into a @code{CHAR}, @code{VARCHAR}, @code{TEXT}, or -@code{BLOB} column that exceeds the column's maximum length. The value is -truncated to the column's maximum length. +When @code{UNIX_TIMESTAMP} is used on a @code{TIMESTAMP} column, the function +will receive the value directly, with no implicit +``string-to-unix-timestamp'' conversion. +If you give @code{UNIX_TIMESTAMP()} a wrong or out-of-range date, it will +return 0. -@item -Inserting a value into a date or time column that is illegal for the column -type. The column is set to the appropriate zero value for the type. -@end itemize +@findex FROM_UNIXTIME() +@item FROM_UNIXTIME(unix_timestamp) +Returns a representation of the @code{unix_timestamp} argument as a value in +@code{'YYYY-MM-DD HH:MM:SS'} or @code{YYYYMMDDHHMMSS} format, depending on +whether the function is used in a string or numeric context: -@findex REPLACE ... SELECT -@findex INSERT ... SELECT -@menu -* INSERT SELECT:: -* INSERT DELAYED:: -@end menu +@example +mysql> select FROM_UNIXTIME(875996580); + -> '1997-10-04 22:23:00' +mysql> select FROM_UNIXTIME(875996580) + 0; + -> 19971004222300 +@end example -@node INSERT SELECT, INSERT DELAYED, INSERT, INSERT -@subsection @code{INSERT ... SELECT} Syntax +@findex FROM_UNIXTIME() +@item FROM_UNIXTIME(unix_timestamp,format) +Returns a string representation of the Unix timestamp, formatted according to +the @code{format} string. @code{format} may contain the same specifiers as +those listed in the entry for the @code{DATE_FORMAT()} function: @example -INSERT [LOW_PRIORITY] [IGNORE] [INTO] tbl_name [(column list)] SELECT ... +mysql> select FROM_UNIXTIME(UNIX_TIMESTAMP(), + '%Y %D %M %h:%i:%s %x'); + -> '1997 23rd December 03:43:30 x' @end example -With @code{INSERT ... SELECT} statement you can quickly insert many rows -into a table from one or many tables. +@findex SEC_TO_TIME() +@item SEC_TO_TIME(seconds) +Returns the @code{seconds} argument, converted to hours, minutes, and seconds, +as a value in @code{'HH:MM:SS'} or @code{HHMMSS} format, depending on whether +the function is used in a string or numeric context: @example -INSERT INTO tblTemp2 (fldID) SELECT tblTemp1.fldOrder_ID FROM tblTemp1 WHERE -tblTemp1.fldOrder_ID > 100; +mysql> select SEC_TO_TIME(2378); + -> '00:39:38' +mysql> select SEC_TO_TIME(2378) + 0; + -> 3938 @end example -The following conditions hold for an @code{INSERT ... SELECT} statement: - -@itemize @minus -@item -The query cannot contain an @code{ORDER BY} clause. +@findex TIME_TO_SEC() +@item TIME_TO_SEC(time) +Returns the @code{time} argument, converted to seconds: -@item -The target table of the @code{INSERT} statement cannot appear in the -@code{FROM} clause of the @code{SELECT} part of the query because it's -forbidden in ANSI SQL to @code{SELECT} from the same table into which you are -inserting. (The problem is that the @code{SELECT} possibly would -find records that were inserted earlier during the same run. When using -sub-select clauses, the situation could easily be very confusing!) +@example +mysql> select TIME_TO_SEC('22:23:00'); + -> 80580 +mysql> select TIME_TO_SEC('00:39:38'); + -> 2378 +@end example +@end table -@item -@code{AUTO_INCREMENT} columns work as usual. +@findex miscellaneous functions +@findex functions, miscellaneous +@node Miscellaneous functions, Group by functions, Date and time functions, Functions +@subsection Miscellaneous Functions -@item -You can use the C API function @code{mysql_info()} to get information about -the query. @xref{INSERT}. +@table @code +@findex DATABASE() +@item DATABASE() +Returns the current database name: -@item -To ensure that the update log/binary log can be used to re-create the -original tables, @strong{MySQL} will not allow concurrent inserts during -@code{INSERT .... SELECT}. -@end itemize +@example +mysql> select DATABASE(); + -> 'test' +@end example -You can of course also use @code{REPLACE} instead of @code{INSERT} to -overwrite old rows. +If there is no current database, @code{DATABASE()} returns the empty string. -@findex INSERT DELAYED -@findex DELAYED -@cindex INSERT DELAYED -@node INSERT DELAYED, , INSERT SELECT, INSERT -@subsection @code{INSERT DELAYED} syntax +@findex USER() +@findex SYSTEM_USER() +@findex SESSION_USER() +@item USER() +@itemx SYSTEM_USER() +@itemx SESSION_USER() +Returns the current @strong{MySQL} user name: @example -INSERT DELAYED ... +mysql> select USER(); + -> 'davida@@localhost' @end example -The @code{DELAYED} option for the @code{INSERT} statement is a -@strong{MySQL}-specific option that is very useful if you have clients -that can't wait for the @code{INSERT} to complete. This is a common -problem when you use @strong{MySQL} for logging and you also -periodically run @code{SELECT} and @code{UPDATE} statements that take a -long time to complete. @code{DELAYED} was introduced in @strong{MySQL} -Version 3.22.15. It is a @strong{MySQL} extension to ANSI SQL92. +In @strong{MySQL} Version 3.22.11 or later, this includes the client hostname +as well as the user name. You can extract just the user name part like this +(which works whether or not the value includes a hostname part): -@code{INSERT DELAYED} only works with @code{ISAM} and @code{MyISAM} -tables. Note that as @code{MyISAM} tables supports concurrent -@code{SELECT} and @code{INSERT}, if there is no free blocks in the -middle of the data file, you very seldom need to use @code{INSERT -DELAYED} with @code{MyISAM}. @xref{MyISAM}. +@example +mysql> select substring_index(USER(),"@@",1); + -> 'davida' +@end example -When you use @code{INSERT DELAYED}, the client will get an OK at once -and the row will be inserted when the table is not in use by any other thread. +@findex PASSWORD() +@item PASSWORD(str) +Calculates a password string from the plaintext password @code{str}. This is +the function that is used for encrypting @strong{MySQL} passwords for storage +in the @code{Password} column of the @code{user} grant table: -Another major benefit of using @code{INSERT DELAYED} is that inserts -from many clients are bundled together and written in one block. This is much -faster than doing many separate inserts. +@example +mysql> select PASSWORD('badpwd'); + -> '7f84554057dd964b' +@end example -Note that currently the queued rows are only stored in memory until they are -inserted into the table. This means that if you kill @code{mysqld} -the hard way (@code{kill -9}) or if @code{mysqld} dies unexpectedly, any -queued rows that weren't written to disk are lost! +@cindex password encryption, reversibility of +@code{PASSWORD()} encryption is non-reversible. -The following describes in detail what happens when you use the -@code{DELAYED} option to @code{INSERT} or @code{REPLACE}. In this -description, the ``thread'' is the thread that received an @code{INSERT -DELAYED} command and ``handler'' is the thread that handles all -@code{INSERT DELAYED} statements for a particular table. +@code{PASSWORD()} does not perform password encryption in the same way that +Unix passwords are encrypted. You should not assume that if your Unix +password and your @strong{MySQL} password are the same, @code{PASSWORD()} +will result in the same encrypted value as is stored in the Unix password +file. See @code{ENCRYPT()}. -@itemize @bullet -@item -When a thread executes a @code{DELAYED} statement for a table, a handler -thread is created to process all @code{DELAYED} statements for the table, if -no such handler already exists. - -@item -The thread checks whether or not the handler has acquired a @code{DELAYED} -lock already; if not, it tells the handler thread to do so. The -@code{DELAYED} lock can be obtained even if other threads have a @code{READ} -or @code{WRITE} lock on the table. However, the handler will wait for all -@code{ALTER TABLE} locks or @code{FLUSH TABLES} to ensure that the table -structure is up to date. - -@item -The thread executes the @code{INSERT} statement, but instead of writing -the row to the table, it puts a copy of the final row into a queue that -is managed by the handler thread. Any syntax errors are noticed by the -thread and reported to the client program. - -@item -The client can't report the number of duplicates or the @code{AUTO_INCREMENT} -value for the resulting row; it can't obtain them from the server, because -the @code{INSERT} returns before the insert operation has been completed. If -you use the C API, the @code{mysql_info()} function doesn't return anything -meaningful, for the same reason. - -@item -The update log is updated by the handler thread when the row is inserted into -the table. In case of multiple-row inserts, the update log is updated when -the first row is inserted. +@findex ENCRYPT() +@item ENCRYPT(str[,salt]) +Encrypt @code{str} using the Unix @code{crypt()} system call. The +@code{salt} argument should be a string with two characters. +(As of @strong{MySQL} Version 3.22.16, @code{salt} may be longer than two characters.): -@item -After every @code{delayed_insert_limit} rows are written, the handler checks -whether or not any @code{SELECT} statements are still pending. If so, it -allows these to execute before continuing. +@example +mysql> select ENCRYPT("hello"); + -> 'VxuFAJXVARROc' +@end example -@cindex delayed_insert_limit -@cindex timeout -@item -When the handler has no more rows in its queue, the table is unlocked. If no -new @code{INSERT DELAYED} commands are received within -@code{delayed_insert_timeout} seconds, the handler terminates. +If @code{crypt()} is not available on your system, @code{ENCRYPT()} always +returns @code{NULL}. -@item -If more than @code{delayed_queue_size} rows are pending already in a -specific handler queue, the thread requesting @code{INSERT DELAYED} -waits until there is room in the queue. This is done to ensure that -the @code{mysqld} server doesn't use all memory for the delayed memory -queue. +@code{ENCRYPT()} ignores all but the first 8 characters of @code{str}, at +least on some systems. This will be determined by the behavior of the +underlying @code{crypt()} system call. -@item -The handler thread will show up in the @strong{MySQL} process list -with @code{delayed_insert} in the @code{Command} column. It will -be killed if you execute a @code{FLUSH TABLES} command or kill it with -@code{KILL thread_id}. However, it will first store all queued rows into the -table before exiting. During this time it will not accept any new -@code{INSERT} commands from another thread. If you execute an @code{INSERT -DELAYED} command after this, a new handler thread will be created. +@findex ENCODE() +@item ENCODE(str,pass_str) +Encrypt @code{str} using @code{pass_str} as the password. +To decrypt the result, use @code{DECODE()}. -@item -Note that the above means that @code{INSERT DELAYED} commands have higher -priority than normal @code{INSERT} commands if there is an @code{INSERT -DELAYED} handler already running! Other update commands will have to wait -until the @code{INSERT DELAYED} queue is empty, someone kills the handler -thread (with @code{KILL thread_id}), or someone executes @code{FLUSH TABLES}. +The results is a binary string of the same length as @code{string}. +If you want to save it in a column, use a @code{BLOB} column type. -@item -The following status variables provide information about @code{INSERT -DELAYED} commands: +@findex DECODE() +@item DECODE(crypt_str,pass_str) +Descrypts the encrypted string @code{crypt_str} using @code{pass_str} as the +password. @code{crypt_str} should be a string returned from +@code{ENCODE()}. -@multitable @columnfractions .35 .65 -@item @strong{Variable} @tab @strong{Meaning} -@item @code{Delayed_insert_threads} @tab Number of handler threads -@item @code{Delayed_writes} @tab Number of rows written with @code{INSERT DELAYED} -@item @code{Not_flushed_delayed_rows} @tab Number of rows waiting to be written -@end multitable +@findex MD5() +@item MD5(string) +Calculates a MD5 checksum for the string. Value is returned as a 32 long +hex number that may, for example, be used as a hash key: -You can view these variables by issuing a @code{SHOW STATUS} statement or -by executing a @code{mysqladmin extended-status} command. -@end itemize +@example +mysql> select MD5("testing"); + -> 'ae2b1fca515949e5d54fb22b8ed95575' +@end example -Note that @code{INSERT DELAYED} is slower than a normal INSERT if the -table is not in use. There is also the additional overhead for the -server to handle a separate thread for each table on which you use -@code{INSERT DELAYED}. This means that you should only use @code{INSERT -DELAYED} when you are really sure you need it! +This is an "RSA Data Security, Inc. MD5 Message-Digest Algorithm". -@findex REPLACE -@node REPLACE, LOAD DATA, INSERT, Reference -@section @code{REPLACE} Syntax +@findex LAST_INSERT_ID([expr]) +@item LAST_INSERT_ID([expr]) +Returns the last automatically generated value that was inserted into an +@code{AUTO_INCREMENT} column. +@xref{mysql_insert_id,, @code{mysql_insert_id()}}. @example - REPLACE [LOW_PRIORITY | DELAYED] - [INTO] tbl_name [(col_name,...)] - VALUES (expression,...) -or REPLACE [LOW_PRIORITY | DELAYED] - [INTO] tbl_name [(col_name,...)] - SELECT ... -or REPLACE [LOW_PRIORITY | DELAYED] - [INTO] tbl_name - SET col_name=expression, col_name=expression,... +mysql> select LAST_INSERT_ID(); + -> 195 @end example -@code{REPLACE} works exactly like @code{INSERT}, except that if an old -record in the table has the same value as a new record on a unique index, -the old record is deleted before the new record is inserted. -@xref{INSERT, , @code{INSERT}}. - -In other words, you can't access the values of the old row from a -@code{REPLACE} statement. In some old @strong{MySQL} version it looked -like you could do this, but that was a bug that has been corrected. +The last ID that was generated is maintained in the server on a +per-connection basis. It will not be changed by another client. It will not +even be changed if you update another @code{AUTO_INCREMENT} column with a +non-magic value (that is, a value that is not @code{NULL} and not @code{0}). -When one uses a @code{REPLACE} command, @code{mysql_affected_rows()} -will return 2 if the new row replaced and old row. This is because in -this case one row was inserted and then the duplicate was deleted. +If you insert many rows at the same time with an insert statement, +@code{LAST_INSERT_ID()} returns the value for the first inserted row. +The reason for this is so that you it makes it possible to easily reproduce +the same @code{INSERT} statement against some other server. -The above makes it easy to check if @code{REPLACE} added or replaced a -row. +@cindex sequence emulation +If @code{expr} is given as an argument to @code{LAST_INSERT_ID()} in an +@code{UPDATE} clause, then the value of the argument is returned as a +@code{LAST_INSERT_ID()} value. This can be used to simulate sequences. -@findex LOAD DATA INFILE -@node LOAD DATA, UPDATE, REPLACE, Reference -@section @code{LOAD DATA INFILE} Syntax +First create the table: @example -LOAD DATA [LOW_PRIORITY | CONCURRENT] [LOCAL] INFILE 'file_name.txt' - [REPLACE | IGNORE] - INTO TABLE tbl_name - [FIELDS - [TERMINATED BY '\t'] - [[OPTIONALLY] ENCLOSED BY ''] - [ESCAPED BY '\\' ] - ] - [LINES TERMINATED BY '\n'] - [IGNORE number LINES] - [(col_name,...)] +mysql> create table sequence (id int not null); +mysql> insert into sequence values (0); @end example -The @code{LOAD DATA INFILE} statement reads rows from a text file into a -table at a very high speed. If the @code{LOCAL} keyword is specified, the -file is read from the client host. If @code{LOCAL} is not specified, the -file must be located on the server. (@code{LOCAL} is available in -@strong{MySQL} Version 3.22.6 or later.) - -For security reasons, when reading text files located on the server, the -files must either reside in the database directory or be readable by all. -Also, to use @code{LOAD DATA INFILE} on server files, you must have the -@strong{file} privilege on the server host. -@xref{Privileges provided}. +Then the table can be used to generate sequence numbers like this: -If you specify the keyword @code{LOW_PRIORITY}, execution of the -@code{LOAD DATA} statement is delayed until no other clients are reading -from the table. +@example +mysql> update sequence set id=LAST_INSERT_ID(id+1); +@end example -If you specify the keyword @code{CONCURRENT} with a @code{MyISAM} table, -then other threads can retrieve data from the table while @code{LOAD -DATA} is executing. Using this option will of course affect the -performance of @code{LOAD DATA} a bit even if no other thread is using -the table at the same time. +You can generate sequences without calling @code{LAST_INSERT_ID()}, but the +utility of using the function this way is that the ID value is maintained in +the server as the last automatically generated value. You can retrieve the +new ID as you would read any normal @code{AUTO_INCREMENT} value in +@strong{MySQL}. For example, @code{LAST_INSERT_ID()} (without an argument) +will return the new ID. The C API function @code{mysql_insert_id()} +can also be used to get the value. -Using @code{LOCAL} will be a bit slower than letting the server access the -files directly, because the contents of the file must travel from the client -host to the server host. On the other hand, you do not need the -@strong{file} privilege to load local files. +@findex FORMAT() +@item FORMAT(X,D) +Formats the number @code{X} to a format like @code{'#,###,###.##'}, rounded +to @code{D} decimals. If @code{D} is @code{0}, the result will have no +decimal point or fractional part: -@c old version -If you are using @strong{MySQL} before Version 3.23.24 you can't read from a -FIFO with @code{LOAD DATA INFILE}. If you need to read from a FIFO (for -example the output from gunzip), use @code{LOAD DATA LOCAL INFILE} -instead. +@example +mysql> select FORMAT(12332.123456, 4); + -> '12,332.1235' +mysql> select FORMAT(12332.1,4); + -> '12,332.1000' +mysql> select FORMAT(12332.2,0); + -> '12,332' +@end example -@cindex @code{mysqlimport} -You can also load data files by using the @code{mysqlimport} utility; it -operates by sending a @code{LOAD DATA INFILE} command to the server. The -@code{--local} option causes @code{mysqlimport} to read data files from the -client host. You can specify the @code{--compress} option to get better -performance over slow networks if the client and server support the -compressed protocol. +@findex VERSION() +@item VERSION() +Returns a string indicating the @strong{MySQL} server version: -When locating files on the server host, the server uses the following rules: +@example +mysql> select VERSION(); + -> '3.23.13-log' +@end example -@itemize @bullet -@item -If an absolute pathname is given, the server uses the pathname as is. +Note that if your version ends with @code{-log} this means that logging is +enabled. -@item -If a relative pathname with one or more leading components is given, -the server searches for the file relative to the server's data directory. +@findex CONNECTION_ID() +@item CONNECTION_ID() +Returns the connection id (@code{thread_id}) for the connection. +Every connection has its own unique id: -@item -If a filename with no leading components is given, the server looks for -the file in the database directory of the current database. -@end itemize +@example +mysql> select CONNECTION_ID(); + -> 1 +@end example -Note that these rules mean a file given as @file{./myfile.txt} is read from -the server's data directory, whereas a file given as @file{myfile.txt} is -read from the database directory of the current database. For example, -the following @code{LOAD DATA} statement reads the file @file{data.txt} -from the database directory for @code{db1} because @code{db1} is the current -database, even though the statement explicitly loads the file into a -table in the @code{db2} database: +@cindex timeout +@findex GET_LOCK() +@item GET_LOCK(str,timeout) +Tries to obtain a lock with a name given by the string @code{str}, with a +timeout of @code{timeout} seconds. Returns @code{1} if the lock was obtained +successfully, @code{0} if the attempt timed out, or @code{NULL} if an error +occurred (such as running out of memory or the thread was killed with +@code{mysqladmin kill}). A lock is released when you execute +@code{RELEASE_LOCK()}, execute a new @code{GET_LOCK()}, or the thread +terminates. This function can be used to implement application locks or to +simulate record locks. It blocks requests by other clients for locks with +the same name; clients that agree on a given lock string name can use the +string to perform cooperative advisory locking: @example -mysql> USE db1; -mysql> LOAD DATA INFILE "data.txt" INTO TABLE db2.my_table; +mysql> select GET_LOCK("lock1",10); + -> 1 +mysql> select GET_LOCK("lock2",10); + -> 1 +mysql> select RELEASE_LOCK("lock2"); + -> 1 +mysql> select RELEASE_LOCK("lock1"); + -> NULL @end example -The @code{REPLACE} and @code{IGNORE} keywords control handling of input -records that duplicate existing records on unique key values. If you specify -@code{REPLACE}, new rows replace existing rows that have the same unique key -value. If you specify @code{IGNORE}, input rows that duplicate an existing -row on a unique key value are skipped. If you don't specify either option, an -error occurs when a duplicate key value is found, and the rest of the text -file is ignored. +Note that the second @code{RELEASE_LOCK()} call returns @code{NULL} because +the lock @code{"lock1"} was automatically released by the second +@code{GET_LOCK()} call. -If you load data from a local file using the @code{LOCAL} keyword, the server -has no way to stop transmission of the file in the middle of the operation, -so the default bahavior is the same as if @code{IGNORE} is specified. +@findex RELEASE_LOCK() +@item RELEASE_LOCK(str) +Releases the lock named by the string @code{str} that was obtained with +@code{GET_LOCK()}. Returns @code{1} if the lock was released, @code{0} if the +lock wasn't locked by this thread (in which case the lock is not released), +and @code{NULL} if the named lock didn't exist. The lock will not exist if +it was never obtained by a call to @code{GET_LOCK()} or if it already has +been released. -If you use @code{LOAD DATA INFILE} on an empty @code{MyISAM} table, -all non-unique indexes are created in a separate batch (like in @code{REPAIR}). -This normally makes @code{LOAD DATA INFILE} much faster when you have many -indexes. +@findex BENCHMARK() +@item BENCHMARK(count,expr) +The @code{BENCHMARK()} function executes the expression @code{expr} +repeatedly @code{count} times. It may be used to time how fast @strong{MySQL} +processes the expression. The result value is always @code{0}. The intended +use is in the @code{mysql} client, which reports query execution times: -@code{LOAD DATA INFILE} is the complement of @code{SELECT ... INTO OUTFILE}. -@xref{SELECT, , @code{SELECT}}. -To write data from a database to a file, use @code{SELECT ... INTO OUTFILE}. -To read the file back into the database, use @code{LOAD DATA INFILE}. -The syntax of the @code{FIELDS} and @code{LINES} clauses is the same for -both commands. Both clauses are optional, but @code{FIELDS} -must precede @code{LINES} if both are specified. +@example +mysql> select BENCHMARK(1000000,encode("hello","goodbye")); ++----------------------------------------------+ +| BENCHMARK(1000000,encode("hello","goodbye")) | ++----------------------------------------------+ +| 0 | ++----------------------------------------------+ +1 row in set (4.74 sec) +@end example -If you specify a @code{FIELDS} clause, -each of its subclauses (@code{TERMINATED BY}, @code{[OPTIONALLY] ENCLOSED -BY}, and @code{ESCAPED BY}) is also optional, except that you must -specify at least one of them. +The time reported is elapsed time on the client end, not CPU time on the +server end. It may be advisable to execute @code{BENCHMARK()} several +times, and interpret the result with regard to how heavily loaded the +server machine is. -If you don't specify a @code{FIELDS} clause, the defaults are the -same as if you had written this: +@findex INET_NTOA() +@item INET_NTOA(expr) +Returns the network address (4 or 8 byte) for the numeric expression: @example -FIELDS TERMINATED BY '\t' ENCLOSED BY '' ESCAPED BY '\\' +mysql> select INET_NTOA(3520061480); + -> "209.207.224.40" @end example -If you don't specify a @code{LINES} clause, the default -is the same as if you had written this: +@findex INET_ATON() +@item INET_ATON(expr) +Returns an integer that represents the numeric value for a network address. +Addresses may be 4 or 8 byte addresses: @example -LINES TERMINATED BY '\n' +mysql> select INET_ATON("209.207.224.40"); + -> 3520061480 @end example -In other words, the defaults cause @code{LOAD DATA INFILE} to act as follows -when reading input: - -@itemize @bullet -@item -Look for line boundaries at newlines. - -@item -Break lines into fields at tabs. - -@item -Do not expect fields to be enclosed within any quoting characters. +The generated number is always in network byte order; For example the +above number is calculated as @code{209*255^3 + 207*255^2 + 224*255 +40}. -@item -Interpret occurrences of tab, newline, or @samp{\} preceded by -@samp{\} as literal characters that are part of field values. -@end itemize +@findex MASTER_POS_WAIT() +@item MASTER_POS_WAIT(log_name, log_pos) +Blocks until the slave reaches the specified position in the master log during +replication. If master information is not initialized, returns NULL. If the +slave is not running, will block and wait until it is started and goes to or +past +the specified position. If the slave is already past the specified position, +returns immediately. The return value is the number of log events it had to +wait to get to the specified position, or NULL in case of error. Useful for +control of master-slave synchronization, but was originally written to +facilitate replication testing. +@end table -Conversely, the defaults cause @code{SELECT ... INTO OUTFILE} to act as -follows when writing output: +@findex GROUP BY functions +@findex functions, GROUP BY +@node Group by functions, , Miscellaneous functions, Functions +@subsection Functions for Use with @code{GROUP BY} Clauses -@itemize @bullet -@item -Write tabs between fields. +If you use a group function in a statement containing no @code{GROUP BY} +clause, it is equivalent to grouping on all rows. -@item -Do not enclose fields within any quoting characters. +@table @code +@findex COUNT() +@item COUNT(expr) +Returns a count of the number of non-@code{NULL} values in the rows +retrieved by a @code{SELECT} statement: -@item -Use @samp{\} to escape instances of tab, newline or @samp{\} that occur -within field values. +@example +mysql> select student.student_name,COUNT(*) + from student,course + where student.student_id=course.student_id + GROUP BY student_name; -@item -Write newlines at the ends of lines. -@end itemize +@end example -Note that to write @code{FIELDS ESCAPED BY '\\'}, you must specify two -backslashes for the value to be read as a single backslash. +@code{COUNT(*)} is somewhat different in that it returns a count of +the number of rows retrieved, whether or not they contain @code{NULL} +values. -The @code{IGNORE number LINES} option can be used to ignore a header of -column names at the start of the file: +@code{COUNT(*)} is optimized to +return very quickly if the @code{SELECT} retrieves from one table, no +other columns are retrieved, and there is no @code{WHERE} clause. +For example: @example -mysql> LOAD DATA INFILE "/tmp/file_name" into table test IGNORE 1 LINES; +mysql> select COUNT(*) from student; @end example -When you use @code{SELECT ... INTO OUTFILE} in tandem with @code{LOAD -DATA INFILE} to write data from a database into a file and then read -the file back into the database later, the field and line handling -options for both commands must match. Otherwise, @code{LOAD DATA -INFILE} will not interpret the contents of the file properly. Suppose -you use @code{SELECT ... INTO OUTFILE} to write a file with -fields delimited by commas: +@findex COUNT(DISTINCT) +@findex DISTINCT +@item COUNT(DISTINCT expr,[expr...]) +Returns a count of the number of different non-@code{NULL} values: @example -mysql> SELECT * INTO OUTFILE 'data.txt' - FIELDS TERMINATED BY ',' - FROM ...; +mysql> select COUNT(DISTINCT results) from student; @end example -To read the comma-delimited file back in, the correct statement would be: - -@example -mysql> LOAD DATA INFILE 'data.txt' INTO TABLE table2 - FIELDS TERMINATED BY ','; -@end example +In @strong{MySQL} you can get the number of distinct expression +combinations that don't contain NULL by giving a list of expressions. +In ANSI SQL you would have to do a concatenation of all expressions +inside @code{CODE(DISTINCT ..)}. -If instead you tried to read in the file with the statement shown below, it -wouldn't work because it instructs @code{LOAD DATA INFILE} to look for -tabs between fields: +@findex AVG() +@item AVG(expr) +Returns the average value of @code{expr}: @example -mysql> LOAD DATA INFILE 'data.txt' INTO TABLE table2 - FIELDS TERMINATED BY '\t'; +mysql> select student_name, AVG(test_score) + from student + GROUP BY student_name; @end example -The likely result is that each input line would be interpreted as -a single field. - -@code{LOAD DATA INFILE} can be used to read files obtained from -external sources, too. For example, a file in dBASE format will have -fields separated by commas and enclosed in double quotes. If lines in -the file are terminated by newlines, the command shown below -illustrates the field and line handling options you would use to load -the file: +@findex MIN() +@findex MAX() +@item MIN(expr) +@itemx MAX(expr) +Returns the minimum or maximum value of @code{expr}. @code{MIN()} and +@code{MAX()} may take a string argument; in such cases they return the +minimum or maximum string value. @xref{MySQL indexes}. @example -mysql> LOAD DATA INFILE 'data.txt' INTO TABLE tbl_name - FIELDS TERMINATED BY ',' ENCLOSED BY '"' - LINES TERMINATED BY '\n'; +mysql> select student_name, MIN(test_score), MAX(test_score) + from student + GROUP BY student_name; @end example -Any of the field or line handling options may specify an empty string -(@code{''}). If not empty, the @code{FIELDS [OPTIONALLY] ENCLOSED BY} -and @code{FIELDS ESCAPED BY} values must be a single character. The -@code{FIELDS TERMINATED BY} and @code{LINES TERMINATED BY} values may -be more than one character. For example, to write lines that are -terminated by carriage return-linefeed pairs, or to read a file -containing such lines, specify a @code{LINES TERMINATED BY '\r\n'} -clause. +@findex SUM() +@item SUM(expr) +Returns the sum of @code{expr}. Note that if the return set has no rows, +it returns NULL! -For example, to read a file of jokes, that are separated with a line -of @code{%%}, into a SQL table you can do: +@findex STD() +@findex STDDEV() +@cindex Oracle compatibility +@cindex compatibility, with Oracle +@item STD(expr) +@itemx STDDEV(expr) +Returns the standard deviation of @code{expr}. This is an extension to +ANSI SQL. The @code{STDDEV()} form of this function is provided for Oracle +compatibility. + +@findex BIT_OR() +@item BIT_OR(expr) +Returns the bitwise @code{OR} of all bits in @code{expr}. The calculation is +performed with 64-bit (@code{BIGINT}) precision. + +@findex BIT_AND() +@item BIT_AND(expr) +Returns the bitwise @code{AND} of all bits in @code{expr}. The calculation is +performed with 64-bit (@code{BIGINT}) precision. +@end table + +@cindex @code{GROUP BY}, extensions to ANSI SQL +@strong{MySQL} has extended the use of @code{GROUP BY}. You can use columns or +calculations in the @code{SELECT} expressions that don't appear in +the @code{GROUP BY} part. This stands for @emph{any possible value for this +group}. You can use this to get better performance by avoiding sorting and +grouping on unnecessary items. For example, you don't need to group on +@code{customer.name} in the following query: @example -create table jokes (a int not null auto_increment primary key, joke text -not null); -load data infile "/tmp/jokes.txt" into table jokes fields terminated by "" -lines terminated by "\n%%\n" (joke); +mysql> select order.custid,customer.name,max(payments) + from order,customer + where order.custid = customer.custid + GROUP BY order.custid; @end example -@code{FIELDS [OPTIONALLY] ENCLOSED BY} controls quoting of fields. For -output (@code{SELECT ... INTO OUTFILE}), if you omit the word -@code{OPTIONALLY}, all fields are enclosed by the @code{ENCLOSED BY} -character. An example of such output (using a comma as the field -delimiter) is shown below: +In ANSI SQL, you would have to add @code{customer.name} to the @code{GROUP +BY} clause. In @strong{MySQL}, the name is redundant if you don't run in +ANSI mode. + +@strong{Don't use this feature} if the columns you omit from the +@code{GROUP BY} part aren't unique in the group! You will get +unpredictable results. + +In some cases, you can use @code{MIN()} and @code{MAX()} to obtain a specific +column value even if it isn't unique. The following gives the value of +@code{column} from the row containing the smallest value in the @code{sort} +column: @example -"1","a string","100.20" -"2","a string containing a , comma","102.20" -"3","a string containing a \" quote","102.20" -"4","a string containing a \", quote and comma","102.20" +substr(MIN(concat(rpad(sort,6,' '),column)),7) @end example -If you specify @code{OPTIONALLY}, the @code{ENCLOSED BY} character is -used only to enclose @code{CHAR} and @code{VARCHAR} fields: +@xref{example-Maximum-column-group-row}. + +@cindex @code{ORDER BY}, aliases in +@cindex aliases, in @code{ORDER BY} clauses +@cindex @code{GROUP BY}, aliases in +@cindex aliases, in @code{GROUP BY} clauses +@cindex expression aliases +@cindex aliases, for expressions +Note that if you are using @strong{MySQL} Version 3.22 (or earlier) or if +you are trying to follow ANSI SQL, you can't use expressions in @code{GROUP +BY} or @code{ORDER BY} clauses. You can work around this limitation by +using an alias for the expression: @example -1,"a string",100.20 -2,"a string containing a , comma",102.20 -3,"a string containing a \" quote",102.20 -4,"a string containing a \", quote and comma",102.20 +mysql> SELECT id,FLOOR(value/100) AS val FROM tbl_name + GROUP BY id,val ORDER BY val; @end example -Note that occurrences of the @code{ENCLOSED BY} character within a -field value are escaped by prefixing them with the @code{ESCAPED BY} -character. Also note that if you specify an empty @code{ESCAPED BY} -value, it is possible to generate output that cannot be read properly by -@code{LOAD DATA INFILE}. For example, the output just shown above would -appear as shown below if the escape character is empty. Observe that the -second field in the fourth line contains a comma following the quote, which -(erroneously) appears to terminate the field: +In @strong{MySQL} Version 3.23 you can do: @example -1,"a string",100.20 -2,"a string containing a , comma",102.20 -3,"a string containing a " quote",102.20 -4,"a string containing a ", quote and comma",102.20 +mysql> SELECT id,FLOOR(value/100) FROM tbl_name ORDER BY RAND(); @end example -For input, the @code{ENCLOSED BY} character, if present, is stripped from the -ends of field values. (This is true whether or not @code{OPTIONALLY} is -specified; @code{OPTIONALLY} has no effect on input interpretation.) -Occurrences of the @code{ENCLOSED BY} character preceded by the -@code{ESCAPED BY} character are interpreted as part of the current field -value. In addition, duplicated @code{ENCLOSED BY} characters occurring -within fields are interpreted as single @code{ENCLOSED BY} characters if the -field itself starts with that character. For example, if @code{ENCLOSED BY -'"'} is specified, quotes are handled as shown below: +@findex CREATE DATABASE +@node CREATE DATABASE, DROP DATABASE, Functions, Reference +@section @code{CREATE DATABASE} Syntax @example -"The ""BIG"" boss" -> The "BIG" boss -The "BIG" boss -> The "BIG" boss -The ""BIG"" boss -> The ""BIG"" boss +CREATE DATABASE [IF NOT EXISTS] db_name @end example -@code{FIELDS ESCAPED BY} controls how to write or read special characters. -If the @code{FIELDS ESCAPED BY} character is not empty, it is used to prefix -the following characters on output: -@itemize @bullet -@item -The @code{FIELDS ESCAPED BY} character -@item -The @code{FIELDS [OPTIONALLY] ENCLOSED BY} character -@item -The first character of the @code{FIELDS TERMINATED BY} and -@code{LINES TERMINATED BY} values -@item -ASCII @code{0} (what is actually written following the escape character is -ASCII @code{'0'}, not a zero-valued byte) -@end itemize +@code{CREATE DATABASE} creates a database with the given name. Rules for +allowable database names are given in @ref{Legal names}. An error occurs if +the database already exists and you didn't specify @code{IF NOT EXISTS}. -If the @code{FIELDS ESCAPED BY} character is empty, no characters are escaped. -It is probably not a good idea to specify an empty escape character, -particularly if field values in your data contain any of the characters in -the list just given. +Databases in @strong{MySQL} are implemented as directories containing files +that correspond to tables in the database. Because there are no tables in a +database when it is initially created, the @code{CREATE DATABASE} statement +only creates a directory under the @strong{MySQL} data directory. -For input, if the @code{FIELDS ESCAPED BY} character is not empty, occurrences -of that character are stripped and the following character is taken literally -as part of a field value. The exceptions are an escaped @samp{0} or -@samp{N} (for example, @code{\0} or @code{\N} if the escape character is -@samp{\}). These sequences are interpreted as ASCII @code{0} (a zero-valued -byte) and @code{NULL}. See below for the rules on @code{NULL} handling. +@cindex @code{mysqladmin} +You can also create databases with @code{mysqladmin}. +@xref{Programs}. -For more information about @samp{\}-escape syntax, -see @ref{Literals}. +@findex DROP DATABASE +@node DROP DATABASE, CREATE TABLE, CREATE DATABASE, Reference +@section @code{DROP DATABASE} Syntax -In certain cases, field and line handling options interact: +@example +DROP DATABASE [IF EXISTS] db_name +@end example -@itemize @bullet -@item -If @code{LINES TERMINATED BY} is an empty string and @code{FIELDS -TERMINATED BY} is non-empty, lines are also terminated with -@code{FIELDS TERMINATED BY}. -@item -If the @code{FIELDS TERMINATED BY} and @code{FIELDS ENCLOSED BY} values are -both empty (@code{''}), a fixed-row (non-delimited) format is used. With -fixed-row format, no delimiters are used between fields. Instead, column -values are written and read using the ``display'' widths of the columns. For -example, if a column is declared as @code{INT(7)}, values for the column are -written using 7-character fields. On input, values for the column are -obtained by reading 7 characters. Fixed-row format also affects handling of -@code{NULL} values; see below. Note that fixed-size format will not work -if you are using a multi-byte character set. -@end itemize +@code{DROP DATABASE} drops all tables in the database and deletes the +database. If you do a @code{DROP DATABASE} on a symbolic linked +database, both the link and the original database is deleted. @strong{Be +VERY careful with this command!} -Handling of @code{NULL} values varies, depending on the @code{FIELDS} and -@code{LINES} options you use: +@code{DROP DATABASE} returns the number of files that were removed from +the database directory. Normally, this is three times the number of +tables, because normally each table corresponds to a @file{.MYD} file, a +@file{.MYI} file, and a @file{.frm} file. -@itemize @bullet -@item -For the default @code{FIELDS} and @code{LINES} values, -@code{NULL} is written as @code{\N} for output and @code{\N} is read -as @code{NULL} for input (assuming the @code{ESCAPED BY} character -is @samp{\}). +The @code{DROP DATABASE} command removes from the given database +directory all files with the following extensions: -@item -If @code{FIELDS ENCLOSED BY} is not empty, a field containing the literal -word @code{NULL} as its value is read as a @code{NULL} value (this differs -from the word @code{NULL} enclosed within @code{FIELDS ENCLOSED BY} -characters, which is read as the string @code{'NULL'}). +@multitable @columnfractions .25 .25 .25 .25 +@item .BAK @tab .DAT @tab .HSH @tab .ISD +@item .ISM @tab .ISM @tab .MRG @tab .MYD +@item .MYI @tab .db @tab .frm +@end multitable -@item -If @code{FIELDS ESCAPED BY} is empty, @code{NULL} is written as the word -@code{NULL}. +All subdirectories that consists of 2 digits (@code{RAID} directories) +are also removed. -@item -With fixed-row format (which happens when @code{FIELDS TERMINATED BY} and -@code{FIELDS ENCLOSED BY} are both empty), @code{NULL} is written as an empty -string. Note that this causes both @code{NULL} values and empty strings in -the table to be indistinguishable when written to the file because they are -both written as empty strings. If you need to be able to tell the two apart -when reading the file back in, you should not use fixed-row format. -@end itemize +In @strong{MySQL} Version 3.22 or later, you can use the keywords +@code{IF EXISTS} to prevent an error from occurring if the database doesn't +exist. -Some cases are not supported by @code{LOAD DATA INFILE}: -@itemize @bullet +@cindex @code{mysqladmin} +You can also drop databases with @code{mysqladmin}. @xref{Programs}. -@item -Fixed-size rows (@code{FIELDS TERMINATED BY} and @code{FIELDS ENCLOSED -BY} both empty) and @code{BLOB} or @code{TEXT} columns. +@findex CREATE TABLE +@node CREATE TABLE, ALTER TABLE, DROP DATABASE, Reference +@section @code{CREATE TABLE} Syntax -@item -If you specify one separator that is the same as or a prefix of another, -@code{LOAD DATA INFILE} won't be able to interpret the input properly. -For example, the following @code{FIELDS} clause would cause problems: +@menu +* Silent column changes:: Silent column changes +@end menu @example -FIELDS TERMINATED BY '"' ENCLOSED BY '"' -@end example - -@item -If @code{FIELDS ESCAPED BY} is empty, a field value that contains an occurrence -of @code{FIELDS ENCLOSED BY} or @code{LINES TERMINATED BY} -followed by the @code{FIELDS TERMINATED BY} value will cause @code{LOAD -DATA INFILE} to stop reading a field or line too early. -This happens because @code{LOAD DATA INFILE} cannot properly determine -where the field or line value ends. -@end itemize +CREATE [TEMPORARY] TABLE [IF NOT EXISTS] tbl_name [(create_definition,...)] +[table_options] [select_statement] -The following example loads all columns of the @code{persondata} table: +create_definition: + col_name type [NOT NULL | NULL] [DEFAULT default_value] [AUTO_INCREMENT] + [PRIMARY KEY] [reference_definition] + or PRIMARY KEY (index_col_name,...) + or KEY [index_name] (index_col_name,...) + or INDEX [index_name] (index_col_name,...) + or UNIQUE [INDEX] [index_name] (index_col_name,...) + or FULLTEXT [INDEX] [index_name] (index_col_name,...) + or [CONSTRAINT symbol] FOREIGN KEY index_name (index_col_name,...) + [reference_definition] + or CHECK (expr) -@example -mysql> LOAD DATA INFILE 'persondata.txt' INTO TABLE persondata; -@end example +type: + TINYINT[(length)] [UNSIGNED] [ZEROFILL] + or SMALLINT[(length)] [UNSIGNED] [ZEROFILL] + or MEDIUMINT[(length)] [UNSIGNED] [ZEROFILL] + or INT[(length)] [UNSIGNED] [ZEROFILL] + or INTEGER[(length)] [UNSIGNED] [ZEROFILL] + or BIGINT[(length)] [UNSIGNED] [ZEROFILL] + or REAL[(length,decimals)] [UNSIGNED] [ZEROFILL] + or DOUBLE[(length,decimals)] [UNSIGNED] [ZEROFILL] + or FLOAT[(length,decimals)] [UNSIGNED] [ZEROFILL] + or DECIMAL(length,decimals) [UNSIGNED] [ZEROFILL] + or NUMERIC(length,decimals) [UNSIGNED] [ZEROFILL] + or CHAR(length) [BINARY] + or VARCHAR(length) [BINARY] + or DATE + or TIME + or TIMESTAMP + or DATETIME + or TINYBLOB + or BLOB + or MEDIUMBLOB + or LONGBLOB + or TINYTEXT + or TEXT + or MEDIUMTEXT + or LONGTEXT + or ENUM(value1,value2,value3,...) + or SET(value1,value2,value3,...) -No field list is specified, so @code{LOAD DATA INFILE} expects input rows -to contain a field for each table column. The default @code{FIELDS} and -@code{LINES} values are used. +index_col_name: + col_name [(length)] -If you wish to load only some of a table's columns, specify a field list: +reference_definition: + REFERENCES tbl_name [(index_col_name,...)] + [MATCH FULL | MATCH PARTIAL] + [ON DELETE reference_option] + [ON UPDATE reference_option] -@example -mysql> LOAD DATA INFILE 'persondata.txt' - INTO TABLE persondata (col1,col2,...); -@end example +reference_option: + RESTRICT | CASCADE | SET NULL | NO ACTION | SET DEFAULT -You must also specify a field list if the order of the fields in the input -file differs from the order of the columns in the table. Otherwise, -@strong{MySQL} cannot tell how to match up input fields with table columns. +table_options: + TYPE = @{BDB | HEAP | ISAM | InnoDB | MERGE | MYISAM @} +or AUTO_INCREMENT = # +or AVG_ROW_LENGTH = # +or CHECKSUM = @{0 | 1@} +or COMMENT = "string" +or MAX_ROWS = # +or MIN_ROWS = # +or PACK_KEYS = @{0 | 1@} +or PASSWORD = "string" +or DELAY_KEY_WRITE = @{0 | 1@} +or ROW_FORMAT= @{ default | dynamic | fixed | compressed @} +or RAID_TYPE= @{1 | STRIPED | RAID0 @} RAID_CHUNKS=# RAID_CHUNKSIZE=# +or UNION = (table_name,[table_name...]) +or DATA DIRECTORY="directory" +or INDEX DIRECTORY="directory" -If a row has too few fields, the columns for which no input field is present -are set to default values. Default value assignment is described in -@ref{CREATE TABLE, , @code{CREATE TABLE}}. +select_statement: + [IGNORE | REPLACE] SELECT ... (Some legal select statement) +@end example -An empty field value is interpreted differently than if the field value -is missing: +@code{CREATE TABLE} +creates a table with the given name in the current database. Rules for +allowable table names are given in @ref{Legal names}. An error occurs if +there is no current database or if the table already exists. -@itemize @bullet -@item -For string types, the column is set to the empty string. +In @strong{MySQL} Version 3.22 or later, the table name can be specified as +@code{db_name.tbl_name}. This works whether or not there is a current +database. -@item -For numeric types, the column is set to @code{0}. +In @strong{MySQL} Version 3.23, you can use the @code{TEMPORARY} keyword when +you create a table. A temporary table will automatically be deleted if a +connection dies and the name is per connection. This means that two different +connections can both use the same temporary table name without conflicting +with each other or with an existing table of the same name. (The existing table +is hidden until the temporary table is deleted). -@item -For date and time types, the column is set to the appropriate ``zero'' -value for the type. -@xref{Date and time types}. -@end itemize +In @strong{MySQL} Version 3.23 or later, you can use the keywords +@code{IF NOT EXISTS} so that an error does not occur if the table already +exists. Note that there is no verification that the table structures are +identical. -Note that these are the same values that result if you assign an empty -string explicitly to a string, numeric, or date or time type explicitly -in an @code{INSERT} or @code{UPDATE} statement. +Each table @code{tbl_name} is represented by some files in the database +directory. In the case of MyISAM-type tables you will get: -@code{TIMESTAMP} columns are only set to the current date and time if there -is a @code{NULL} value for the column, or (for the first @code{TIMESTAMP} -column only) if the @code{TIMESTAMP} column is left out from the field list -when a field list is specified. +@multitable @columnfractions .2 .8 +@item @strong{File} @tab @strong{Purpose} +@item @code{tbl_name.frm} @tab Table definition (form) file +@item @code{tbl_name.MYD} @tab Data file +@item @code{tbl_name.MYI} @tab Index file +@end multitable -If an input row has too many fields, the extra fields are ignored and -the number of warnings is incremented. +For more information on the properties of the various column types, see +@ref{Column types}: -@code{LOAD DATA INFILE} regards all input as strings, so you can't use -numeric values for @code{ENUM} or @code{SET} columns the way you can with -@code{INSERT} statements. All @code{ENUM} and @code{SET} values must be -specified as strings! +@itemize @bullet +@item +If neither @code{NULL} nor @code{NOT NULL} is specified, the column +is treated as though @code{NULL} had been specified. -@findex mysql_info() -If you are using the C API, you can get information about the query by -calling the API function @code{mysql_info()} when the @code{LOAD DATA INFILE} -query finishes. The format of the information string is shown below: +@item +An integer column may have the additional attribute @code{AUTO_INCREMENT}. +When you insert a value of @code{NULL} (recommended) or @code{0} into an +@code{AUTO_INCREMENT} column, the column is set to @code{value+1}, where +@code{value} is the largest value for the column currently in the table. +@code{AUTO_INCREMENT} sequences begin with @code{1}. +@xref{mysql_insert_id, , @code{mysql_insert_id()}}. -@example -Records: 1 Deleted: 0 Skipped: 0 Warnings: 0 -@end example +If you delete the row containing the maximum value for an +@code{AUTO_INCREMENT} column, the value will be reused with an +@code{ISAM}, or @code{BDB} table but not with a +@code{MyISAM} or @code{InnoDB} table. If you delete all rows in the table +with @code{DELETE FROM table_name} (without a @code{WHERE}) in +@code{AUTOCOMMIT} mode, the sequence starts over for all table types. -Warnings occur under the same circumstances as when values are inserted -via the @code{INSERT} statement (@pxref{INSERT, , @code{INSERT}}), except -that @code{LOAD DATA INFILE} also generates warnings when there are too few -or too many fields in the input row. The warnings are not stored anywhere; -the number of warnings can only be used as an indication if everything went -well. If you get warnings and want to know exactly why you got them, one way -to do this is to use @code{SELECT ... INTO OUTFILE} into another file and -compare this to your original input file. +@strong{NOTE:} There can be only one @code{AUTO_INCREMENT} column per +table, and it must be indexed. @strong{MySQL} Version 3.23 will also only +work properly if the auto_increment column only has positive +values. Inserting a negative number is regarded as inserting a very large +positive number. This is done to avoid precision problems when +numbers 'wrap' over from positive to negative and also to ensure that one +doesn't accidentally get an auto_increment column that contains 0. -If you need @code{LOAD DATA} to read from a pipe, you can use the -following trick: +@cindex ODBC compatibility +@cindex compatibility, with ODBC +To make @strong{MySQL} compatible with some ODBC applications, you can find +the last inserted row with the following query: @example -mkfifo /mysql/db/x/x -chmod 666 /mysql/db/x/x -cat < /dev/tcp/10.1.1.12/4711 > /nt/mysql/db/x/x -mysql -e "LOAD DATA INFILE 'x' INTO TABLE x" x +SELECT * FROM tbl_name WHERE auto_col IS NULL @end example -If you are using a version of @strong{MySQL} older than 3.23.25 -you can only do the above with @code{LOAD DATA LOCAL INFILE}. - -For more information about the efficiency of @code{INSERT} versus -@code{LOAD DATA INFILE} and speeding up @code{LOAD DATA INFILE}, -@xref{Insert speed}. +@item +@code{NULL} values are handled differently for @code{TIMESTAMP} columns than +for other column types. You cannot store a literal @code{NULL} in a +@code{TIMESTAMP} column; setting the column to @code{NULL} sets it to the +current date and time. Because @code{TIMESTAMP} columns behave this way, the +@code{NULL} and @code{NOT NULL} attributes do not apply in the normal way and +are ignored if you specify them. -@findex UPDATE -@node UPDATE, USE, LOAD DATA, Reference -@section @code{UPDATE} Syntax +On the other hand, to make it easier for @strong{MySQL} clients to use +@code{TIMESTAMP} columns, the server reports that such columns may be +assigned @code{NULL} values (which is true), even though @code{TIMESTAMP} +never actually will contain a @code{NULL} value. You can see this when you +use @code{DESCRIBE tbl_name} to get a description of your table. -@example -UPDATE [LOW_PRIORITY] [IGNORE] tbl_name - SET col_name1=expr1, [col_name2=expr2, ...] - [WHERE where_definition] - [LIMIT #] -@end example +Note that setting a @code{TIMESTAMP} column to @code{0} is not the same +as setting it to @code{NULL}, because @code{0} is a valid @code{TIMESTAMP} +value. -@code{UPDATE} updates columns in existing table rows with new values. -The @code{SET} clause indicates which columns to modify and the values -they should be given. The @code{WHERE} clause, if given, specifies -which rows should be updated. Otherwise all rows are updated. If the -@code{ORDER BY} clause is specified, the rows will be updated in the -order that is specified. +@item +If no @code{DEFAULT} value is specified for a column, @strong{MySQL} +automatically assigns one. -If you specify the keyword @code{LOW_PRIORITY}, execution of the -@code{UPDATE} is delayed until no other clients are reading from the table. +If the column may take @code{NULL} as a value, the default value is +@code{NULL}. -If you specify the keyword @code{IGNORE}, the update statement will not -abort even if we get duplicate key errors during the update. Rows that -would cause conflicts will not be updated. +If the column is declared as @code{NOT NULL}, the default value depends on +the column type: -If you access a column from @code{tbl_name} in an expression, -@code{UPDATE} uses the current value of the column. For example, the -following statement sets the @code{age} column to one more than its -current value: +@itemize @minus +@item +For numeric types other than those declared with the @code{AUTO_INCREMENT} +attribute, the default is @code{0}. For an @code{AUTO_INCREMENT} column, the +default value is the next value in the sequence. -@example -mysql> UPDATE persondata SET age=age+1; -@end example +@item +For date and time types other than @code{TIMESTAMP}, the default is the +appropriate zero value for the type. For the first @code{TIMESTAMP} +column in a table, the default value is the current date and time. +@xref{Date and time types}. -@code{UPDATE} assignments are evaluated from left to right. For example, the -following statement doubles the @code{age} column, then increments it: +@item +For string types other than @code{ENUM}, the default value is the empty string. +For @code{ENUM}, the default is the first enumeration value. +@end itemize -@example -mysql> UPDATE persondata SET age=age*2, age=age+1; -@end example +Default values must be constants. This means, for example, that you cannot +set the default for a date column to be the value of a function such as +@code{NOW()} or @code{CURRENT_DATE}. -If you set a column to the value it currently has, @strong{MySQL} notices -this and doesn't update it. +@item +@code{KEY} is a synonym for @code{INDEX}. -@findex mysql_info() -@code{UPDATE} returns the number of rows that were actually changed. -In @strong{MySQL} Version 3.22 or later, the C API function @code{mysql_info()} -returns the number of rows that were matched and updated and the number of -warnings that occurred during the @code{UPDATE}. +@item +In @strong{MySQL}, a @code{UNIQUE} key can have only distinct values. An +error occurs if you try to add a new row with a key that matches an existing +row. -In @strong{MySQL} Version 3.23, you can use @code{LIMIT #} to ensure that -only a given number of rows are changed. +@item +@tindex PRIMARY KEY +A @code{PRIMARY KEY} is a unique @code{KEY} with the extra constraint +that all key columns must be defined as @code{NOT NULL}. In @strong{MySQL} +the key is named @code{PRIMARY}. A table can have only one @code{PRIMARY KEY}. +If you don't have a @code{PRIMARY KEY} and some applications ask for the +@code{PRIMARY KEY} in your tables, @strong{MySQL} will return the first +@code{UNIQUE} key, which doesn't have any @code{NULL} columns, as the +@code{PRIMARY KEY}. -@findex USE -@node USE, FLUSH, UPDATE, Reference -@section @code{USE} Syntax +@item +A @code{PRIMARY KEY} can be a multiple-column index. However, you cannot +create a multiple-column index using the @code{PRIMARY KEY} key attibute in a +column specification. Doing so will mark only that single column as primary. +You must use the @code{PRIMARY KEY(index_col_name, ...)} syntax. -@example -USE db_name -@end example +@item +If the @code{PRIMARY} or @code{UNIQUE} key consists of only one column and this +is of type integer, you can also refer to it as @code{_rowid} +(new in Version 3.23.11). -The @code{USE db_name} statement tells @strong{MySQL} to use the @code{db_name} -database as the default database for subsequent queries. The database remains -current until the end of the session or until another @code{USE} statement -is issued: +@item +If you don't assign a name to an index, the index will be assigned the same +name as the first @code{index_col_name}, with an optional suffix (@code{_2}, +@code{_3}, @code{...}) to make it unique. You can see index names for a +table using @code{SHOW INDEX FROM tbl_name}. +@xref{SHOW, , @code{SHOW}}. -@example -mysql> USE db1; -mysql> SELECT count(*) FROM mytable; # selects from db1.mytable -mysql> USE db2; -mysql> SELECT count(*) FROM mytable; # selects from db2.mytable -@end example +@item +@cindex @code{NULL} values, and indexes +@cindex indexes, and @code{NULL} values +Only the @code{MyISAM} table type supports indexes on columns that can have +@code{NULL} values. In other cases you must declare such columns +@code{NOT NULL} or an error results. -Making a particular database current by means of the @code{USE} statement -does not preclude you from accessing tables in other databases. The example -below accesses the @code{author} table from the @code{db1} database and the -@code{editor} table from the @code{db2} database: +@item +With @code{col_name(length)} syntax, you can specify an index that +uses only a part of a @code{CHAR} or @code{VARCHAR} column. This can +make the index file much smaller. +@xref{Indexes}. +@item +@cindex @code{BLOB} columns, indexing +@cindex indexes, and @code{BLOB} columns +@cindex @code{TEXT} columns, indexing +@cindex indexes, and @code{TEXT} columns +Only the @code{MyISAM} table type supports indexing on @code{BLOB} and +@code{TEXT} columns. When putting an index on a @code{BLOB} or @code{TEXT} +column you MUST always specify the length of the index: @example -mysql> USE db1; -mysql> SELECT author_name,editor_name FROM author,db2.editor - WHERE author.editor_id = db2.editor.editor_id; +CREATE TABLE test (blob_col BLOB, index(blob_col(10))); @end example -@cindex Sybase compatibility -@cindex compatibility, with Sybase -The @code{USE} statement is provided for Sybase compatibility. +@item +When you use @code{ORDER BY} or @code{GROUP BY} with a @code{TEXT} or +@code{BLOB} column, only the first @code{max_sort_length} bytes are used. +@xref{BLOB, , @code{BLOB}}. -@cindex @code{mysqladmin} -@cindex clearing, caches -@cindex caches, clearing -@findex FLUSH -@node FLUSH, KILL, USE, Reference -@section @code{FLUSH} Syntax +@item +In @strong{MySQL} Version 3.23.23 or later, you can also create special +@strong{FULLTEXT} indexes. They are used for full-text search. Only the +@code{MyISAM} table type supports @code{FULLTEXT} indexes. They can be created +only from @code{VARCHAR} and @code{TEXT} columns. +Indexing always happens over the entire column, partial indexing is not +supported. See @ref{Fulltext Search} for details of operation. -@example -FLUSH flush_option [,flush_option] -@end example +@item +The @code{FOREIGN KEY}, @code{CHECK}, and @code{REFERENCES} clauses don't +actually do anything. The syntax for them is provided only for compatibility, +to make it easier to port code from other SQL servers and to run applications +that create tables with references. +@xref{Missing functions}. -You should use the @code{FLUSH} command if you want to clear some of the -internal caches @strong{MySQL} uses. To execute @code{FLUSH}, you must have -the @strong{RELOAD} privilege. +@item +Each @code{NULL} column takes one bit extra, rounded up to the nearest byte. -@code{flush_option} can be any of the following: +@item +The maximum record length in bytes can be calculated as follows: -@multitable @columnfractions .15 .85 -@item @code{HOSTS} @tab Empties the host cache tables. You should flush the -host tables if some of your hosts change IP number or if you get the -error message @code{Host ... is blocked}. When more than -@code{max_connect_errors} errors occur in a row for a given host while -connection to the @strong{MySQL} server, @strong{MySQL} assumes -something is wrong and blocks the host from further connection requests. -Flushing the host tables allows the host to attempt to connect -again. @xref{Blocked host}.) You can start @code{mysqld} with -@code{-O max_connection_errors=999999999} to avoid this error message. +@example +row length = 1 + + (sum of column lengths) + + (number of NULL columns + 7)/8 + + (number of variable-length columns) +@end example -@item @code{LOGS} @tab Closes and reopens all log files. -If you have specified the update log file or a binary log file without -an extension, the extension number of the log file will be incremented -by one relative to the previous file. If you have used an extension in -the file name, @strong{MySQL} will close and reopen the update log file. -@xref{Update log}. This is the same thing as sending the @code{SIGHUP} -signal to the @code{mysqld} server. +@item +The @code{table_options} and @code{SELECT} options are only +implemented in @strong{MySQL} Version 3.23 and above. -@item @code{PRIVILEGES} @tab Reloads the privileges from the grant tables in -the @code{mysql} database. +The different table types are: -@item @code{TABLES} @tab Closes all open tables and force all tables in use to be closed. +@multitable @columnfractions .20 .80 +@item BDB or Berkeley_db @tab Transaction-safe tables with page locking. @xref{BDB}. +@item HEAP @tab The data for this table is only stored in memory. @xref{HEAP}. +@item ISAM @tab The original table handler. @xref{ISAM}. +@item InnoDB @tab Transaction-safe tables with row locking. @xref{InnoDB}. +@item MERGE @tab A collection of MyISAM tables used as one table. @xref{MERGE}. +@item MyISAM @tab The new binary portable table handler that is replacing ISAM. @xref{MyISAM}. +@end multitable +@xref{Table types}. -@item @code{[TABLE | TABLES] table_name [,table_name...]} @tab Flushes only the given tables. +If a table type is specified, and that particular type is not available, +@strong{MySQL} will choose the closest table type to the one that you have +specified. For example, if @code{TYPE=BDB} is specified, and that distribution +of @strong{MySQL} does not support @code{BDB} tables, the table will be created +as @code{MyISAM} instead. -@item @code{TABLES WITH READ LOCK} @tab Closes all open tables and locks all tables for all databases with a read until one executes @code{UNLOCK TABLES}. This is very convenient way to get backups if you have a file system, like Veritas,that can take snapshots in time. +The other table options are used to optimize the behavior of the +table. In most cases, you don't have to specify any of them. +The options work for all table types, if not otherwise indicated: -@item @code{STATUS} @tab Resets most status variables to zero. This is something one should only use when debugging a query. +@multitable @columnfractions .20 .80 +@item @code{AUTO_INCREMENT} @tab The next auto_increment value you want to set for your table (MyISAM). +@item @code{AVG_ROW_LENGTH} @tab An approximation of the average row length for your table. You only need to set this for large tables with variable size records. +@item @code{CHECKSUM} @tab Set this to 1 if you want @strong{MySQL} to maintain a checksum for all rows (makes the table a little slower to update but makes it easier to find corrupted tables) (MyISAM). +@item @code{COMMENT} @tab A 60-character comment for your table. +@item @code{MAX_ROWS} @tab Max number of rows you plan to store in the table. +@item @code{MIN_ROWS} @tab Minimum number of rows you plan to store in the table. +@item @code{PACK_KEYS} @tab Set this to 1 if you want to have a smaller index. This usually makes updates slower and reads faster (MyISAM, ISAM). +@item @code{PASSWORD} @tab Encrypt the @code{.frm} file with a password. This option doesn't do anything in the standard @strong{MySQL} version. +@item @code{DELAY_KEY_WRITE} @tab Set this to 1 if want to delay key table updates until the table is closed (MyISAM). +@item @code{ROW_FORMAT} @tab Defines how the rows should be stored. Currently you can only use the DYNAMIC and STATIC options for MyISAM tables. @end multitable -You can also access each of the commands shown above with the @code{mysqladmin} -utility, using the @code{flush-hosts}, @code{flush-logs}, @code{reload}, -or @code{flush-tables} commands. +When you use a @code{MyISAM} table, @strong{MySQL} uses the product of +@code{max_rows * avg_row_length} to decide how big the resulting table +will be. If you don't specify any of the above options, the maximum size +for a table will be 4G (or 2G if your operating systems only supports 2G +tables). The reason for this is just to keep down the pointer sizes +to make the index smaller and faster if you don't really need big files. -Take also a look at the @code{RESET} command used with -replication. @xref{Replication SQL}. +If you don't use @code{PACK_KEYS}, the default is to only pack strings, +not numbers. If you use @code{PACK_KEYS=1}, numbers will be packed as well. -@cindex @code{mysqladmin} -@findex KILL -@node KILL, SHOW, FLUSH, Reference -@section @code{KILL} Syntax +When packing binary number keys, @strong{MySQL} will use prefix compression. +This means that you will only get a big benefit of this if you have +many numbers that are the same. Prefix compression means that every +key needs one extra byte to indicate how many bytes of the previous key are +the same for the next key (note that the pointer to the row is stored +in high-byte-first-order directly after the key, to improve +compression.) This means that if you have many equal keys on two rows +in a row, all following 'same' keys will usually only take 2 bytes +(including the pointer to the row). Compare this to the ordinary case +where the following keys will take storage_size_for_key + +pointer_size (usually 4). On the other hand, if all keys are +totally different, you will lose 1 byte per key, if the key isn't a +key that can have @code{NULL} values (In this case the packed key length will +be stored in the same byte that is used to mark if a key is @code{NULL}.) + +@item +If you specify a @code{SELECT} after the @code{CREATE} statement, +@strong{MySQL} will create new fields for all elements in the +@code{SELECT}. For example: @example -KILL thread_id +mysql> CREATE TABLE test (a int not null auto_increment, + primary key (a), key(b)) + TYPE=MyISAM SELECT b,c from test2; @end example -Each connection to @code{mysqld} runs in a separate thread. You can see -which threads are running with the @code{SHOW PROCESSLIST} command and kill -a thread with the @code{KILL thread_id} command. +This will create a @code{MyISAM} table with three columns, a, b, and c. +Notice that the columns from the @code{SELECT} statement are appended to +the right side of the table, not overlapped onto it. Take the following +example: -If you have the @strong{process} privilege, you can see and kill all threads. -Otherwise, you can see and kill only your own threads. +@example +mysql> select * from foo; ++---+ +| n | ++---+ +| 1 | ++---+ -You can also use the @code{mysqladmin processlist} and @code{mysqladmin kill} -commands to examine and kill threads. +mysql> create table bar (m int) select n from foo; +Query OK, 1 row affected (0.02 sec) +Records: 1 Duplicates: 0 Warnings: 0 + +mysql> select * from bar; ++------+---+ +| m | n | ++------+---+ +| NULL | 1 | ++------+---+ +1 row in set (0.00 sec) +@end example -When you do a @code{KILL}, a thread specific @code{kill flag} is set for -the thread. +For each row in table @code{foo}, a row is inserted in @code{bar} with +the values from @code{foo} and default values for the new columns. -In most cases it may take some time for the thread to die as the kill -flag is only checked at specific intervals. +If any errors occur while copying the data to the table, it will +automatically be deleted. -@itemize @bullet +To ensure that the update log/binary log can be used to re-create the +original tables, @strong{MySQL} will not allow concurrent inserts during +@code{CREATE TABLE .... SELECT}. @item -In @code{SELECT}, @code{ORDER BY} and @code{GROUP BY} loops, the flag is -checked after reading a block of rows. If the kill flag is set the -statement is aborted +The @code{RAID_TYPE} option will help you to break the 2G/4G limit for +the MyISAM data file (not the index file) on +operating systems that don't support big files. You can get also more speed +from the I/O bottleneck by putting @code{RAID} directories on different +physical disks. @code{RAID_TYPE} will work on any OS, as long as you have +configured @strong{MySQL} with @code{--with-raid}. For now the only allowed +@code{RAID_TYPE} is @code{STRIPED} (@code{1} and @code{RAID0} are aliases +for this). + +If you specify @code{RAID_TYPE=STRIPED} for a @code{MyISAM} table, +@code{MyISAM} will create @code{RAID_CHUNKS} subdirectories named 00, +01, 02 in the database directory. In each of these directories +@code{MyISAM} will create a @code{table_name.MYD}. When writing data +to the data file, the @code{RAID} handler will map the first +@code{RAID_CHUNKSIZE} *1024 bytes to the first file, the next +@code{RAID_CHUNKSIZE} *1024 bytes to the next file and so on. @item -When doing an @code{ALTER TABLE} the kill flag is checked before each block of -rows are read from the original table. If the kill flag was set the command -is aborted and the temporary table is deleted. +@code{UNION} is used when you want to use a collection of identical +tables as one. This only works with MERGE tables. @xref{MERGE}. + +For the moment you need to have @code{SELECT}, @code{UPDATE}, and +@code{DELETE} privileges on the tables you map to a @code{MERGE} table. +All mapped tables must be in the same database as the @code{MERGE} table. @item -When doing an @code{UPDATE TABLE} and @code{DELETE TABLE}, the kill flag -is checked after each block read and after each updated or delete -row. If the kill flag is set the statement is aborted. Note that if you -are not using transactions, the changes will not be rolled back! +In the created table the @code{PRIMARY} key will be placed first, followed +by all @code{UNIQUE} keys and then the normal keys. This helps the +@strong{MySQL} optimizer to prioritize which key to use and also more quickly +detect duplicated @code{UNIQUE} keys. + @item -@code{GET_LOCK()} will abort with @code{NULL}. +By using @code{DATA DIRECTORY="directory"} or @code{INDEX +DIRECTORY="directory"} you can specify where the table handler should +put it's table and index files. This only works for @code{MyISAM} tables +in @code{MySQL} 4.0, when you are not using the @code{--skip-symlink} +option. @xref{Symbolic links to tables}. + +@end itemize + +@cindex silent column changes +@node Silent column changes, , CREATE TABLE, CREATE TABLE +@subsection Silent Column Specification Changes + +In some cases, @strong{MySQL} silently changes a column specification from +that given in a @code{CREATE TABLE} statement. (This may also occur with +@code{ALTER TABLE}.): + +@itemize @bullet @item -An @code{INSERT DELAYED} thread will quickly flush all rows it has in -memory and die. +@code{VARCHAR} columns with a length less than four are changed to +@code{CHAR}. + @item -If the thread is in the table lock handler (state: @code{Locked}), -the table lock will be quickly aborted. +If any column in a table has a variable length, the entire row is +variable-length as a result. Therefore, if a table contains any +variable-length columns (@code{VARCHAR}, @code{TEXT}, or @code{BLOB}), +all @code{CHAR} columns longer than three characters are changed to +@code{VARCHAR} columns. This doesn't affect how you use the columns in +any way; in @strong{MySQL}, @code{VARCHAR} is just a different way to +store characters. @strong{MySQL} performs this conversion because it +saves space and makes table operations faster. @xref{Table types}. + @item -If the thread is waiting for free disk space in a @code{write} call, the -write is aborted with an disk full error message. +@code{TIMESTAMP} display sizes must be even and in the range from 2 to 14. +If you specify a display size of 0 or greater than 14, the size is coerced +to 14. Odd-valued sizes in the range from 1 to 13 are coerced +to the next higher even number. + +@item +You cannot store a literal @code{NULL} in a @code{TIMESTAMP} column; setting +it to @code{NULL} sets it to the current date and time. Because +@code{TIMESTAMP} columns behave this way, the @code{NULL} and @code{NOT NULL} +attributes do not apply in the normal way and are ignored if you specify +them. @code{DESCRIBE tbl_name} always reports that a @code{TIMESTAMP} +column may be assigned @code{NULL} values. + +@item +@strong{MySQL} maps certain column types used by other SQL database vendors +to @strong{MySQL} types. @xref{Other-vendor column types}. @end itemize -@findex SHOW DATABASES -@findex SHOW TABLES -@findex SHOW COLUMNS -@findex SHOW FIELDS -@findex SHOW INDEX -@findex SHOW KEYS -@findex SHOW STATUS -@findex SHOW VARIABLES -@findex SHOW PROCESSLIST -@findex SHOW TABLE STATUS -@findex SHOW GRANTS -@findex SHOW CREATE TABLE -@findex SHOW MASTER STATUS -@findex SHOW MASTER LOGS -@findex SHOW SLAVE STATUS -@node SHOW, EXPLAIN, KILL, Reference -@section @code{SHOW} Syntax +If you want to see whether or not @strong{MySQL} used a column type other +than the one you specified, issue a @code{DESCRIBE tbl_name} statement after +creating or altering your table. + +@cindex @code{myisampack} +Certain other column type changes may occur if you compress a table +using @code{myisampack}. @xref{Compressed format}. + +@findex ALTER TABLE +@node ALTER TABLE, RENAME TABLE, CREATE TABLE, Reference +@section @code{ALTER TABLE} Syntax @example - SHOW DATABASES [LIKE wild] -or SHOW [OPEN] TABLES [FROM db_name] [LIKE wild] -or SHOW [FULL] COLUMNS FROM tbl_name [FROM db_name] [LIKE wild] -or SHOW INDEX FROM tbl_name [FROM db_name] -or SHOW TABLE STATUS [FROM db_name] [LIKE wild] -or SHOW STATUS [LIKE wild] -or SHOW VARIABLES [LIKE wild] -or SHOW LOGS -or SHOW [FULL] PROCESSLIST -or SHOW GRANTS FOR user -or SHOW CREATE TABLE table_name -or SHOW MASTER STATUS -or SHOW MASTER LOGS -or SHOW SLAVE STATUS -@end example +ALTER [IGNORE] TABLE tbl_name alter_spec [, alter_spec ...] -@code{SHOW} provides information about databases, tables, columns, or -status information about the server. If the @code{LIKE wild} part is -used, the @code{wild} string can be a string that uses the SQL @samp{%} -and @samp{_} wild-card characters. +alter_specification: + ADD [COLUMN] create_definition [FIRST | AFTER column_name ] + or ADD [COLUMN] (create_definition, create_definition,...) + or ADD INDEX [index_name] (index_col_name,...) + or ADD PRIMARY KEY (index_col_name,...) + or ADD UNIQUE [index_name] (index_col_name,...) + or ADD FULLTEXT [index_name] (index_col_name,...) + or ADD [CONSTRAINT symbol] FOREIGN KEY index_name (index_col_name,...) + [reference_definition] + or ALTER [COLUMN] col_name @{SET DEFAULT literal | DROP DEFAULT@} + or CHANGE [COLUMN] old_col_name create_definition + or MODIFY [COLUMN] create_definition + or DROP [COLUMN] col_name + or DROP PRIMARY KEY + or DROP INDEX index_name + or RENAME [TO] new_tbl_name + or ORDER BY col + or table_options +@end example -@findex SHOW DATABASES -@findex SHOW TABLES -@findex SHOW COLUMNS -@findex SHOW FIELDS -@findex SHOW INDEX -@findex SHOW KEYS -@menu -* SHOW DATABASE INFO:: -* SHOW TABLE STATUS:: -* SHOW STATUS:: -* SHOW VARIABLES:: -* SHOW LOGS:: -* SHOW PROCESSLIST:: -* SHOW GRANTS:: -* SHOW CREATE TABLE:: -@end menu +@code{ALTER TABLE} allows you to change the structure of an existing table. +For example, you can add or delete columns, create or destroy indexes, change +the type of existing columns, or rename columns or the table itself. You can +also change the comment for the table and type of the table. +@xref{CREATE TABLE, , @code{CREATE TABLE}}. -@cindex displaying, information, @code{SHOW} -@node SHOW DATABASE INFO, SHOW TABLE STATUS, SHOW, SHOW -@subsection @code{SHOW} Information About Databases, Tables, Columns, and Indexes +If you use @code{ALTER TABLE} to change a column specification but +@code{DESCRIBE tbl_name} indicates that your column was not changed, it is +possible that @strong{MySQL} ignored your modification for one of the reasons +described in @ref{Silent column changes}. For example, if you try to change +a @code{VARCHAR} column to @code{CHAR}, @strong{MySQL} will still use +@code{VARCHAR} if the table contains other variable-length columns. -You can use @code{db_name.tbl_name} as an alternative to the @code{tbl_name -FROM db_name} syntax. These two statements are equivalent: +@code{ALTER TABLE} works by making a temporary copy of the original table. +The alteration is performed on the copy, then the original table is +deleted and the new one is renamed. This is done in such a way that +all updates are automatically redirected to the new table without +any failed updates. While @code{ALTER TABLE} is executing, the original +table is readable by other clients. Updates and writes to the table +are stalled until the new table is ready. -@example -mysql> SHOW INDEX FROM mytable FROM mydb; -mysql> SHOW INDEX FROM mydb.mytable; -@end example +Note that if you use any other option to @code{ALTER TABLE} than +@code{RENAME}, @strong{MySQL} will always create a temporary table, even +if the data wouldn't strictly need to be copied (like when you change the +name of a column). We plan to fix this in the future, but as one doesn't +normally do @code{ALTER TABLE} that often this isn't that high on our TODO. -@code{SHOW DATABASES} lists the databases on the @strong{MySQL} server -host. You can also get this list using the @code{mysqlshow} command. +@itemize @bullet +@item +To use @code{ALTER TABLE}, you need @strong{ALTER}, @strong{INSERT}, +and @strong{CREATE} privileges on the table. -@code{SHOW TABLES} lists the tables in a given database. You can also -get this list using the @code{mysqlshow db_name} command. +@item +@code{IGNORE} is a @strong{MySQL} extension to ANSI SQL92. +It controls how @code{ALTER TABLE} works if there are duplicates on +unique keys in the new table. +If @code{IGNORE} isn't specified, the copy is aborted and rolled back. +If @code{IGNORE} is specified, then for rows with duplicates on a unique +key, only the first row is used; the others are deleted. -@strong{NOTE:} If a user doesn't have any privileges for a table, the table -will not show up in the output from @code{SHOW TABLES} or @code{mysqlshow -db_name}. +@item +You can issue multiple @code{ADD}, @code{ALTER}, @code{DROP}, and +@code{CHANGE} clauses in a single @code{ALTER TABLE} statement. This is a +@strong{MySQL} extension to ANSI SQL92, which allows only one of each clause +per @code{ALTER TABLE} statement. -@code{SHOW OPEN TABLES} lists the tables that are currently open in -the table cache. @xref{Table cache}. The @code{Comment} field tells -how many times the table is @code{cached} and @code{in_use}. +@item +@code{CHANGE col_name}, @code{DROP col_name}, and @code{DROP +INDEX} are @strong{MySQL} extensions to ANSI SQL92. -@code{SHOW COLUMNS} lists the columns in a given table. If you specify -the @code{FULL} option, you will also get the privileges you have for -each column. If the column types are different than you expect them to -be based on a @code{CREATE TABLE} statement, note that @strong{MySQL} -sometimes changes column types. @xref{Silent column changes}. +@item +@code{MODIFY} is an Oracle extension to @code{ALTER TABLE}. -The @code{DESCRIBE} statement provides information similar to -@code{SHOW COLUMNS}. -@xref{DESCRIBE, , @code{DESCRIBE}}. +@item +The optional word @code{COLUMN} is a pure noise word and can be omitted. -@code{SHOW FIELDS} is a synonym for @code{SHOW COLUMNS}, and -@code{SHOW KEYS} is a synonym for @code{SHOW INDEX}. You can also -list a table's columns or indexes with @code{mysqlshow db_name tbl_name} -or @code{mysqlshow -k db_name tbl_name}. +@item +If you use @code{ALTER TABLE tbl_name RENAME TO new_name} without any other +options, @strong{MySQL} simply renames the files that correspond to the table +@code{tbl_name}. There is no need to create the temporary table. +@xref{RENAME TABLE,, @code{RENAME TABLE}}. -@code{SHOW INDEX} returns the index information in a format that closely -resembles the @code{SQLStatistics} call in ODBC. The following columns -are returned: +@item +@code{create_definition} clauses use the same syntax for @code{ADD} and +@code{CHANGE} as for @code{CREATE TABLE}. Note that this syntax includes +the column name, not just the column type. +@xref{CREATE TABLE, , @code{CREATE TABLE}}. -@multitable @columnfractions .35 .65 -@item @strong{Column} @tab @strong{Meaning} -@item @code{Table} @tab Name of the table. -@item @code{Non_unique} @tab 0 if the index can't contain duplicates. -@item @code{Key_name} @tab Name of the index. -@item @code{Seq_in_index} @tab Column sequence number in index, - starting with 1. -@item @code{Column_name} @tab Column name. -@item @code{Collation} @tab How the column is sorted in the index. - In @strong{MySQL}, this can have values - @samp{A} (Ascending) or @code{NULL} (Not - sorted). -@item @code{Cardinality} @tab Number of unique values in the index. - This is updated by running - @code{isamchk -a}. -@item @code{Sub_part} @tab Number of indexed characters if the - column is only partly indexed. - @code{NULL} if the entire key is indexed. -@item @code{Comment} @tab Various remarks. For now, it tells - whether index is FULLTEXT or not. -@end multitable +@item +You can rename a column using a @code{CHANGE old_col_name create_definition} +clause. To do so, specify the old and new column names and the type that +the column currently has. For example, to rename an @code{INTEGER} column +from @code{a} to @code{b}, you can do this: -Note that as the @code{Cardinality} is counted based on statistics -stored as integers, it's not necessarily accurate for small tables. +@example +mysql> ALTER TABLE t1 CHANGE a b INTEGER; +@end example -@cindex displaying, table status -@cindex tables, displaying status -@cindex status, tables -@node SHOW TABLE STATUS, SHOW STATUS, SHOW DATABASE INFO, SHOW -@subsection @code{SHOW TABLE STATUS} +If you want to change a column's type but not the name, @code{CHANGE} +syntax still requires two column names even if they are the same. For +example: @example -SHOW TABLE STATUS [FROM db_name] [LIKE wild] +mysql> ALTER TABLE t1 CHANGE b b BIGINT NOT NULL; @end example -@code{SHOW TABLE STATUS} (new in Version 3.23) works likes @code{SHOW -STATUS}, but provides a lot of information about each table. You can -also get this list using the @code{mysqlshow --status db_name} command. -The following columns are returned: +However, as of @strong{MySQL} Version 3.22.16a, you can also use @code{MODIFY} +to change a column's type without renaming it: -@multitable @columnfractions .30 .70 -@item @strong{Column} @tab @strong{Meaning} -@item @code{Name} @tab Name of the table. -@item @code{Type} @tab Type of table. @xref{Table types}. -@item @code{Row_format} @tab The row storage format (Fixed, Dynamic, or Compressed). -@item @code{Rows} @tab Number of rows. -@item @code{Avg_row_length} @tab Average row length. -@item @code{Data_length} @tab Length of the data file. -@item @code{Max_data_length} @tab Max length of the data file. -@item @code{Index_length} @tab Length of the index file. -@item @code{Data_free} @tab Number of allocated but not used bytes. -@item @code{Auto_increment} @tab Next autoincrement value. -@item @code{Create_time} @tab When the table was created. -@item @code{Update_time} @tab When the data file was last updated. -@item @code{Check_time} @tab When the table was last checked. -@item @code{Create_options} @tab Extra options used with @code{CREATE TABLE}. -@item @code{Comment} @tab The comment used when creating the table (or some information why @strong{MySQL} couldn't access the table information). -@end multitable +@example +mysql> ALTER TABLE t1 MODIFY b BIGINT NOT NULL; +@end example -@code{InnoDB} tables will report the free space in the tablespace -in the table comment. +@item +If you use @code{CHANGE} or @code{MODIFY} to shorten a column for which +an index exists on part of the column (for instance, if you have an index +on the first 10 characters of a @code{VARCHAR} column), you cannot make +the column shorter than the number of characters that are indexed. -@node SHOW STATUS, SHOW VARIABLES, SHOW TABLE STATUS, SHOW -@subsection @code{SHOW STATUS} +@item +When you change a column type using @code{CHANGE} or @code{MODIFY}, +@strong{MySQL} tries to convert data to the new type as well as possible. -@cindex @code{mysqladmin} -@code{SHOW STATUS} provides server status information -(like @code{mysqladmin extended-status}). The output resembles that shown -below, though the format and numbers probably differ: +@item +In @strong{MySQL} Version 3.22 or later, you can use @code{FIRST} or +@code{ADD ... AFTER col_name} to add a column at a specific position within +a table row. The default is to add the column last. -@example -+--------------------------+------------+ -| Variable_name | Value | -+--------------------------+------------+ -| Aborted_clients | 0 | -| Aborted_connects | 0 | -| Bytes_received | 155372598 | -| Bytes_sent | 1176560426 | -| Connections | 30023 | -| Created_tmp_disk_tables | 0 | -| Created_tmp_tables | 8340 | -| Created_tmp_files | 60 | -| Delayed_insert_threads | 0 | -| Delayed_writes | 0 | -| Delayed_errors | 0 | -| Flush_commands | 1 | -| Handler_delete | 462604 | -| Handler_read_first | 105881 | -| Handler_read_key | 27820558 | -| Handler_read_next | 390681754 | -| Handler_read_prev | 6022500 | -| Handler_read_rnd | 30546748 | -| Handler_read_rnd_next | 246216530 | -| Handler_update | 16945404 | -| Handler_write | 60356676 | -| Key_blocks_used | 14955 | -| Key_read_requests | 96854827 | -| Key_reads | 162040 | -| Key_write_requests | 7589728 | -| Key_writes | 3813196 | -| Max_used_connections | 0 | -| Not_flushed_key_blocks | 0 | -| Not_flushed_delayed_rows | 0 | -| Open_tables | 1 | -| Open_files | 2 | -| Open_streams | 0 | -| Opened_tables | 44600 | -| Questions | 2026873 | -| Select_full_join | 0 | -| Select_full_range_join | 0 | -| Select_range | 99646 | -| Select_range_check | 0 | -| Select_scan | 30802 | -| Slave_running | OFF | -| Slave_open_temp_tables | 0 | -| Slow_launch_threads | 0 | -| Slow_queries | 0 | -| Sort_merge_passes | 30 | -| Sort_range | 500 | -| Sort_rows | 30296250 | -| Sort_scan | 4650 | -| Table_locks_immediate | 1920382 | -| Table_locks_waited | 0 | -| Threads_cached | 0 | -| Threads_created | 30022 | -| Threads_connected | 1 | -| Threads_running | 1 | -| Uptime | 80380 | -+--------------------------+------------+ -@end example +@findex ALTER COLUMN +@item +@code{ALTER COLUMN} specifies a new default value for a column +or removes the old default value. +If the old default is removed and the column can be @code{NULL}, the new +default is @code{NULL}. If the column cannot be @code{NULL}, @strong{MySQL} +assigns a default value, as described in +@ref{CREATE TABLE, , @code{CREATE TABLE}}. -@cindex variables, status -The status variables listed above have the following meaning: +@findex DROP INDEX +@item +@code{DROP INDEX} removes an index. This is a @strong{MySQL} extension to +ANSI SQL92. @xref{DROP INDEX}. -@multitable @columnfractions .35 .65 -@item @strong{Variable} @tab @strong{Meaning} -@item @code{Aborted_clients} @tab Number of connections aborted because the client died without closing the connection properly. @xref{Communication errors}. -@item @code{Aborted_connects} @tab Number of tries to connect to the @strong{MySQL} server that failed. @xref{Communication errors}. -@item @code{Bytes_received} @tab Number of bytes received from all clients. -@item @code{Bytes_sent} @tab Number of bytes sent to all clients. -@item @code{Connections} @tab Number of connection attempts to the @strong{MySQL} server. -@item @code{Created_tmp_disk_tables} @tab Number of implicit temporary tables on disk created while executing statements. -@item @code{Created_tmp_tables} @tab Number of implicit temporary tables in memory created while executing statements. -@item @code{Created_tmp_files} @tab How many temporary files @code{mysqld} have created. -@item @code{Delayed_insert_threads} @tab Number of delayed insert handler threads in use. -@item @code{Delayed_writes} @tab Number of rows written with @code{INSERT DELAYED}. -@item @code{Delayed_errors} @tab Number of rows written with @code{INSERT DELAYED} for which some error occurred (probably @code{duplicate key}). -@item @code{Flush_commands} @tab Number of executed @code{FLUSH} commands. -@item @code{Handler_delete} @tab Number of times a row was deleted from a table. -@item @code{Handler_read_first} @tab Number of times the first entry was read from an index. -If this is high, it suggests that the server is doing a lot of full index scans, for example, -@code{SELECT col1 FROM foo}, assuming that col1 is indexed. -@item @code{Handler_read_key} @tab Number of requests to read a row based on a key. If this -is high, it is a good indication that your queries and tables are properly indexed. -@item @code{Handler_read_next} @tab Number of requests to read next row in key order. This -will be incremented if you are querying an index column with a range constraint. This also -will be incremented if you are doing an index scan. -@item @code{Handler_read_rnd} @tab Number of requests to read a row based on a fixed position. -This will be high if you are doing a lot of queries that require sorting of the result. -@item @code{Handler_read_rnd_next} @tab Number of requests to read the next row in the datafile. -This will be high if you are doing a lot of table scans. Generally this suggests that your tables -are not properly indexed or that your queries are not written to take advantage of the indexes you -have. -@item @code{Handler_update} @tab Number of requests to update a row in a table. -@item @code{Handler_write} @tab Number of requests to insert a row in a table. -@item @code{Key_blocks_used} @tab The number of used blocks in the key cache. -@item @code{Key_read_requests} @tab The number of requests to read a key block from the cache. -@item @code{Key_reads} @tab The number of physical reads of a key block from disk. -@item @code{Key_write_requests} @tab The number of requests to write a key block to the cache. -@item @code{Key_writes} @tab The number of physical writes of a key block to disk. -@item @code{Max_used_connections} @tab The maximum number of connections in use simultaneously. -@item @code{Not_flushed_key_blocks} @tab Keys blocks in the key cache that has changed but hasn't yet been flushed to disk. -@item @code{Not_flushed_delayed_rows} @tab Number of rows waiting to be written in @code{INSERT DELAY} queues. -@item @code{Open_tables} @tab Number of tables that are open. -@item @code{Open_files} @tab Number of files that are open. -@item @code{Open_streams} @tab Number of streams that are open (used mainly for logging). -@item @code{Opened_tables} @tab Number of tables that have been opened. -@item @code{Select_full_join} @tab Number of joins without keys (Should be 0). -@item @code{Select_full_range_join} @tab Number of joins where we used a range search on reference table. -@item @code{Select_range} @tab Number of joins where we used ranges on the first table. (It's normally not critical even if this is big.) -@item @code{Select_scan} @tab Number of joins where we scanned the first table. -@item @code{Select_range_check} @tab Number of joins without keys where we check for key usage after each row (Should be 0). -@item @code{Questions} @tab Number of queries sent to the server. -@item @code{Slave_open_temp_tables} @tab Number of temporary tables currently -open by the slave thread -@item @code{Slow_launch_threads} @tab Number of threads that have taken more than @code{slow_launch_time} to connect. -@item @code{Slow_queries} @tab Number of queries that have taken more than @code{long_query_time}. @xref{Slow query log}. -@item @code{Sort_merge_passes} @tab Number of merges the sort has to do. If this value is large you should consider increasing @code{sort_buffer}. -@item @code{Sort_range} @tab Number of sorts that where done with ranges. -@item @code{Sort_rows} @tab Number of sorted rows. -@item @code{Sort_scan} @tab Number of sorts that where done by scanning the table. -@item @code{Table_locks_immediate} @tab Number of times a table lock was -acquired immediately. Available after 3.23.33. -@item @code{Table_locks_waited} @tab Number of times a table lock could not -be acquired immediately and a wait was needed. If this is high, and you -have performance problems, you should first optimize your queries, and then -either split your table(s) or use replication. Available after 3.23.33. -@item @code{Threads_cached} @tab Number of threads in the thread cache. -@item @code{Threads_connected} @tab Number of currently open connections. -@item @code{Threads_created} @tab Number of threads created to handle connections. -@item @code{Threads_running} @tab Number of threads that are not sleeping. -@item @code{Uptime} @tab How many seconds the server has been up. -@end multitable +@item +If columns are dropped from a table, the columns are also removed from any +index of which they are a part. If all columns that make up an index are +dropped, the index is dropped as well. -Some comments about the above: +@item +If a table contains only one column, the column cannot be dropped. +If what you intend is to remove the table, use @code{DROP TABLE} instead. -@itemize @bullet +@findex DROP PRIMARY KEY @item -If @code{Opened_tables} is big, then your @code{table_cache} -variable is probably too small. +@code{DROP PRIMARY KEY} drops the primary index. If no such +index exists, it drops the first @code{UNIQUE} index in the table. +(@strong{MySQL} marks the first @code{UNIQUE} key as the @code{PRIMARY KEY} +if no @code{PRIMARY KEY} was specified explicitly.) + +@findex ORDER BY @item -If @code{key_reads} is big, then your @code{key_cache} is probably too -small. The cache hit rate can be calculated with -@code{key_reads}/@code{key_read_requests}. +@code{ORDER BY} allows you to create the new table with the rows in a +specific order. Note that the table will not remain in this order after +inserts and deletes. In some cases, it may make sorting easier for +@strong{MySQL} if the table is in order by the column that you wish to +order it by later. This option is mainly useful when you know that you +are mostly going to query the rows in a certain order; By using this +option after big changes to the table, you may be able to get higher +performance. + +@findex ALTER TABLE @item -If @code{Handler_read_rnd} is big, then you probably have a lot of -queries that require @strong{MySQL} to scan whole tables or you have -joins that don't use keys properly. +If you use @code{ALTER TABLE} on a @code{MyISAM} table, all non-unique +indexes are created in a separate batch (like in @code{REPAIR}). +This should make @code{ALTER TABLE} much faster when you have many indexes. + @item -If @code{Threads_created} is big, you may want to increase the -@code{thread_cache_size} variable. +@findex mysql_info() +With the C API function @code{mysql_info()}, you can find out how many +records were copied, and (when @code{IGNORE} is used) how many records were +deleted due to duplication of unique key values. + +@item +@cindex foreign keys +@cindex references +The @code{FOREIGN KEY}, @code{CHECK}, and @code{REFERENCES} clauses don't +actually do anything. The syntax for them is provided only for compatibility, +to make it easier to port code from other SQL servers and to run applications +that create tables with references. +@xref{Missing functions}. @end itemize -@node SHOW VARIABLES, SHOW LOGS, SHOW STATUS, SHOW -@subsection @code{SHOW VARIABLES} +Here is an example that shows some of the uses of @code{ALTER TABLE}. We +begin with a table @code{t1} that is created as shown below: @example -SHOW VARIABLES [LIKE wild] +mysql> CREATE TABLE t1 (a INTEGER,b CHAR(10)); @end example -@code{SHOW VARIABLES} shows the values of some @strong{MySQL} system -variables. You can also get this information using the @code{mysqladmin -variables} command. If the default values are unsuitable, you can set most -of these variables using command-line options when @code{mysqld} starts up. -@xref{Command-line options}. - -The output resembles that shown below, though the format and numbers may -differ somewhat: +To rename the table from @code{t1} to @code{t2}: @example -+-------------------------+---------------------------+ -| Variable_name | Value | -+-------------------------+---------------------------+ -| ansi_mode | OFF | -| back_log | 50 | -| basedir | /my/monty/ | -| bdb_cache_size | 16777216 | -| bdb_log_buffer_size | 32768 | -| bdb_home | /my/monty/data/ | -| bdb_max_lock | 10000 | -| bdb_logdir | | -| bdb_shared_data | OFF | -| bdb_tmpdir | /tmp/ | -| binlog_cache_size | 32768 | -| concurrent_insert | ON | -| connect_timeout | 5 | -| datadir | /my/monty/data/ | -| delay_key_write | ON | -| delayed_insert_limit | 100 | -| delayed_insert_timeout | 300 | -| delayed_queue_size | 1000 | -| flush | OFF | -| flush_time | 0 | -| have_bdb | YES | -| have_innodb | YES | -| have_raid | YES | -| have_ssl | NO | -| init_file | | -| interactive_timeout | 28800 | -| join_buffer_size | 131072 | -| key_buffer_size | 16776192 | -| language | /my/monty/share/english/ | -| large_files_support | ON | -| log | OFF | -| log_update | OFF | -| log_bin | OFF | -| log_slave_updates | OFF | -| long_query_time | 10 | -| low_priority_updates | OFF | -| lower_case_table_names | 0 | -| max_allowed_packet | 1048576 | -| max_binlog_cache_size | 4294967295 | -| max_connections | 100 | -| max_connect_errors | 10 | -| max_delayed_threads | 20 | -| max_heap_table_size | 16777216 | -| max_join_size | 4294967295 | -| max_sort_length | 1024 | -| max_tmp_tables | 32 | -| max_write_lock_count | 4294967295 | -| myisam_recover_options | DEFAULT | -| myisam_sort_buffer_size | 8388608 | -| net_buffer_length | 16384 | -| net_read_timeout | 30 | -| net_retry_count | 10 | -| net_write_timeout | 60 | -| open_files_limit | 0 | -| pid_file | /my/monty/data/donna.pid | -| port | 3306 | -| protocol_version | 10 | -| record_buffer | 131072 | -| query_buffer_size | 0 | -| safe_show_database | OFF | -| server_id | 0 | -| skip_locking | ON | -| skip_networking | OFF | -| skip_show_database | OFF | -| slow_launch_time | 2 | -| socket | /tmp/mysql.sock | -| sort_buffer | 2097116 | -| table_cache | 64 | -| table_type | MYISAM | -| thread_cache_size | 4 | -| thread_stack | 65536 | -| tmp_table_size | 1048576 | -| tmpdir | /tmp/ | -| version | 3.23.29a-gamma-debug | -| wait_timeout | 28800 | -+-------------------------+---------------------------+ +mysql> ALTER TABLE t1 RENAME t2; @end example -Each option is described below. Values for buffer sizes, lengths, and stack -sizes are given in bytes. You can specify values with a suffix of @samp{K} -or @samp{M} to indicate kilobytes or megabytes. For example, @code{16M} -indicates 16 megabytes. The case of suffix letters does not matter; -@code{16M} and @code{16m} are equivalent: +To change column @code{a} from @code{INTEGER} to @code{TINYINT NOT NULL} +(leaving the name the same), and to change column @code{b} from +@code{CHAR(10)} to @code{CHAR(20)} as well as renaming it from @code{b} to +@code{c}: -@cindex variables, values -@table @code -@item @code{ansi_mode}. -Is @code{ON} if @code{mysqld} was started with @code{--ansi}. -@xref{ANSI mode}. +@example +mysql> ALTER TABLE t2 MODIFY a TINYINT NOT NULL, CHANGE b c CHAR(20); +@end example -@item @code{back_log} -The number of outstanding connection requests @strong{MySQL} can have. This -comes into play when the main @strong{MySQL} thread gets @strong{VERY} -many connection requests in a very short time. It then takes some time -(although very little) for the main thread to check the connection and start -a new thread. The @code{back_log} value indicates how many requests can be -stacked during this short time before @strong{MySQL} momentarily stops -answering new requests. You need to increase this only if you expect a large -number of connections in a short period of time. +To add a new @code{TIMESTAMP} column named @code{d}: -In other words, this value is the size of the listen queue for incoming -TCP/IP connections. Your operating system has its own limit on the size -of this queue. The manual page for the Unix @code{listen(2)} system -call should have more details. Check your OS documentation for the -maximum value for this variable. Attempting to set @code{back_log} -higher than your operating system limit will be ineffective. +@example +mysql> ALTER TABLE t2 ADD d TIMESTAMP; +@end example -@item @code{basedir} -The value of the @code{--basedir} option. +To add an index on column @code{d}, and make column @code{a} the primary key: -@item @code{bdb_cache_size} -The buffer that is allocated to cache index and rows for @code{BDB} -tables. If you don't use @code{BDB} tables, you should start -@code{mysqld} with @code{--skip-bdb} to not waste memory for this -cache. +@example +mysql> ALTER TABLE t2 ADD INDEX (d), ADD PRIMARY KEY (a); +@end example -@item @code{bdb_log_buffer_size} -The buffer that is allocated to cache index and rows for @code{BDB} -tables. If you don't use @code{BDB} tables, you should set this to 0 or -start @code{mysqld} with @code{--skip-bdb} to not waste memory for this -cache. +To remove column @code{c}: -@item @code{bdb_home} -The value of the @code{--bdb-home} option. +@example +mysql> ALTER TABLE t2 DROP COLUMN c; +@end example -@item @code{bdb_max_lock} -The maximum number of locks (1000 by default) you can have active on a -BDB table. You should increase this if you get errors of type @code{bdb: -Lock table is out of available locks} or @code{Got error 12 from ...} -when you have do long transactions or when @code{mysqld} has to examine -a lot of rows to calculate the query. +To add a new @code{AUTO_INCREMENT} integer column named @code{c}: -@item @code{bdb_logdir} -The value of the @code{--bdb-logdir} option. +@example +mysql> ALTER TABLE t2 ADD c INT UNSIGNED NOT NULL AUTO_INCREMENT, + ADD INDEX (c); +@end example -@item @code{bdb_shared_data} -Is @code{ON} if you are using @code{--bdb-shared-data}. +Note that we indexed @code{c}, because @code{AUTO_INCREMENT} columns must be +indexed, and also that we declare @code{c} as @code{NOT NULL}, because +indexed columns cannot be @code{NULL}. -@item @code{bdb_tmpdir} -The value of the @code{--bdb-tmpdir} option. +When you add an @code{AUTO_INCREMENT} column, column values are filled in +with sequence numbers for you automatically. You can set the first +sequence number by executing @code{SET INSERT_ID=#} before +@code{ALTER TABLE} or using the @code{AUTO_INCREMENT = #} table option. +@xref{SET OPTION}. -@item @code{binlog_cache_size}. The size of the cache to hold the SQL -statements for the binary log during a transaction. If you often use -big, multi-statement transactions you can increase this to get more -performance. @xref{COMMIT}. +With MyISAM tables, if you don't change the @code{AUTO_INCREMENT} +column, the sequence number will not be affected. If you drop an +@code{AUTO_INCREMENT} column and then add another @code{AUTO_INCREMENT} +column, the numbers will start from 1 again. -@item @code{character_set} -The default character set. +@xref{ALTER TABLE problems}. -@item @code{character_sets} -The supported character sets. +@findex RENAME TABLE +@node RENAME TABLE, DROP TABLE, ALTER TABLE, Reference +@section @code{RENAME TABLE} Syntax -@item @code{concurrent_inserts} -If @code{ON} (the default), @strong{MySQL} will allow you to use @code{INSERT} -on @code{MyISAM} tables at the same time as you run @code{SELECT} queries -on them. You can turn this option off by starting @code{mysqld} with @code{--safe} -or @code{--skip-new}. +@example +RENAME TABLE tbl_name TO new_table_name[, tbl_name2 TO new_table_name2,...] +@end example -@cindex timeout -@item @code{connect_timeout} -The number of seconds the @code{mysqld} server is waiting for a connect -packet before responding with @code{Bad handshake}. +The rename is done atomically, which means that no other thread can +access any of the tables while the rename is running. This makes it +possible to replace a table with an empty one: -@item @code{datadir} -The value of the @code{--datadir} option. +@example +CREATE TABLE new_table (...); +RENAME TABLE old_table TO backup_table, new_table TO old_table; +@end example -@item @code{delay_key_write} -If enabled (is on by default), @strong{MySQL} will honor the -@code{delay_key_write} option @code{CREATE TABLE}. This means that the -key buffer for tables with this option will not get flushed on every -index update, but only when a table is closed. This will speed up -writes on keys a lot, but you should add automatic checking of all tables -with @code{myisamchk --fast --force} if you use this. Note that if you -start @code{mysqld} with the @code{--delay-key-write-for-all-tables} -option this means that all tables will be treated as if they were -created with the @code{delay_key_write} option. You can clear this flag -by starting @code{mysqld} with @code{--skip-new} or @code{--safe-mode}. +The rename is done from left to right, which means that if you want to +swap two tables names, you have to: -@item @code{delayed_insert_limit} -After inserting @code{delayed_insert_limit} rows, the @code{INSERT -DELAYED} handler will check if there are any @code{SELECT} statements -pending. If so, it allows these to execute before continuing. - -@item @code{delayed_insert_timeout} -How long a @code{INSERT DELAYED} thread should wait for @code{INSERT} -statements before terminating. +@example +RENAME TABLE old_table TO backup_table, + new_table TO old_table, + backup_table TO new_table; +@end example -@item @code{delayed_queue_size} -What size queue (in rows) should be allocated for handling @code{INSERT -DELAYED}. If the queue becomes full, any client that does @code{INSERT -DELAYED} will wait until there is room in the queue again. +As long as two databases are on the same disk you can also rename +from one database to another: -@item @code{flush} -This is @code{ON} if you have started @strong{MySQL} with the @code{--flush} -option. +@example +RENAME TABLE current_database.table_name TO other_database.table_name; +@end example -@item @code{flush_time} -If this is set to a non-zero value, then every @code{flush_time} seconds all -tables will be closed (to free up resources and sync things to disk). We -only recommend this option on Win95, Win98, or on systems where you have -very little resources. +When you execute @code{RENAME}, you can't have any locked tables or +active transactions. You must also have the @code{ALTER} and @code{DROP} +privilege on the original table and @code{CREATE} and @code{INSERT} +privilege on the new table. -@item @code{have_bdb} -@code{YES} if @code{mysqld} supports Berkeley DB tables. @code{DISABLED} -if @code{--skip-bdb} is used. -@item @code{have_innodb} -@code{YES} if @code{mysqld} supports InnoDB tables. @code{DISABLED} -if @code{--skip-innodb} is used. -@item @code{have_raid} -@code{YES} if @code{mysqld} supports the @code{RAID} option. -@item @code{have_ssl} -@code{YES} if @code{mysqld} supports SSL (encryption) on the client/server -protocol. +If @strong{MySQL} encounters any errors in a multiple table rename, it +will do a reverse rename for all renamed tables to get everything back +to the original state. -@item @code{init_file} -The name of the file specified with the @code{--init-file} option when -you start the server. This is a file of SQL statements you want the -server to execute when it starts. +@findex DROP TABLE +@node DROP TABLE, OPTIMIZE TABLE, RENAME TABLE, Reference +@section @code{DROP TABLE} Syntax -@item @code{interactive_timeout} -The number of seconds the server waits for activity on an interactive -connection before closing it. An interactive client is defined as a -client that uses the @code{CLIENT_INTERACTIVE} option to -@code{mysql_real_connect()}. See also @code{wait_timeout}. +@example +DROP TABLE [IF EXISTS] tbl_name [, tbl_name,...] [RESTRICT | CASCADE] +@end example -@item @code{join_buffer_size} -The size of the buffer that is used for full joins (joins that do not -use indexes). The buffer is allocated one time for each full join -between two tables. Increase this value to get a faster full join when -adding indexes is not possible. (Normally the best way to get fast joins -is to add indexes.) +@code{DROP TABLE} removes one or more tables. All table data and the table +definition are @emph{removed}, so @strong{be careful} with this command! -@c Make texi2html support index @anchor{Index cache size}. Then change -@c some xrefs to point here -@cindex indexes, block size -@item @code{key_buffer_size} -Index blocks are buffered and are shared by all threads. -@code{key_buffer_size} is the size of the buffer used for index blocks. +In @strong{MySQL} Version 3.22 or later, you can use the keywords +@code{IF EXISTS} to prevent an error from occurring for tables that don't +exist. -Increase this to get better index handling (for all reads and multiple -writes) to as much as you can afford; 64M on a 256M machine that mainly -runs @strong{MySQL} is quite common. If you, however, make this too big -(more than 50% of your total memory?) your system may start to page and -become REALLY slow. Remember that because @strong{MySQL} does not cache -data read, that you will have to leave some room for the OS filesystem -cache. +@code{RESTRICT} and @code{CASCADE} are allowed to make porting easier. +For the moment they don't do anything. -You can check the performance of the key buffer by doing @code{show -status} and examine the variables @code{Key_read_requests}, -@code{Key_reads}, @code{Key_write_requests}, and @code{Key_writes}. The -@code{Key_reads/Key_read_request} ratio should normally be < 0.01. -The @code{Key_write/Key_write_requests} is usually near 1 if you are -using mostly updates/deletes but may be much smaller if you tend to -do updates that affect many at the same time or if you are -using @code{delay_key_write}. @xref{SHOW}. +@strong{NOTE}: @code{DROP TABLE} is not transaction-safe and will +automatically commit any active transactions. -To get even more speed when writing many rows at the same time, use -@code{LOCK TABLES}. @xref{LOCK TABLES, , @code{LOCK TABLES}}. +@cindex tables, defragment +@cindex tables, fragmentation +@findex OPTIMIZE TABLE +@node OPTIMIZE TABLE, CHECK TABLE, DROP TABLE, Reference +@section @code{OPTIMIZE TABLE} Syntax -@item @code{language} -The language used for error messages. +@example +OPTIMIZE TABLE tbl_name[,tbl_name]... +@end example -@item @code{large_file_support} -If @code{mysqld} was compiled with options for big file support. +@code{OPTIMIZE TABLE} should be used if you have deleted a large part of a +table or if you have made many changes to a table with variable-length rows +(tables that have @code{VARCHAR}, @code{BLOB}, or @code{TEXT} columns). +Deleted records are maintained in a linked list and subsequent @code{INSERT} +operations reuse old record positions. You can use @code{OPTIMIZE TABLE} to +reclaim the unused space and to defragment the data file. -@item @code{locked_in_memory} -If @code{mysqld} was locked in memory with @code{--memlock} +For the moment @code{OPTIMIZE TABLE} only works on @strong{MyISAM} and +@code{BDB} tables. For @code{BDB} tables, @code{OPTIMIZE TABLE} is +currently mapped to @code{ANALYZE TABLE}. @xref{ANALYZE TABLE}. -@item @code{log} -If logging of all queries is enabled. +You can get optimize table to work on other table types by starting +@code{mysqld} with @code{--skip-new} or @code{--safe-mode}, but in this +case @code{OPTIMIZE TABLE} is just mapped to @code{ALTER TABLE}. -@item @code{log_update} -If the update log is enabled. +@code{OPTIMIZE TABLE} works the following way: +@itemize @bullet +@item +If the table has deleted or split rows, repair the table. +@item +If the index pages are not sorted, sort them. +@item +If the statistics are not up to date (and the repair couldn't be done +by sorting the index), update them. +@end itemize -@item @code{log_bin} -If the binary log is enabled. +@code{OPTIMIZE TABLE} for @code{MyISAM} tables is equvialent of running +@code{myisamchk --quick --check-changed-tables --sort-index --analyze} +on the table. -@item @code{log_slave_updates} -If the updates from the slave should be logged. +Note that the table is locked during the time @code{OPTIMIZE TABLE} is +running! -@item @code{long_query_time} -If a query takes longer than this (in seconds), the @code{Slow_queries} counter -will be incremented. If you are using @code{--log-slow-queries}, the query -will be logged to the slow query logfile. @xref{Slow query log}. +@findex CHECK TABLE +@node CHECK TABLE, BACKUP TABLE, OPTIMIZE TABLE, Reference +@section @code{CHECK TABLE} Syntax -@item @code{lower_case_table_names} -If set to 1 table names are stored in lowercase on disk. This will enable -you to access the table names case-insensitive also on Unix. -@xref{Name case sensitivity}. +@example +CHECK TABLE tbl_name[,tbl_name...] [option [option...]] -@item @code{max_allowed_packet} -The maximum size of one packet. The message buffer is initialized to -@code{net_buffer_length} bytes, but can grow up to @code{max_allowed_packet} -bytes when needed. This value by default is small, to catch big (possibly -wrong) packets. You must increase this value if you are using big -@code{BLOB} columns. It should be as big as the biggest @code{BLOB} you want -to use. The current protocol limits @code{max_allowed_packet} to 16M. +option = QUICK | FAST | MEDIUM | EXTEND | CHANGED +@end example -@item @code{max_binlog_cache_size} -If a multi-statement transaction requires more than this amount of memory, -one will get the error "Multi-statement transaction required more than -'max_binlog_cache_size' bytes of storage". +@code{CHECK TABLE} only works on @code{MyISAM} tables. On +@code{MyISAM} tables it's the same thing as running @code{myisamchk -m +table_name} on the table. -@item @code{max_binlog_size} -Available after 3.23.33. If a write to the binary (replication) log exceeds -the given value, rotate the logs. You cannot set it to less than 1024 bytes, -or more than 1 GB. Default is 1 GB. +If you don't specify any option @code{MEDIUM} is used. -@item @code{max_connections} -The number of simultaneous clients allowed. Increasing this value increases -the number of file descriptors that @code{mysqld} requires. See below for -comments on file descriptor limits. @xref{Too many connections}. +Checks the table(s) for errors. For @code{MyISAM} tables the key statistics +is updated. The command returns a table with the following columns: -@item @code{max_connect_errors} -If there is more than this number of interrupted connections from a host -this host will be blocked from further connections. You can unblock a host -with the command @code{FLUSH HOSTS}. +@multitable @columnfractions .35 .65 +@item @strong{Column} @tab @strong{Value} +@item Table @tab Table name. +@item Op @tab Always ``check''. +@item Msg_type @tab One of @code{status}, @code{error}, @code{info}, or @code{warning}. +@item Msg_text @tab The message. +@end multitable -@item @code{max_delayed_threads} -Don't start more than this number of threads to handle @code{INSERT DELAYED} -statements. If you try to insert data into a new table after all @code{INSERT -DELAYED} threads are in use, the row will be inserted as if the -@code{DELAYED} attribute wasn't specified. +Note that you can get many rows of information for each checked +table. The last row will be of @code{Msg_type status} and should +normally be @code{OK}. If you don't get @code{OK}, or @code{Not +checked} you should normally run a repair of the table. @xref{Table +maintenance}. @code{Not checked} means that the table the given @code{TYPE} +told @strong{MySQL} that there wasn't any need to check the table. -@item @code{max_heap_table_size} -Don't allow creation of heap tables bigger than this. +The different check types stand for the following: -@item @code{max_join_size} -Joins that are probably going to read more than @code{max_join_size} -records return an error. Set this value if your users tend to perform joins -that lack a @code{WHERE} clause, that take a long time, and that return -millions of rows. +@multitable @columnfractions .20 .80 +@item @strong{Type} @tab @strong{Meaning} +@item @code{QUICK} @tab Don't scan the rows to check for wrong links. +@item @code{FAST} @tab Only check tables which haven't been closed properly. +@item @code{CHANGED} @tab Only check tables which have been changed since last check or haven't been closed properly. +@item @code{MEDIUM} @tab Scan rows to verify that deleted links are ok. This also calculates a key checksum for the rows and verifies this with a calcualted checksum for the keys. +@item @code{EXTENDED} @tab Do a full key lookup for all keys for each row. This ensures that the table is 100 % consistent, but will take a long time! +@end multitable -@item @code{max_sort_length} -The number of bytes to use when sorting @code{BLOB} or @code{TEXT} -values (only the first @code{max_sort_length} bytes of each value -are used; the rest are ignored). +For dynamic sized @code{MyISAM} tables a started check will always +do a @code{MEDIUM} check. For static size rows we skip the row scan +for @code{QUICK} and @code{FAST} as the rows are very seldom corrupted. -@item @code{max_user_connections} -The maximum number of active connections for a single user (0 = no limit). +You can combine check options as in: -@item @code{max_tmp_tables} -(This option doesn't yet do anything.) -Maximum number of temporary tables a client can keep open at the same time. +@example +CHECK TABLE test_table FAST QUICK; +@end example -@item @code{max_write_lock_count} -After this many write locks, allow some read locks to run in between. +Which only would do a quick check on the table if it wasn't closed properly. -@item @code{myisam_recover_options} -The value of the @code{--myisam-recover} option. +@strong{NOTE:} that in some case @code{CHECK TABLE} will change the +table! This happens if the table is marked as 'corrupted' or 'not +closed properly' but @code{CHECK TABLE} didn't find any problems in the +table. In this case @code{CHECK TABLE} will mark the table as ok. -@item @code{myisam_sort_buffer_size} -The buffer that is allocated when sorting the index when doing a -@code{REPAIR} or when creating indexes with @code{CREATE INDEX} or -@code{ALTER TABLE}. +If a table is corrupted, then it's most likely that the problem is in +the indexes and not in the data part. All of the above check types +checks the indexes throughly and should thus find most errors. -@item @code{myisam_max_extra_sort_file_size}. -If the creating of the temporary file for fast index creation would be -this much bigger than using the key cache, then prefer the key cache -method. This is mainly used to force long character keys in large -tables to use the slower key cache method to create the index. -@strong{NOTE} that this parameter is given in megabytes! +If you just want to check a table that you assume is ok, you should use +no check options or the @code{QUICK} option. The later should be used +when you are in a hurry and can take the very small risk that +@code{QUICK} didn't find an error in the data file (In most cases +@strong{MySQL} should find, under normal usage, any error in the data +file. If this happens then the table will be marked as 'corrupted', +in which case the table can't be used until it's repaired). -@item @code{myisam_max_sort_file_size} -The maximum size of the temporary file @strong{MySQL} is allowed to use -while recreating the index (during @code{REPAIR}, @code{ALTER TABLE} -or @code{LOAD DATA INFILE}. If the file size would be bigger than this, -the index will be created through the key cache (which is slower). -@strong{NOTE} that this parameter is given in megabytes! +@code{FAST} and @code{CHANGED} are mostly intended to be used from a +script (for example to be executed from cron) if you want to check your +table from time to time. In most cases you @code{FAST} is to be prefered +over @code{CHANGED}. (The only case when it isn't is when you suspect a +bug you have found a bug in the @code{MyISAM} code.). -@item @code{net_buffer_length} -The communication buffer is reset to this size between queries. This -should not normally be changed, but if you have very little memory, you -can set it to the expected size of a query. (That is, the expected length of -SQL statements sent by clients. If statements exceed this length, the buffer -is automatically enlarged, up to @code{max_allowed_packet} bytes.) +@code{EXTENDED} is only to be used after you have run a normal check but +still get strange errors from a table when @strong{MySQL} tries to +update a row or find a row by key (this is VERY unlikely to happen if a +normal check has succeeded!). -@item @code{net_read_timeout} -Number of seconds to wait for more data from a connection before aborting -the read. Note that when we don't expect data from a connection, the timeout -is defined by @code{write_timeout}. See also @code{slave_read_timeout}. +Some things reported by check table, can't be corrected automatically: -@item @code{net_retry_count} -If a read on a communication port is interrupted, retry this many times -before giving up. This value should be quite high on @code{FreeBSD} as -internal interrupts are sent to all threads. +@itemize @bullet +@item +@code{Found row where the auto_increment column has the value 0}. -@item @code{net_write_timeout} -Number of seconds to wait for a block to be written to a connection before -aborting the write. +This means that you have in the table a row where the +@code{auto_increment} index column contains the value 0. +(It's possible to create a row where the auto_increment column is 0 by +explicitely setting the column to 0 with an @code{UPDATE} statement) -@item @code{open_files_limit} -If this is not 0, then @code{mysqld} will use this value to reserve file -descriptors to use with @code{setrlimit()}. If this value is 0 then -@code{mysqld} will reserve @code{max_connections*5} or -@code{max_connections + table_cache*2} (whichever is larger) number of -files. You should try increasing this if @code{mysqld} gives you the -error 'Too many open files'. +This isn't an error in itself, but could cause trouble if you decide to +dump the table and restore it or do an @code{ALTER TABLE} on the +table. In this case the auto_increment column will change value, +according to the rules of auto_increment columns, which could cause +problems like a duplicate key error. -@item @code{pid_file} -The value of the @code{--pid-file} option. +To get rid of the warning, just execute an @code{UPDATE} statement +to set the column to some other value than 0. +@end itemize -@item @code{port} -The value of the @code{--port} option. -@item @code{protocol_version} -The protocol version used by the @strong{MySQL} server. +@findex BACKUP TABLE +@node BACKUP TABLE, RESTORE TABLE, CHECK TABLE, Reference +@section @code{BACKUP TABLE} Syntax -@item @code{record_buffer} -Each thread that does a sequential scan allocates a buffer of this -size for each table it scans. If you do many sequential scans, you may -want to increase this value. +@example +BACKUP TABLE tbl_name[,tbl_name...] TO '/path/to/backup/directory' +@end example -@item @code{query_buffer_size} -The initial allocation of the query buffer. If most of your queries are -long (like when inserting blobs), you should increase this! +Make a copy of all the table files to the backup directory that are the +minimum needed to restore it. Currenlty only works for @code{MyISAM} +tables. For @code{MyISAM} table, copies @code{.frm} (definition) and +@code{.MYD} (data) files. The index file can be rebuilt from those two. -@item @code{safe_show_databases} -Don't show databases for which the user doesn't have any database or -table privileges. This can improve security if you're concerned about -people being able to see what databases other users have. See also -@code{skip_show_databases}. +Before using this command, please see @xref{Backup}. -@item @code{server_id} -The value of the @code{--server-id} option. +During the backup, read lock will be held for each table, one at time, +as they are being backed up. If you want to backup several tables as +a snapshot, you must first issue @code{LOCK TABLES} obtaining a read +lock for each table in the group. -@item @code{skip_locking} -Is OFF if @code{mysqld} uses external locking. +The command returns a table with the following columns: -@item @code{skip_networking} -Is ON if we only allow local (socket) connections. +@multitable @columnfractions .35 .65 +@item @strong{Column} @tab @strong{Value} +@item Table @tab Table name +@item Op @tab Always ``backup'' +@item Msg_type @tab One of @code{status}, @code{error}, @code{info} or @code{warning}. +@item Msg_text @tab The message. +@end multitable -@item @code{skip_show_databases} -This prevents people from doing @code{SHOW DATABASES} if they don't have -the @code{PROCESS_PRIV} privilege. This can improve security if you're -concerned about people being able to see what databases other users -have. See also @code{safe_show_databases}. +Note that @code{BACKUP TABLE} is only available in @strong{MySQL} +version 3.23.25 and later. -@item @code{slave_read_timeout} -Number of seconds to wait for more data from a master/slave connection -before aborting the read. +@findex RESTORE TABLE +@node RESTORE TABLE, ANALYZE TABLE, BACKUP TABLE, Reference +@section @code{RESTORE TABLE} Syntax -@item @code{slow_launch_time} -If creating the thread takes longer than this value (in seconds), the -@code{Slow_launch_threads} counter will be incremented. +@example +RESTORE TABLE tbl_name[,tbl_name...] FROM '/path/to/backup/directory' +@end example -@item @code{socket} -The Unix socket used by the server. +Restores the table(s) from the backup that was made with +@code{BACKUP TABLE}. Existing tables will not be overwritten - if you +try to restore over an existing table, you will get an error. Restore +will take longer than BACKUP due to the need to rebuilt the index. The +more keys you have, the longer it is going to take. Just as +@code{BACKUP TABLE}, currently only works of @code{MyISAM} tables. -@item @code{sort_buffer} -Each thread that needs to do a sort allocates a buffer of this -size. Increase this value for faster @code{ORDER BY} or @code{GROUP BY} -operations. -@xref{Temporary files}. -@item @code{table_cache} -The number of open tables for all threads. Increasing this value -increases the number of file descriptors that @code{mysqld} requires. -@strong{MySQL} needs two file descriptors for each unique open table. -See below for comments on file descriptor limits. You can check if you -need to increase the table cache by checking the @code{Opened_tables} -variable. @xref{SHOW}. If this variable is big and you don't do -@code{FLUSH TABLES} a lot (which just forces all tables to be closed and -reopenend), then you should increase the value of this variable. +The command returns a table with the following columns: -Make sure that your operating system can handle the number of open file -descriptors implied by the @code{table_cache} setting. If @code{table_cache} -is set too high, @strong{MySQL} may run out of file descriptors and refuse -connections, fail to perform queries, and be very unreliable. +@multitable @columnfractions .35 .65 +@item @strong{Column} @tab @strong{Value} +@item Table @tab Table name +@item Op @tab Always ``restore'' +@item Msg_type @tab One of @code{status}, @code{error}, @code{info} or @code{warning}. +@item Msg_text @tab The message. +@end multitable -For information about how the table cache works, see @ref{Table cache}. -@item @code{table_type} -The default table type +@findex ANALYZE TABLE +@node ANALYZE TABLE, REPAIR TABLE, RESTORE TABLE, Reference +@section @code{ANALYZE TABLE} Syntax -@item @code{thread_cache_size} -How many threads we should keep in a cache for reuse. When a -client disconnects, the client's threads are put in the cache if there -aren't more than @code{thread_cache_size} threads from before. All new -threads are first taken from the cache, and only when the cache is empty -is a new thread created. This variable can be increased to improve -performance if you have a lot of new connections. (Normally this doesn't -give a notable performance improvement if you have a good -thread implementation.) By examing the difference between -the @code{Connections} and @code{Threads_created} you can see how efficient -the current thread cache is for you. +@example +ANALYZE TABLE tbl_name[,tbl_name...] +@end example -@item @code{thread_concurrency} -On Solaris, @code{mysqld} will call @code{thr_setconcurrency()} with -this value. @code{thr_setconcurrency()} permits the application to give -the threads system a hint for the desired number of threads that should -be run at the same time. +Analyze and store the key distribution for the table. During the +analyze the table is locked with a read lock. This works on +@code{MyISAM} and @code{BDB} tables. -@item @code{thread_stack} -The stack size for each thread. Many of the limits detected by the -@code{crash-me} test are dependent on this value. The default is -large enough for normal operation. @xref{Benchmarks}. +This is equivalent to running @code{myisamchk -a} on the table. -@item @code{timezone} -The timezone for the server. +@strong{MySQL} uses the stored key distribution to decide in which order +tables should be joined when one does a join on something else than a +constant. -@item @code{tmp_table_size} -If an in-memory temporary table exceeds this size, @strong{MySQL} -will automatically convert it to an on-disk @code{MyISAM} table. -Increase the value of @code{tmp_table_size} if you do many advanced -@code{GROUP BY} queries and you have lots of memory. +The command returns a table with the following columns: -@item @code{tmpdir} -The directory used for temporary files and temporary tables. +@multitable @columnfractions .35 .65 +@item @strong{Column} @tab @strong{Value} +@item Table @tab Table name +@item Op @tab Always ``analyze'' +@item Msg_type @tab One of @code{status}, @code{error}, @code{info} or @code{warning}. +@item Msg_text @tab The message. +@end multitable -@item @code{version} -The version number for the server. +You can check the stored key distribution with the @code{SHOW INDEX} command. +@xref{SHOW DATABASE INFO}. -@item @code{wait_timeout} -The number of seconds the server waits for activity on a connection before -closing it. See also @code{interactive_timeout}. -@end table +If the table hasn't changed since the last @code{ANALYZE TABLE} command, +the table will not be analyzed again. -The manual section that describes tuning @strong{MySQL} contains some -information of how to tune the above variables. @xref{Server parameters}. +@findex REPAIR TABLE +@node REPAIR TABLE, DELETE, ANALYZE TABLE, Reference +@section @code{REPAIR TABLE} Syntax -@node SHOW LOGS, SHOW PROCESSLIST, SHOW VARIABLES, SHOW -@subsection @code{SHOW LOGS} -@code{SHOW LOGS} shows you status information about existing log -files. It currently only displays information about Berkeley DB log -files. +@example +REPAIR TABLE tbl_name[,tbl_name...] [QUICK] [EXTENDED] +@end example -@itemize @bullet -@item @code{File} shows the full path to the log file -@item @code{Type} shows the type of the log file (@code{BDB} for Berkeley -DB log files) -@item @code{Status} shows the status of the log file (@code{FREE} if the -file can be removed, or @code{IN USE} if the file is needed by the transaction -subsystem) -@end itemize +@code{REPAIR TABLE} only works on @code{MyISAM} tables and is the same +as running @code{myisamchk -r table_name} on the table. -@cindex threads, display -@cindex processes, display -@findex threads -@findex PROCESSLIST -@node SHOW PROCESSLIST, SHOW GRANTS, SHOW LOGS, SHOW -@subsection @code{SHOW PROCESSLIST} +Normally you should never have to run this command, but if disaster strikes +you are very likely to get back all your data from a MyISAM table with +@code{REPAIR TABLE}. If your tables get corrupted a lot you should +try to find the reason for this! @xref{Crashing}. @xref{MyISAM table problems}. -@code{SHOW PROCESSLIST} shows you which threads are running. You can -also get this information using the @code{mysqladmin processlist} -command. If you have the @strong{process} privilege, you can see all -threads. Otherwise, you can see only your own threads. @xref{KILL, , -@code{KILL}}. If you don't use the @code{FULL} option, then only -the first 100 characters of each query will be shown. +@code{REPAIR TABLE} repairs a possible corrupted table. The command returns a +table with the following columns: -This command is very useful if you get the 'too many connections' error -message and want to find out what's going on. @strong{MySQL} reserves -one extra connection for a client with the @code{Process_priv} privilege -to ensure that you should always be able to login and check the system -(assuming you are not giving this privilege to all your users). +@multitable @columnfractions .35 .65 +@item @strong{Column} @tab @strong{Value} +@item Table @tab Table name +@item Op @tab Always ``repair'' +@item Msg_type @tab One of @code{status}, @code{error}, @code{info} or @code{warning}. +@item Msg_text @tab The message. +@end multitable -@cindex privileges, display -@node SHOW GRANTS, SHOW CREATE TABLE, SHOW PROCESSLIST, SHOW -@subsection @code{SHOW GRANTS} +Note that you can get many rows of information for each repaired +table. The last one row will be of @code{Msg_type status} and should +normally be @code{OK}. If you don't get @code{OK}, you should try +repairing the table with @code{myisamchk -o}, as @code{REPAIR TABLE} +does not yet implement all the options of @code{myisamchk}. In the near +future, we will make it more flexible. -@code{SHOW GRANTS FOR user} lists the grant commands that must be issued to -duplicate the grants for a user. +If @code{QUICK} is given then @strong{MySQL} will try to do a +@code{REPAIR} of only the index tree. + +If you use @code{EXTENDED} then @strong{MySQL} will create the index row +by row instead of creating one index at a time with sorting; This may be +better than sorting on fixed-length keys if you have long @code{char()} +keys that compress very good. + +@findex DELETE +@node DELETE, TRUNCATE, REPAIR TABLE, Reference +@section @code{DELETE} Syntax @example -mysql> SHOW GRANTS FOR root@@localhost; -+---------------------------------------------------------------------+ -| Grants for root@@localhost | -+---------------------------------------------------------------------+ -| GRANT ALL PRIVILEGES ON *.* TO 'root'@@'localhost' WITH GRANT OPTION | -+---------------------------------------------------------------------+ +DELETE [LOW_PRIORITY] FROM tbl_name + [WHERE where_definition] + [LIMIT rows] @end example -@node SHOW CREATE TABLE, , SHOW GRANTS, SHOW -@subsection @code{SHOW CREATE TABLE} +@code{DELETE} deletes rows from @code{tbl_name} that satisfy the condition +given by @code{where_definition}, and returns the number of records deleted. -Shows a @code{CREATE TABLE} statement that will create the given table: +@c If an @code{ORDER BY} clause is used, the rows will be deleted in that order. +@c This is really only useful in conjunction with @code{LIMIT}. For example: -@example -mysql> show create table t\G -*************************** 1. row *************************** - Table: t -Create Table: CREATE TABLE t ( - id int(11) default NULL auto_increment, - s char(60) default NULL, - PRIMARY KEY (id) -) TYPE=MyISAM +@c @example +@c DELETE FROM somelog +@c WHERE user = 'jcole' +@c ORDER BY timestamp +@c LIMIT 1 +@c @end example -@end example +@c This will delete the oldest entry (by @code{timestamp}) where the row matches +@c the @code{WHERE} clause. -@code{SHOW CREATE TABLE} will quote table and column names according to -@code{SQL_QUOTE_SHOW_CREATE} option. -@ref{SET OPTION, , @code{SET OPTION SQL_QUOTE_SHOW_CREATE}}. +If you issue a @code{DELETE} with no @code{WHERE} clause, all rows are +deleted. If you do this in @code{AUTOCOMMIT} mode, this works as +@code{TRUNCATE}. @xref{TRUNCATE}. One problem with this is that +@code{DELETE} will return zero as the number of affected records, but +this will be fixed in 4.0. -@findex EXPLAIN -@findex SELECT, optimizing -@node EXPLAIN, DESCRIBE, SHOW, Reference -@section @code{EXPLAIN} Syntax (Get Information About a @code{SELECT}) +If you really want to know how many records are deleted when you are deleting +all rows, and are willing to suffer a speed penalty, you can use a +@code{DELETE} statement of this form: @example - EXPLAIN tbl_name -or EXPLAIN SELECT select_options +mysql> DELETE FROM tbl_name WHERE 1>0; @end example -@code{EXPLAIN tbl_name} is a synonym for @code{DESCRIBE tbl_name} or -@code{SHOW COLUMNS FROM tbl_name}. - -When you precede a @code{SELECT} statement with the keyword @code{EXPLAIN}, -@strong{MySQL} explains how it would process the @code{SELECT}, providing -information about how tables are joined and in which order. - -With the help of @code{EXPLAIN}, you can see when you must add indexes -to tables to get a faster @code{SELECT} that uses indexes to find the -records. You can also see if the optimizer joins the tables in an optimal -order. To force the optimizer to use a specific join order for a -@code{SELECT} statement, add a @code{STRAIGHT_JOIN} clause. +Note that this is MUCH slower than @code{DELETE FROM tbl_name} with no +@code{WHERE} clause, because it deletes rows one at a time. -For non-simple joins, @code{EXPLAIN} returns a row of information for each -table used in the @code{SELECT} statement. The tables are listed in the order -they would be read. @strong{MySQL} resolves all joins using a single-sweep -multi-join method. This means that @strong{MySQL} reads a row from the first -table, then finds a matching row in the second table, then in the third table -and so on. When all tables are processed, it outputs the selected columns and -backtracks through the table list until a table is found for which there are -more matching rows. The next row is read from this table and the process -continues with the next table. +If you specify the keyword @code{LOW_PRIORITY}, execution of the +@code{DELETE} is delayed until no other clients are reading from the table. -Output from @code{EXPLAIN} includes the following columns: +Deleted records are maintained in a linked list and subsequent @code{INSERT} +operations reuse old record positions. To reclaim unused space and reduce +file sizes, use the @code{OPTIMIZE TABLE} statement or the @code{myisamchk} +utility to reorganize tables. @code{OPTIMIZE TABLE} is easier, but +@code{myisamchk} is faster. +See @ref{OPTIMIZE TABLE, , @code{OPTIMIZE TABLE}} and @ref{Optimization}. -@table @code -@item table -The table to which the row of output refers. +The @strong{MySQL}-specific @code{LIMIT rows} option to @code{DELETE} tells +the server the maximum number of rows to be deleted before control is +returned to the client. This can be used to ensure that a specific +@code{DELETE} command doesn't take too much time. You can simply repeat +the @code{DELETE} command until the number of affected rows is less than +the @code{LIMIT} value. -@item type -The join type. Information about the various types is given below. +@findex TRUNCATE +@node TRUNCATE, SELECT, DELETE, Reference +@section @code{TRUNCATE} Syntax -@item possible_keys -The @code{possible_keys} column indicates which indexes @strong{MySQL} -could use to find the rows in this table. Note that this column is -totally independent of the order of the tables. That means that some of -the keys in possible_keys may not be usable in practice with the -generated table order. +@example +TRUNCATE TABLE table_name +@end example -If this column is empty, there are no relevant indexes. In this case, -you may be able to improve the performance of your query by examining -the @code{WHERE} clause to see if it refers to some column or columns -that would be suitable for indexing. If so, create an appropriate index -and check the query with @code{EXPLAIN} again. @xref{ALTER TABLE}. +Is in 3.23 and the same thing as @code{DELETE FROM table_name}. @xref{DELETE}. +The differences are: -To see what indexes a table has, use @code{SHOW INDEX FROM tbl_name}. +@itemize @bullet +@item +Implemented as a drop and re-create of the table, which makes this +much faster when deleting many rows. +@item +Not transaction-safe; @code{TRUNCATE TABLE} will automatically end the current +transaction as if @code{COMMIT} would have been called. +@item +Doesn't return the number of deleted rows. +@item +As long as the table definition file @file{table_name.frm} is +valid, the table can be re-created this way, even if the data or index +files have become corrupted. +@end itemize -@item key -The @code{key} column indicates the key that @strong{MySQL} actually -decided to use. The key is @code{NULL} if no index was chosen. If -@strong{MySQL} chooses the wrong index, you can probably force -@strong{MySQL} to use another index by using @code{myisamchk --analyze}, -@xref{myisamchk syntax}, or by using @code{USE INDEX/IGNORE INDEX}. -@xref{JOIN}. +@code{TRUNCATE} is an Oracle SQL extension. -@item key_len -The @code{key_len} column indicates the length of the key that -@strong{MySQL} decided to use. The length is @code{NULL} if the -@code{key} is @code{NULL}. Note that this tells us how many parts of a -multi-part key @strong{MySQL} will actually use. +@findex SELECT +@node SELECT, JOIN, TRUNCATE, Reference +@section @code{SELECT} Syntax -@item ref -The @code{ref} column shows which columns or constants are used with the -@code{key} to select rows from the table. +@c help SELECT +@example +SELECT [STRAIGHT_JOIN] [SQL_SMALL_RESULT] [SQL_BIG_RESULT] [SQL_BUFFER_RESULT] + [HIGH_PRIORITY] + [DISTINCT | DISTINCTROW | ALL] + select_expression,... + [INTO @{OUTFILE | DUMPFILE@} 'file_name' export_options] + [FROM table_references + [WHERE where_definition] + [GROUP BY @{unsigned_integer | col_name | formula@} [ASC | DESC], ...] + [HAVING where_definition] + [ORDER BY @{unsigned_integer | col_name | formula@} [ASC | DESC] ,...] + [LIMIT [offset,] rows] + [PROCEDURE procedure_name] + [FOR UPDATE | LOCK IN SHARE MODE]] +@end example +@c help end -@item rows -The @code{rows} column indicates the number of rows @strong{MySQL} -believes it must examine to execute the query. +@code{SELECT} is used to retrieve rows selected from one or more tables. +@code{select_expression} indicates the columns you want to retrieve. +@code{SELECT} may also be used to retrieve rows computed without reference to +any table. For example: -@item Extra -This column contains additional information of how @strong{MySQL} will -resolve the query. Here is an explanation of the different text -strings that can be found in this column: +@example +mysql> SELECT 1 + 1; + -> 2 +@end example -@table @code -@item Distinct -@strong{MySQL} will not continue searching for more rows for the current row -combination after it has found the first matching row. +All keywords used must be given in exactly the order shown above. For example, +a @code{HAVING} clause must come after any @code{GROUP BY} clause and before +any @code{ORDER BY} clause. -@item Not exists -@strong{MySQL} was able to do a @code{LEFT JOIN} optimization on the -query and will not examine more rows in this table for the previous row -combination after it finds one row that matches the @code{LEFT JOIN} criteria. +@itemize @bullet -Here is an example for this: +@item +@cindex aliases, on expressions +@cindex expression aliases +A @code{SELECT} expression may be given an alias using @code{AS}. The alias +is used as the expression's column name and can be used with +@code{ORDER BY} or @code{HAVING} clauses. For example: @example -SELECT * FROM t1 LEFT JOIN t2 ON t1.id=t2.id WHERE t2.id IS NULL; +mysql> select concat(last_name,', ',first_name) AS full_name + from mytable ORDER BY full_name; @end example -Assume that @code{t2.id} is defined with @code{NOT NULL}. In this case -@strong{MySQL} will scan @code{t1} and look up the rows in @code{t2} -through @code{t1.id}. If @strong{MySQL} finds a matching row in -@code{t2}, it knows that @code{t2.id} can never be @code{NULL}, and will -not scan through the rest of the rows in @code{t2} that has the same -@code{id}. In other words, for each row in @code{t1}, @strong{MySQL} -only needs to do a single lookup in @code{t2}, independent of how many -matching rows there are in @code{t2}. +@item +The @code{FROM table_references} clause indicates the tables from which to +retrieve rows. If you name more than one table, you are performing a +join. For information on join syntax, see @ref{JOIN, , @code{JOIN}}. -@item @code{range checked for each record (index map: #)} -@strong{MySQL} didn't find a real good index to use. It will, instead, for -each row combination in the preceding tables, do a check on which index to -use (if any), and use this index to retrieve the rows from the table. This -isn't very fast but is faster than having to do a join without -an index. +@item +You can refer to a column as @code{col_name}, @code{tbl_name.col_name}, or +@code{db_name.tbl_name.col_name}. You need not specify a @code{tbl_name} or +@code{db_name.tbl_name} prefix for a column reference in a @code{SELECT} +statement unless the reference would be ambiguous. See @ref{Legal names}, +for examples of ambiguity that require the more explicit column reference +forms. -@item Using filesort -@strong{MySQL} will need to do an extra pass to find out how to retrieve -the rows in sorted order. The sort is done by going through all rows -according to the @code{join type} and storing the sort key + pointer to -the row for all rows that match the @code{WHERE}. Then the keys are -sorted. Finally the rows are retrieved in sorted order. +@item +@cindex aliases, for tables +@cindex table aliases +A table reference may be aliased using @code{tbl_name [AS] alias_name}: -@item Using index -The column information is retrieved from the table using only -information in the index tree without having to do an additional seek to -read the actual row. This can be done when all the used columns for -the table are part of the same index. +@example +mysql> select t1.name, t2.salary from employee AS t1, info AS t2 + where t1.name = t2.name; +mysql> select t1.name, t2.salary from employee t1, info t2 + where t1.name = t2.name; +@end example -@item Using temporary -To resolve the query @strong{MySQL} will need to create a -temporary table to hold the result. This typically happens if you do an -@code{ORDER BY} on a different column set than you did a @code{GROUP -BY} on. +@item +Columns selected for output may be referred to in @code{ORDER BY} and +@code{GROUP BY} clauses using column names, column aliases, or column +positions. Column positions begin with 1: -@item Where used -A @code{WHERE} clause will be used to restrict which rows will be -matched against the next table or sent to the client. If you don't have -this information and the table is of type @code{ALL} or @code{index}, -you may have something wrong in your query (if you don't intend to -fetch/examine all rows from the table). -@end table +@example +mysql> select college, region, seed from tournament + ORDER BY region, seed; +mysql> select college, region AS r, seed AS s from tournament + ORDER BY r, s; +mysql> select college, region, seed from tournament + ORDER BY 2, 3; +@end example -If you want to get your queries as fast as possible, you should look out for -@code{Using filesort} and @code{Using temporary}. -@end table +To sort in reverse order, add the @code{DESC} (descending) keyword to the +name of the column in the @code{ORDER BY} clause that you are sorting by. +The default is ascending order; this may be specified explicitly using +the @code{ASC} keyword. -The different join types are listed below, ordered from best to worst type: +@item +You can in the @code{WHERE} clause use any of the functions that +@strong{MySQL} support. @xref{Functions}. -@cindex system table -@cindex tables, system -@table @code -@item system -The table has only one row (= system table). This is a special case of -the @code{const} join type. +@item +The @code{HAVING} clause can refer to any column or alias named in the +@code{select_expression}. It is applied last, just before items are sent to +the client, with no optimization. Don't use @code{HAVING} for items that +should be in the @code{WHERE} clause. For example, do not write this: -@cindex constant table -@cindex tables, constant -@item const -The table has at most one matching row, which will be read at the start -of the query. Because there is only one row, values from the column in -this row can be regarded as constants by the rest of the -optimizer. @code{const} tables are very fast as they are read only once! +@example +mysql> select col_name from tbl_name HAVING col_name > 0; +@end example -@item eq_ref -One row will be read from this table for each combination of rows from -the previous tables. This is the best possible join type, other than the -@code{const} types. It is used when all parts of an index are used by -the join and the index is @code{UNIQUE} or a @code{PRIMARY KEY}. +Write this instead: -@item ref -All rows with matching index values will be read from this table for each -combination of rows from the previous tables. @code{ref} is used if the join -uses only a leftmost prefix of the key, or if the key is not @code{UNIQUE} -or a @code{PRIMARY KEY} (in other words, if the join cannot select a single -row based on the key value). If the key that is used matches only a few rows, -this join type is good. +@example +mysql> select col_name from tbl_name WHERE col_name > 0; +@end example -@item range -Only rows that are in a given range will be retrieved, using an index to -select the rows. The @code{key} column indicates which index is used. -The @code{key_len} contains the longest key part that was used. -The @code{ref} column will be NULL for this type. +In @strong{MySQL} Version 3.22.5 or later, you can also write queries like this: -@item index -This is the same as @code{ALL}, except that only the index tree is -scanned. This is usually faster than @code{ALL}, as the index file is usually -smaller than the data file. +@example +mysql> select user,max(salary) from users + group by user HAVING max(salary)>10; +@end example -@item ALL -A full table scan will be done for each combination of rows from the -previous tables. This is normally not good if the table is the first -table not marked @code{const}, and usually @strong{very} bad in all other -cases. You normally can avoid @code{ALL} by adding more indexes, so that -the row can be retrieved based on constant values or column values from -earlier tables. -@end table +In older @strong{MySQL} versions, you can write this instead: -You can get a good indication of how good a join is by multiplying all values -in the @code{rows} column of the @code{EXPLAIN} output. This should tell you -roughly how many rows @strong{MySQL} must examine to execute the query. This -number is also used when you restrict queries with the @code{max_join_size} -variable. -@xref{Server parameters}. +@example +mysql> select user,max(salary) AS sum from users + group by user HAVING sum>10; +@end example -The following example shows how a @code{JOIN} can be optimized progressively -using the information provided by @code{EXPLAIN}. +@item +@code{SQL_SMALL_RESULT}, @code{SQL_BIG_RESULT}, @code{SQL_BUFFER_RESULT}, +@code{STRAIGHT_JOIN}, and @code{HIGH_PRIORITY} are @strong{MySQL} extensions +to ANSI SQL92. -Suppose you have the @code{SELECT} statement shown below, that you examine -using @code{EXPLAIN}: +@item +@code{HIGH_PRIORITY} will give the @code{SELECT} higher priority than +a statement that updates a table. You should only use this for queries +that are very fast and must be done at once. A @code{SELECT HIGH_PRIORITY} +query will run if the table is locked for read even if there is an update +statement that is waiting for the table to be free. + +@item +@code{SQL_BIG_RESULT} can be used with @code{GROUP BY} or @code{DISTINCT} +to tell the optimizer that the result set will have many rows. In this case, +@strong{MySQL} will directly use disk-based temporary tables if needed. +@strong{MySQL} will also, in this case, prefer sorting to doing a +temporary table with a key on the @code{GROUP BY} elements. + +@item +@cindex @code{GROUP BY}, extensions to ANSI SQL +If you use @code{GROUP BY}, the output rows will be sorted according to the +@code{GROUP BY} as if you would have had an @code{ORDER BY} over all the fields +in the @code{GROUP BY}. @strong{MySQL} has extended the @code{GROUP BY} so that +you can also specify @code{ASC} and @code{DESC} to @code{GROUP BY}: @example -EXPLAIN SELECT tt.TicketNumber, tt.TimeIn, - tt.ProjectReference, tt.EstimatedShipDate, - tt.ActualShipDate, tt.ClientID, - tt.ServiceCodes, tt.RepetitiveID, - tt.CurrentProcess, tt.CurrentDPPerson, - tt.RecordVolume, tt.DPPrinted, et.COUNTRY, - et_1.COUNTRY, do.CUSTNAME - FROM tt, et, et AS et_1, do - WHERE tt.SubmitTime IS NULL - AND tt.ActualPC = et.EMPLOYID - AND tt.AssignedPC = et_1.EMPLOYID - AND tt.ClientID = do.CUSTNMBR; +SELECT a,COUNT(b) FROM test_table GROUP BY a DESC @end example -For this example, assume that: - -@itemize @bullet @item -The columns being compared have been declared as follows: +@strong{MySQL} has extended the use of @code{GROUP BY} to allow you to +select fields which are not mentioned in the @code{GROUP BY} clause. +If you are not getting the results you expect from your query, please +read the @code{GROUP BY} description. +@xref{Group by functions}. -@multitable @columnfractions .1 .2 .7 -@item @strong{Table} @tab @strong{Column} @tab @strong{Column type} -@item @code{tt} @tab @code{ActualPC} @tab @code{CHAR(10)} -@item @code{tt} @tab @code{AssignedPC} @tab @code{CHAR(10)} -@item @code{tt} @tab @code{ClientID} @tab @code{CHAR(10)} -@item @code{et} @tab @code{EMPLOYID} @tab @code{CHAR(15)} -@item @code{do} @tab @code{CUSTNMBR} @tab @code{CHAR(15)} -@end multitable +@item +@code{SQL_BUFFER_RESULT} will force the result to be put into a temporary +table. This will help @strong{MySQL} free the table locks early and will help +in cases where it takes a long time to send the result set to the client. @item -The tables have the indexes shown below: +@code{SQL_SMALL_RESULT}, a @strong{MySQL}-specific option, can be used +with @code{GROUP BY} or @code{DISTINCT} to tell the optimizer that the +result set will be small. In this case, @strong{MySQL} will use fast +temporary tables to store the resulting table instead of using sorting. In +@strong{MySQL} Version 3.23 this shouldn't normally be needed. -@multitable @columnfractions .1 .9 -@item @strong{Table} @tab @strong{Index} -@item @code{tt} @tab @code{ActualPC} -@item @code{tt} @tab @code{AssignedPC} -@item @code{tt} @tab @code{ClientID} -@item @code{et} @tab @code{EMPLOYID} (primary key) -@item @code{do} @tab @code{CUSTNMBR} (primary key) -@end multitable +@item +@code{STRAIGHT_JOIN} forces the optimizer to join the tables in the order in +which they are listed in the @code{FROM} clause. You can use this to speed up +a query if the optimizer joins the tables in non-optimal order. +@xref{EXPLAIN, , @code{EXPLAIN}}. @item -The @code{tt.ActualPC} values aren't evenly distributed. -@end itemize +The @code{LIMIT} clause can be used to constrain the number of rows returned +by the @code{SELECT} statement. @code{LIMIT} takes one or two numeric +arguments. -Initially, before any optimizations have been performed, the @code{EXPLAIN} -statement produces the following information: +If two arguments are given, the first specifies the offset of the first row to +return, the second specifies the maximum number of rows to return. +The offset of the initial row is 0 (not 1): @example -table type possible_keys key key_len ref rows Extra -et ALL PRIMARY NULL NULL NULL 74 -do ALL PRIMARY NULL NULL NULL 2135 -et_1 ALL PRIMARY NULL NULL NULL 74 -tt ALL AssignedPC,ClientID,ActualPC NULL NULL NULL 3872 - range checked for each record (key map: 35) +mysql> select * from table LIMIT 5,10; # Retrieve rows 6-15 @end example -Because @code{type} is @code{ALL} for each table, this output indicates that -@strong{MySQL} is doing a full join for all tables! This will take quite a -long time, as the product of the number of rows in each table must be -examined! For the case at hand, this is @code{74 * 2135 * 74 * 3872 = -45,268,558,720} rows. If the tables were bigger, you can only imagine how -long it would take. - -One problem here is that @strong{MySQL} can't (yet) use indexes on columns -efficiently if they are declared differently. In this context, -@code{VARCHAR} and @code{CHAR} are the same unless they are declared as -different lengths. Because @code{tt.ActualPC} is declared as @code{CHAR(10)} -and @code{et.EMPLOYID} is declared as @code{CHAR(15)}, there is a length -mismatch. - -To fix this disparity between column lengths, use @code{ALTER TABLE} to -lengthen @code{ActualPC} from 10 characters to 15 characters: +If one argument is given, it indicates the maximum number of rows to return: @example -mysql> ALTER TABLE tt MODIFY ActualPC VARCHAR(15); +mysql> select * from table LIMIT 5; # Retrieve first 5 rows @end example -Now @code{tt.ActualPC} and @code{et.EMPLOYID} are both @code{VARCHAR(15)}. -Executing the @code{EXPLAIN} statement again produces this result: +In other words, @code{LIMIT n} is equivalent to @code{LIMIT 0,n}. -@example -table type possible_keys key key_len ref rows Extra -tt ALL AssignedPC,ClientID,ActualPC NULL NULL NULL 3872 where used -do ALL PRIMARY NULL NULL NULL 2135 - range checked for each record (key map: 1) -et_1 ALL PRIMARY NULL NULL NULL 74 - range checked for each record (key map: 1) -et eq_ref PRIMARY PRIMARY 15 tt.ActualPC 1 -@end example +@item +@tindex /etc/passwd +The @code{SELECT ... INTO OUTFILE 'file_name'} form of @code{SELECT} writes +the selected rows to a file. The file is created on the server host and +cannot already exist (among other things, this prevents database tables and +files such as @file{/etc/passwd} from being destroyed). You must have the +@strong{file} privilege on the server host to use this form of @code{SELECT}. -This is not perfect, but is much better (the product of the @code{rows} -values is now less by a factor of 74). This version is executed in a couple -of seconds. +@code{SELECT ... INTO OUTFILE} is mainly intended to let you very +quickly dump a table on the server machine. If you want to create the +resulting file on some other host than the server host you can't use +@code{SELECT ... INTO OUTFILE}. In this case you should instead use some +client program like @code{mysqldump --tab} or @code{mysql -e "SELECT +..." > outfile} to generate the file. -A second alteration can be made to eliminate the column length mismatches -for the @code{tt.AssignedPC = et_1.EMPLOYID} and @code{tt.ClientID = -do.CUSTNMBR} comparisons: +@code{SELECT ... INTO OUTFILE} is the complement of @code{LOAD DATA +INFILE}; the syntax for the @code{export_options} part of the statement +consists of the same @code{FIELDS} and @code{LINES} clauses that are used +with the @code{LOAD DATA INFILE} statement. +@xref{LOAD DATA, , @code{LOAD DATA}}. -@example -mysql> ALTER TABLE tt MODIFY AssignedPC VARCHAR(15), - MODIFY ClientID VARCHAR(15); -@end example +In the resulting text file, only the following characters are escaped by +the @code{ESCAPED BY} character: -Now @code{EXPLAIN} produces the output shown below: +@itemize @bullet +@item The @code{ESCAPED BY} character +@item The first character in @code{FIELDS TERMINATED BY} +@item The first character in @code{LINES TERMINATED BY} +@end itemize -@example -table type possible_keys key key_len ref rows Extra -et ALL PRIMARY NULL NULL NULL 74 -tt ref AssignedPC,ClientID,ActualPC ActualPC 15 et.EMPLOYID 52 where used -et_1 eq_ref PRIMARY PRIMARY 15 tt.AssignedPC 1 -do eq_ref PRIMARY PRIMARY 15 tt.ClientID 1 -@end example +Additionally, @code{ASCII 0} is converted to @code{ESCAPED BY} followed by 0 +(@code{ASCII 48}). -This is almost as good as it can get. +The reason for the above is that you MUST escape any @code{FIELDS +TERMINATED BY}, @code{ESCAPED BY}, or @code{LINES TERMINATED BY} +characters to reliably be able to read the file back. @code{ASCII 0} is +escaped to make it easier to view with some pagers. -The remaining problem is that, by default, @strong{MySQL} assumes that values -in the @code{tt.ActualPC} column are evenly distributed, and that isn't the -case for the @code{tt} table. Fortunately, it is easy to tell @strong{MySQL} -about this: +As the resulting file doesn't have to conform to the SQL syntax, nothing +else need be escaped. + +Here follows an example of getting a file in the format used by many +old programs. @example -shell> myisamchk --analyze PATH_TO_MYSQL_DATABASE/tt -shell> mysqladmin refresh +SELECT a,b,a+b INTO OUTFILE "/tmp/result.text" +FIELDS TERMINATED BY ',' OPTIONALLY ENCLOSED BY '"' +LINES TERMINATED BY "\n" +FROM test_table; @end example -Now the join is perfect, and @code{EXPLAIN} produces this result: +@item +@findex DUMPFILE +If you use @code{INTO DUMPFILE} instead of @code{INTO OUTFILE}, @strong{MySQL} +will only write one row into the file, without any column or line +terminations and without any escaping. This is useful if you want to +store a blob in a file. -@example -table type possible_keys key key_len ref rows Extra -tt ALL AssignedPC,ClientID,ActualPC NULL NULL NULL 3872 where used -et eq_ref PRIMARY PRIMARY 15 tt.ActualPC 1 -et_1 eq_ref PRIMARY PRIMARY 15 tt.AssignedPC 1 -do eq_ref PRIMARY PRIMARY 15 tt.ClientID 1 -@end example +@item +Note that any file created by @code{INTO OUTFILE} and @code{INTO +DUMPFILE} is going to be readable for all users! The reason is that the +@strong{MySQL} server can't create a file that is owned by anyone else +than the user it's running as (you should never run @code{mysqld} as root), +the file has to be word readable so that you can retrieve the rows. -Note that the @code{rows} column in the output from @code{EXPLAIN} is an -educated guess from the @strong{MySQL} join optimizer. To optimize a -query, you should check if the numbers are even close to the truth. If not, -you may get better performance by using @code{STRAIGHT_JOIN} in your -@code{SELECT} statement and trying to list the tables in a different order in -the @code{FROM} clause. +@item +If you are using @code{FOR UPDATE} on a table handler with page/row locks, +the examined rows will be write locked. +@end itemize -@findex DESC -@findex DESCRIBE -@node DESCRIBE, COMMIT, EXPLAIN, Reference -@section @code{DESCRIBE} Syntax (Get Information About Columns) +@findex JOIN +@findex INNER JOIN +@findex CROSS JOIN +@findex LEFT JOIN +@findex LEFT OUTER JOIN +@findex NATURAL LEFT JOIN +@findex NATURAL LEFT OUTER JOIN +@findex RIGHT JOIN +@findex RIGHT OUTER JOIN +@findex NATURAL RIGHT JOIN +@findex NATURAL RIGHT OUTER JOIN +@findex STRAIGHT_JOIN +@node JOIN, INSERT, SELECT, Reference +@section @code{JOIN} Syntax + +@strong{MySQL} supports the following @code{JOIN} syntaxes for use in +@code{SELECT} statements: @example -@{DESCRIBE | DESC@} tbl_name @{col_name | wild@} +table_reference, table_reference +table_reference [CROSS] JOIN table_reference +table_reference INNER JOIN table_reference join_condition +table_reference STRAIGHT_JOIN table_reference +table_reference LEFT [OUTER] JOIN table_reference join_condition +table_reference LEFT [OUTER] JOIN table_reference +table_reference NATURAL [LEFT [OUTER]] JOIN table_reference +@{ oj table_reference LEFT OUTER JOIN table_reference ON conditional_expr @} +table_reference RIGHT [OUTER] JOIN table_reference join_condition +table_reference RIGHT [OUTER] JOIN table_reference +table_reference NATURAL [RIGHT [OUTER]] JOIN table_reference @end example -@code{DESCRIBE} is a shortcut for @code{SHOW COLUMNS FROM}. -@xref{SHOW DATABASE INFO}. - -@code{DESCRIBE} provides information about a table's columns. @code{col_name} -may be a column name or a string containing the SQL @samp{%} and @samp{_} -wild-card characters. - -If the column types are different than you expect them to be based on a -@code{CREATE TABLE} statement, note that @strong{MySQL} sometimes -changes column types. @xref{Silent column changes}. +Where @code{table_reference} is defined as: +@example +table_name [[AS] alias] [USE INDEX (key_list)] [IGNORE INDEX (key_list)] +@end example -@cindex Oracle compatibility -@cindex compatibility, with Oracle -This statement is provided for Oracle compatibility. +and @code{join_condition} is defined as: -The @code{SHOW} statement provides similar information. -@xref{SHOW, , @code{SHOW}}. +@example +ON conditional_expr | +USING (column_list) +@end example -@findex BEGIN -@findex COMMIT -@findex ROLLBACK -@node COMMIT, LOCK TABLES, DESCRIBE, Reference -@section @code{BEGIN/COMMIT/ROLLBACK} Syntax +Note that in versions before Version 3.23.16, the @code{INNER JOIN} didn't take +a join condition! -By default, @strong{MySQL} runs in @code{autocommit} mode. This means that -as soon as you execute an update, @strong{MySQL} will store the update on -disk. +@cindex ODBC compatibility +@cindex compatibility, with ODBC +The last @code{LEFT OUTER JOIN} syntax shown above exists only for +compatibility with ODBC: -If you are using transactions safe tables (like @code{BDB}, -@code{InnoDB}, you can put @strong{MySQL} into -non-@code{autocommit} mode with the following command: +@itemize @bullet +@item +A table reference may be aliased using @code{tbl_name AS alias_name} or +@code{tbl_name alias_name}: @example -SET AUTOCOMMIT=0 +mysql> select t1.name, t2.salary from employee AS t1, info AS t2 + where t1.name = t2.name; @end example -After this you must use @code{COMMIT} to store your changes to disk or -@code{ROLLBACK} if you want to ignore the changes you have made since -the beginning of your transaction. +@item +@code{INNER JOIN} and @code{,} (comma) are semantically equivalent. +Both do a full join between the tables used. Normally, you specify how +the tables should be linked in the @code{WHERE} condition. -If you want to switch from @code{AUTOCOMMIT} mode for one series of -statements, you can use the @code{BEGIN} or @code{BEGIN WORK} statement: +@item +The @code{ON} conditional is any conditional of the form that may be used in +a @code{WHERE} clause. + +@item +If there is no matching record for the right table in the @code{ON} or +@code{USING} part in a @code{LEFT JOIN}, a row with all columns set to +@code{NULL} is used for the right table. You can use this fact to find +records in a table that have no counterpart in another table: @example -BEGIN; -SELECT @@A:=SUM(salary) FROM table1 WHERE type=1; -UPDATE table2 SET summmary=@@A WHERE type=1; -COMMIT; +mysql> select table1.* from table1 + LEFT JOIN table2 ON table1.id=table2.id + where table2.id is NULL; @end example -Note that if you are using non-transaction-safe tables, the changes will be -stored at once, independent of the status of the @code{autocommit} mode. - -If you do a @code{ROLLBACK} when you have updated a non-transactional -table you will get an error (@code{ER_WARNING_NOT_COMPLETE_ROLLBACK}) as -a warning. All transactional safe tables will be restored but any -non-transactional table will not change. - -If you are using @code{BEGIN} or @code{SET AUTOCOMMIT=0}, you -should use the @strong{MySQL} binary log for backups instead of the -older update log. Transactions are stored in the binary log -in one chunk, upon @code{COMMIT}, to ensure that transactions which are -rolled back are not stored. @xref{Binary log}. - -The following commands automatically end a transaction (as if you had done -a @code{COMMIT} before executing the command): +This example finds all rows in @code{table1} with an @code{id} value that is +not present in @code{table2} (that is, all rows in @code{table1} with no +corresponding row in @code{table2}). This assumes that @code{table2.id} is +declared @code{NOT NULL}, of course. @xref{LEFT JOIN optimization}. -@multitable @columnfractions .33 .33 .33 -@item @code{ALTER TABLE} @tab @code{BEGIN} @tab @code{CREATE INDEX} -@item @code{DROP DATABASE} @tab @code{DROP TABLE} @tab @code{RENAME TABLE} -@item @code{TRUNCATE} -@end multitable +@item +The @code{USING} @code{(column_list)} clause names a list of columns that must +exist in both tables. A @code{USING} clause such as: -You can change the isolation level for transactions with -@code{SET TRANSACTION ISOLATION LEVEL ...}. @xref{SET TRANSACTION}. +@example +A LEFT JOIN B USING (C1,C2,C3,...) +@end example -@findex LOCK TABLES -@findex UNLOCK TABLES -@node LOCK TABLES, SET OPTION, COMMIT, Reference -@section @code{LOCK TABLES/UNLOCK TABLES} Syntax +is defined to be semantically identical to an @code{ON} expression like +this: @example -LOCK TABLES tbl_name [AS alias] @{READ | [READ LOCAL] | [LOW_PRIORITY] WRITE@} - [, tbl_name @{READ | [LOW_PRIORITY] WRITE@} ...] -... -UNLOCK TABLES +A.C1=B.C1 AND A.C2=B.C2 AND A.C3=B.C3,... @end example -@code{LOCK TABLES} locks tables for the current thread. @code{UNLOCK -TABLES} releases any locks held by the current thread. All tables that -are locked by the current thread are automatically unlocked when the -thread issues another @code{LOCK TABLES}, or when the connection to the -server is closed. +@item +The @code{NATURAL [LEFT] JOIN} of two tables is defined to be +semantically equivalent to an @code{INNER JOIN} or a @code{LEFT JOIN} +with a @code{USING} clause that names all columns that exist in both +tables. -The main reasons to use @code{LOCK TABLES} are: +@item +@code{RIGHT JOIN} works analogously as @code{LEFT JOIN}. To keep code +portable across databases, it's recommended to use @code{LEFT JOIN} +instead of @code{RIGHT JOIN}. -@itemize @bullet @item -Emulate transactions with not transaction safe tables. +@code{STRAIGHT_JOIN} is identical to @code{JOIN}, except that the left table +is always read before the right table. This can be used for those (few) +cases where the join optimizer puts the tables in the wrong order. + @item -To get more speed with @code{MyISAM} tables when inserting/updating data -over many statements. The main reason this gives more speed is that -@strong{MySQL} will not flush the key cache for the locked tables until -@code{UNLOCK TABLES} is called. +As of @strong{MySQL} Version 3.23.12, you can give hints about which +index @strong{MySQL} should use when retrieving information from a +table. This is useful if @code{EXPLAIN} shows that @strong{MySQL} is +using the wrong index. By specifying @code{USE INDEX (key_list)}, you +can tell @strong{MySQL} to use only one of the specified indexes to +find rows in the table. The alternative syntax @code{IGNORE INDEX +(key_list)} can be used to tell @strong{MySQL} to not use some +particular index. @end itemize -If a thread obtains a @code{READ} lock on a table, that thread (and all other -threads) can only read from the table. If a thread obtains a @code{WRITE} -lock on a table, then only the thread holding the lock can @code{READ} from -or @code{WRITE} to the table. Other threads are blocked. +Some examples: -The difference between @code{READ LOCAL} and @code{READ} is that -@code{READ LOCAL} allows non-conflicting @code{INSERT} statements to -execute while the lock is held. This can't however be used if you are -going to manipulate the database files outside @strong{MySQL} while you -hold the lock. +@example +mysql> select * from table1,table2 where table1.id=table2.id; +mysql> select * from table1 LEFT JOIN table2 ON table1.id=table2.id; +mysql> select * from table1 LEFT JOIN table2 USING (id); +mysql> select * from table1 LEFT JOIN table2 ON table1.id=table2.id + LEFT JOIN table3 ON table2.id=table3.id; +mysql> select * from table1 USE INDEX (key1,key2) WHERE key1=1 and key2=2 AND + key3=3; +mysql> select * from table1 IGNORE INDEX (key3) WHERE key1=1 and key2=2 AND + key3=3; +@end example -Each thread waits (without timing out) until it obtains all the locks it has -requested. +@xref{LEFT JOIN optimization, , @code{LEFT JOIN} optimization}. -@code{WRITE} locks normally have higher priority than @code{READ} locks, to -ensure that updates are processed as soon as possible. This means that if one -thread obtains a @code{READ} lock and then another thread requests a -@code{WRITE} lock, subsequent @code{READ} lock requests will wait until the -@code{WRITE} thread has gotten the lock and released it. You can use -@code{LOW_PRIORITY WRITE} locks to allow other threads to obtain @code{READ} -locks while the thread is waiting for the @code{WRITE} lock. You should only -use @code{LOW_PRIORITY WRITE} locks if you are sure that there will -eventually be a time when no threads will have a @code{READ} lock. +@findex INSERT +@node INSERT, REPLACE, JOIN, Reference +@section @code{INSERT} Syntax -@code{LOCK TABLES} and @code{UNLOCK TABLES} both commits any active -transactions. +@example + INSERT [LOW_PRIORITY | DELAYED] [IGNORE] + [INTO] tbl_name [(col_name,...)] + VALUES (expression,...),(...),... +or INSERT [LOW_PRIORITY | DELAYED] [IGNORE] + [INTO] tbl_name [(col_name,...)] + SELECT ... +or INSERT [LOW_PRIORITY | DELAYED] [IGNORE] + [INTO] tbl_name + SET col_name=expression, col_name=expression, ... -When you use @code{LOCK TABLES}, you must lock all tables that you are -going to use and you must use the same alias that you are going to use -in your queries! If you are using a table multiple times in a query -(with aliases), you must get a lock for each alias! This policy ensures -that table locking is deadlock free and makes the locking code smaller, -simpler and much faster. +@end example -Note that you should @strong{NOT} lock any tables that you are using with -@code{INSERT DELAYED}. This is because that in this case the @code{INSERT} -is done by a separate thread. +@code{INSERT} inserts new rows into an existing table. The @code{INSERT +... VALUES} form of the statement inserts rows based on explicitly +specified values. The @code{INSERT ... SELECT} form inserts rows +selected from another table or tables. The @code{INSERT ... VALUES} +form with multiple value lists is supported in @strong{MySQL} Version +3.22.5 or later. The @code{col_name=expression} syntax is supported in +@strong{MySQL} Version 3.22.10 or later. -Normally, you don't have to lock tables, as all single @code{UPDATE} statements -are atomic; no other thread can interfere with any other currently executing -SQL statement. There are a few cases when you would like to lock tables -anyway: +@code{tbl_name} is the table into which rows should be inserted. The column +name list or the @code{SET} clause indicates which columns the statement +specifies values for: @itemize @bullet @item -If you are going to run many operations on a bunch of tables, it's much -faster to lock the tables you are going to use. The downside is, of course, -that no other thread can update a @code{READ}-locked table and no other -thread can read a @code{WRITE}-locked table. +If you specify no column list for @code{INSERT ... VALUES} or @code{INSERT +... SELECT}, values for all columns must be provided in the +@code{VALUES()} list or by the @code{SELECT}. If you don't know the order of +the columns in the table, use @code{DESCRIBE tbl_name} to find out. + @item -If you are using a table handler in @strong{MySQL} that doesn't support -transactions, you must use @code{LOCK TABLES} if you want to ensure that -no other thread comes between a @code{SELECT} and an @code{UPDATE}. The -example shown below requires @code{LOCK TABLES} in order to execute safely: +Any column not explicitly given a value is set to its default value. For +example, if you specify a column list that doesn't name all the columns in +the table, unnamed columns are set to their default values. Default value +assignment is described in @ref{CREATE TABLE, , @code{CREATE TABLE}}. + +@item +An @code{expression} may refer to any column that was set earlier in a value +list. For example, you can say this: @example -mysql> LOCK TABLES trans READ, customer WRITE; -mysql> select sum(value) from trans where customer_id= some_id; -mysql> update customer set total_value=sum_from_previous_statement - where customer_id=some_id; -mysql> UNLOCK TABLES; +mysql> INSERT INTO tbl_name (col1,col2) VALUES(15,col1*2); @end example -Without @code{LOCK TABLES}, there is a chance that another thread might -insert a new row in the @code{trans} table between execution of the -@code{SELECT} and @code{UPDATE} statements. -@end itemize +But not this: -By using incremental updates (@code{UPDATE customer SET -value=value+new_value}) or the @code{LAST_INSERT_ID()} function, you can -avoid using @code{LOCK TABLES} in many cases. +@example +mysql> INSERT INTO tbl_name (col1,col2) VALUES(col2*2,15); +@end example -You can also solve some cases by using the user-level lock functions -@code{GET_LOCK()} and @code{RELEASE_LOCK()}. These locks are saved in a hash -table in the server and implemented with @code{pthread_mutex_lock()} and -@code{pthread_mutex_unlock()} for high speed. -@xref{Miscellaneous functions}. +@item +If you specify the keyword @code{LOW_PRIORITY}, execution of the +@code{INSERT} is delayed until no other clients are reading from the +table. In this case the client has to wait until the insert statement +is completed, which may take a long time if the table is in heavy +use. This is in contrast to @code{INSERT DELAYED}, which lets the client +continue at once. @xref{INSERT DELAYED}. Note that @code{LOW_PRIORITY} +should normally not be used with @code{MyISAM} tables as this disables +concurrent inserts. @xref{MyISAM}. -See @ref{Internal locking}, for more information on locking policy. +@item +If you specify the keyword @code{IGNORE} in an @code{INSERT} with many value +rows, any rows that duplicate an existing @code{PRIMARY} or @code{UNIQUE} +key in the table are ignored and are not inserted. If you do not specify +@code{IGNORE}, the insert is aborted if there is any row that duplicates an +existing key value. You can determine with the C API function +@code{mysql_info()} how many rows were inserted into the table. -You can also lock all tables in all databases with read locks with the -@code{FLUSH TABLES WITH READ LOCK} command. @xref{FLUSH}. This is very -convenient way to get backups if you have a file system, like Veritas, -that can take snapshots in time. +@item +If @strong{MySQL} was configured using the @code{DONT_USE_DEFAULT_FIELDS} +option, @code{INSERT} statements generate an error unless you explicitly +specify values for all columns that require a non-@code{NULL} value. +@xref{configure options, , @code{configure} options}. -@strong{NOTE}: @code{LOCK TABLES} is not transaction-safe and will -automatically commit any active transactions before attempting to lock the -tables. +@item +You can find the value used for an @code{AUTO_INCREMENT} column +with the @code{mysql_insert_id} function. +@xref{mysql_insert_id, , @code{mysql_insert_id()}}. +@end itemize -@findex SET OPTION -@node SET OPTION, SET TRANSACTION, LOCK TABLES, Reference -@section @code{SET} Syntax +@findex mysql_info() +If you use @code{INSERT ... SELECT} or an @code{INSERT ... VALUES} +statement with multiple value lists, you can use the C API function +@code{mysql_info()} to get information about the query. The format of the +information string is shown below: @example -SET [OPTION] SQL_VALUE_OPTION= value, ... +Records: 100 Duplicates: 0 Warnings: 0 @end example -@code{SET OPTION} sets various options that affect the operation of the -server or your client. Any option you set remains in effect until the -current session ends, or until you set the option to a different value. - -@table @code -@item CHARACTER SET character_set_name | DEFAULT -This maps all strings from and to the client with the given mapping. -Currently the only option for @code{character_set_name} is -@code{cp1251_koi8}, but you can easily add new mappings by editing the -@file{sql/convert.cc} file in the @strong{MySQL} source distribution. The -default mapping can be restored by using a @code{character_set_name} value of -@code{DEFAULT}. +@code{Duplicates} indicates the number of rows that couldn't be inserted +because they would duplicate some existing unique index value. +@code{Warnings} indicates the number of attempts to insert column values that +were problematic in some way. Warnings can occur under any of the following +conditions: -Note that the syntax for setting the @code{CHARACTER SET} option differs -from the syntax for setting the other options. +@itemize @bullet +@item +Inserting @code{NULL} into a column that has been declared @code{NOT NULL}. +The column is set to its default value. -@item PASSWORD = PASSWORD('some password') -@cindex passwords, setting -Set the password for the current user. Any non-anonymous user can change his -own password! +@item +Setting a numeric column to a value that lies outside the column's range. +The value is clipped to the appropriate endpoint of the range. -@item PASSWORD FOR user = PASSWORD('some password') -Set the password for a specific user on the current server host. Only a user -with access to the @code{mysql} database can do this. The user should be -given in @code{user@@hostname} format, where @code{user} and @code{hostname} -are exactly as they are listed in the @code{User} and @code{Host} columns of -the @code{mysql.user} table entry. For example, if you had an entry with -@code{User} and @code{Host} fields of @code{'bob'} and @code{'%.loc.gov'}, -you would write: +@item +Setting a numeric column to a value such as @code{'10.34 a'}. The trailing +garbage is stripped and the remaining numeric part is inserted. If the value +doesn't make sense as a number at all, the column is set to @code{0}. -@example -mysql> SET PASSWORD FOR bob@@"%.loc.gov" = PASSWORD("newpass"); +@item +Inserting a string into a @code{CHAR}, @code{VARCHAR}, @code{TEXT}, or +@code{BLOB} column that exceeds the column's maximum length. The value is +truncated to the column's maximum length. -or +@item +Inserting a value into a date or time column that is illegal for the column +type. The column is set to the appropriate zero value for the type. +@end itemize -mysql> UPDATE mysql.user SET password=PASSWORD("newpass") where user="bob' and host="%.loc.gov"; -@end example +@findex REPLACE ... SELECT +@findex INSERT ... SELECT +@menu +* INSERT SELECT:: +* INSERT DELAYED:: +@end menu -@item SQL_AUTO_IS_NULL = 0 | 1 -If set to @code{1} (default) then one can find the last inserted row -for a table with an auto_increment row with the following construct: -@code{WHERE auto_increment_column IS NULL}. This is used by some -ODBC programs like Access. +@node INSERT SELECT, INSERT DELAYED, INSERT, INSERT +@subsection @code{INSERT ... SELECT} Syntax -@item AUTOCOMMIT= 0 | 1 -If set to @code{1} all changes to a table will be done at once. To start -a multi-command transaction, you have to use the @code{BEGIN} -statement. @xref{COMMIT}. If set to @code{0} you have to use @code{COMMIT} / -@code{ROLLBACK} to accept/revoke that transaction. @xref{COMMIT}. Note -that when you change from not @code{AUTOCOMMIT} mode to -@code{AUTOCOMMIT} mode, @strong{MySQL} will do an automatic -@code{COMMIT} on any open transactions. +@example +INSERT [LOW_PRIORITY] [IGNORE] [INTO] tbl_name [(column list)] SELECT ... +@end example -@item SQL_BIG_TABLES = 0 | 1 -@cindex table is full -If set to @code{1}, all temporary tables are stored on disk rather than in -memory. This will be a little slower, but you will not get the error -@code{The table tbl_name is full} for big @code{SELECT} operations that -require a large temporary table. The default value for a new connection is -@code{0} (that is, use in-memory temporary tables). +With @code{INSERT ... SELECT} statement you can quickly insert many rows +into a table from one or many tables. -@item SQL_BIG_SELECTS = 0 | 1 -If set to @code{0}, @strong{MySQL} will abort if a @code{SELECT} is attempted -that probably will take a very long time. This is useful when an inadvisable -@code{WHERE} statement has been issued. A big query is defined as a -@code{SELECT} that probably will have to examine more than -@code{max_join_size} rows. The default value for a new connection is -@code{1} (which will allow all @code{SELECT} statements). +@example +INSERT INTO tblTemp2 (fldID) SELECT tblTemp1.fldOrder_ID FROM tblTemp1 WHERE +tblTemp1.fldOrder_ID > 100; +@end example -@item SQL_BUFFER_RESULT = 0 | 1 -@code{SQL_BUFFER_RESULT} will force the result from @code{SELECT}'s -to be put into a temporary table. This will help @strong{MySQL} free the -table locks early and will help in cases where it takes a long time to -send the result set to the client. +The following conditions hold for an @code{INSERT ... SELECT} statement: -@item SQL_LOW_PRIORITY_UPDATES = 0 | 1 -If set to @code{1}, all @code{INSERT}, @code{UPDATE}, @code{DELETE}, and -and @code{LOCK TABLE WRITE} statements wait until there is no pending -@code{SELECT} or @code{LOCK TABLE READ} on the affected table. +@itemize @minus +@item +The query cannot contain an @code{ORDER BY} clause. -@item SQL_MAX_JOIN_SIZE = value | DEFAULT -Don't allow @code{SELECT}s that will probably need to examine more than -@code{value} row combinations. By setting this value, you can catch -@code{SELECT}s where keys are not used properly and that would probably -take a long time. Setting this to a value other than @code{DEFAULT} will reset -the @code{SQL_BIG_SELECTS} flag. If you set the @code{SQL_BIG_SELECTS} -flag again, the @code{SQL_MAX_JOIN_SIZE} variable will be ignored. -You can set a default value for this variable by starting @code{mysqld} with -@code{-O max_join_size=#}. +@item +The target table of the @code{INSERT} statement cannot appear in the +@code{FROM} clause of the @code{SELECT} part of the query because it's +forbidden in ANSI SQL to @code{SELECT} from the same table into which you are +inserting. (The problem is that the @code{SELECT} possibly would +find records that were inserted earlier during the same run. When using +sub-select clauses, the situation could easily be very confusing!) -@item SQL_SAFE_MODE = 0 | 1 -If set to @code{1}, @strong{MySQL} will abort if an @code{UPDATE} or -@code{DELETE} is attempted that doesn't use a key or @code{LIMIT} in the -@code{WHERE} clause. This makes it possible to catch wrong updates -when creating SQL commands by hand. +@item +@code{AUTO_INCREMENT} columns work as usual. -@item SQL_SELECT_LIMIT = value | DEFAULT -The maximum number of records to return from @code{SELECT} statements. If -a @code{SELECT} has a @code{LIMIT} clause, the @code{LIMIT} takes precedence -over the value of @code{SQL_SELECT_LIMIT}. The default value for a new -connection is ``unlimited.'' If you have changed the limit, the default value -can be restored by using a @code{SQL_SELECT_LIMIT} value of @code{DEFAULT}. +@item +You can use the C API function @code{mysql_info()} to get information about +the query. @xref{INSERT}. -@item SQL_LOG_OFF = 0 | 1 -If set to @code{1}, no logging will be done to the standard log for this -client, if the client has the @strong{process} privilege. This does not -affect the update log! +@item +To ensure that the update log/binary log can be used to re-create the +original tables, @strong{MySQL} will not allow concurrent inserts during +@code{INSERT .... SELECT}. +@end itemize -@item SQL_LOG_UPDATE = 0 | 1 -If set to @code{0}, no logging will be done to the update log for the client, -if the client has the @strong{process} privilege. This does not affect the -standard log! +You can of course also use @code{REPLACE} instead of @code{INSERT} to +overwrite old rows. -@item SQL_QUOTE_SHOW_CREATE = 0 | 1 -If set to @code{1}, @code{SHOW CREATE TABLE} will quote -table and column names. This is @strong{on} by default, -for replication of tables with fancy column names to work. -@ref{SHOW CREATE TABLE, , @code{SHOW CREATE TABLE}}. +@findex INSERT DELAYED +@findex DELAYED +@cindex INSERT DELAYED +@node INSERT DELAYED, , INSERT SELECT, INSERT +@subsection @code{INSERT DELAYED} syntax -@item TIMESTAMP = timestamp_value | DEFAULT -Set the time for this client. This is used to get the original timestamp if -you use the update log to restore rows. @code{timestamp_value} should be a -UNIX Epoch timestamp, not a @strong{MySQL} timestamp. +@example +INSERT DELAYED ... +@end example -@item LAST_INSERT_ID = # -Set the value to be returned from @code{LAST_INSERT_ID()}. This is stored in -the update log when you use @code{LAST_INSERT_ID()} in a command that updates -a table. +The @code{DELAYED} option for the @code{INSERT} statement is a +@strong{MySQL}-specific option that is very useful if you have clients +that can't wait for the @code{INSERT} to complete. This is a common +problem when you use @strong{MySQL} for logging and you also +periodically run @code{SELECT} and @code{UPDATE} statements that take a +long time to complete. @code{DELAYED} was introduced in @strong{MySQL} +Version 3.22.15. It is a @strong{MySQL} extension to ANSI SQL92. -@item INSERT_ID = # -Set the value to be used by the following @code{INSERT} or @code{ALTER TABLE} -command when inserting an @code{AUTO_INCREMENT} value. This is mainly used -with the update log. -@end table +@code{INSERT DELAYED} only works with @code{ISAM} and @code{MyISAM} +tables. Note that as @code{MyISAM} tables supports concurrent +@code{SELECT} and @code{INSERT}, if there is no free blocks in the +middle of the data file, you very seldom need to use @code{INSERT +DELAYED} with @code{MyISAM}. @xref{MyISAM}. -@findex ISOLATION LEVEL -@node SET TRANSACTION, GRANT, SET OPTION, Reference -@section @code{SET TRANSACTION} Syntax +When you use @code{INSERT DELAYED}, the client will get an OK at once +and the row will be inserted when the table is not in use by any other thread. -@example -SET [GLOBAL | SESSION] TRANSACTION ISOLATION LEVEL -[READ UNCOMMITTED | READ COMMITTED | REPEATABLE READ | SERIALIZABLE] -@end example +Another major benefit of using @code{INSERT DELAYED} is that inserts +from many clients are bundled together and written in one block. This is much +faster than doing many separate inserts. -Sets the transaction isolation level for the global, whole session or -the next transaction. +Note that currently the queued rows are only stored in memory until they are +inserted into the table. This means that if you kill @code{mysqld} +the hard way (@code{kill -9}) or if @code{mysqld} dies unexpectedly, any +queued rows that weren't written to disk are lost! -The default behavior is to set the isolation level for the next (not started) -transaction. +The following describes in detail what happens when you use the +@code{DELAYED} option to @code{INSERT} or @code{REPLACE}. In this +description, the ``thread'' is the thread that received an @code{INSERT +DELAYED} command and ``handler'' is the thread that handles all +@code{INSERT DELAYED} statements for a particular table. -If you set the @code{GLOBAL} privilege it will affect all new created threads. -You will need the @code{PROCESS} privilege to do do this. +@itemize @bullet +@item +When a thread executes a @code{DELAYED} statement for a table, a handler +thread is created to process all @code{DELAYED} statements for the table, if +no such handler already exists. -Setting the @code{SESSION} privilege will affect the following and all -future transactions. +@item +The thread checks whether or not the handler has acquired a @code{DELAYED} +lock already; if not, it tells the handler thread to do so. The +@code{DELAYED} lock can be obtained even if other threads have a @code{READ} +or @code{WRITE} lock on the table. However, the handler will wait for all +@code{ALTER TABLE} locks or @code{FLUSH TABLES} to ensure that the table +structure is up to date. -You can set the default isolation level for @code{mysqld} with -@code{--transaction-isolation=...}. @xref{Command-line options}. +@item +The thread executes the @code{INSERT} statement, but instead of writing +the row to the table, it puts a copy of the final row into a queue that +is managed by the handler thread. Any syntax errors are noticed by the +thread and reported to the client program. -@cindex privileges, granting -@cindex privileges, revoking -@cindex global privileges -@cindex revoking, privleges -@cindex granting, privleges +@item +The client can't report the number of duplicates or the @code{AUTO_INCREMENT} +value for the resulting row; it can't obtain them from the server, because +the @code{INSERT} returns before the insert operation has been completed. If +you use the C API, the @code{mysql_info()} function doesn't return anything +meaningful, for the same reason. -@findex GRANT -@findex REVOKE -@node GRANT, CREATE INDEX, SET TRANSACTION, Reference -@section @code{GRANT} and @code{REVOKE} Syntax +@item +The update log is updated by the handler thread when the row is inserted into +the table. In case of multiple-row inserts, the update log is updated when +the first row is inserted. -@example -GRANT priv_type [(column_list)] [, priv_type [(column_list)] ...] - ON @{tbl_name | * | *.* | db_name.*@} - TO user_name [IDENTIFIED BY 'password'] - [, user_name [IDENTIFIED BY 'password'] ...] - [WITH GRANT OPTION] +@item +After every @code{delayed_insert_limit} rows are written, the handler checks +whether or not any @code{SELECT} statements are still pending. If so, it +allows these to execute before continuing. -REVOKE priv_type [(column_list)] [, priv_type [(column_list)] ...] - ON @{tbl_name | * | *.* | db_name.*@} - FROM user_name [, user_name ...] -@end example +@cindex delayed_insert_limit +@cindex timeout +@item +When the handler has no more rows in its queue, the table is unlocked. If no +new @code{INSERT DELAYED} commands are received within +@code{delayed_insert_timeout} seconds, the handler terminates. -@code{GRANT} is implemented in @strong{MySQL} Version 3.22.11 or later. For -earlier @strong{MySQL} versions, the @code{GRANT} statement does nothing. +@item +If more than @code{delayed_queue_size} rows are pending already in a +specific handler queue, the thread requesting @code{INSERT DELAYED} +waits until there is room in the queue. This is done to ensure that +the @code{mysqld} server doesn't use all memory for the delayed memory +queue. -The @code{GRANT} and @code{REVOKE} commands allow system administrators -to create users and grant and revoke rights to @strong{MySQL} users at -four privilege levels: +@item +The handler thread will show up in the @strong{MySQL} process list +with @code{delayed_insert} in the @code{Command} column. It will +be killed if you execute a @code{FLUSH TABLES} command or kill it with +@code{KILL thread_id}. However, it will first store all queued rows into the +table before exiting. During this time it will not accept any new +@code{INSERT} commands from another thread. If you execute an @code{INSERT +DELAYED} command after this, a new handler thread will be created. -@table @strong -@item Global level -Global privileges apply to all databases on a given server. These privileges -are stored in the @code{mysql.user} table. +@item +Note that the above means that @code{INSERT DELAYED} commands have higher +priority than normal @code{INSERT} commands if there is an @code{INSERT +DELAYED} handler already running! Other update commands will have to wait +until the @code{INSERT DELAYED} queue is empty, someone kills the handler +thread (with @code{KILL thread_id}), or someone executes @code{FLUSH TABLES}. -@item Database level -Database privileges apply to all tables in a given database. These privileges -are stored in the @code{mysql.db} and @code{mysql.host} tables. +@item +The following status variables provide information about @code{INSERT +DELAYED} commands: -@item Table level -Table privileges apply to all columns in a given table. These privileges are -stored in the @code{mysql.tables_priv} table. +@multitable @columnfractions .35 .65 +@item @strong{Variable} @tab @strong{Meaning} +@item @code{Delayed_insert_threads} @tab Number of handler threads +@item @code{Delayed_writes} @tab Number of rows written with @code{INSERT DELAYED} +@item @code{Not_flushed_delayed_rows} @tab Number of rows waiting to be written +@end multitable -@item Column level -Column privileges apply to single columns in a given table. These privileges are -stored in the @code{mysql.columns_priv} table. -@end table +You can view these variables by issuing a @code{SHOW STATUS} statement or +by executing a @code{mysqladmin extended-status} command. +@end itemize -If you give a grant for a users that doesn't exists, that user is created. -For examples of how @code{GRANT} works, see @ref{Adding users}. +Note that @code{INSERT DELAYED} is slower than a normal INSERT if the +table is not in use. There is also the additional overhead for the +server to handle a separate thread for each table on which you use +@code{INSERT DELAYED}. This means that you should only use @code{INSERT +DELAYED} when you are really sure you need it! -For the @code{GRANT} and @code{REVOKE} statements, @code{priv_type} may be -specified as any of the following: +@findex REPLACE +@node REPLACE, LOAD DATA, INSERT, Reference +@section @code{REPLACE} Syntax @example -ALL PRIVILEGES FILE RELOAD -ALTER INDEX SELECT -CREATE INSERT SHUTDOWN -DELETE PROCESS UPDATE -DROP REFERENCES USAGE + REPLACE [LOW_PRIORITY | DELAYED] + [INTO] tbl_name [(col_name,...)] + VALUES (expression,...) +or REPLACE [LOW_PRIORITY | DELAYED] + [INTO] tbl_name [(col_name,...)] + SELECT ... +or REPLACE [LOW_PRIORITY | DELAYED] + [INTO] tbl_name + SET col_name=expression, col_name=expression,... @end example -@code{ALL} is a synonym for @code{ALL PRIVILEGES}. @code{REFERENCES} is not -yet implemented. @code{USAGE} is currently a synonym for ``no privileges.'' -It can be used when you want to create a user that has no privileges. +@code{REPLACE} works exactly like @code{INSERT}, except that if an old +record in the table has the same value as a new record on a unique index, +the old record is deleted before the new record is inserted. +@xref{INSERT, , @code{INSERT}}. -To revoke the @strong{grant} privilege from a user, use a @code{priv_type} -value of @code{GRANT OPTION}: +In other words, you can't access the values of the old row from a +@code{REPLACE} statement. In some old @strong{MySQL} version it looked +like you could do this, but that was a bug that has been corrected. + +When one uses a @code{REPLACE} command, @code{mysql_affected_rows()} +will return 2 if the new row replaced and old row. This is because in +this case one row was inserted and then the duplicate was deleted. + +The above makes it easy to check if @code{REPLACE} added or replaced a +row. + +@findex LOAD DATA INFILE +@node LOAD DATA, UPDATE, REPLACE, Reference +@section @code{LOAD DATA INFILE} Syntax @example -REVOKE GRANT OPTION ON ... FROM ...; +LOAD DATA [LOW_PRIORITY | CONCURRENT] [LOCAL] INFILE 'file_name.txt' + [REPLACE | IGNORE] + INTO TABLE tbl_name + [FIELDS + [TERMINATED BY '\t'] + [[OPTIONALLY] ENCLOSED BY ''] + [ESCAPED BY '\\' ] + ] + [LINES TERMINATED BY '\n'] + [IGNORE number LINES] + [(col_name,...)] @end example -The only @code{priv_type} values you can specify for a table are @code{SELECT}, -@code{INSERT}, @code{UPDATE}, @code{DELETE}, @code{CREATE}, @code{DROP}, -@code{GRANT}, @code{INDEX}, and @code{ALTER}. +The @code{LOAD DATA INFILE} statement reads rows from a text file into a +table at a very high speed. If the @code{LOCAL} keyword is specified, the +file is read from the client host. If @code{LOCAL} is not specified, the +file must be located on the server. (@code{LOCAL} is available in +@strong{MySQL} Version 3.22.6 or later.) -The only @code{priv_type} values you can specify for a column (that is, when -you use a @code{column_list} clause) are @code{SELECT}, @code{INSERT}, and -@code{UPDATE}. +For security reasons, when reading text files located on the server, the +files must either reside in the database directory or be readable by all. +Also, to use @code{LOAD DATA INFILE} on server files, you must have the +@strong{file} privilege on the server host. +@xref{Privileges provided}. -You can set global privileges by using @code{ON *.*} syntax. You can set -database privileges by using @code{ON db_name.*} syntax. If you specify -@code{ON *} and you have a current database, you will set the privileges for -that database. (@strong{WARNING:} If you specify @code{ON *} and you -@emph{don't} have a current database, you will affect the global privileges!) +If you specify the keyword @code{LOW_PRIORITY}, execution of the +@code{LOAD DATA} statement is delayed until no other clients are reading +from the table. -In order to accommodate granting rights to users from arbitrary hosts, -@strong{MySQL} supports specifying the @code{user_name} value in the form -@code{user@@host}. If you want to specify a @code{user} string -containing special characters (such as @samp{-}), or a @code{host} string -containing special characters or wild-card characters (such as @samp{%}), you -can quote the user or host name (for example, @code{'test-user'@@'test-hostname'}). +If you specify the keyword @code{CONCURRENT} with a @code{MyISAM} table, +then other threads can retrieve data from the table while @code{LOAD +DATA} is executing. Using this option will of course affect the +performance of @code{LOAD DATA} a bit even if no other thread is using +the table at the same time. -You can specify wild cards in the hostname. For example, -@code{user@@"%.loc.gov"} applies to @code{user} for any host in the -@code{loc.gov} domain, and @code{user@@"144.155.166.%"} applies to @code{user} -for any host in the @code{144.155.166} class C subnet. +Using @code{LOCAL} will be a bit slower than letting the server access the +files directly, because the contents of the file must travel from the client +host to the server host. On the other hand, you do not need the +@strong{file} privilege to load local files. -The simple form @code{user} is a synonym for @code{user@@"%"}. -@strong{NOTE:} If you allow anonymous users to connect to the @strong{MySQL} -server (which is the default), you should also add all local users as -@code{user@@localhost} because otherwise the anonymous user entry for the -local host in the @code{mysql.user} table will be used when the user tries to -log into the @strong{MySQL} server from the local machine! Anonymous users -are defined by inserting entries with @code{User=''} into the -@code{mysql.user} table. You can verify if this applies to you by executing -this query: +@c old version +If you are using @strong{MySQL} before Version 3.23.24 you can't read from a +FIFO with @code{LOAD DATA INFILE}. If you need to read from a FIFO (for +example the output from gunzip), use @code{LOAD DATA LOCAL INFILE} +instead. -@example -mysql> SELECT Host,User FROM mysql.user WHERE User=''; -@end example +@cindex @code{mysqlimport} +You can also load data files by using the @code{mysqlimport} utility; it +operates by sending a @code{LOAD DATA INFILE} command to the server. The +@code{--local} option causes @code{mysqlimport} to read data files from the +client host. You can specify the @code{--compress} option to get better +performance over slow networks if the client and server support the +compressed protocol. -For the moment, @code{GRANT} only supports host, table, database, and -column names up to 60 characters long. A user name can be up to 16 -characters. +When locating files on the server host, the server uses the following rules: -The privileges for a table or column are formed from the -logical OR of the privileges at each of the four privilege -levels. For example, if the @code{mysql.user} table specifies that a -user has a global @strong{select} privilege, this can't be denied by an -entry at the database, table, or column level. +@itemize @bullet +@item +If an absolute pathname is given, the server uses the pathname as is. -The privileges for a column can be calculated as follows: +@item +If a relative pathname with one or more leading components is given, +the server searches for the file relative to the server's data directory. + +@item +If a filename with no leading components is given, the server looks for +the file in the database directory of the current database. +@end itemize + +Note that these rules mean a file given as @file{./myfile.txt} is read from +the server's data directory, whereas a file given as @file{myfile.txt} is +read from the database directory of the current database. For example, +the following @code{LOAD DATA} statement reads the file @file{data.txt} +from the database directory for @code{db1} because @code{db1} is the current +database, even though the statement explicitly loads the file into a +table in the @code{db2} database: @example -global privileges -OR (database privileges AND host privileges) -OR table privileges -OR column privileges +mysql> USE db1; +mysql> LOAD DATA INFILE "data.txt" INTO TABLE db2.my_table; @end example -In most cases, you grant rights to a user at only one of the privilege -levels, so life isn't normally as complicated as above. The details of the -privilege-checking procedure are presented in -@ref{Privilege system}. +The @code{REPLACE} and @code{IGNORE} keywords control handling of input +records that duplicate existing records on unique key values. If you specify +@code{REPLACE}, new rows replace existing rows that have the same unique key +value. If you specify @code{IGNORE}, input rows that duplicate an existing +row on a unique key value are skipped. If you don't specify either option, an +error occurs when a duplicate key value is found, and the rest of the text +file is ignored. -If you grant privileges for a user/hostname combination that does not exist -in the @code{mysql.user} table, an entry is added and remains there until -deleted with a @code{DELETE} command. In other words, @code{GRANT} may -create @code{user} table entries, but @code{REVOKE} will not remove them; -you must do that explicitly using @code{DELETE}. +If you load data from a local file using the @code{LOCAL} keyword, the server +has no way to stop transmission of the file in the middle of the operation, +so the default bahavior is the same as if @code{IGNORE} is specified. -@cindex passwords, setting -In @strong{MySQL} Version 3.22.12 or later, -if a new user is created or if you have global grant privileges, the user's -password will be set to the password specified by the @code{IDENTIFIED BY} -clause, if one is given. If the user already had a password, it is replaced -by the new one. +If you use @code{LOAD DATA INFILE} on an empty @code{MyISAM} table, +all non-unique indexes are created in a separate batch (like in @code{REPAIR}). +This normally makes @code{LOAD DATA INFILE} much faster when you have many +indexes. -@strong{WARNING:} If you create a new user but do not specify an -@code{IDENTIFIED BY} clause, the user has no password. This is insecure. +@code{LOAD DATA INFILE} is the complement of @code{SELECT ... INTO OUTFILE}. +@xref{SELECT, , @code{SELECT}}. +To write data from a database to a file, use @code{SELECT ... INTO OUTFILE}. +To read the file back into the database, use @code{LOAD DATA INFILE}. +The syntax of the @code{FIELDS} and @code{LINES} clauses is the same for +both commands. Both clauses are optional, but @code{FIELDS} +must precede @code{LINES} if both are specified. -Passwords can also be set with the @code{SET PASSWORD} command. -@xref{SET OPTION, , @code{SET OPTION}}. +If you specify a @code{FIELDS} clause, +each of its subclauses (@code{TERMINATED BY}, @code{[OPTIONALLY] ENCLOSED +BY}, and @code{ESCAPED BY}) is also optional, except that you must +specify at least one of them. -If you grant privileges for a database, an entry in the @code{mysql.db} -table is created if needed. When all privileges for the database have been -removed with @code{REVOKE}, this entry is deleted. +If you don't specify a @code{FIELDS} clause, the defaults are the +same as if you had written this: -If a user doesn't have any privileges on a table, the table is not displayed -when the user requests a list of tables (for example, with a @code{SHOW TABLES} -statement). +@example +FIELDS TERMINATED BY '\t' ENCLOSED BY '' ESCAPED BY '\\' +@end example -The @code{WITH GRANT OPTION} clause gives the user the ability to give -to other users any privileges the user has at the specified privilege level. -You should be careful to whom you give the @strong{grant} privilege, as two -users with different privileges may be able to join privileges! +If you don't specify a @code{LINES} clause, the default +is the same as if you had written this: -You cannot grant another user a privilege you don't have yourself; -the @strong{grant} privilege allows you to give away only those privileges -you possess. +@example +LINES TERMINATED BY '\n' +@end example -Be aware that when you grant a user the @strong{grant} privilege at a -particular privilege level, any privileges the user already possesses (or -is given in the future!) at that level are also grantable by that user. -Suppose you grant a user the @strong{insert} privilege on a database. If -you then grant the @strong{select} privilege on the database and specify -@code{WITH GRANT OPTION}, the user can give away not only the @strong{select} -privilege, but also @strong{insert}. If you then grant the @strong{update} -privilege to the user on the database, the user can give away the -@strong{insert}, @strong{select} and @strong{update}. - -You should not grant @strong{alter} privileges to a normal user. If you -do that, the user can try to subvert the privilege system by renaming -tables! +In other words, the defaults cause @code{LOAD DATA INFILE} to act as follows +when reading input: -Note that if you are using table or column privileges for even one user, the -server examines table and column privileges for all users and this will slow -down @strong{MySQL} a bit. +@itemize @bullet +@item +Look for line boundaries at newlines. -When @code{mysqld} starts, all privileges are read into memory. -Database, table, and column privileges take effect at once, and -user-level privileges take effect the next time the user connects. -Modifications to the grant tables that you perform using @code{GRANT} or -@code{REVOKE} are noticed by the server immediately. -If you modify the grant tables manually (using @code{INSERT}, @code{UPDATE}, -etc.), you should execute a @code{FLUSH PRIVILEGES} statement or run -@code{mysqladmin flush-privileges} to tell the server to reload the grant -tables. -@xref{Privilege changes}. +@item +Break lines into fields at tabs. -@cindex ANSI SQL, differences from -The biggest differences between the ANSI SQL and @strong{MySQL} versions of -@code{GRANT} are: +@item +Do not expect fields to be enclosed within any quoting characters. -@itemize @bullet @item -In @strong{MySQL} privileges are given for an username + hostname combination -and not only for an username. +Interpret occurrences of tab, newline, or @samp{\} preceded by +@samp{\} as literal characters that are part of field values. +@end itemize + +Conversely, the defaults cause @code{SELECT ... INTO OUTFILE} to act as +follows when writing output: +@itemize @bullet @item -ANSI SQL doesn't have global or database-level privileges, and ANSI SQL -doesn't support all privilege types that @strong{MySQL} supports. -@strong{MySQL} doesn't support the ANSI SQL @code{TRIGGER}, @code{EXECUTE} or -@code{UNDER} privileges. +Write tabs between fields. @item -ANSI SQL privileges are structured in a hierarchal manner. If you remove -an user, all privileges the user has granted are revoked. In -@strong{MySQL} the granted privileges are not automatically revoked, but -you have to revoke these yourself if needed. +Do not enclose fields within any quoting characters. @item -If you in @strong{MySQL} have the @code{INSERT} grant on only part of the -columns in a table, you can execute @code{INSERT} statements on the -table; The columns for which you don't have the @code{INSERT} privilege -will set to their default values. ANSI SQL requires you to have the -@code{INSERT} privilege on all columns. +Use @samp{\} to escape instances of tab, newline or @samp{\} that occur +within field values. @item -When you drop a table in ANSI SQL, all privileges for the table are revoked. -If you revoke a privilege in ANSI SQL, all privileges that were granted based -on this privilege are also revoked. In @strong{MySQL}, privileges can be -dropped only with explicit @code{REVOKE} commands or by manipulating the -@strong{MySQL} grant tables. +Write newlines at the ends of lines. @end itemize -@cindex indexes -@cindex indexes, multi-part -@cindex multi-part index -@findex CREATE INDEX -@node CREATE INDEX, DROP INDEX, GRANT, Reference -@section @code{CREATE INDEX} Syntax +Note that to write @code{FIELDS ESCAPED BY '\\'}, you must specify two +backslashes for the value to be read as a single backslash. + +The @code{IGNORE number LINES} option can be used to ignore a header of +column names at the start of the file: @example -CREATE [UNIQUE|FULLTEXT] INDEX index_name ON tbl_name (col_name[(length)],... ) +mysql> LOAD DATA INFILE "/tmp/file_name" into table test IGNORE 1 LINES; @end example -The @code{CREATE INDEX} statement doesn't do anything in @strong{MySQL} prior -to Version 3.22. In Version 3.22 or later, @code{CREATE INDEX} is mapped to an -@code{ALTER TABLE} statement to create indexes. -@xref{ALTER TABLE, , @code{ALTER TABLE}}. +When you use @code{SELECT ... INTO OUTFILE} in tandem with @code{LOAD +DATA INFILE} to write data from a database into a file and then read +the file back into the database later, the field and line handling +options for both commands must match. Otherwise, @code{LOAD DATA +INFILE} will not interpret the contents of the file properly. Suppose +you use @code{SELECT ... INTO OUTFILE} to write a file with +fields delimited by commas: -Normally, you create all indexes on a table at the time the table itself -is created with @code{CREATE TABLE}. -@xref{CREATE TABLE, , @code{CREATE TABLE}}. -@code{CREATE INDEX} allows you to add indexes to existing tables. +@example +mysql> SELECT * INTO OUTFILE 'data.txt' + FIELDS TERMINATED BY ',' + FROM ...; +@end example -A column list of the form @code{(col1,col2,...)} creates a multiple-column -index. Index values are formed by concatenating the values of the given -columns. +To read the comma-delimited file back in, the correct statement would be: -For @code{CHAR} and @code{VARCHAR} columns, indexes can be created that -use only part of a column, using @code{col_name(length)} syntax. (On -@code{BLOB} and @code{TEXT} columns the length is required). The -statement shown below creates an index using the first 10 characters of -the @code{name} column: +@example +mysql> LOAD DATA INFILE 'data.txt' INTO TABLE table2 + FIELDS TERMINATED BY ','; +@end example + +If instead you tried to read in the file with the statement shown below, it +wouldn't work because it instructs @code{LOAD DATA INFILE} to look for +tabs between fields: @example -mysql> CREATE INDEX part_of_name ON customer (name(10)); +mysql> LOAD DATA INFILE 'data.txt' INTO TABLE table2 + FIELDS TERMINATED BY '\t'; @end example -Because most names usually differ in the first 10 characters, this index should -not be much slower than an index created from the entire @code{name} column. -Also, using partial columns for indexes can make the index file much smaller, -which could save a lot of disk space and might also speed up @code{INSERT} -operations! +The likely result is that each input line would be interpreted as +a single field. -Note that you can only add an index on a column that can have @code{NULL} -values or on a @code{BLOB}/@code{TEXT} column if you are using -@strong{MySQL} Version 3.23.2 or newer and are using the @code{MyISAM} -table type. +@code{LOAD DATA INFILE} can be used to read files obtained from +external sources, too. For example, a file in dBASE format will have +fields separated by commas and enclosed in double quotes. If lines in +the file are terminated by newlines, the command shown below +illustrates the field and line handling options you would use to load +the file: -For more information about how @strong{MySQL} uses indexes, see -@ref{MySQL indexes, , @strong{MySQL} indexes}. +@example +mysql> LOAD DATA INFILE 'data.txt' INTO TABLE tbl_name + FIELDS TERMINATED BY ',' ENCLOSED BY '"' + LINES TERMINATED BY '\n'; +@end example -@code{FULLTEXT} indexes can index only @code{VARCHAR} and -@code{TEXT} columns, and only in @code{MyISAM} tables. @code{FULLTEXT} indexes -are available in @strong{MySQL} Version 3.23.23 and later. -@ref{Fulltext Search}. +Any of the field or line handling options may specify an empty string +(@code{''}). If not empty, the @code{FIELDS [OPTIONALLY] ENCLOSED BY} +and @code{FIELDS ESCAPED BY} values must be a single character. The +@code{FIELDS TERMINATED BY} and @code{LINES TERMINATED BY} values may +be more than one character. For example, to write lines that are +terminated by carriage return-linefeed pairs, or to read a file +containing such lines, specify a @code{LINES TERMINATED BY '\r\n'} +clause. -@findex DROP INDEX -@node DROP INDEX, Comments, CREATE INDEX, Reference -@section @code{DROP INDEX} Syntax +For example, to read a file of jokes, that are separated with a line +of @code{%%}, into a SQL table you can do: @example -DROP INDEX index_name ON tbl_name +create table jokes (a int not null auto_increment primary key, joke text +not null); +load data infile "/tmp/jokes.txt" into table jokes fields terminated by "" +lines terminated by "\n%%\n" (joke); @end example -@code{DROP INDEX} drops the index named @code{index_name} from the table -@code{tbl_name}. @code{DROP INDEX} doesn't do anything in @strong{MySQL} -prior to Version 3.22. In Version 3.22 or later, @code{DROP INDEX} is mapped to an -@code{ALTER TABLE} statement to drop the index. -@xref{ALTER TABLE, , @code{ALTER TABLE}}. +@code{FIELDS [OPTIONALLY] ENCLOSED BY} controls quoting of fields. For +output (@code{SELECT ... INTO OUTFILE}), if you omit the word +@code{OPTIONALLY}, all fields are enclosed by the @code{ENCLOSED BY} +character. An example of such output (using a comma as the field +delimiter) is shown below: -@findex Comment syntax -@cindex comments, adding -@node Comments, CREATE FUNCTION, DROP INDEX, Reference -@section Comment Syntax +@example +"1","a string","100.20" +"2","a string containing a , comma","102.20" +"3","a string containing a \" quote","102.20" +"4","a string containing a \", quote and comma","102.20" +@end example -The @strong{MySQL} server supports the @code{# to end of line}, @code{-- -to end of line} and @code{/* in-line or multiple-line */} comment -styles: +If you specify @code{OPTIONALLY}, the @code{ENCLOSED BY} character is +used only to enclose @code{CHAR} and @code{VARCHAR} fields: @example -mysql> select 1+1; # This comment continues to the end of line -mysql> select 1+1; -- This comment continues to the end of line -mysql> select 1 /* this is an in-line comment */ + 1; -mysql> select 1+ -/* -this is a -multiple-line comment -*/ -1; +1,"a string",100.20 +2,"a string containing a , comma",102.20 +3,"a string containing a \" quote",102.20 +4,"a string containing a \", quote and comma",102.20 @end example -Note that the @code{--} comment style requires you to have at least one space -after the @code{--}! +Note that occurrences of the @code{ENCLOSED BY} character within a +field value are escaped by prefixing them with the @code{ESCAPED BY} +character. Also note that if you specify an empty @code{ESCAPED BY} +value, it is possible to generate output that cannot be read properly by +@code{LOAD DATA INFILE}. For example, the output just shown above would +appear as shown below if the escape character is empty. Observe that the +second field in the fourth line contains a comma following the quote, which +(erroneously) appears to terminate the field: -Although the server understands the comment syntax just described, -there are some limitations on the way that the @code{mysql} client -parses @code{/* ... */} comments: +@example +1,"a string",100.20 +2,"a string containing a , comma",102.20 +3,"a string containing a " quote",102.20 +4,"a string containing a ", quote and comma",102.20 +@end example + +For input, the @code{ENCLOSED BY} character, if present, is stripped from the +ends of field values. (This is true whether or not @code{OPTIONALLY} is +specified; @code{OPTIONALLY} has no effect on input interpretation.) +Occurrences of the @code{ENCLOSED BY} character preceded by the +@code{ESCAPED BY} character are interpreted as part of the current field +value. In addition, duplicated @code{ENCLOSED BY} characters occurring +within fields are interpreted as single @code{ENCLOSED BY} characters if the +field itself starts with that character. For example, if @code{ENCLOSED BY +'"'} is specified, quotes are handled as shown below: + +@example +"The ""BIG"" boss" -> The "BIG" boss +The "BIG" boss -> The "BIG" boss +The ""BIG"" boss -> The ""BIG"" boss +@end example +@code{FIELDS ESCAPED BY} controls how to write or read special characters. +If the @code{FIELDS ESCAPED BY} character is not empty, it is used to prefix +the following characters on output: @itemize @bullet @item -Single-quote and double-quote characters are taken to indicate the beginning -of a quoted string, even within a comment. If the quote is not matched by a -second quote within the comment, the parser doesn't realize the comment has -ended. If you are running @code{mysql} interactively, you can tell that it -has gotten confused like this because the prompt changes from @code{mysql>} -to @code{'>} or @code{">}. - +The @code{FIELDS ESCAPED BY} character @item -A semicolon is taken to indicate the end of the current SQL statement -and anything following it to indicate the beginning of the next statement. +The @code{FIELDS [OPTIONALLY] ENCLOSED BY} character +@item +The first character of the @code{FIELDS TERMINATED BY} and +@code{LINES TERMINATED BY} values +@item +ASCII @code{0} (what is actually written following the escape character is +ASCII @code{'0'}, not a zero-valued byte) @end itemize -These limitations apply both when you run @code{mysql} interactively -and when you put commands in a file and tell @code{mysql} to read its -input from that file with @code{mysql < some-file}. +If the @code{FIELDS ESCAPED BY} character is empty, no characters are escaped. +It is probably not a good idea to specify an empty escape character, +particularly if field values in your data contain any of the characters in +the list just given. -@strong{MySQL} doesn't support the @samp{--} ANSI SQL comment style. -@xref{Missing comments}. +For input, if the @code{FIELDS ESCAPED BY} character is not empty, occurrences +of that character are stripped and the following character is taken literally +as part of a field value. The exceptions are an escaped @samp{0} or +@samp{N} (for example, @code{\0} or @code{\N} if the escape character is +@samp{\}). These sequences are interpreted as ASCII @code{0} (a zero-valued +byte) and @code{NULL}. See below for the rules on @code{NULL} handling. -@findex CREATE FUNCTION -@findex DROP FUNCTION -@findex UDF functions -@findex User-defined functions -@findex Functions, user-defined -@node CREATE FUNCTION, Reserved words, Comments, Reference -@section @code{CREATE FUNCTION/DROP FUNCTION} Syntax +For more information about @samp{\}-escape syntax, +see @ref{Literals}. -@example -CREATE [AGGREGATE] FUNCTION function_name RETURNS @{STRING|REAL|INTEGER@} - SONAME shared_library_name +In certain cases, field and line handling options interact: -DROP FUNCTION function_name -@end example +@itemize @bullet +@item +If @code{LINES TERMINATED BY} is an empty string and @code{FIELDS +TERMINATED BY} is non-empty, lines are also terminated with +@code{FIELDS TERMINATED BY}. +@item +If the @code{FIELDS TERMINATED BY} and @code{FIELDS ENCLOSED BY} values are +both empty (@code{''}), a fixed-row (non-delimited) format is used. With +fixed-row format, no delimiters are used between fields. Instead, column +values are written and read using the ``display'' widths of the columns. For +example, if a column is declared as @code{INT(7)}, values for the column are +written using 7-character fields. On input, values for the column are +obtained by reading 7 characters. Fixed-row format also affects handling of +@code{NULL} values; see below. Note that fixed-size format will not work +if you are using a multi-byte character set. +@end itemize -A user-definable function (UDF) is a way to extend @strong{MySQL} with a new -function that works like native (built in) @strong{MySQL} functions such as -@code{ABS()} and @code{CONCAT()}. +Handling of @code{NULL} values varies, depending on the @code{FIELDS} and +@code{LINES} options you use: -@code{AGGREGATE} is a new option for @strong{MySQL} Version 3.23. An -@code{AGGREGATE} function works exactly like a native @strong{MySQL} -@code{GROUP} function like @code{SUM} or @code{COUNT()}. +@itemize @bullet +@item +For the default @code{FIELDS} and @code{LINES} values, +@code{NULL} is written as @code{\N} for output and @code{\N} is read +as @code{NULL} for input (assuming the @code{ESCAPED BY} character +is @samp{\}). -@code{CREATE FUNCTION} saves the function's name, type, and shared library -name in the @code{mysql.func} system table. You must have the -@strong{insert} and @strong{delete} privileges for the @code{mysql} database -to create and drop functions. +@item +If @code{FIELDS ENCLOSED BY} is not empty, a field containing the literal +word @code{NULL} as its value is read as a @code{NULL} value (this differs +from the word @code{NULL} enclosed within @code{FIELDS ENCLOSED BY} +characters, which is read as the string @code{'NULL'}). -All active functions are reloaded each time the server starts, unless -you start @code{mysqld} with the @code{--skip-grant-tables} option. In -this case, UDF initialization is skipped and UDFs are unavailable. -(An active function is one that has been loaded with @code{CREATE FUNCTION} -and not removed with @code{DROP FUNCTION}.) +@item +If @code{FIELDS ESCAPED BY} is empty, @code{NULL} is written as the word +@code{NULL}. -For instructions on writing user-definable functions, see @ref{Adding -functions}. For the UDF mechanism to work, functions must be written in C or -C++, your operating system must support dynamic loading and you must have -compiled @code{mysqld} dynamically (not statically). +@item +With fixed-row format (which happens when @code{FIELDS TERMINATED BY} and +@code{FIELDS ENCLOSED BY} are both empty), @code{NULL} is written as an empty +string. Note that this causes both @code{NULL} values and empty strings in +the table to be indistinguishable when written to the file because they are +both written as empty strings. If you need to be able to tell the two apart +when reading the file back in, you should not use fixed-row format. +@end itemize -@cindex keywords -@cindex reserved words, exceptions -@node Reserved words, , CREATE FUNCTION, Reference -@section Is MySQL Picky About Reserved Words? +Some cases are not supported by @code{LOAD DATA INFILE}: +@itemize @bullet -A common problem stems from trying to create a table with column names that -use the names of datatypes or functions built into @strong{MySQL}, such as -@code{TIMESTAMP} or @code{GROUP}. You're allowed to do it (for example, -@code{ABS} is an allowed column name), but whitespace is not allowed between -a function name and the @samp{(} when using functions whose names are also -column names. +@item +Fixed-size rows (@code{FIELDS TERMINATED BY} and @code{FIELDS ENCLOSED +BY} both empty) and @code{BLOB} or @code{TEXT} columns. -The following words are explicitly reserved in @strong{MySQL}. Most of -them are forbidden by ANSI SQL92 as column and/or table names -(for example, @code{group}). -A few are reserved because @strong{MySQL} needs them and is -(currently) using a @code{yacc} parser: +@item +If you specify one separator that is the same as or a prefix of another, +@code{LOAD DATA INFILE} won't be able to interpret the input properly. +For example, the following @code{FIELDS} clause would cause problems: -@c This is fixed by including the symbols table from lex.h here and then running -@c fix-mysql-reserved-words in emacs (or let David do it): -@c (defun fix-mysql-reserved-words () -@c (interactive) -@c (let ((cnt 0)) -@c (insert "\n@item ") -@c (while (looking-at "[ \t]*{ +\"\\([^\"]+\\)\"[ \t]*,.*\n") -@c (replace-match "@code{\\1}") -@c (incf cnt) -@c (if (> cnt 3) -@c (progn -@c (setf cnt 0) -@c (insert "\n@item ")) -@c (insert " @tab "))))) -@c But remove the non alphanumeric entries by hand first. -@c Updated after 3.23.4 990928 by David +@example +FIELDS TERMINATED BY '"' ENCLOSED BY '"' +@end example -@multitable @columnfractions .25 .25 .25 .25 -@item @code{action} @tab @code{add} @tab @code{aggregate} @tab @code{all} -@item @code{alter} @tab @code{after} @tab @code{and} @tab @code{as} -@item @code{asc} @tab @code{avg} @tab @code{avg_row_length} @tab @code{auto_increment} -@item @code{between} @tab @code{bigint} @tab @code{bit} @tab @code{binary} -@item @code{blob} @tab @code{bool} @tab @code{both} @tab @code{by} -@item @code{cascade} @tab @code{case} @tab @code{char} @tab @code{character} -@item @code{change} @tab @code{check} @tab @code{checksum} @tab @code{column} -@item @code{columns} @tab @code{comment} @tab @code{constraint} @tab @code{create} -@item @code{cross} @tab @code{current_date} @tab @code{current_time} @tab @code{current_timestamp} -@item @code{data} @tab @code{database} @tab @code{databases} @tab @code{date} -@item @code{datetime} @tab @code{day} @tab @code{day_hour} @tab @code{day_minute} -@item @code{day_second} @tab @code{dayofmonth} @tab @code{dayofweek} @tab @code{dayofyear} -@item @code{dec} @tab @code{decimal} @tab @code{default} @tab @code{delayed} -@item @code{delay_key_write} @tab @code{delete} @tab @code{desc} @tab @code{describe} -@item @code{distinct} @tab @code{distinctrow} @tab @code{double} @tab @code{drop} -@item @code{end} @tab @code{else} @tab @code{escape} @tab @code{escaped} -@item @code{enclosed} @tab @code{enum} @tab @code{explain} @tab @code{exists} -@item @code{fields} @tab @code{file} @tab @code{first} @tab @code{float} -@item @code{float4} @tab @code{float8} @tab @code{flush} @tab @code{foreign} -@item @code{from} @tab @code{for} @tab @code{full} @tab @code{function} -@item @code{global} @tab @code{grant} @tab @code{grants} @tab @code{group} -@item @code{having} @tab @code{heap} @tab @code{high_priority} @tab @code{hour} -@item @code{hour_minute} @tab @code{hour_second} @tab @code{hosts} @tab @code{identified} -@item @code{ignore} @tab @code{in} @tab @code{index} @tab @code{infile} -@item @code{inner} @tab @code{insert} @tab @code{insert_id} @tab @code{int} -@item @code{integer} @tab @code{interval} @tab @code{int1} @tab @code{int2} -@item @code{int3} @tab @code{int4} @tab @code{int8} @tab @code{into} -@item @code{if} @tab @code{is} @tab @code{isam} @tab @code{join} -@item @code{key} @tab @code{keys} @tab @code{kill} @tab @code{last_insert_id} -@item @code{leading} @tab @code{left} @tab @code{length} @tab @code{like} -@item @code{lines} @tab @code{limit} @tab @code{load} @tab @code{local} -@item @code{lock} @tab @code{logs} @tab @code{long} @tab @code{longblob} -@item @code{longtext} @tab @code{low_priority} @tab @code{max} @tab @code{max_rows} -@item @code{match} @tab @code{mediumblob} @tab @code{mediumtext} @tab @code{mediumint} -@item @code{middleint} @tab @code{min_rows} @tab @code{minute} @tab @code{minute_second} -@item @code{modify} @tab @code{month} @tab @code{monthname} @tab @code{myisam} -@item @code{natural} @tab @code{numeric} @tab @code{no} @tab @code{not} -@item @code{null} @tab @code{on} @tab @code{optimize} @tab @code{option} -@item @code{optionally} @tab @code{or} @tab @code{order} @tab @code{outer} -@item @code{outfile} @tab @code{pack_keys} @tab @code{partial} @tab @code{password} -@item @code{precision} @tab @code{primary} @tab @code{procedure} @tab @code{process} -@item @code{processlist} @tab @code{privileges} @tab @code{read} @tab @code{real} -@item @code{references} @tab @code{reload} @tab @code{regexp} @tab @code{rename} -@item @code{replace} @tab @code{restrict} @tab @code{returns} @tab @code{revoke} -@item @code{rlike} @tab @code{row} @tab @code{rows} @tab @code{second} -@item @code{select} @tab @code{set} @tab @code{show} @tab @code{shutdown} -@item @code{smallint} @tab @code{soname} @tab @code{sql_big_tables} @tab @code{sql_big_selects} -@item @code{sql_low_priority_updates} @tab @code{sql_log_off} @tab @code{sql_log_update} @tab @code{sql_select_limit} -@item @code{sql_small_result} @tab @code{sql_big_result} @tab @code{sql_warnings} @tab @code{straight_join} -@item @code{starting} @tab @code{status} @tab @code{string} @tab @code{table} -@item @code{tables} @tab @code{temporary} @tab @code{terminated} @tab @code{text} -@item @code{then} @tab @code{time} @tab @code{timestamp} @tab @code{tinyblob} -@item @code{tinytext} @tab @code{tinyint} @tab @code{trailing} @tab @code{to} -@item @code{type} @tab @code{use} @tab @code{using} @tab @code{unique} -@item @code{unlock} @tab @code{unsigned} @tab @code{update} @tab @code{usage} -@item @code{values} @tab @code{varchar} @tab @code{variables} @tab @code{varying} -@item @code{varbinary} @tab @code{with} @tab @code{write} @tab @code{when} -@item @code{where} @tab @code{year} @tab @code{year_month} @tab @code{zerofill} -@end multitable +@item +If @code{FIELDS ESCAPED BY} is empty, a field value that contains an occurrence +of @code{FIELDS ENCLOSED BY} or @code{LINES TERMINATED BY} +followed by the @code{FIELDS TERMINATED BY} value will cause @code{LOAD +DATA INFILE} to stop reading a field or line too early. +This happens because @code{LOAD DATA INFILE} cannot properly determine +where the field or line value ends. +@end itemize -The following symbols (from the table above) are disallowed by ANSI SQL -but allowed by @strong{MySQL} as column/table names. This is because some -of these names are very natural names and a lot of people have already -used them. +The following example loads all columns of the @code{persondata} table: -@itemize @bullet -@item @code{ACTION} -@item @code{BIT} -@item @code{DATE} -@item @code{ENUM} -@item @code{NO} -@item @code{TEXT} -@item @code{TIME} -@item @code{TIMESTAMP} -@end itemize +@example +mysql> LOAD DATA INFILE 'persondata.txt' INTO TABLE persondata; +@end example -@cindex table types, choosing -@cindex @code{BDB} table type -@cindex @code{Berkeley_db} table type -@cindex @code{HEAP} table type -@cindex @code{ISAM} table type -@cindex @code{InnoDB} table type -@cindex @code{MERGE} table type -@cindex MySQL table types -@cindex @code{MyISAM} table type -@cindex types, of tables -@node Table types, Tutorial, Reference, Top -@chapter MySQL Table Types +No field list is specified, so @code{LOAD DATA INFILE} expects input rows +to contain a field for each table column. The default @code{FIELDS} and +@code{LINES} values are used. -As of @strong{MySQL} Version 3.23.6, you can choose between three basic -table formats (@code{ISAM}, @code{HEAP} and @code{MyISAM}. Newer -@strong{MySQL} may support additional table type (@code{BDB}, -or @code{InnoDB}), depending on how you compile it. - -When you create a new table, you can tell @strong{MySQL} which table -type it should use for the table. @strong{MySQL} will always create a -@code{.frm} file to hold the table and column definitions. Depending on -the table type, the index and data will be stored in other files. +If you wish to load only some of a table's columns, specify a field list: -Note that to use @code{InnoDB} tables you have to use at least -the @code{innodb_data_file_path} startup option. @xref{InnoDB start}. +@example +mysql> LOAD DATA INFILE 'persondata.txt' + INTO TABLE persondata (col1,col2,...); +@end example -The default table type in @strong{MySQL} is @code{MyISAM}. If you are -trying to use a table type that is not compiled-in or activated, -@strong{MySQL} will instead create a table of type @code{MyISAM}. This -is a very useful feature when you want to copy tables between different -SQL servers that supports different table types (like copying tables to -a slave that is optimized for speed by not having transactional tables). -This automatic table changing can however also be very confusing for new -@strong{MySQL} users. We plan to fix this by introducing warnings in -@strong{MySQL} 4.0 and giving a warning when a table type is automatically -changed. - -You can convert tables between different types with the @code{ALTER -TABLE} statement. @xref{ALTER TABLE, , @code{ALTER TABLE}}. +You must also specify a field list if the order of the fields in the input +file differs from the order of the columns in the table. Otherwise, +@strong{MySQL} cannot tell how to match up input fields with table columns. -Note that @strong{MySQL} supports two different kinds of -tables. Transaction-safe tables (@code{BDB}, @code{InnoDB} -and not transaction-safe tables (@code{HEAP}, @code{ISAM}, -@code{MERGE}, and @code{MyISAM}). +If a row has too few fields, the columns for which no input field is present +are set to default values. Default value assignment is described in +@ref{CREATE TABLE, , @code{CREATE TABLE}}. -Advantages of transaction-safe tables (TST): +An empty field value is interpreted differently than if the field value +is missing: @itemize @bullet @item -Safer. Even if @strong{MySQL} crashes or you get hardware problems, you -can get your data back, either by automatic recovery or from a backup -+ the transaction log. -@item -You can combine many statements and accept these all in one go with -the @code{COMMIT} command. -@item -You can execute @code{ROLLBACK} to ignore your changes (if you are not -running in auto commit mode). -@item -If an update fails, all your changes will be restored. (With NTST tables all -changes that have taken place are permanent) -@end itemize - -Advantages of not transaction-safe tables (NTST): +For string types, the column is set to the empty string. -@itemize @bullet -@item -Much faster as there is no transaction overhead. @item -Will use less disk space as there is no overhead of transactions. +For numeric types, the column is set to @code{0}. + @item -Will use less memory to do updates. +For date and time types, the column is set to the appropriate ``zero'' +value for the type. +@xref{Date and time types}. @end itemize -You can combine TST and NTST tables in the same statements to get the best -of both worlds. +Note that these are the same values that result if you assign an empty +string explicitly to a string, numeric, or date or time type explicitly +in an @code{INSERT} or @code{UPDATE} statement. -@menu -* MyISAM:: MyISAM tables -* MERGE:: MERGE tables -* ISAM:: ISAM tables -* HEAP:: HEAP tables -* InnoDB:: InnoDB tables -* BDB:: BDB or Berkeley_db tables -@end menu +@code{TIMESTAMP} columns are only set to the current date and time if there +is a @code{NULL} value for the column, or (for the first @code{TIMESTAMP} +column only) if the @code{TIMESTAMP} column is left out from the field list +when a field list is specified. -@node MyISAM, MERGE, Table types, Table types -@section MyISAM Tables +If an input row has too many fields, the extra fields are ignored and +the number of warnings is incremented. -@code{MyISAM} is the default table type in @strong{MySQL} Version 3.23. It's -based on the @code{ISAM} code and has a lot of useful extensions. +@code{LOAD DATA INFILE} regards all input as strings, so you can't use +numeric values for @code{ENUM} or @code{SET} columns the way you can with +@code{INSERT} statements. All @code{ENUM} and @code{SET} values must be +specified as strings! -The index is stored in a file with the @code{.MYI} (MYIndex) extension, -and the data is stored in a file with the @code{.MYD} (MYData) extension. -You can check/repair @code{MyISAM} tables with the @code{myisamchk} -utility. @xref{Crash recovery}. You can compress @code{MyISAM} tables with -@code{myisampack} to take up much less space. @xref{myisampack}. +@findex mysql_info() +If you are using the C API, you can get information about the query by +calling the API function @code{mysql_info()} when the @code{LOAD DATA INFILE} +query finishes. The format of the information string is shown below: -The following is new in @code{MyISAM}: +@example +Records: 1 Deleted: 0 Skipped: 0 Warnings: 0 +@end example -@itemize @bullet -@item -There is a flag in the @code{MyISAM} file that indicates whether or not -the table was closed correctly. If @code{mysqld} is started with -@code{--myisam-recover}, @code{MyISAM} tables will automatically be -checked and/or repaired on open if the table wasn't closed properly. -@item -You can @code{INSERT} new rows in a table that doesn't have free blocks -in the middle of the data file, at the same time other threads are -reading from the table (concurrent insert). An free block can come from -an update of a dynamic length row with much data to a row with less data -or when deleting rows. When all free blocks are used up, all future -inserts will be concurrent again. -@item -Support for big files (63-bit) on filesystems/operating systems that -support big files. -@item -All data is stored with the low byte first. This makes the data machine -and OS independent. The only requirement is that the machine uses -two's-complement signed integers (as every machine for the last 20 years -has) and IEEE floating-point format (also totally dominant among -mainstream machines). The only area of machines that may not support -binary compatibility are embedded systems (because they sometimes have -peculiar processors). +Warnings occur under the same circumstances as when values are inserted +via the @code{INSERT} statement (@pxref{INSERT, , @code{INSERT}}), except +that @code{LOAD DATA INFILE} also generates warnings when there are too few +or too many fields in the input row. The warnings are not stored anywhere; +the number of warnings can only be used as an indication if everything went +well. If you get warnings and want to know exactly why you got them, one way +to do this is to use @code{SELECT ... INTO OUTFILE} into another file and +compare this to your original input file. -There is no big speed penalty in storing data low byte first; The bytes -in a table row is normally unaligned and it doesn't take that much more -power to read an unaligned byte in order than in reverse order. The -actual fetch-column-value code is also not time critical compared to -other code. -@item -All number keys are stored with high byte first to give better index -compression. -@item -Internal handling of one @code{AUTO_INCREMENT} column. @code{MyISAM} -will automatically update this on @code{INSERT/UPDATE}. The -@code{AUTO_INCREMENT} value can be reset with @code{myisamchk}. This -will make @code{AUTO_INCREMENT} columns faster (at least 10 %) and old -numbers will not be reused as with the old @code{ISAM}. Note that when an -@code{AUTO_INCREMENT} is defined on the end of a multi-part-key the old -behavior is still present. -@item -When inserted in sorted order (as when you are using an @code{AUTO_INCREMENT} -column) the key tree will be split so that the high node only contains one -key. This will improve the space utilization in the key tree. -@item -@code{BLOB} and @code{TEXT} columns can be indexed. -@item -@code{NULL} values are allowed in indexed columns. This takes 0-1 -bytes/key. -@item -Maximum key length is 500 bytes by default (can be changed by -recompiling). In cases of keys longer than 250 bytes, a bigger key -block size than the default of 1024 bytes is used for this key. -@item -Maximum number of keys/table is 32 as default. This can be enlarged to 64 -without having to recompile @code{myisamchk}. -@item -@code{myisamchk} will mark tables as checked if one runs it with -@code{--update-state}. @code{myisamchk --fast} will only check those -tables that don't have this mark. -@item -@code{myisamchk -a} stores statistics for key parts (and not only for -whole keys as in @code{ISAM}). -@item -Dynamic size rows will now be much less fragmented when mixing deletes -with updates and inserts. This is done by automatically combining adjacent -deleted blocks and by extending blocks if the next block is deleted. -@item -@code{myisampack} can pack @code{BLOB} and @code{VARCHAR} columns. -@item -You can use put the datafile and index file on different directories -to get more speed (with the @code{DATA/INDEX DIRECTORY="path"} option to -@code{CREATE TABLE}). @xref{CREATE TABLE}. -@end itemize +If you need @code{LOAD DATA} to read from a pipe, you can use the +following trick: -@code{MyISAM} also supports the following things, which @strong{MySQL} -will be able to use in the near future: +@example +mkfifo /mysql/db/x/x +chmod 666 /mysql/db/x/x +cat < /dev/tcp/10.1.1.12/4711 > /nt/mysql/db/x/x +mysql -e "LOAD DATA INFILE 'x' INTO TABLE x" x +@end example -@itemize @bullet -@item -Support for a true @code{VARCHAR} type; A @code{VARCHAR} column starts -with a length stored in 2 bytes. -@item -Tables with @code{VARCHAR} may have fixed or dynamic record length. -@item -@code{VARCHAR} and @code{CHAR} may be up to 64K. -All key segments have their own language definition. This will enable -@strong{MySQL} to have different language definitions per column. -@item -A hashed computed index can be used for @code{UNIQUE}. This will allow -you to have @code{UNIQUE} on any combination of columns in a table. (You -can't search on a @code{UNIQUE} computed index, however.) -@end itemize +If you are using a version of @strong{MySQL} older than 3.23.25 +you can only do the above with @code{LOAD DATA LOCAL INFILE}. -Note that index files are usually much smaller with @code{MyISAM} than with -@code{ISAM}. This means that @code{MyISAM} will normally use less -system resources than @code{ISAM}, but will need more CPU when inserting -data into a compressed index. +For more information about the efficiency of @code{INSERT} versus +@code{LOAD DATA INFILE} and speeding up @code{LOAD DATA INFILE}, +@xref{Insert speed}. -The following options to @code{mysqld} can be used to change the behavior of -@code{MyISAM} tables. @xref{SHOW VARIABLES}. +@findex UPDATE +@node UPDATE, USE, LOAD DATA, Reference +@section @code{UPDATE} Syntax -@multitable @columnfractions .40 .60 -@item @strong{Option} @tab @strong{Meaning} -@item @code{--myisam-recover=#} @tab Automatic recover of crashed tables. -@item @code{-O myisam_sort_buffer_size=#} @tab Buffer used when recovering tables. -@item @code{--delay-key-write-for-all-tables} @tab Don't flush key buffers between writes for any MyISAM table -@item @code{-O myisam_max_extra_sort_file_size=#} @tab Used to help @strong{MySQL} to decide when to use the slow but safe key cache index create method. @strong{NOTE} that this parameter is given in megabytes! -@item @code{-O myisam_max_sort_file_size=#} @tab Don't use the fast sort index method to created index if the temporary file would get bigger than this. -@strong{NOTE} that this paramter is given in megabytes! -@end multitable +@example +UPDATE [LOW_PRIORITY] [IGNORE] tbl_name + SET col_name1=expr1, [col_name2=expr2, ...] + [WHERE where_definition] + [LIMIT #] +@end example -The automatic recovery is activated if you start @code{mysqld} with -@code{--myisam-recover=#}. @xref{Command-line options}. -On open, the table is checked if it's marked as crashed or if the open -count variable for the table is not 0 and you are running with -@code{--skip-locking}. If either of the above is true the following happens. +@code{UPDATE} updates columns in existing table rows with new values. +The @code{SET} clause indicates which columns to modify and the values +they should be given. The @code{WHERE} clause, if given, specifies +which rows should be updated. Otherwise all rows are updated. If the +@code{ORDER BY} clause is specified, the rows will be updated in the +order that is specified. -@itemize @bullet -@item -The table is checked for errors. -@item -If we found an error, try to do a fast repair (with sorting and without -re-creating the data file) of the table. -@item -If the repair fails because of an error in the data file (for example a -duplicate key error), we try again, but this time we re-create the data file. -@item -If the repair fails, retry once more with the old repair option method -(write row by row without sorting) which should be able to repair any -type of error with little disk requirements.. -@end itemize +If you specify the keyword @code{LOW_PRIORITY}, execution of the +@code{UPDATE} is delayed until no other clients are reading from the table. -If the recover wouldn't be able to recover all rows from a previous -completed statement and you didn't specify @code{FORCE} as an option to -@code{myisam-recover}, then the automatic repair will abort with an error -message in the error file: +If you specify the keyword @code{IGNORE}, the update statement will not +abort even if we get duplicate key errors during the update. Rows that +would cause conflicts will not be updated. + +If you access a column from @code{tbl_name} in an expression, +@code{UPDATE} uses the current value of the column. For example, the +following statement sets the @code{age} column to one more than its +current value: @example -Error: Couldn't repair table: test.g00pages +mysql> UPDATE persondata SET age=age+1; @end example -If you in this case had used the @code{FORCE} option you would instead have got -a warning in the error file: +@code{UPDATE} assignments are evaluated from left to right. For example, the +following statement doubles the @code{age} column, then increments it: @example -Warning: Found 344 of 354 rows when repairing ./test/g00pages +mysql> UPDATE persondata SET age=age*2, age=age+1; @end example -Note that if you run automatic recover with the @code{BACKUP} option, -you should have a cron script that automatically moves file with names -like @file{tablename-datetime.BAK} from the database directories to a -backup media. +If you set a column to the value it currently has, @strong{MySQL} notices +this and doesn't update it. -@xref{Command-line options}. +@findex mysql_info() +@code{UPDATE} returns the number of rows that were actually changed. +In @strong{MySQL} Version 3.22 or later, the C API function @code{mysql_info()} +returns the number of rows that were matched and updated and the number of +warnings that occurred during the @code{UPDATE}. -@menu -* Key space:: Space needed for keys -* MyISAM table formats:: MyISAM table formats -* MyISAM table problems:: -@end menu +In @strong{MySQL} Version 3.23, you can use @code{LIMIT #} to ensure that +only a given number of rows are changed. -@cindex key space, MyISAM -@node Key space, MyISAM table formats, MyISAM, MyISAM -@subsection Space Needed for Keys +@findex USE +@node USE, FLUSH, UPDATE, Reference +@section @code{USE} Syntax -@strong{MySQL} can support different index types, but the normal type is -ISAM or MyISAM. These use a B-tree index, and you can roughly calculate -the size for the index file as @code{(key_length+4)/0.67}, summed over -all keys. (This is for the worst case when all keys are inserted in -sorted order and we don't have any compressed keys.) +@example +USE db_name +@end example -String indexes are space compressed. If the first index part is a -string, it will also be prefix compressed. Space compression makes the -index file smaller than the above figures if the string column has a lot -of trailing space or is a @code{VARCHAR} column that is not always used -to the full length. Prefix compression is used on keys that start -with a string. Prefix compression helps if there are many strings -with an identical prefix. +The @code{USE db_name} statement tells @strong{MySQL} to use the @code{db_name} +database as the default database for subsequent queries. The database remains +current until the end of the session or until another @code{USE} statement +is issued: -In @code{MyISAM} tables, you can also prefix compress numbers by specifying -@code{PACK_KEYS=1} when you create the table. This helps when you have -many integer keys that have an identical prefix when the numbers are stored -high-byte first. +@example +mysql> USE db1; +mysql> SELECT count(*) FROM mytable; # selects from db1.mytable +mysql> USE db2; +mysql> SELECT count(*) FROM mytable; # selects from db2.mytable +@end example -@node MyISAM table formats, MyISAM table problems, Key space, MyISAM -@subsection MyISAM Table Formats +Making a particular database current by means of the @code{USE} statement +does not preclude you from accessing tables in other databases. The example +below accesses the @code{author} table from the @code{db1} database and the +@code{editor} table from the @code{db2} database: -@strong{MyISAM} supports 3 different table types. Two of them are chosen -automatically depending on the type of columns you are using. The third, -compressed tables, can only be created with the @code{myisampack} tool. +@example +mysql> USE db1; +mysql> SELECT author_name,editor_name FROM author,db2.editor + WHERE author.editor_id = db2.editor.editor_id; +@end example -@menu -* Static format:: Static (Fixed-length) table characteristics -* Dynamic format:: Dynamic table characteristics -* Compressed format:: Compressed table characteristics -@end menu +@cindex Sybase compatibility +@cindex compatibility, with Sybase +The @code{USE} statement is provided for Sybase compatibility. -@node Static format, Dynamic format, MyISAM table formats, MyISAM table formats -@subsubsection Static (Fixed-length) Table Characteristics +@cindex @code{mysqladmin} +@cindex clearing, caches +@cindex caches, clearing +@findex FLUSH +@node FLUSH, KILL, USE, Reference +@section @code{FLUSH} Syntax -This is the default format. It's used when the table contains no -@code{VARCHAR}, @code{BLOB}, or @code{TEXT} columns. +@example +FLUSH flush_option [,flush_option] +@end example -This format is the simplest and most secure format. It is also the -fastest of the on-disk formats. The speed comes from the easy way data -can be found on disk. When looking up something with an index and static -format it is very simple. Just multiply the row number by the row length. +You should use the @code{FLUSH} command if you want to clear some of the +internal caches @strong{MySQL} uses. To execute @code{FLUSH}, you must have +the @strong{RELOAD} privilege. -Also, when scanning a table it is very easy to read a constant number of -records with each disk read. +@code{flush_option} can be any of the following: -The security is evidenced if your computer crashes when writing to a -fixed-size MyISAM file, in which case @code{myisamchk} can easily figure out where each -row starts and ends. So it can usually reclaim all records except the -partially written one. Note that in @strong{MySQL} all indexes can always be -reconstructed: +@multitable @columnfractions .15 .85 +@item @code{HOSTS} @tab Empties the host cache tables. You should flush the +host tables if some of your hosts change IP number or if you get the +error message @code{Host ... is blocked}. When more than +@code{max_connect_errors} errors occur in a row for a given host while +connection to the @strong{MySQL} server, @strong{MySQL} assumes +something is wrong and blocks the host from further connection requests. +Flushing the host tables allows the host to attempt to connect +again. @xref{Blocked host}.) You can start @code{mysqld} with +@code{-O max_connection_errors=999999999} to avoid this error message. -@itemize @bullet -@item -All @code{CHAR}, @code{NUMERIC}, and @code{DECIMAL} columns are space-padded -to the column width. -@item -Very quick. -@item -Easy to cache. -@item -Easy to reconstruct after a crash, because records are located in fixed -positions. -@item -Doesn't have to be reorganized (with @code{myisamchk}) unless a huge number of -records are deleted and you want to return free disk space to the operating -system. -@item -Usually requires more disk space than dynamic tables. -@end itemize +@item @code{LOGS} @tab Closes and reopens all log files. +If you have specified the update log file or a binary log file without +an extension, the extension number of the log file will be incremented +by one relative to the previous file. If you have used an extension in +the file name, @strong{MySQL} will close and reopen the update log file. +@xref{Update log}. This is the same thing as sending the @code{SIGHUP} +signal to the @code{mysqld} server. -@cindex dynamic table characteristics -@cindex tables, dynamic -@node Dynamic format, Compressed format, Static format, MyISAM table formats -@subsubsection Dynamic Table Characteristics +@item @code{PRIVILEGES} @tab Reloads the privileges from the grant tables in +the @code{mysql} database. -This format is used if the table contains any @code{VARCHAR}, @code{BLOB}, -or @code{TEXT} columns or if the table was created with -@code{ROW_FORMAT=dynamic}. +@item @code{TABLES} @tab Closes all open tables and force all tables in use to be closed. -This format is a little more complex because each row has to have a header -that says how long it is. One record can also end up at more than one -location when it is made longer at an update. +@item @code{[TABLE | TABLES] table_name [,table_name...]} @tab Flushes only the given tables. -@cindex tables, defragment -You can use @code{OPTIMIZE table} or @code{myisamchk} to defragment a -table. If you have static data that you access/change a lot in the same -table as some @code{VARCHAR} or @code{BLOB} columns, it might be a good -idea to move the dynamic columns to other tables just to avoid -fragmentation: +@item @code{TABLES WITH READ LOCK} @tab Closes all open tables and locks all tables for all databases with a read until one executes @code{UNLOCK TABLES}. This is very convenient way to get backups if you have a file system, like Veritas,that can take snapshots in time. -@itemize @bullet -@item -All string columns are dynamic (except those with a length less than 4). -@item -Each record is preceded by a bitmap indicating which columns are empty -(@code{''}) for string columns, or zero for numeric columns. (This isn't -the same as columns containing @code{NULL} values.) If a string column -has a length of zero after removal of trailing spaces, or a numeric -column has a value of zero, it is marked in the bit map and not saved to -disk. Non-empty strings are saved as a length byte plus the string -contents. -@item -Usually takes much less disk space than fixed-length tables. -@item -Each record uses only as much space as is required. If a record becomes -larger, it is split into as many pieces as are required. This results in record -fragmentation. -@item -If you update a row with information that extends the row length, the -row will be fragmented. In this case, you may have to run @code{myisamchk --r} from time to time to get better performance. Use @code{myisamchk -ei -tbl_name} for some statistics. -@item -Not as easy to reconstruct after a crash, because a record may be fragmented -into many pieces and a link (fragment) may be missing. -@item -The expected row length for dynamic sized records is: +@item @code{STATUS} @tab Resets most status variables to zero. This is something one should only use when debugging a query. +@end multitable + +You can also access each of the commands shown above with the @code{mysqladmin} +utility, using the @code{flush-hosts}, @code{flush-logs}, @code{reload}, +or @code{flush-tables} commands. + +Take also a look at the @code{RESET} command used with +replication. @xref{Replication SQL}. + +@cindex @code{mysqladmin} +@findex KILL +@node KILL, SHOW, FLUSH, Reference +@section @code{KILL} Syntax @example -3 -+ (number of columns + 7) / 8 -+ (number of char columns) -+ packed size of numeric columns -+ length of strings -+ (number of NULL columns + 7) / 8 +KILL thread_id @end example -There is a penalty of 6 bytes for each link. A dynamic record is linked -whenever an update causes an enlargement of the record. Each new link will be -at least 20 bytes, so the next enlargement will probably go in the same link. -If not, there will be another link. You may check how many links there are -with @code{myisamchk -ed}. All links may be removed with @code{myisamchk -r}. -@end itemize +Each connection to @code{mysqld} runs in a separate thread. You can see +which threads are running with the @code{SHOW PROCESSLIST} command and kill +a thread with the @code{KILL thread_id} command. -@cindex tables, compressed format -@node Compressed format, , Dynamic format, MyISAM table formats -@subsubsection Compressed Table Characteristics +If you have the @strong{process} privilege, you can see and kill all threads. +Otherwise, you can see and kill only your own threads. -This is a read-only type that is generated with the optional -@code{myisampack} tool (@code{pack_isam} for @code{ISAM} tables): +You can also use the @code{mysqladmin processlist} and @code{mysqladmin kill} +commands to examine and kill threads. + +When you do a @code{KILL}, a thread specific @code{kill flag} is set for +the thread. + +In most cases it may take some time for the thread to die as the kill +flag is only checked at specific intervals. @itemize @bullet @item -All @strong{MySQL} distributions, even those that existed before @strong{MySQL} -went GPL, can read tables that were compressed with @code{myisampack}. +In @code{SELECT}, @code{ORDER BY} and @code{GROUP BY} loops, the flag is +checked after reading a block of rows. If the kill flag is set the +statement is aborted @item -Compressed tables take very little disk space. This minimizes disk usage, which -is very nice when using slow disks (like CD-ROMs). -@item -Each record is compressed separately (very little access overhead). The -header for a record is fixed (1-3 bytes) depending on the biggest record in the -table. Each column is compressed differently. Some of the compression types -are: -@itemize @minus -@item -There is usually a different Huffman table for each column. -@item -Suffix space compression. -@item -Prefix space compression. -@item -Numbers with value @code{0} are stored using 1 bit. +When doing an @code{ALTER TABLE} the kill flag is checked before each block of +rows are read from the original table. If the kill flag was set the command +is aborted and the temporary table is deleted. @item -If values in an integer column have a small range, the column is stored using -the smallest possible type. For example, a @code{BIGINT} column (8 bytes) may -be stored as a @code{TINYINT} column (1 byte) if all values are in the range -@code{0} to @code{255}. +When doing an @code{UPDATE TABLE} and @code{DELETE TABLE}, the kill flag +is checked after each block read and after each updated or delete +row. If the kill flag is set the statement is aborted. Note that if you +are not using transactions, the changes will not be rolled back! @item -If a column has only a small set of possible values, the column type is -converted to @code{ENUM}. +@code{GET_LOCK()} will abort with @code{NULL}. @item -A column may use a combination of the above compressions. -@end itemize +An @code{INSERT DELAYED} thread will quickly flush all rows it has in +memory and die. @item -Can handle fixed- or dynamic-length records, but not @code{BLOB} or @code{TEXT} -columns. +If the thread is in the table lock handler (state: @code{Locked}), +the table lock will be quickly aborted. @item -Can be uncompressed with @code{myisamchk}. +If the thread is waiting for free disk space in a @code{write} call, the +write is aborted with an disk full error message. @end itemize -@node MyISAM table problems, , MyISAM table formats, MyISAM -@subsection MyISAM table problems. +@findex SHOW DATABASES +@findex SHOW TABLES +@findex SHOW COLUMNS +@findex SHOW FIELDS +@findex SHOW INDEX +@findex SHOW KEYS +@findex SHOW STATUS +@findex SHOW VARIABLES +@findex SHOW PROCESSLIST +@findex SHOW TABLE STATUS +@findex SHOW GRANTS +@findex SHOW CREATE TABLE +@findex SHOW MASTER STATUS +@findex SHOW MASTER LOGS +@findex SHOW SLAVE STATUS +@node SHOW, EXPLAIN, KILL, Reference +@section @code{SHOW} Syntax -The file format that @strong{MySQL} uses to store data has been extensively -tested, but there are always circumstances that may cause database tables -to become corrupted. +@example + SHOW DATABASES [LIKE wild] +or SHOW [OPEN] TABLES [FROM db_name] [LIKE wild] +or SHOW [FULL] COLUMNS FROM tbl_name [FROM db_name] [LIKE wild] +or SHOW INDEX FROM tbl_name [FROM db_name] +or SHOW TABLE STATUS [FROM db_name] [LIKE wild] +or SHOW STATUS [LIKE wild] +or SHOW VARIABLES [LIKE wild] +or SHOW LOGS +or SHOW [FULL] PROCESSLIST +or SHOW GRANTS FOR user +or SHOW CREATE TABLE table_name +or SHOW MASTER STATUS +or SHOW MASTER LOGS +or SHOW SLAVE STATUS +@end example + +@code{SHOW} provides information about databases, tables, columns, or +status information about the server. If the @code{LIKE wild} part is +used, the @code{wild} string can be a string that uses the SQL @samp{%} +and @samp{_} wild-card characters. +@findex SHOW DATABASES +@findex SHOW TABLES +@findex SHOW COLUMNS +@findex SHOW FIELDS +@findex SHOW INDEX +@findex SHOW KEYS @menu -* Corrupted MyISAM tables:: -* MyISAM table close:: +* SHOW DATABASE INFO:: +* SHOW TABLE STATUS:: +* SHOW STATUS:: +* SHOW VARIABLES:: +* SHOW LOGS:: +* SHOW PROCESSLIST:: +* SHOW GRANTS:: +* SHOW CREATE TABLE:: @end menu -@node Corrupted MyISAM tables, MyISAM table close, MyISAM table problems, MyISAM table problems -@subsubsection Corrupted MyISAM tables. - -Even if the MyISAM table format is very reliable (all changes to a table -is written before the SQL statements returns) , you can still get -corrupted tables if some of the following things happens: - -@itemize @bullet -@item -The @code{mysqld} process being killed in the middle of a write. -@item -Unexpected shutdown of the computer (for example, if the computer is turned -off). -@item -A hardware error. -@item -You are using an external program (like myisamchk) on a live table. -@item -A software bug in the MySQL or MyISAM code. -@end itemize +@cindex displaying, information, @code{SHOW} +@node SHOW DATABASE INFO, SHOW TABLE STATUS, SHOW, SHOW +@subsection @code{SHOW} Information About Databases, Tables, Columns, and Indexes -Typial typical symptoms for a corrupt table is: +You can use @code{db_name.tbl_name} as an alternative to the @code{tbl_name +FROM db_name} syntax. These two statements are equivalent: -@itemize @bullet -@item -You get the error @code{Incorrect key file for table: '...'. Try to repair it} -while selecting data from the table. -@item -Queries doesn't find rows in the table or returns incomplete data. -@end itemize +@example +mysql> SHOW INDEX FROM mytable FROM mydb; +mysql> SHOW INDEX FROM mydb.mytable; +@end example -You can check if a table is ok with the command @code{CHECK -TABLE}. @xref{CHECK TABLE}. +@code{SHOW DATABASES} lists the databases on the @strong{MySQL} server +host. You can also get this list using the @code{mysqlshow} command. -You can repair a corrupted table with @code{REPAIR TABLE}. @xref{REPAIR TABLE}. -You can also repair a table, when @code{mysqld} is not running with -the @code{myisamchk} command. @code{myisamchk syntax}. +@code{SHOW TABLES} lists the tables in a given database. You can also +get this list using the @code{mysqlshow db_name} command. -If your tables get corrupted a lot you should try to find the reason for -this! @xref{Crashing}. +@strong{NOTE:} If a user doesn't have any privileges for a table, the table +will not show up in the output from @code{SHOW TABLES} or @code{mysqlshow +db_name}. -In this case the most important thing to know is if the table got -corrupted if the @code{mysqld} died (one can easily verify this by -checking if there is a recent row @code{restarted mysqld} in the mysqld -error file). If this isn't the case, then you should try to make a test -case of this. @xref{Reproduceable test case}. +@code{SHOW OPEN TABLES} lists the tables that are currently open in +the table cache. @xref{Table cache}. The @code{Comment} field tells +how many times the table is @code{cached} and @code{in_use}. -@node MyISAM table close, , Corrupted MyISAM tables, MyISAM table problems -@subsubsection Clients is using or hasn't closed the table properly +@code{SHOW COLUMNS} lists the columns in a given table. If you specify +the @code{FULL} option, you will also get the privileges you have for +each column. If the column types are different than you expect them to +be based on a @code{CREATE TABLE} statement, note that @strong{MySQL} +sometimes changes column types. @xref{Silent column changes}. -Each @code{MyISAM} @code{.MYI} file has in the header a counter that can -be used to check if a table has been closed properly. +The @code{DESCRIBE} statement provides information similar to +@code{SHOW COLUMNS}. +@xref{DESCRIBE, , @code{DESCRIBE}}. -If you get the following warning from @code{CHECK TABLE} or @code{myisamchk}: +@code{SHOW FIELDS} is a synonym for @code{SHOW COLUMNS}, and +@code{SHOW KEYS} is a synonym for @code{SHOW INDEX}. You can also +list a table's columns or indexes with @code{mysqlshow db_name tbl_name} +or @code{mysqlshow -k db_name tbl_name}. -@example -# clients is using or hasn't closed the table properly -@end example +@code{SHOW INDEX} returns the index information in a format that closely +resembles the @code{SQLStatistics} call in ODBC. The following columns +are returned: -this means that this counter has come out of sync. This doesn't mean -that the table is corrupted, but means that you should at least do a -check on the table to verify that it's ok. +@multitable @columnfractions .35 .65 +@item @strong{Column} @tab @strong{Meaning} +@item @code{Table} @tab Name of the table. +@item @code{Non_unique} @tab 0 if the index can't contain duplicates. +@item @code{Key_name} @tab Name of the index. +@item @code{Seq_in_index} @tab Column sequence number in index, + starting with 1. +@item @code{Column_name} @tab Column name. +@item @code{Collation} @tab How the column is sorted in the index. + In @strong{MySQL}, this can have values + @samp{A} (Ascending) or @code{NULL} (Not + sorted). +@item @code{Cardinality} @tab Number of unique values in the index. + This is updated by running + @code{isamchk -a}. +@item @code{Sub_part} @tab Number of indexed characters if the + column is only partly indexed. + @code{NULL} if the entire key is indexed. +@item @code{Comment} @tab Various remarks. For now, it tells + whether index is FULLTEXT or not. +@end multitable -The counter works as follows: +Note that as the @code{Cardinality} is counted based on statistics +stored as integers, it's not necessarily accurate for small tables. -@itemize @bullet -@item -The first time a table is updated in @strong{MySQL}, a counter in the -header of the index files is incremented. -@item -The counter is not changed during further updates. -@item -When the last instance of a table is closed (because of a @code{FLUSH} or -because there isn't room in the table cache) the counter is -decremented if the table has been updated at any point. -@item -When you repair the table or check the table and it was ok, -the counter is reset to 0. -@item -To avoid problems with interaction with other processes that may do a -check on the table, the counter is not decremented on close if it was 0. -@end itemize +@cindex displaying, table status +@cindex tables, displaying status +@cindex status, tables +@node SHOW TABLE STATUS, SHOW STATUS, SHOW DATABASE INFO, SHOW +@subsection @code{SHOW TABLE STATUS} -In other words, the only ways this can go out of sync are: +@example +SHOW TABLE STATUS [FROM db_name] [LIKE wild] +@end example -@itemize @bullet -@item -The @code{MyISAM} tables are copied without a @code{LOCK} and -@code{FLUSH TABLES}. -@item -@strong{MySQL} has crashed between an update and the final close -(Note that the table may still be ok, as @strong{MySQL} always issues writes -for everything between each statement). -@item -Someone has done a @code{myisamchk --repair} or @code{myisamchk ---update-state}on a table that was in use by @code{mysqld}. -@item -Many @code{mysqld} servers are using the table and one has done a -@code{REPAIR} or @code{CHECK} of the table while it was in use by -another server. In this setup the @code{CHECK} is safe to do (even if -you will get the warning from other servers), but @code{REPAIR} should -be avoided as it currently replaces the data file with a new one, which -is not signaled to the other servers. -@end itemize +@code{SHOW TABLE STATUS} (new in Version 3.23) works likes @code{SHOW +STATUS}, but provides a lot of information about each table. You can +also get this list using the @code{mysqlshow --status db_name} command. +The following columns are returned: -@cindex tables, merging -@cindex MERGE tables, defined -@node MERGE, ISAM, MyISAM, Table types -@section MERGE Tables +@multitable @columnfractions .30 .70 +@item @strong{Column} @tab @strong{Meaning} +@item @code{Name} @tab Name of the table. +@item @code{Type} @tab Type of table. @xref{Table types}. +@item @code{Row_format} @tab The row storage format (Fixed, Dynamic, or Compressed). +@item @code{Rows} @tab Number of rows. +@item @code{Avg_row_length} @tab Average row length. +@item @code{Data_length} @tab Length of the data file. +@item @code{Max_data_length} @tab Max length of the data file. +@item @code{Index_length} @tab Length of the index file. +@item @code{Data_free} @tab Number of allocated but not used bytes. +@item @code{Auto_increment} @tab Next autoincrement value. +@item @code{Create_time} @tab When the table was created. +@item @code{Update_time} @tab When the data file was last updated. +@item @code{Check_time} @tab When the table was last checked. +@item @code{Create_options} @tab Extra options used with @code{CREATE TABLE}. +@item @code{Comment} @tab The comment used when creating the table (or some information why @strong{MySQL} couldn't access the table information). +@end multitable -@code{MERGE} tables are new in @strong{MySQL} Version 3.23.25. The code -is still in gamma, but should be resonable stable. +@code{InnoDB} tables will report the free space in the tablespace +in the table comment. -A @code{MERGE} table is a collection of identical @code{MyISAM} tables -that can be used as one. You can only @code{SELECT}, @code{DELETE}, and -@code{UPDATE} from the collection of tables. If you @code{DROP} the -@code{MERGE} table, you are only dropping the @code{MERGE} -specification. +@node SHOW STATUS, SHOW VARIABLES, SHOW TABLE STATUS, SHOW +@subsection @code{SHOW STATUS} -Note that @code{DELETE FROM merge_table} used without a @code{WHERE} -will only clear the mapping for the table, not delete everything in the -mapped tables. (We plan to fix this in 4.0). +@cindex @code{mysqladmin} +@code{SHOW STATUS} provides server status information +(like @code{mysqladmin extended-status}). The output resembles that shown +below, though the format and numbers probably differ: -With identical tables we mean that all tables are created with identical -column and key information. You can't put a MERGE over tables where the -columns are packed differently or doesn't have exactly the same columns. -Some of the tables can however be compressed with @code{myisampack}. -@xref{myisampack}. +@example ++--------------------------+------------+ +| Variable_name | Value | ++--------------------------+------------+ +| Aborted_clients | 0 | +| Aborted_connects | 0 | +| Bytes_received | 155372598 | +| Bytes_sent | 1176560426 | +| Connections | 30023 | +| Created_tmp_disk_tables | 0 | +| Created_tmp_tables | 8340 | +| Created_tmp_files | 60 | +| Delayed_insert_threads | 0 | +| Delayed_writes | 0 | +| Delayed_errors | 0 | +| Flush_commands | 1 | +| Handler_delete | 462604 | +| Handler_read_first | 105881 | +| Handler_read_key | 27820558 | +| Handler_read_next | 390681754 | +| Handler_read_prev | 6022500 | +| Handler_read_rnd | 30546748 | +| Handler_read_rnd_next | 246216530 | +| Handler_update | 16945404 | +| Handler_write | 60356676 | +| Key_blocks_used | 14955 | +| Key_read_requests | 96854827 | +| Key_reads | 162040 | +| Key_write_requests | 7589728 | +| Key_writes | 3813196 | +| Max_used_connections | 0 | +| Not_flushed_key_blocks | 0 | +| Not_flushed_delayed_rows | 0 | +| Open_tables | 1 | +| Open_files | 2 | +| Open_streams | 0 | +| Opened_tables | 44600 | +| Questions | 2026873 | +| Select_full_join | 0 | +| Select_full_range_join | 0 | +| Select_range | 99646 | +| Select_range_check | 0 | +| Select_scan | 30802 | +| Slave_running | OFF | +| Slave_open_temp_tables | 0 | +| Slow_launch_threads | 0 | +| Slow_queries | 0 | +| Sort_merge_passes | 30 | +| Sort_range | 500 | +| Sort_rows | 30296250 | +| Sort_scan | 4650 | +| Table_locks_immediate | 1920382 | +| Table_locks_waited | 0 | +| Threads_cached | 0 | +| Threads_created | 30022 | +| Threads_connected | 1 | +| Threads_running | 1 | +| Uptime | 80380 | ++--------------------------+------------+ +@end example -When you create a @code{MERGE} table, you will get a @code{.frm} table -definition file and a @code{.MRG} table list file. The @code{.MRG} just -contains a list of the index files (@code{.MYI} files) that should -be used as one. +@cindex variables, status +The status variables listed above have the following meaning: -For the moment you need to have @code{SELECT}, @code{UPDATE}, and -@code{DELETE} privileges on the tables you map to a @code{MERGE} table. +@multitable @columnfractions .35 .65 +@item @strong{Variable} @tab @strong{Meaning} +@item @code{Aborted_clients} @tab Number of connections aborted because the client died without closing the connection properly. @xref{Communication errors}. +@item @code{Aborted_connects} @tab Number of tries to connect to the @strong{MySQL} server that failed. @xref{Communication errors}. +@item @code{Bytes_received} @tab Number of bytes received from all clients. +@item @code{Bytes_sent} @tab Number of bytes sent to all clients. +@item @code{Connections} @tab Number of connection attempts to the @strong{MySQL} server. +@item @code{Created_tmp_disk_tables} @tab Number of implicit temporary tables on disk created while executing statements. +@item @code{Created_tmp_tables} @tab Number of implicit temporary tables in memory created while executing statements. +@item @code{Created_tmp_files} @tab How many temporary files @code{mysqld} have created. +@item @code{Delayed_insert_threads} @tab Number of delayed insert handler threads in use. +@item @code{Delayed_writes} @tab Number of rows written with @code{INSERT DELAYED}. +@item @code{Delayed_errors} @tab Number of rows written with @code{INSERT DELAYED} for which some error occurred (probably @code{duplicate key}). +@item @code{Flush_commands} @tab Number of executed @code{FLUSH} commands. +@item @code{Handler_delete} @tab Number of times a row was deleted from a table. +@item @code{Handler_read_first} @tab Number of times the first entry was read from an index. +If this is high, it suggests that the server is doing a lot of full index scans, for example, +@code{SELECT col1 FROM foo}, assuming that col1 is indexed. +@item @code{Handler_read_key} @tab Number of requests to read a row based on a key. If this +is high, it is a good indication that your queries and tables are properly indexed. +@item @code{Handler_read_next} @tab Number of requests to read next row in key order. This +will be incremented if you are querying an index column with a range constraint. This also +will be incremented if you are doing an index scan. +@item @code{Handler_read_rnd} @tab Number of requests to read a row based on a fixed position. +This will be high if you are doing a lot of queries that require sorting of the result. +@item @code{Handler_read_rnd_next} @tab Number of requests to read the next row in the datafile. +This will be high if you are doing a lot of table scans. Generally this suggests that your tables +are not properly indexed or that your queries are not written to take advantage of the indexes you +have. +@item @code{Handler_update} @tab Number of requests to update a row in a table. +@item @code{Handler_write} @tab Number of requests to insert a row in a table. +@item @code{Key_blocks_used} @tab The number of used blocks in the key cache. +@item @code{Key_read_requests} @tab The number of requests to read a key block from the cache. +@item @code{Key_reads} @tab The number of physical reads of a key block from disk. +@item @code{Key_write_requests} @tab The number of requests to write a key block to the cache. +@item @code{Key_writes} @tab The number of physical writes of a key block to disk. +@item @code{Max_used_connections} @tab The maximum number of connections in use simultaneously. +@item @code{Not_flushed_key_blocks} @tab Keys blocks in the key cache that has changed but hasn't yet been flushed to disk. +@item @code{Not_flushed_delayed_rows} @tab Number of rows waiting to be written in @code{INSERT DELAY} queues. +@item @code{Open_tables} @tab Number of tables that are open. +@item @code{Open_files} @tab Number of files that are open. +@item @code{Open_streams} @tab Number of streams that are open (used mainly for logging). +@item @code{Opened_tables} @tab Number of tables that have been opened. +@item @code{Select_full_join} @tab Number of joins without keys (Should be 0). +@item @code{Select_full_range_join} @tab Number of joins where we used a range search on reference table. +@item @code{Select_range} @tab Number of joins where we used ranges on the first table. (It's normally not critical even if this is big.) +@item @code{Select_scan} @tab Number of joins where we scanned the first table. +@item @code{Select_range_check} @tab Number of joins without keys where we check for key usage after each row (Should be 0). +@item @code{Questions} @tab Number of queries sent to the server. +@item @code{Slave_open_temp_tables} @tab Number of temporary tables currently +open by the slave thread +@item @code{Slow_launch_threads} @tab Number of threads that have taken more than @code{slow_launch_time} to connect. +@item @code{Slow_queries} @tab Number of queries that have taken more than @code{long_query_time}. @xref{Slow query log}. +@item @code{Sort_merge_passes} @tab Number of merges the sort has to do. If this value is large you should consider increasing @code{sort_buffer}. +@item @code{Sort_range} @tab Number of sorts that where done with ranges. +@item @code{Sort_rows} @tab Number of sorted rows. +@item @code{Sort_scan} @tab Number of sorts that where done by scanning the table. +@item @code{Table_locks_immediate} @tab Number of times a table lock was +acquired immediately. Available after 3.23.33. +@item @code{Table_locks_waited} @tab Number of times a table lock could not +be acquired immediately and a wait was needed. If this is high, and you +have performance problems, you should first optimize your queries, and then +either split your table(s) or use replication. Available after 3.23.33. +@item @code{Threads_cached} @tab Number of threads in the thread cache. +@item @code{Threads_connected} @tab Number of currently open connections. +@item @code{Threads_created} @tab Number of threads created to handle connections. +@item @code{Threads_running} @tab Number of threads that are not sleeping. +@item @code{Uptime} @tab How many seconds the server has been up. +@end multitable -@code{MERGE} tables can help you solve the following problems: +Some comments about the above: @itemize @bullet @item -Easily manage a set of log tables. For example, you can put data from -different months into separate files, compress some of them with -@code{myisampack}, and then create a @code{MERGE} to use these as one. -@item -Give you more speed. You can split a big read-only table based on some -criteria and then put the different table part on different disks. -A @code{MERGE} table on this could be much faster than using -the big table. (You can, of course, also use a RAID to get the same -kind of benefits.) -@item -Do more efficient searches. If you know exactly what you are looking -after, you can search in just one of the split tables for some queries -and use @strong{MERGE} table for others. You can even have many -different @code{MERGE} tables active, with possible overlapping files. -@item -More efficient repairs. It's easier to repair the individual files that -are mapped to a @code{MERGE} file than trying to repair a real big file. -@item -Instant mapping of many files as one. A @code{MERGE} table uses the -index of the individual tables. It doesn't need to maintain an index of -its one. This makes @code{MERGE} table collections VERY fast to make or -remap. Note that you must specify the key definitions when you create -a @code{MERGE} table!. +If @code{Opened_tables} is big, then your @code{table_cache} +variable is probably too small. @item -If you have a set of tables that you join to a big table on demand or -batch, you should instead create a @code{MERGE} table on them on demand. -This is much faster and will save a lot of disk space. +If @code{key_reads} is big, then your @code{key_cache} is probably too +small. The cache hit rate can be calculated with +@code{key_reads}/@code{key_read_requests}. @item -Go around the file size limit for the operating system. +If @code{Handler_read_rnd} is big, then you probably have a lot of +queries that require @strong{MySQL} to scan whole tables or you have +joins that don't use keys properly. @item -You can create an alias/synonym for a table by just using MERGE over one -table. There shouldn't be any really notable performance impacts of doing this -(only a couple of indirect calls and memcpy's for each read). +If @code{Threads_created} is big, you may want to increase the +@code{thread_cache_size} variable. @end itemize -The disadvantages with @code{MERGE} tables are: - -@itemize @bullet -@item -You can't use @code{INSERT} on @code{MERGE} tables, as @strong{MySQL} -can't know in which of the tables we should insert the row. -@item -You can only use identical @code{MyISAM} tables for a @code{MERGE} table. -@item -@code{MERGE} tables uses more file descriptors. If you are using a -@strong{MERGE} that maps over 10 tables and 10 users are using this, you -are using 10*10 + 10 file descriptors. (10 data files for 10 users -and 10 shared index files.) -@item -Key reads are slower. When you do a read on a key, the @code{MERGE} -handler will need to issue a read on all underlying tables to check -which one most closely matches the given key. If you then do a 'read-next' -then the merge table handler will need to search the read buffers -to find the next key. Only when one key buffer is used up, the handler -will need to read the next key block. This makes @code{MERGE} keys much slower -on @code{eq_ref} searches, but not much slower on @code{ref} searches. -@xref{EXPLAIN}. -@item -You can't do @code{DROP TABLE}, @code{ALTER TABLE} or @code{DELETE FROM -table_name} without a @code{WHERE} clause on any of the table that is -mapped by a @code{MERGE} table that is 'open'. If you do this, the -@code{MERGE} table may still refer to the original table and you will -get unexpected results. -@end itemize - -The following example shows you how to use @code{MERGE} tables: +@node SHOW VARIABLES, SHOW LOGS, SHOW STATUS, SHOW +@subsection @code{SHOW VARIABLES} @example -CREATE TABLE t1 (a INT AUTO_INCREMENT PRIMARY KEY, message CHAR(20)); -CREATE TABLE t2 (a INT AUTO_INCREMENT PRIMARY KEY, message CHAR(20)); -INSERT INTO t1 (message) VALUES ("Testing"),("table"),("t1"); -INSERT INTO t2 (message) VALUES ("Testing"),("table"),("t2"); -CREATE TABLE total (a INT NOT NULL, message CHAR(20), KEY(a)) TYPE=MERGE UNION=(t1,t2); +SHOW VARIABLES [LIKE wild] @end example -Note that we didn't create a @code{UNIQUE} or @code{PRIMARY KEY} in the -@code{total} table as the key isn't going to be unique in the @code{total} -table. +@code{SHOW VARIABLES} shows the values of some @strong{MySQL} system +variables. You can also get this information using the @code{mysqladmin +variables} command. If the default values are unsuitable, you can set most +of these variables using command-line options when @code{mysqld} starts up. +@xref{Command-line options}. -Note that you can also manipulate the @code{.MRG} file directly from -the outside of the @strong{MySQL} server: +The output resembles that shown below, though the format and numbers may +differ somewhat: @example -shell> cd /mysql-data-directory/current-database -shell> ls -1 t1.MYI t2.MYI > total.MRG -shell> mysqladmin flush-tables ++-------------------------+---------------------------+ +| Variable_name | Value | ++-------------------------+---------------------------+ +| ansi_mode | OFF | +| back_log | 50 | +| basedir | /my/monty/ | +| bdb_cache_size | 16777216 | +| bdb_log_buffer_size | 32768 | +| bdb_home | /my/monty/data/ | +| bdb_max_lock | 10000 | +| bdb_logdir | | +| bdb_shared_data | OFF | +| bdb_tmpdir | /tmp/ | +| binlog_cache_size | 32768 | +| concurrent_insert | ON | +| connect_timeout | 5 | +| datadir | /my/monty/data/ | +| delay_key_write | ON | +| delayed_insert_limit | 100 | +| delayed_insert_timeout | 300 | +| delayed_queue_size | 1000 | +| flush | OFF | +| flush_time | 0 | +| have_bdb | YES | +| have_innodb | YES | +| have_raid | YES | +| have_ssl | NO | +| init_file | | +| interactive_timeout | 28800 | +| join_buffer_size | 131072 | +| key_buffer_size | 16776192 | +| language | /my/monty/share/english/ | +| large_files_support | ON | +| log | OFF | +| log_update | OFF | +| log_bin | OFF | +| log_slave_updates | OFF | +| long_query_time | 10 | +| low_priority_updates | OFF | +| lower_case_table_names | 0 | +| max_allowed_packet | 1048576 | +| max_binlog_cache_size | 4294967295 | +| max_connections | 100 | +| max_connect_errors | 10 | +| max_delayed_threads | 20 | +| max_heap_table_size | 16777216 | +| max_join_size | 4294967295 | +| max_sort_length | 1024 | +| max_tmp_tables | 32 | +| max_write_lock_count | 4294967295 | +| myisam_recover_options | DEFAULT | +| myisam_sort_buffer_size | 8388608 | +| net_buffer_length | 16384 | +| net_read_timeout | 30 | +| net_retry_count | 10 | +| net_write_timeout | 60 | +| open_files_limit | 0 | +| pid_file | /my/monty/data/donna.pid | +| port | 3306 | +| protocol_version | 10 | +| record_buffer | 131072 | +| query_buffer_size | 0 | +| safe_show_database | OFF | +| server_id | 0 | +| skip_locking | ON | +| skip_networking | OFF | +| skip_show_database | OFF | +| slow_launch_time | 2 | +| socket | /tmp/mysql.sock | +| sort_buffer | 2097116 | +| table_cache | 64 | +| table_type | MYISAM | +| thread_cache_size | 4 | +| thread_stack | 65536 | +| tmp_table_size | 1048576 | +| tmpdir | /tmp/ | +| version | 3.23.29a-gamma-debug | +| wait_timeout | 28800 | ++-------------------------+---------------------------+ @end example -Now you can do things like: +Each option is described below. Values for buffer sizes, lengths, and stack +sizes are given in bytes. You can specify values with a suffix of @samp{K} +or @samp{M} to indicate kilobytes or megabytes. For example, @code{16M} +indicates 16 megabytes. The case of suffix letters does not matter; +@code{16M} and @code{16m} are equivalent: -@example -mysql> select * from total; -+---+---------+ -| a | message | -+---+---------+ -| 1 | Testing | -| 2 | table | -| 3 | t1 | -| 1 | Testing | -| 2 | table | -| 3 | t2 | -+---+---------+ -@end example +@cindex variables, values +@table @code +@item @code{ansi_mode}. +Is @code{ON} if @code{mysqld} was started with @code{--ansi}. +@xref{ANSI mode}. -To remap a @code{MERGE} table you can do one of the following: +@item @code{back_log} +The number of outstanding connection requests @strong{MySQL} can have. This +comes into play when the main @strong{MySQL} thread gets @strong{VERY} +many connection requests in a very short time. It then takes some time +(although very little) for the main thread to check the connection and start +a new thread. The @code{back_log} value indicates how many requests can be +stacked during this short time before @strong{MySQL} momentarily stops +answering new requests. You need to increase this only if you expect a large +number of connections in a short period of time. -@itemize @bullet -@item -@code{DROP} the table and re-create it -@item -Use @code{ALTER TABLE table_name UNION(...)} -@item -Change the @code{.MRG} file and issue a @code{FLUSH TABLE} on the -@code{MERGE} table and all underlying tables to force the handler to -read the new definition file. -@end itemize +In other words, this value is the size of the listen queue for incoming +TCP/IP connections. Your operating system has its own limit on the size +of this queue. The manual page for the Unix @code{listen(2)} system +call should have more details. Check your OS documentation for the +maximum value for this variable. Attempting to set @code{back_log} +higher than your operating system limit will be ineffective. -@cindex tables, ISAM -@node ISAM, HEAP, MERGE, Table types -@section ISAM Tables +@item @code{basedir} +The value of the @code{--basedir} option. -You can also use the deprecated ISAM table type. This will disappear -rather soon because @code{MyISAM} is a better implementation of the same -thing. ISAM uses a @code{B-tree} index. The index is stored in a file -with the @code{.ISM} extension, and the data is stored in a file with the -@code{.ISD} extension. You can check/repair ISAM tables with the -@code{isamchk} utility. @xref{Crash recovery}. +@item @code{bdb_cache_size} +The buffer that is allocated to cache index and rows for @code{BDB} +tables. If you don't use @code{BDB} tables, you should start +@code{mysqld} with @code{--skip-bdb} to not waste memory for this +cache. -@code{ISAM} has the following features/properties: +@item @code{bdb_log_buffer_size} +The buffer that is allocated to cache index and rows for @code{BDB} +tables. If you don't use @code{BDB} tables, you should set this to 0 or +start @code{mysqld} with @code{--skip-bdb} to not waste memory for this +cache. -@itemize @bullet -@item Compressed and fixed-length keys -@item Fixed and dynamic record length -@item 16 keys with 16 key parts/key -@item Max key length 256 (default) -@item Data is stored in machine format; this is fast, but is machine/OS dependent. -@end itemize +@item @code{bdb_home} +The value of the @code{--bdb-home} option. -Most of the things true for @code{MyISAM} tables are also true for @code{ISAM} -tables. @xref{MyISAM}. The major differences compared to @code{MyISAM} -tables are: +@item @code{bdb_max_lock} +The maximum number of locks (1000 by default) you can have active on a +BDB table. You should increase this if you get errors of type @code{bdb: +Lock table is out of available locks} or @code{Got error 12 from ...} +when you have do long transactions or when @code{mysqld} has to examine +a lot of rows to calculate the query. -@itemize @bullet -@item ISAM tables are not binary portable across OS/Platforms. -@item Can't handle tables > 4G. -@item Only support prefix compression on strings. -@item Smaller key limits. -@item Dynamic tables get more fragmented. -@item Tables are compressed with @code{pack_isam} rather than with @code{myisampack}. -@end itemize +@item @code{bdb_logdir} +The value of the @code{--bdb-logdir} option. -If you want to convert an @code{ISAM} table to a @code{MyISAM} table so -that you can use utilities such as @code{mysqlcheck}, use an @code{ALTER -TABLE} statement: +@item @code{bdb_shared_data} +Is @code{ON} if you are using @code{--bdb-shared-data}. -@example -mysql> ALTER TABLE tbl_name TYPE = MYISAM; -@end example +@item @code{bdb_tmpdir} +The value of the @code{--bdb-tmpdir} option. -@cindex tables, @code{HEAP} -@node HEAP, InnoDB, ISAM, Table types -@section HEAP Tables +@item @code{binlog_cache_size}. The size of the cache to hold the SQL +statements for the binary log during a transaction. If you often use +big, multi-statement transactions you can increase this to get more +performance. @xref{COMMIT}. -@code{HEAP} tables use a hashed index and are stored in memory. This -makes them very fast, but if @strong{MySQL} crashes you will lose all -data stored in them. @code{HEAP} is very useful for temporary tables! +@item @code{character_set} +The default character set. -The @strong{MySQL} internal HEAP tables use 100% dynamic hashing -without overflow areas. There is no extra space needed for free lists. -@code{HEAP} tables also don't have problems with delete + inserts, which -normally is common with hashed tables: +@item @code{character_sets} +The supported character sets. -@example -mysql> CREATE TABLE test TYPE=HEAP SELECT ip,SUM(downloads) as down - FROM log_table GROUP BY ip; -mysql> SELECT COUNT(ip),AVG(down) FROM test; -mysql> DROP TABLE test; -@end example +@item @code{concurrent_inserts} +If @code{ON} (the default), @strong{MySQL} will allow you to use @code{INSERT} +on @code{MyISAM} tables at the same time as you run @code{SELECT} queries +on them. You can turn this option off by starting @code{mysqld} with @code{--safe} +or @code{--skip-new}. -Here are some things you should consider when you use @code{HEAP} tables: +@cindex timeout +@item @code{connect_timeout} +The number of seconds the @code{mysqld} server is waiting for a connect +packet before responding with @code{Bad handshake}. -@itemize @bullet -@item -You should always use specify @code{MAX_ROWS} in the @code{CREATE} statement -to ensure that you accidentally do not use all memory. -@item -Indexes will only be used with @code{=} and @code{<=>} (but are VERY fast). -@item -@code{HEAP} tables can only use whole keys to search for a row; compare this -to @code{MyISAM} tables where any prefix of the key can be used to find rows. -@item -@code{HEAP} tables use a fixed record length format. -@item -@code{HEAP} doesn't support @code{BLOB}/@code{TEXT} columns. -@item -@code{HEAP} doesn't support @code{AUTO_INCREMENT} columns. -@item -@code{HEAP} doesn't support an index on a @code{NULL} column. -@item -You can have non-unique keys in a @code{HEAP} table (this isn't common for -hashed tables). -@item -@code{HEAP} tables are shared between all clients (just like any other -table). -@item -You can't search for the next entry in order (that is, to use the index -to do an @code{ORDER BY}). -@item -Data for @code{HEAP} tables are allocated in small blocks. The tables -are 100% dynamic (on inserting). No overflow areas and no extra key -space are needed. Deleted rows are put in a linked list and are -reused when you insert new data into the table. -@item -You need enough extra memory for all HEAP tables that you want to use at -the same time. -@item -To free memory, you should execute @code{DELETE FROM heap_table}, -@code{TRUNCATE heap_table} or @code{DROP TABLE heap_table}. -@item -@strong{MySQL} cannot find out approximately how many rows there -are between two values (this is used by the range optimizer to decide which -index to use). This may affect some queries if you change a @code{MyISAM} -table to a @code{HEAP} table. -@item -To ensure that you accidentally don't do anything foolish, you can't create -@code{HEAP} tables bigger than @code{max_heap_table_size}. -@end itemize +@item @code{datadir} +The value of the @code{--datadir} option. -The memory needed for one row in a @code{HEAP} table is: +@item @code{delay_key_write} +If enabled (is on by default), @strong{MySQL} will honor the +@code{delay_key_write} option @code{CREATE TABLE}. This means that the +key buffer for tables with this option will not get flushed on every +index update, but only when a table is closed. This will speed up +writes on keys a lot, but you should add automatic checking of all tables +with @code{myisamchk --fast --force} if you use this. Note that if you +start @code{mysqld} with the @code{--delay-key-write-for-all-tables} +option this means that all tables will be treated as if they were +created with the @code{delay_key_write} option. You can clear this flag +by starting @code{mysqld} with @code{--skip-new} or @code{--safe-mode}. -@example -SUM_OVER_ALL_KEYS(max_length_of_key + sizeof(char*) * 2) -+ ALIGN(length_of_row+1, sizeof(char*)) -@end example +@item @code{delayed_insert_limit} +After inserting @code{delayed_insert_limit} rows, the @code{INSERT +DELAYED} handler will check if there are any @code{SELECT} statements +pending. If so, it allows these to execute before continuing. -@code{sizeof(char*)} is 4 on 32-bit machines and 8 on 64-bit machines. +@item @code{delayed_insert_timeout} +How long a @code{INSERT DELAYED} thread should wait for @code{INSERT} +statements before terminating. -@node InnoDB, BDB, HEAP, Table types -@section InnoDB Tables +@item @code{delayed_queue_size} +What size queue (in rows) should be allocated for handling @code{INSERT +DELAYED}. If the queue becomes full, any client that does @code{INSERT +DELAYED} will wait until there is room in the queue again. -@menu -* InnoDB overview:: InnoDB tables overview -* InnoDB start:: InnoDB startup options -* InnoDB init:: Creating InnoDB table space. -* Using InnoDB tables:: Creating InnoDB tables -* Adding and removing:: Adding and removing InnoDB data and log files -* Backing up:: Backing up and recovering an InnoDB database -* Moving:: Moving an InnoDB database to another machine -* InnoDB transaction model:: InnoDB transaction model. -* Implementation:: Implementation of multiversioning -* Table and index:: Table and index structures -* File space management:: File space management and disk i/o -* Error handling:: Error handling -* InnoDB restrictions:: Some restrictions on InnoDB tables -* InnoDB contact information:: InnoDB contact information. -@end menu +@item @code{flush} +This is @code{ON} if you have started @strong{MySQL} with the @code{--flush} +option. -@node InnoDB overview, InnoDB start, InnoDB, InnoDB -@subsection InnoDB tables overview +@item @code{flush_time} +If this is set to a non-zero value, then every @code{flush_time} seconds all +tables will be closed (to free up resources and sync things to disk). We +only recommend this option on Win95, Win98, or on systems where you have +very little resources. -InnoDB tables are included in the @strong{MySQL} source distribution -starting from 3.23.34a and are activated in the @strong{MySQL -max} -binary. +@item @code{have_bdb} +@code{YES} if @code{mysqld} supports Berkeley DB tables. @code{DISABLED} +if @code{--skip-bdb} is used. +@item @code{have_innodb} +@code{YES} if @code{mysqld} supports InnoDB tables. @code{DISABLED} +if @code{--skip-innodb} is used. +@item @code{have_raid} +@code{YES} if @code{mysqld} supports the @code{RAID} option. +@item @code{have_ssl} +@code{YES} if @code{mysqld} supports SSL (encryption) on the client/server +protocol. -If you have downloaded a binary version of @strong{MySQL} that includes -support for InnoDB (mysqld-max), simply follow the instructions for -installing a binary version of @strong{MySQL}. @xref{Installing binary}. -@xref{mysqld-max, , @code{mysqld-max}}. +@item @code{init_file} +The name of the file specified with the @code{--init-file} option when +you start the server. This is a file of SQL statements you want the +server to execute when it starts. -To compile @strong{MySQL} with InnoDB support, download MySQL-3.23.37 or newer -and configure @code{MySQL} with the @code{--with-innodb} option. -@xref{Installing source}. +@item @code{interactive_timeout} +The number of seconds the server waits for activity on an interactive +connection before closing it. An interactive client is defined as a +client that uses the @code{CLIENT_INTERACTIVE} option to +@code{mysql_real_connect()}. See also @code{wait_timeout}. -@example -cd /path/to/source/of/mysql-3.23.37 -./configure --with-innodb -@end example +@item @code{join_buffer_size} +The size of the buffer that is used for full joins (joins that do not +use indexes). The buffer is allocated one time for each full join +between two tables. Increase this value to get a faster full join when +adding indexes is not possible. (Normally the best way to get fast joins +is to add indexes.) -InnoDB provides @strong{MySQL} with a transaction-safe table handler with -commit, rollback, and crash recovery capabilities. InnoDB does -locking on row level, and also provides an Oracle-style consistent -non-locking read in @code{SELECTS}, which increases transaction -concurrency. There is not need for lock escalation in InnoDB, -because row level locks in InnoDB fit in very small space. +@c Make texi2html support index @anchor{Index cache size}. Then change +@c some xrefs to point here +@cindex indexes, block size +@item @code{key_buffer_size} +Index blocks are buffered and are shared by all threads. +@code{key_buffer_size} is the size of the buffer used for index blocks. -InnoDB has been designed for maximum performance when processing -large data volumes. Its CPU efficiency is probably not -matched by any other disk-based relational database engine. +Increase this to get better index handling (for all reads and multiple +writes) to as much as you can afford; 64M on a 256M machine that mainly +runs @strong{MySQL} is quite common. If you, however, make this too big +(more than 50% of your total memory?) your system may start to page and +become REALLY slow. Remember that because @strong{MySQL} does not cache +data read, that you will have to leave some room for the OS filesystem +cache. -You can find the latest information about InnoDB at -@uref{http://www.innodb.com}. The most up-to-date version of the -InnoDB manual is always placed there, and you can also order commercial -support for InnoDB. +You can check the performance of the key buffer by doing @code{show +status} and examine the variables @code{Key_read_requests}, +@code{Key_reads}, @code{Key_write_requests}, and @code{Key_writes}. The +@code{Key_reads/Key_read_request} ratio should normally be < 0.01. +The @code{Key_write/Key_write_requests} is usually near 1 if you are +using mostly updates/deletes but may be much smaller if you tend to +do updates that affect many at the same time or if you are +using @code{delay_key_write}. @xref{SHOW}. -Technically, InnoDB is a database backend placed under @strong{MySQL}. InnoDB -has its own buffer pool for caching data and indexes in main -memory. InnoDB stores its tables and indexes in a tablespace, which -may consist of several files. This is different from, for example, -@code{MyISAM} tables where each table is stored as a separate file. +To get even more speed when writing many rows at the same time, use +@code{LOCK TABLES}. @xref{LOCK TABLES, , @code{LOCK TABLES}}. -InnoDB is distributed under the GNU GPL License Version 2 (of June 1991). -In the source distribution of @strong{MySQL}, InnoDB appears as a subdirectory. +@item @code{language} +The language used for error messages. -@node InnoDB start, InnoDB init, InnoDB overview, InnoDB -@subsection InnoDB startup options +@item @code{large_file_support} +If @code{mysqld} was compiled with options for big file support. -Beginning from @strong{MySQL}-3.23.37 the prefix of the options is changed -from @code{innobase_...} to @code{innodb_...}. +@item @code{locked_in_memory} +If @code{mysqld} was locked in memory with @code{--memlock} -To use InnoDB tables you @strong{MUST} specify configuration parameters -in the @strong{MySQL} configuration file in the @code{[mysqld]} section of -the configuration file @file{my.cnf}. @xref{Option files}. +@item @code{log} +If logging of all queries is enabled. -The only required parameter to use InnoDB is @code{innodb_data_file_path}, -but you should set others if you want to get a better performance. +@item @code{log_update} +If the update log is enabled. -Suppose you have a Windows NT machine with 128 MB RAM and a single 10 GB -hard disk. Below is an example of possible configuration parameters in -@file{my.cnf} for InnoDB: +@item @code{log_bin} +If the binary log is enabled. -@example -innodb_data_file_path = ibdata1:2000M;ibdata2:2000M -innodb_data_home_dir = c:\ibdata -set-variable = innodb_mirrored_log_groups=1 -innodb_log_group_home_dir = c:\iblogs -set-variable = innodb_log_files_in_group=3 -set-variable = innodb_log_file_size=30M -set-variable = innodb_log_buffer_size=8M -innodb_flush_log_at_trx_commit=1 -innodb_log_arch_dir = c:\iblogs -innodb_log_archive=0 -set-variable = innodb_buffer_pool_size=80M -set-variable = innodb_additional_mem_pool_size=10M -set-variable = innodb_file_io_threads=4 -set-variable = innodb_lock_wait_timeout=50 -@end example +@item @code{log_slave_updates} +If the updates from the slave should be logged. -Suppose you have a Linux machine with 512 MB RAM and -three 20 GB hard disks (at directory paths @file{/}, -@file{/dr2} and @file{/dr3}). -Below is an example of possible configuration parameters in @file{my.cnf} for -InnoDB: +@item @code{long_query_time} +If a query takes longer than this (in seconds), the @code{Slow_queries} counter +will be incremented. If you are using @code{--log-slow-queries}, the query +will be logged to the slow query logfile. @xref{Slow query log}. -@example -innodb_data_file_path = ibdata/ibdata1:2000M;dr2/ibdata/ibdata2:2000M -innodb_data_home_dir = / -set-variable = innodb_mirrored_log_groups=1 -innodb_log_group_home_dir = /dr3 -set-variable = innodb_log_files_in_group=3 -set-variable = innodb_log_file_size=50M -set-variable = innodb_log_buffer_size=8M -innodb_flush_log_at_trx_commit=1 -innodb_log_arch_dir = /dr3/iblogs -innodb_log_archive=0 -set-variable = innodb_buffer_pool_size=400M -set-variable = innodb_additional_mem_pool_size=20M -set-variable = innodb_file_io_threads=4 -set-variable = innodb_lock_wait_timeout=50 -@end example +@item @code{lower_case_table_names} +If set to 1 table names are stored in lowercase on disk. This will enable +you to access the table names case-insensitive also on Unix. +@xref{Name case sensitivity}. -Note that we have placed the two data files on different disks. -The reason for the name @code{innodb_data_file_path} is that -you can also specify paths to your data files, and -@code{innodb_data_home_dir} is just textually catenated -before your data file paths, adding a possible slash or -backslash in between. InnoDB will fill the tablespace -formed by the data files from bottom up. In some cases it will -improve the performance of the database if all data is not placed -on the same physical disk. Putting log files on a different disk from -data is very often beneficial for performance. +@item @code{max_allowed_packet} +The maximum size of one packet. The message buffer is initialized to +@code{net_buffer_length} bytes, but can grow up to @code{max_allowed_packet} +bytes when needed. This value by default is small, to catch big (possibly +wrong) packets. You must increase this value if you are using big +@code{BLOB} columns. It should be as big as the biggest @code{BLOB} you want +to use. The current protocol limits @code{max_allowed_packet} to 16M. -The meanings of the configuration parameters are the following: +@item @code{max_binlog_cache_size} +If a multi-statement transaction requires more than this amount of memory, +one will get the error "Multi-statement transaction required more than +'max_binlog_cache_size' bytes of storage". -@multitable @columnfractions .30 .70 -@item @code{innodb_data_home_dir} @tab -The common part of the directory path for all InnoDB data files. -@item @code{innodb_data_file_path} @tab -Paths to individual data files and their sizes. The full directory path -to each data file is acquired by concatenating innodb_data_home_dir to -the paths specified here. The file sizes are specified in megabytes, -hence the 'M' after the size specification above. Do not set a file size -bigger than 4000M, and on most operating systems not bigger than 2000M. -InnoDB also understands the abbreviation 'G', 1G meaning 1024M. -The sum of the sizes of the files must be at least 10 MB. -@item @code{innodb_mirrored_log_groups} @tab -Number of identical copies of log groups we -keep for the database. Currently this should be set to 1. -@item @code{innodb_log_group_home_dir} @tab -Directory path to InnoDB log files. -@item @code{innodb_log_files_in_group} @tab -Number of log files in the log group. InnoDB writes to the files in a -circular fashion. Value 3 is recommended here. -@item @code{innodb_log_file_size} @tab -Size of each log file in a log group in megabytes. Sensible values range -from 1M to the size of the buffer pool specified below. The bigger the -value, the less checkpoint flush activity is needed in the buffer pool, -saving disk i/o. But bigger log files also mean that recovery will be -slower in case of a crash. File size restriction as for a data file. -@item @code{innodb_log_buffer_size} @tab -The size of the buffer which InnoDB uses to write log to the log files -on disk. Sensible values range from 1M to half the combined size of log -files. A big log buffer allows large transactions to run without a need -to write the log to disk until the transaction commit. Thus, if you have -big transactions, making the log buffer big will save disk i/o. -@item @code{innodb_flush_log_at_trx_commit} @tab -Normally this is set to 1, meaning that at a transaction commit the log -is flushed to disk, and the modifications made by the transaction become -permanent, and survive a database crash. If you are willing to -compromise this safety, and you are running small transactions, you may -set this to 0 to reduce disk i/o to the logs. -@item @code{innodb_log_arch_dir} @tab -The directory where fully written log files would be archived if we used -log archiving. The value of this parameter should currently be set the -same as @code{innodb_log_group_home_dir}. -@item @code{innodb_log_archive} @tab -This value should currently be set to 0. As recovery from a backup is -done by @strong{MySQL} using its own log files, there is currently no need to -archive InnoDB log files. -@item @code{innodb_buffer_pool_size} @tab -The size of the memory buffer InnoDB uses to cache data and indexes of -its tables. The bigger you set this the less disk i/o is needed to -access data in tables. On a dedicated database server you may set this -parameter up to 90 % of the machine physical memory size. Do not set it -too large, though, because competition of the physical memory may cause -paging in the operating system. -@item @code{innodb_additional_mem_pool_size} @tab -Size of a memory pool InnoDB uses to store data dictionary information -and other internal data structures. A sensible value for this might be -2M, but the more tables you have in your application the more you will -need to allocate here. If InnoDB runs out of memory in this pool, it -will start to allocate memory from the operating system, and write -warning messages to the @strong{MySQL} error log. -@item @code{innodb_file_io_threads} @tab -Number of file i/o threads in InnoDB. Normally, this should be 4, but -on Windows NT disk i/o may benefit from a larger number. -@item @code{innodb_lock_wait_timeout} @tab -Timeout in seconds an InnoDB transaction may wait for a lock before -being rolled back. InnoDB automatically detects transaction deadlocks -in its own lock table and rolls back the transaction. If you use -@code{LOCK TABLES} command, or other transaction-safe table handlers -than InnoDB in the same transaction, then a deadlock may arise which -InnoDB cannot notice. In cases like this the timeout is useful to -resolve the situation. -@item @code{innodb_flush_method} @tab -(Available from 3.23.40 up.) -The default value for this is @code{fdatasync}. -Another option is @code{O_DSYNC}. -Options @code{littlesync} and @code{nosync} have the -risk that in an operating system crash or a power outage you may easily -end up with a half-written database page, and you have to do a recovery -from a backup. See the section "InnoDB performance tuning", item 6, below -for tips on how to set this parameter. If you are happy with your database -performance it is wisest not to specify this parameter at all, in which -case it will get the default value. -@end multitable - -@node InnoDB init, Using InnoDB tables, InnoDB start, InnoDB -@subsection Creating InnoDB table space +@item @code{max_binlog_size} +Available after 3.23.33. If a write to the binary (replication) log exceeds +the given value, rotate the logs. You cannot set it to less than 1024 bytes, +or more than 1 GB. Default is 1 GB. -Suppose you have installed @strong{MySQL} and have edited @file{my.cnf} so that -it contains the necessary InnoDB configuration parameters. -Before starting @strong{MySQL} you should check that the directories you have -specified for InnoDB data files and log files exist and that you have -access rights to those directories. InnoDB -cannot create directories, only files. Check also you have enough disk space -for the data and log files. +@item @code{max_connections} +The number of simultaneous clients allowed. Increasing this value increases +the number of file descriptors that @code{mysqld} requires. See below for +comments on file descriptor limits. @xref{Too many connections}. -When you now start @strong{MySQL}, InnoDB will start creating your data files -and log files. InnoDB will print something like the following: +@item @code{max_connect_errors} +If there is more than this number of interrupted connections from a host +this host will be blocked from further connections. You can unblock a host +with the command @code{FLUSH HOSTS}. -@example -~/mysqlm/sql > mysqld -InnoDB: The first specified data file /home/heikki/data/ibdata1 did not exist: -InnoDB: a new database to be created! -InnoDB: Setting file /home/heikki/data/ibdata1 size to 134217728 -InnoDB: Database physically writes the file full: wait... -InnoDB: Data file /home/heikki/data/ibdata2 did not exist: new to be created -InnoDB: Setting file /home/heikki/data/ibdata2 size to 262144000 -InnoDB: Database physically writes the file full: wait... -InnoDB: Log file /home/heikki/data/logs/ib_logfile0 did not exist: new to be c -reated -InnoDB: Setting log file /home/heikki/data/logs/ib_logfile0 size to 5242880 -InnoDB: Log file /home/heikki/data/logs/ib_logfile1 did not exist: new to be c -reated -InnoDB: Setting log file /home/heikki/data/logs/ib_logfile1 size to 5242880 -InnoDB: Log file /home/heikki/data/logs/ib_logfile2 did not exist: new to be c -reated -InnoDB: Setting log file /home/heikki/data/logs/ib_logfile2 size to 5242880 -InnoDB: Started -mysqld: ready for connections -@end example +@item @code{max_delayed_threads} +Don't start more than this number of threads to handle @code{INSERT DELAYED} +statements. If you try to insert data into a new table after all @code{INSERT +DELAYED} threads are in use, the row will be inserted as if the +@code{DELAYED} attribute wasn't specified. -A new InnoDB database has now been created. You can connect to the @strong{MySQL} -server with the usual @strong{MySQL} client programs like @code{mysql}. -When you shut down the @strong{MySQL} server with @file{mysqladmin shutdown}, -InnoDB output will be like the following: +@item @code{max_heap_table_size} +Don't allow creation of heap tables bigger than this. -@example -010321 18:33:34 mysqld: Normal shutdown -010321 18:33:34 mysqld: Shutdown Complete -InnoDB: Starting shutdown... -InnoDB: Shutdown completed -@end example +@item @code{max_join_size} +Joins that are probably going to read more than @code{max_join_size} +records return an error. Set this value if your users tend to perform joins +that lack a @code{WHERE} clause, that take a long time, and that return +millions of rows. -You can now look at the data files and logs directories and you -will see the files created. The log directory will also contain -a small file named @file{ib_arch_log_0000000000}. That file -resulted from the database creation, after which InnoDB switched off -log archiving. -When @strong{MySQL} is again started, the output will be like the following: +@item @code{max_sort_length} +The number of bytes to use when sorting @code{BLOB} or @code{TEXT} +values (only the first @code{max_sort_length} bytes of each value +are used; the rest are ignored). -@example -~/mysqlm/sql > mysqld -InnoDB: Started -mysqld: ready for connections -@end example +@item @code{max_user_connections} +The maximum number of active connections for a single user (0 = no limit). -@menu -* Error creating InnoDB:: -@end menu +@item @code{max_tmp_tables} +(This option doesn't yet do anything.) +Maximum number of temporary tables a client can keep open at the same time. -@node Error creating InnoDB, , InnoDB init, InnoDB init -@subsubsection If something goes wrong in database creation +@item @code{max_write_lock_count} +After this many write locks, allow some read locks to run in between. -If something goes wrong in an InnoDB database creation, you should -delete all files created by InnoDB. This means all data files, all log -files, the small archived log file, and in the case you already did -create some InnoDB tables, delete also the corresponding @file{.frm} -files for these tables from the @strong{MySQL} database -directories. Then you can try the InnoDB database creation again. +@item @code{myisam_recover_options} +The value of the @code{--myisam-recover} option. -@node Using InnoDB tables, Adding and removing, InnoDB init, InnoDB -@subsection Creating InnoDB tables +@item @code{myisam_sort_buffer_size} +The buffer that is allocated when sorting the index when doing a +@code{REPAIR} or when creating indexes with @code{CREATE INDEX} or +@code{ALTER TABLE}. -Suppose you have started the @strong{MySQL} client with the command -@code{mysql test}. -To create a table in the InnoDB format you must specify -@code{TYPE = InnoDB} in the table creation SQL command: +@item @code{myisam_max_extra_sort_file_size}. +If the creating of the temporary file for fast index creation would be +this much bigger than using the key cache, then prefer the key cache +method. This is mainly used to force long character keys in large +tables to use the slower key cache method to create the index. +@strong{NOTE} that this parameter is given in megabytes! -@example -CREATE TABLE CUSTOMER (A INT, B CHAR (20), INDEX (A)) TYPE = InnoDB; -@end example +@item @code{myisam_max_sort_file_size} +The maximum size of the temporary file @strong{MySQL} is allowed to use +while recreating the index (during @code{REPAIR}, @code{ALTER TABLE} +or @code{LOAD DATA INFILE}. If the file size would be bigger than this, +the index will be created through the key cache (which is slower). +@strong{NOTE} that this parameter is given in megabytes! -This SQL command will create a table and an index on column @code{A} -into the InnoDB tablespace consisting of the data files you specified -in @file{my.cnf}. In addition @strong{MySQL} will create a file -@file{CUSTOMER.frm} to the @strong{MySQL} database directory @file{test}. -Internally, InnoDB will add to its own data dictionary an entry -for table @code{'test/CUSTOMER'}. Thus you can create a table -of the same name @code{CUSTOMER} in another database of @strong{MySQL}, and -the table names will not collide inside InnoDB. +@item @code{net_buffer_length} +The communication buffer is reset to this size between queries. This +should not normally be changed, but if you have very little memory, you +can set it to the expected size of a query. (That is, the expected length of +SQL statements sent by clients. If statements exceed this length, the buffer +is automatically enlarged, up to @code{max_allowed_packet} bytes.) -You can query the amount of free space in the InnoDB tablespace -by issuing the table status command of @strong{MySQL} for any table you have -created with @code{TYPE = InnoDB}. Then the amount of free -space in the tablespace appears in the table comment section in the -output of @code{SHOW}. An example: +@item @code{net_read_timeout} +Number of seconds to wait for more data from a connection before aborting +the read. Note that when we don't expect data from a connection, the timeout +is defined by @code{write_timeout}. See also @code{slave_read_timeout}. -@example -SHOW TABLE STATUS FROM test LIKE 'CUSTOMER' -@end example +@item @code{net_retry_count} +If a read on a communication port is interrupted, retry this many times +before giving up. This value should be quite high on @code{FreeBSD} as +internal interrupts are sent to all threads. -Note that the statistics @code{SHOW} gives about InnoDB tables -are only approximate: they are used in SQL optimization. Table and -index reserved sizes in bytes are accurate, though. +@item @code{net_write_timeout} +Number of seconds to wait for a block to be written to a connection before +aborting the write. -NOTE: @code{DROP DATABASE} does not currently work for InnoDB tables! -You must drop the tables individually. Also take care not to delete or -add @file{.frm} files to your InnoDB database manually: use -@code{CREATE TABLE} and @code{DROP TABLE} commands. -InnoDB has its own internal data dictionary, and you will get problems -if the @strong{MySQL} @file{.frm} files are out of 'sync' with the InnoDB -internal data dictionary. +@item @code{open_files_limit} +If this is not 0, then @code{mysqld} will use this value to reserve file +descriptors to use with @code{setrlimit()}. If this value is 0 then +@code{mysqld} will reserve @code{max_connections*5} or +@code{max_connections + table_cache*2} (whichever is larger) number of +files. You should try increasing this if @code{mysqld} gives you the +error 'Too many open files'. -@subsubsection Converting MyISAM tables to InnoDB +@item @code{pid_file} +The value of the @code{--pid-file} option. -InnoDB does not have a special optimization for separate index creation. -Therefore it does not pay to export and import the table and create indexes -afterwards. -The fastest way to alter a table to InnoDB is to do the inserts -directly to an InnoDB table, that is, use @code{ALTER TABLE ... TYPE=INNODB}, -or create an empty InnoDB table with identical definitions and insert -the rows with @code{INSERT INTO ... SELECT * FROM ...}. +@item @code{port} +The value of the @code{--port} option. -To get better control over the insertion process, it may be good to insert -big tables in pieces: +@item @code{protocol_version} +The protocol version used by the @strong{MySQL} server. -@example -INSERT INTO newtable SELECT * FROM oldtable WHERE yourkey > something - AND yourkey <= somethingelse; -@end example +@item @code{record_buffer} +Each thread that does a sequential scan allocates a buffer of this +size for each table it scans. If you do many sequential scans, you may +want to increase this value. -After all data has been inserted you can rename the tables. +@item @code{query_buffer_size} +The initial allocation of the query buffer. If most of your queries are +long (like when inserting blobs), you should increase this! -During the conversion of big tables you should set the InnoDB -buffer pool size big -to reduce disk i/o. Not bigger than 80 % of the physical memory, though. -You should set InnoDB log files big, and also the log buffer large. +@item @code{safe_show_databases} +Don't show databases for which the user doesn't have any database or +table privileges. This can improve security if you're concerned about +people being able to see what databases other users have. See also +@code{skip_show_databases}. -Make sure you do not run out of tablespace: InnoDB tables take a lot -more space than MyISAM tables. If an @code{ALTER TABLE} runs out -of space, it will start a rollback, and that can take hours if it is -disk-bound. -In inserts InnoDB uses the insert buffer to merge secondary index records -to indexes in batches. That saves a lot of disk i/o. In rollback no such -mechanism is used, and the rollback can take 30 times longer than the -insertion. +@item @code{server_id} +The value of the @code{--server-id} option. -In the case of a runaway rollback, if you do not have valuable data in your -database, -it is better that you kill the database process and delete all InnoDB data -and log files and all InnoDB table @file{.frm} files, and start -your job again, rather than wait for millions of disk i/os to complete. +@item @code{skip_locking} +Is OFF if @code{mysqld} uses external locking. -@node Adding and removing, Backing up, Using InnoDB tables, InnoDB -@subsection Adding and removing InnoDB data and log files +@item @code{skip_networking} +Is ON if we only allow local (socket) connections. -You cannot increase the size of an InnoDB data file. To add more into -your tablespace you have to add a new data file. To do this you have to -shut down your @strong{MySQL} database, edit the @file{my.cnf} file, adding a -new file to @code{innodb_data_file_path}, and then start @strong{MySQL} -again. +@item @code{skip_show_databases} +This prevents people from doing @code{SHOW DATABASES} if they don't have +the @code{PROCESS_PRIV} privilege. This can improve security if you're +concerned about people being able to see what databases other users +have. See also @code{safe_show_databases}. -Currently you cannot remove a data file from InnoDB. To decrease the -size of your database you have to use @code{mysqldump} to dump -all your tables, create a new database, and import your tables to the -new database. +@item @code{slave_read_timeout} +Number of seconds to wait for more data from a master/slave connection +before aborting the read. -If you want to change the number or the size of your InnoDB log files, -you have to shut down @strong{MySQL} and make sure that it shuts down without errors. -Then copy the old log files into a safe place just in case something -went wrong in the shutdown and you will need them to recover the -database. Delete then the old log files from the log file directory, -edit @file{my.cnf}, and start @strong{MySQL} again. InnoDB will tell -you at the startup that it is creating new log files. +@item @code{slow_launch_time} +If creating the thread takes longer than this value (in seconds), the +@code{Slow_launch_threads} counter will be incremented. -@node Backing up, Moving, Adding and removing, InnoDB -@subsection Backing up and recovering an InnoDB database +@item @code{socket} +The Unix socket used by the server. -The key to safe database management is taking regular backups. -To take a 'binary' backup of your database you have to do the following: +@item @code{sort_buffer} +Each thread that needs to do a sort allocates a buffer of this +size. Increase this value for faster @code{ORDER BY} or @code{GROUP BY} +operations. +@xref{Temporary files}. -@itemize @bullet -@item -Shut down your @strong{MySQL} database and make sure it shuts down without errors. -@item -Copy all your data files into a safe place. -@item -Copy all your InnoDB log files to a safe place. -@item -Copy your @file{my.cnf} configuration file(s) to a safe place. -@item -Copy all the @file{.frm} files for your InnoDB tables into a -safe place. -@end itemize +@item @code{table_cache} +The number of open tables for all threads. Increasing this value +increases the number of file descriptors that @code{mysqld} requires. +@strong{MySQL} needs two file descriptors for each unique open table. +See below for comments on file descriptor limits. You can check if you +need to increase the table cache by checking the @code{Opened_tables} +variable. @xref{SHOW}. If this variable is big and you don't do +@code{FLUSH TABLES} a lot (which just forces all tables to be closed and +reopenend), then you should increase the value of this variable. -There is currently no on-line or incremental backup tool available for -InnoDB, though they are in the TODO list. +Make sure that your operating system can handle the number of open file +descriptors implied by the @code{table_cache} setting. If @code{table_cache} +is set too high, @strong{MySQL} may run out of file descriptors and refuse +connections, fail to perform queries, and be very unreliable. -In addition to taking the binary backups described above, -you should also regularly take dumps of your tables with -@file{mysqldump}. The reason to this is that a binary file -may be corrupted without you noticing it. Dumped tables are stored -into text files which are human-readable and much simpler than -database binary files. Seeing table corruption from dumped files -is easier, and since their format is simpler, the chance for -serious data corruption in them is smaller. +For information about how the table cache works, see @ref{Table cache}. -A good idea is to take the dumps at the same time you take a binary -backup of your database. You have to shut out all clients from your -database to get a consistent snapshot of all your tables into your -dumps. Then you can take the binary backup, and you will then have -a consistent snapshot of your database in two formats. +@item @code{table_type} +The default table type -To be able to recover your InnoDB database to the present from the -binary backup described above, you have to run your @strong{MySQL} database -with the general logging and log archiving of @strong{MySQL} switched on. Here -by the general logging we mean the logging mechanism of the @strong{MySQL} server -which is independent of InnoDB logs. +@item @code{thread_cache_size} +How many threads we should keep in a cache for reuse. When a +client disconnects, the client's threads are put in the cache if there +aren't more than @code{thread_cache_size} threads from before. All new +threads are first taken from the cache, and only when the cache is empty +is a new thread created. This variable can be increased to improve +performance if you have a lot of new connections. (Normally this doesn't +give a notable performance improvement if you have a good +thread implementation.) By examing the difference between +the @code{Connections} and @code{Threads_created} you can see how efficient +the current thread cache is for you. -To recover from a crash of your @strong{MySQL} server process, the only thing -you have to do is to restart it. InnoDB will automatically check the -logs and perform a roll-forward of the database to the present. -InnoDB will automatically roll back uncommitted transactions which were -present at the time of the crash. During recovery, InnoDB will print -out something like the following: +@item @code{thread_concurrency} +On Solaris, @code{mysqld} will call @code{thr_setconcurrency()} with +this value. @code{thr_setconcurrency()} permits the application to give +the threads system a hint for the desired number of threads that should +be run at the same time. -@example -~/mysqlm/sql > mysqld -InnoDB: Database was not shut down normally. -InnoDB: Starting recovery from log files... -InnoDB: Starting log scan based on checkpoint at -InnoDB: log sequence number 0 13674004 -InnoDB: Doing recovery: scanned up to log sequence number 0 13739520 -InnoDB: Doing recovery: scanned up to log sequence number 0 13805056 -InnoDB: Doing recovery: scanned up to log sequence number 0 13870592 -InnoDB: Doing recovery: scanned up to log sequence number 0 13936128 -... -InnoDB: Doing recovery: scanned up to log sequence number 0 20555264 -InnoDB: Doing recovery: scanned up to log sequence number 0 20620800 -InnoDB: Doing recovery: scanned up to log sequence number 0 20664692 -InnoDB: 1 uncommitted transaction(s) which must be rolled back -InnoDB: Starting rollback of uncommitted transactions -InnoDB: Rolling back trx no 16745 -InnoDB: Rolling back of trx no 16745 completed -InnoDB: Rollback of uncommitted transactions completed -InnoDB: Starting an apply batch of log records to the database... -InnoDB: Apply batch completed -InnoDB: Started -mysqld: ready for connections -@end example +@item @code{thread_stack} +The stack size for each thread. Many of the limits detected by the +@code{crash-me} test are dependent on this value. The default is +large enough for normal operation. @xref{Benchmarks}. -If your database gets corrupted or your disk fails, you have -to do the recovery from a backup. In the case of corruption, you should -first find a backup which is not corrupted. From a backup do the recovery -from the general log files of @strong{MySQL} according to instructions in the -MySQL manual. +@item @code{timezone} +The timezone for the server. -@menu -* InnoDB checkpoints:: -@end menu +@item @code{tmp_table_size} +If an in-memory temporary table exceeds this size, @strong{MySQL} +will automatically convert it to an on-disk @code{MyISAM} table. +Increase the value of @code{tmp_table_size} if you do many advanced +@code{GROUP BY} queries and you have lots of memory. -@node InnoDB checkpoints, , Backing up, Backing up -@subsubsection Checkpoints +@item @code{tmpdir} +The directory used for temporary files and temporary tables. -InnoDB implements a checkpoint mechanism called a fuzzy -checkpoint. InnoDB will flush modified database pages from the buffer -pool in small batches, there is no need to flush the buffer pool -in one single batch, which would in practice stop processing -of user SQL statements for a while. +@item @code{version} +The version number for the server. -In crash recovery InnoDB looks for a checkpoint label written -to the log files. It knows that all modifications to the database -before the label are already present on the disk image of the database. -Then InnoDB scans the log files forward from the place of the checkpoint -applying the logged modifications to the database. +@item @code{wait_timeout} +The number of seconds the server waits for activity on a connection before +closing it. See also @code{interactive_timeout}. +@end table -InnoDB writes to the log files in a circular fashion. -All committed modifications which make the database pages in the buffer -pool different from the images on disk must be available in the log files -in case InnoDB has to do a recovery. This means that when InnoDB starts -to reuse a log file in the circular fashion, it has to make sure that the -database page images on disk already contain the modifications -logged in the log file InnoDB is going to reuse. In other words, InnoDB -has to make a checkpoint and often this involves flushing of -modified database pages to disk. - -The above explains why making your log files very big may save -disk i/o in checkpointing. It can make sense to set -the total size of the log files as big as the buffer pool or even bigger. -The drawback in big log files is that crash recovery can last longer -because there will be more log to apply to the database. - -@node Moving, InnoDB transaction model, Backing up, InnoDB -@subsection Moving an InnoDB database to another machine - -InnoDB data and log files are binary-compatible on all platforms -if the floating point number format on the machines is the same. -You can move an InnoDB database simply by copying all the relevant -files, which we already listed in the previous section on backing up -a database. If the floating point formats on the machines are -different but you have not used @code{FLOAT} or @code{DOUBLE} -data types in your tables then the procedure is the same: just copy -the relevant files. If the formats are different and your tables -contain floating point data, you have to use @file{mysqldump} -and @file{mysqlimport} to move those tables. - -A performance tip is to switch off the auto commit when you import -data into your database, assuming your tablespace has enough space for -the big rollback segment the big import transaction will generate. -Do the commit only after importing a whole table or a segment of -a table. - -@node InnoDB transaction model, Implementation, Moving, InnoDB -@subsection InnoDB transaction model - -In the InnoDB transaction model the goal has been to combine the best -sides of a multiversioning database to traditional two-phase locking. -InnoDB does locking on row level and runs queries by default -as non-locking consistent reads, in the style of Oracle. -The lock table in InnoDB is stored so space-efficiently that lock -escalation is not needed: typically several users are allowed -to lock every row in the database, or any random subset of the rows, -without InnoDB running out of memory. - -In InnoDB all user activity happens inside transactions. If the -auto commit mode is used in @strong{MySQL}, then each SQL statement -will form a single transaction. If the auto commit mode is -switched off, then we can think that a user always has a transaction -open. If he issues -the SQL @code{COMMIT} or @code{ROLLBACK} statement, that -ends the current transaction, and a new starts. Both statements -will release all InnoDB locks that were set during the -current transaction. A @code{COMMIT} means that the -changes made in the current transaction are made permanent -and become visible to other users. A @code{ROLLBACK} -on the other hand cancels all modifications made by the current -transaction. - -@menu -* InnoDB consistent read:: -* InnoDB locking reads:: -* InnoDB Next-key locking:: -* InnoDB Locks set:: -* InnoDB Deadlock detection:: -@end menu - -@node InnoDB consistent read, InnoDB locking reads, InnoDB transaction model, InnoDB transaction model -@subsubsection Consistent read +The manual section that describes tuning @strong{MySQL} contains some +information of how to tune the above variables. @xref{Server parameters}. -A consistent read means that InnoDB uses its multiversioning to -present to a query a snapshot of the database at a point in time. -The query will see the changes made by exactly those transactions that -committed before that point of time, and no changes made by later -or uncommitted transactions. The exception to this rule is that the -query will see the changes made by the transaction itself which issues -the query. +@node SHOW LOGS, SHOW PROCESSLIST, SHOW VARIABLES, SHOW +@subsection @code{SHOW LOGS} +@code{SHOW LOGS} shows you status information about existing log +files. It currently only displays information about Berkeley DB log +files. -When a transaction issues its first consistent read, InnoDB assigns -the snapshot, or the point of time, which all consistent reads in the -same transaction will use. In the snapshot are all transactions that -committed before assigning the snapshot. Thus the consistent reads -within the same transaction will also be consistent with respect to each -other. You can get a fresher snapshot for your queries by committing -the current transaction and after that issuing new queries. +@itemize @bullet +@item @code{File} shows the full path to the log file +@item @code{Type} shows the type of the log file (@code{BDB} for Berkeley +DB log files) +@item @code{Status} shows the status of the log file (@code{FREE} if the +file can be removed, or @code{IN USE} if the file is needed by the transaction +subsystem) +@end itemize -Consistent read is the default mode in which InnoDB processes -@code{SELECT} statements. A consistent read does not set any locks -on the tables it accesses, and therefore other users are free to -modify those tables at the same time a consistent read is being performed -on the table. +@cindex threads, display +@cindex processes, display +@findex threads +@findex PROCESSLIST +@node SHOW PROCESSLIST, SHOW GRANTS, SHOW LOGS, SHOW +@subsection @code{SHOW PROCESSLIST} -@node InnoDB locking reads, InnoDB Next-key locking, InnoDB consistent read, InnoDB transaction model -@subsubsection Locking reads +@code{SHOW PROCESSLIST} shows you which threads are running. You can +also get this information using the @code{mysqladmin processlist} +command. If you have the @strong{process} privilege, you can see all +threads. Otherwise, you can see only your own threads. @xref{KILL, , +@code{KILL}}. If you don't use the @code{FULL} option, then only +the first 100 characters of each query will be shown. -A consistent read is not convenient in some circumstances. -Suppose you want to add a new row into your table @code{CHILD}, -and make sure that the child already has a parent in table -@code{PARENT}. +This command is very useful if you get the 'too many connections' error +message and want to find out what's going on. @strong{MySQL} reserves +one extra connection for a client with the @code{Process_priv} privilege +to ensure that you should always be able to login and check the system +(assuming you are not giving this privilege to all your users). -Suppose you use a consistent read to read the table @code{PARENT} -and indeed see the parent of the child in the table. Can you now safely -add the child row to table @code{CHILD}? No, because it may -happen that meanwhile some other user has deleted the parent row -from the table @code{PARENT}, and you are not aware of that. +@cindex privileges, display +@node SHOW GRANTS, SHOW CREATE TABLE, SHOW PROCESSLIST, SHOW +@subsection @code{SHOW GRANTS} -The solution is to perform the @code{SELECT} in a locking -mode, @code{LOCK IN SHARE MODE}. +@code{SHOW GRANTS FOR user} lists the grant commands that must be issued to +duplicate the grants for a user. @example -SELECT * FROM PARENT WHERE NAME = 'Jones' LOCK IN SHARE MODE; +mysql> SHOW GRANTS FOR root@@localhost; ++---------------------------------------------------------------------+ +| Grants for root@@localhost | ++---------------------------------------------------------------------+ +| GRANT ALL PRIVILEGES ON *.* TO 'root'@@'localhost' WITH GRANT OPTION | ++---------------------------------------------------------------------+ @end example -Performing a read in share mode means that we read the latest -available data, and set a shared mode lock on the rows we read. -If the latest data belongs to a yet uncommitted transaction of another -user, we will wait until that transaction commits. -A shared mode lock prevents others from updating or deleting -the row we have read. After we see that the above query returns -the parent @code{'Jones'}, we can safely add his child -to table @code{CHILD}, and commit our transaction. -This example shows how to implement referential -integrity in your application code. - -Let us look at another example: we have an integer counter field in -a table @code{CHILD_CODES} which we use to assign -a unique identifier to each child we add to table @code{CHILD}. -Obviously, using a consistent read or a shared mode read -to read the present value of the counter is not a good idea, since -then two users of the database may see the same value for the -counter, and we will get a duplicate key error when we add -the two children with the same identifier to the table. +@node SHOW CREATE TABLE, , SHOW GRANTS, SHOW +@subsection @code{SHOW CREATE TABLE} -In this case there are two good ways to implement the -reading and incrementing of the counter: (1) update the counter -first by incrementing it by 1 and only after that read it, -or (2) read the counter first with -a lock mode @code{FOR UPDATE}, and increment after that: +Shows a @code{CREATE TABLE} statement that will create the given table: @example -SELECT COUNTER_FIELD FROM CHILD_CODES FOR UPDATE; -UPDATE CHILD_CODES SET COUNTER_FIELD = COUNTER_FIELD + 1; -@end example - -A @code{SELECT ... FOR UPDATE} will read the latest -available data setting exclusive locks on each row it reads. -Thus it sets the same locks a searched SQL @code{UPDATE} would set -on the rows. +mysql> show create table t\G +*************************** 1. row *************************** + Table: t +Create Table: CREATE TABLE t ( + id int(11) default NULL auto_increment, + s char(60) default NULL, + PRIMARY KEY (id) +) TYPE=MyISAM -@node InnoDB Next-key locking, InnoDB Locks set, InnoDB locking reads, InnoDB transaction model -@subsubsection Next-key locking: avoiding the phantom problem +@end example -In row level locking InnoDB uses an algorithm called next-key locking. -InnoDB does the row level locking so that when it searches or -scans an index of a table, it sets shared or exclusive locks -on the index records in encounters. Thus the row level locks are -more precisely called index record locks. +@code{SHOW CREATE TABLE} will quote table and column names according to +@code{SQL_QUOTE_SHOW_CREATE} option. +@ref{SET OPTION, , @code{SET OPTION SQL_QUOTE_SHOW_CREATE}}. -The locks InnoDB sets on index records also affect the 'gap' -before that index record. If a user has a shared or exclusive -lock on record R in an index, then another user cannot insert -a new index record immediately before R in the index order. -This locking of gaps is done to prevent the so-called phantom -problem. Suppose I want to read and lock all children with identifier -bigger than 100 from table @code{CHILD}, -and update some field in the selected rows. +@findex EXPLAIN +@findex SELECT, optimizing +@node EXPLAIN, DESCRIBE, SHOW, Reference +@section @code{EXPLAIN} Syntax (Get Information About a @code{SELECT}) @example -SELECT * FROM CHILD WHERE ID > 100 FOR UPDATE; + EXPLAIN tbl_name +or EXPLAIN SELECT select_options @end example -Suppose there is an index on table @code{CHILD} on column -@code{ID}. Our query will scan that index starting from -the first record where @code{ID} is bigger than 100. -Now, if the locks set on the index records would not lock out -inserts made in the gaps, a new child might meanwhile be -inserted to the table. If now I in my transaction execute +@code{EXPLAIN tbl_name} is a synonym for @code{DESCRIBE tbl_name} or +@code{SHOW COLUMNS FROM tbl_name}. -@example -SELECT * FROM CHILD WHERE ID > 100 FOR UPDATE; -@end example +When you precede a @code{SELECT} statement with the keyword @code{EXPLAIN}, +@strong{MySQL} explains how it would process the @code{SELECT}, providing +information about how tables are joined and in which order. -again, I will see a new child in the result set the query returns. -This is against the isolation principle of transactions: -a transaction should be able to run so that the data -it has read does not change during the transaction. If we regard -a set of rows as a data item, then the new 'phantom' child would break -this isolation principle. +With the help of @code{EXPLAIN}, you can see when you must add indexes +to tables to get a faster @code{SELECT} that uses indexes to find the +records. You can also see if the optimizer joins the tables in an optimal +order. To force the optimizer to use a specific join order for a +@code{SELECT} statement, add a @code{STRAIGHT_JOIN} clause. -When InnoDB scans an index it can also lock the gap -after the last record in the index. Just that happens in the previous -example: the locks set by InnoDB will prevent any insert to -the table where @code{ID} would be bigger than 100. +For non-simple joins, @code{EXPLAIN} returns a row of information for each +table used in the @code{SELECT} statement. The tables are listed in the order +they would be read. @strong{MySQL} resolves all joins using a single-sweep +multi-join method. This means that @strong{MySQL} reads a row from the first +table, then finds a matching row in the second table, then in the third table +and so on. When all tables are processed, it outputs the selected columns and +backtracks through the table list until a table is found for which there are +more matching rows. The next row is read from this table and the process +continues with the next table. -You can use the next-key locking to implement a uniqueness -check in your application: if you read your data in share mode -and do not see a duplicate for a row you are going to insert, -then you can safely insert your row and know that the next-key -lock set on the successor of your row during the read will prevent -anyone meanwhile inserting a duplicate for your row. Thus the next-key -locking allows you to 'lock' the non-existence of something in your -table. +Output from @code{EXPLAIN} includes the following columns: -@node InnoDB Locks set, InnoDB Deadlock detection, InnoDB Next-key locking, InnoDB transaction model -@subsubsection Locks set by different SQL statements in InnoDB +@table @code +@item table +The table to which the row of output refers. -@itemize @bullet -@item -@code{SELECT ... FROM ...} : this is a consistent read, reading a -snapshot of the database and setting no locks. -@item -@code{SELECT ... FROM ... LOCK IN SHARE MODE} : sets shared next-key locks -on all index records the read encounters. -@item -@code{SELECT ... FROM ... FOR UPDATE} : sets exclusive next-key locks -on all index records the read encounters. -@item -@code{INSERT INTO ... VALUES (...)} : sets an exclusive lock -on the inserted row; note that this lock is not a next-key lock -and does not prevent other users from inserting to the gap before the -inserted row. If a duplicate key error occurs, sets a shared lock -on the duplicate index record. -@item -@code{INSERT INTO T SELECT ... FROM S WHERE ...} sets an exclusive -(non-next-key) lock on each row inserted into @code{T}. Does -the search on @code{S} as a consistent read, but sets shared next-key -locks on @code{S} if the @strong{MySQL} logging is on. InnoDB has to set -locks in the latter case because in roll-forward recovery from a -backup every SQL statement has to be executed in exactly the same -way as it was done originally. -@item -@code{CREATE TABLE ... SELECT ...} performs the @code{SELECT} -as a consistent read or with shared locks, like in the previous -item. -@item -@code{REPLACE} is done like an insert if there is no collision -on a unique key. Otherwise, an exclusive next-key lock is placed -on the row which has to be updated. -@item -@code{UPDATE ... SET ... WHERE ...} : sets an exclusive next-key -lock on every record the search encounters. -@item -@code{DELETE FROM ... WHERE ...} : sets an exclusive next-key -lock on every record the search encounters. -@item -@code{LOCK TABLES ... } : sets table locks. In the implementation -the @strong{MySQL} layer of code sets these locks. The automatic deadlock detection -of InnoDB cannot detect deadlocks where such table locks are involved: -see the next section below. See also section 13 'InnoDB restrictions' -about the following: since @strong{MySQL} does know about row level locks, -it is possible that you -get a table lock on a table where another user currently has row level -locks. But that does not put transaction integerity into danger. -@end itemize +@item type +The join type. Information about the various types is given below. -@node InnoDB Deadlock detection, , InnoDB Locks set, InnoDB transaction model -@subsubsection Deadlock detection and rollback +@item possible_keys +The @code{possible_keys} column indicates which indexes @strong{MySQL} +could use to find the rows in this table. Note that this column is +totally independent of the order of the tables. That means that some of +the keys in possible_keys may not be usable in practice with the +generated table order. -InnoDB automatically detects a deadlock of transactions and rolls -back the transaction whose lock request was the last one to build -a deadlock, that is, a cycle in the waits-for graph of transactions. -InnoDB cannot detect deadlocks where a lock set by a @strong{MySQL} -@code{LOCK TABLES} statement is involved, or if a lock set -in another table handler than InnoDB is involved. You have to resolve -these situations using @code{innodb_lock_wait_timeout} set in -@file{my.cnf}. +If this column is empty, there are no relevant indexes. In this case, +you may be able to improve the performance of your query by examining +the @code{WHERE} clause to see if it refers to some column or columns +that would be suitable for indexing. If so, create an appropriate index +and check the query with @code{EXPLAIN} again. @xref{ALTER TABLE}. -When InnoDB performs a complete rollback of a transaction, all the -locks of the transaction are released. However, if just a single SQL -statement is rolled back as a result of an error, some of the locks -set by the SQL statement may be preserved. This is because InnoDB -stores row locks in a format where it cannot afterwards know which was -set by which SQL statement. +To see what indexes a table has, use @code{SHOW INDEX FROM tbl_name}. -@subsection Performance tuning tips +@item key +The @code{key} column indicates the key that @strong{MySQL} actually +decided to use. The key is @code{NULL} if no index was chosen. If +@strong{MySQL} chooses the wrong index, you can probably force +@strong{MySQL} to use another index by using @code{myisamchk --analyze}, +@xref{myisamchk syntax}, or by using @code{USE INDEX/IGNORE INDEX}. +@xref{JOIN}. -@strong{1.} -If the Unix @file{top} or the Windows @file{Task Manager} shows that -the CPU usage percentage with your workload is less than 70 %, -your workload is probably -disk-bound. Maybe you are making too many transaction commits, or the -buffer pool is too small. -Making the buffer pool bigger can help, but do not set -it bigger than 80 % of physical memory. +@item key_len +The @code{key_len} column indicates the length of the key that +@strong{MySQL} decided to use. The length is @code{NULL} if the +@code{key} is @code{NULL}. Note that this tells us how many parts of a +multi-part key @strong{MySQL} will actually use. -@strong{2.} -Wrap several modifications into one transaction. InnoDB must -flush the log to disk at each transaction commit, if that transaction -made modifications to the database. Since the rotation speed of a disk -is typically -at most 167 revolutions/second, that constrains the number of commits -to the same 167/second if the disk does not fool the operating system. +@item ref +The @code{ref} column shows which columns or constants are used with the +@code{key} to select rows from the table. -@strong{3.} -If you can afford the loss of some latest committed transactions, you can -set the @file{my.cnf} parameter @code{innodb_flush_log_at_trx_commit} -to zero. InnoDB tries to flush the log anyway once in a second, -though the flush is not guaranteed. +@item rows +The @code{rows} column indicates the number of rows @strong{MySQL} +believes it must examine to execute the query. -@strong{4.} -Make your log files big, even as big as the buffer pool. When InnoDB -has written the log files full, it has to write the modified contents -of the buffer pool to disk in a checkpoint. Small log files will cause many -unnecessary disk writes. The drawback in big log files is that recovery -time will be longer. +@item Extra +This column contains additional information of how @strong{MySQL} will +resolve the query. Here is an explanation of the different text +strings that can be found in this column: -@strong{5.} -Also the log buffer should be quite big, say 8 MB. +@table @code +@item Distinct +@strong{MySQL} will not continue searching for more rows for the current row +combination after it has found the first matching row. -@strong{6.} (Relevant from 3.23.39 up.) -In some versions of Linux and Unix, flushing files to disk with the Unix -@code{fdatasync} and other similar methods is surprisingly slow. -The default method InnoDB uses is the @code{fdatasync} function. -If you are not satisfied with the database write performance, you may -try setting @code{innodb_flush_method} in @file{my.cnf} -to @code{O_DSYNC}, though O_DSYNC seems to be slower on most systems. -You can also try setting it to @code{littlesync}, which means that -InnoDB does not call the file flush for every write it does to a -file, but only -in log flush at transaction commits and data file flush at a checkpoint. -The drawback in @code{littlesync} is that if the operating system -crashes, you can easily end up with a half-written database page, -and you have to -do a recovery from a backup. With @code{nosync} you have even less safety: -InnoDB will only flush the database files to disk at database shutdown +@item Not exists +@strong{MySQL} was able to do a @code{LEFT JOIN} optimization on the +query and will not examine more rows in this table for the previous row +combination after it finds one row that matches the @code{LEFT JOIN} criteria. -@strong{7.} In importing data to InnoDB, make sure that MySQL does not have -@code{autocommit=1} on. Then every insert requires a log flush to disk. -Put before your plain SQL import file line +Here is an example for this: @example -set autocommit=0; +SELECT * FROM t1 LEFT JOIN t2 ON t1.id=t2.id WHERE t2.id IS NULL; @end example -and after it +Assume that @code{t2.id} is defined with @code{NOT NULL}. In this case +@strong{MySQL} will scan @code{t1} and look up the rows in @code{t2} +through @code{t1.id}. If @strong{MySQL} finds a matching row in +@code{t2}, it knows that @code{t2.id} can never be @code{NULL}, and will +not scan through the rest of the rows in @code{t2} that has the same +@code{id}. In other words, for each row in @code{t1}, @strong{MySQL} +only needs to do a single lookup in @code{t2}, independent of how many +matching rows there are in @code{t2}. -@example -commit; -@end example +@item @code{range checked for each record (index map: #)} +@strong{MySQL} didn't find a real good index to use. It will, instead, for +each row combination in the preceding tables, do a check on which index to +use (if any), and use this index to retrieve the rows from the table. This +isn't very fast but is faster than having to do a join without +an index. -If you use the @file{mysqldump} option @code{--opt}, you will get dump -files which are fast to import also to an InnoDB table, even without wrapping -them to the above @code{set autocommit=0; ... commit;} wrappers. +@item Using filesort +@strong{MySQL} will need to do an extra pass to find out how to retrieve +the rows in sorted order. The sort is done by going through all rows +according to the @code{join type} and storing the sort key + pointer to +the row for all rows that match the @code{WHERE}. Then the keys are +sorted. Finally the rows are retrieved in sorted order. -@strong{8.} -Beware of big rollbacks of mass inserts: InnoDB uses the insert buffer -to save disk i/o in inserts, but in a corresponding rollback no such -mechanism is used. A disk-bound rollback can take 30 times the time -of the corresponding insert. Killing the database process will not -help because the rollback will start again at the database startup. The -only way to get rid of a runaway rollback is to increase the buffer pool -so that the rollback becomes CPU-bound and runs fast, or delete the whole -InnoDB database. +@item Using index +The column information is retrieved from the table using only +information in the index tree without having to do an additional seek to +read the actual row. This can be done when all the used columns for +the table are part of the same index. -@strong{9.} -Beware also of other big disk-bound operations. -Use @code{DROP TABLE} -or @code{TRUNCATE} (from MySQL-4.0 up) to empty a table, not -@code{DELETE FROM yourtable}. +@item Using temporary +To resolve the query @strong{MySQL} will need to create a +temporary table to hold the result. This typically happens if you do an +@code{ORDER BY} on a different column set than you did a @code{GROUP +BY} on. -@strong{10.} -Use the multi-line @code{INSERT} to reduce -communication overhead between the client and the server if you need -to insert many rows: +@item Where used +A @code{WHERE} clause will be used to restrict which rows will be +matched against the next table or sent to the client. If you don't have +this information and the table is of type @code{ALL} or @code{index}, +you may have something wrong in your query (if you don't intend to +fetch/examine all rows from the table). +@end table -@example -INSERT INTO yourtable VALUES (1, 2), (5, 5); -@end example +If you want to get your queries as fast as possible, you should look out for +@code{Using filesort} and @code{Using temporary}. +@end table -This tip is of course valid for inserts into any table type, not just InnoDB. +The different join types are listed below, ordered from best to worst type: -@node Implementation, Table and index, InnoDB transaction model, InnoDB -@subsection Implementation of multiversioning +@cindex system table +@cindex tables, system +@table @code +@item system +The table has only one row (= system table). This is a special case of +the @code{const} join type. -Since InnoDB is a multiversioned database, it must keep information -of old versions of rows in the tablespace. This information is stored -in a data structure we call a rollback segment after an analogous -data structure in Oracle. +@cindex constant table +@cindex tables, constant +@item const +The table has at most one matching row, which will be read at the start +of the query. Because there is only one row, values from the column in +this row can be regarded as constants by the rest of the +optimizer. @code{const} tables are very fast as they are read only once! -InnoDB internally adds two fields to each row stored in the database. -A 6-byte field tells the transaction identifier for the last -transaction which inserted or updated the row. Also a deletion -is internally treated as an update where a special bit in the row -is set to mark it as deleted. Each row also contains a 7-byte -field called the roll pointer. The roll pointer points to an -undo log record written to the rollback segment. If the row was -updated, then the undo log record contains the information necessary -to rebuild the content of the row before it was updated. +@item eq_ref +One row will be read from this table for each combination of rows from +the previous tables. This is the best possible join type, other than the +@code{const} types. It is used when all parts of an index are used by +the join and the index is @code{UNIQUE} or a @code{PRIMARY KEY}. -InnoDB uses the information in the rollback segment to perform the -undo operations needed in a transaction rollback. It also uses the -information to build earlier versions of a row for a consistent -read. +@item ref +All rows with matching index values will be read from this table for each +combination of rows from the previous tables. @code{ref} is used if the join +uses only a leftmost prefix of the key, or if the key is not @code{UNIQUE} +or a @code{PRIMARY KEY} (in other words, if the join cannot select a single +row based on the key value). If the key that is used matches only a few rows, +this join type is good. -Undo logs in the rollback segment are divided into insert and update -undo logs. Insert undo logs are only needed in transaction rollback -and can be discarded as soon as the transaction commits. Update undo logs -are used also in consistent reads, and they can be discarded only after -there is no transaction present for which InnoDB has assigned -a snapshot that in a consistent read could need the information -in the update undo log to build an earlier version of a database -row. +@item range +Only rows that are in a given range will be retrieved, using an index to +select the rows. The @code{key} column indicates which index is used. +The @code{key_len} contains the longest key part that was used. +The @code{ref} column will be NULL for this type. -You must remember to commit your transactions regularly. Otherwise -InnoDB cannot discard data from the update undo logs, and the -rollback segment may grow too big, filling up your tablespace. +@item index +This is the same as @code{ALL}, except that only the index tree is +scanned. This is usually faster than @code{ALL}, as the index file is usually +smaller than the data file. -The physical size of an undo log record in the rollback segment -is typically smaller than the corresponding inserted or updated -row. You can use this information to calculate the space need -for your rollback segment. +@item ALL +A full table scan will be done for each combination of rows from the +previous tables. This is normally not good if the table is the first +table not marked @code{const}, and usually @strong{very} bad in all other +cases. You normally can avoid @code{ALL} by adding more indexes, so that +the row can be retrieved based on constant values or column values from +earlier tables. +@end table -In our multiversioning scheme a row is not physically removed from -the database immediately when you delete it with an SQL statement. -Only when InnoDB can discard the update undo log record written for -the deletion, it can also physically remove the corresponding row and -its index records from the database. This removal operation is -called a purge, and it is quite fast, usually taking the same order of -time as the SQL statement which did the deletion. +You can get a good indication of how good a join is by multiplying all values +in the @code{rows} column of the @code{EXPLAIN} output. This should tell you +roughly how many rows @strong{MySQL} must examine to execute the query. This +number is also used when you restrict queries with the @code{max_join_size} +variable. +@xref{Server parameters}. -@node Table and index, File space management, Implementation, InnoDB -@subsection Table and index structures +The following example shows how a @code{JOIN} can be optimized progressively +using the information provided by @code{EXPLAIN}. -Every InnoDB table has a special index called the clustered index -where the data of the rows is stored. If you define a -@code{PRIMARY KEY} on your table, then the index of the primary key -will be the clustered index. +Suppose you have the @code{SELECT} statement shown below, that you examine +using @code{EXPLAIN}: -If you do not define a primary key for -your table, InnoDB will internally generate a clustered index -where the rows are ordered by the row id InnoDB assigns -to the rows in such a table. The row id is a 6-byte field which -monotonically increases as new rows are inserted. Thus the rows -ordered by the row id will be physically in the insertion order. +@example +EXPLAIN SELECT tt.TicketNumber, tt.TimeIn, + tt.ProjectReference, tt.EstimatedShipDate, + tt.ActualShipDate, tt.ClientID, + tt.ServiceCodes, tt.RepetitiveID, + tt.CurrentProcess, tt.CurrentDPPerson, + tt.RecordVolume, tt.DPPrinted, et.COUNTRY, + et_1.COUNTRY, do.CUSTNAME + FROM tt, et, et AS et_1, do + WHERE tt.SubmitTime IS NULL + AND tt.ActualPC = et.EMPLOYID + AND tt.AssignedPC = et_1.EMPLOYID + AND tt.ClientID = do.CUSTNMBR; +@end example -Accessing a row through the clustered index is fast, because -the row data will be on the same page where the index search -leads us. In many databases the data is traditionally stored on a different -page from the index record. If a table is large, the clustered -index architecture often saves a disk i/o when compared to the -traditional solution. +For this example, assume that: -The records in non-clustered indexes (we also call them secondary indexes), -in InnoDB contain the primary key value for the row. InnoDB -uses this primary key value to search for the row from the clustered -index. Note that if the primary key is long, the secondary indexes -will use more space. +@itemize @bullet +@item +The columns being compared have been declared as follows: -@menu -* InnoDB physical structure:: -* InnoDB Insert buffering:: -* InnoDB Adaptive hash:: -* InnoDB Physical record:: -@end menu +@multitable @columnfractions .1 .2 .7 +@item @strong{Table} @tab @strong{Column} @tab @strong{Column type} +@item @code{tt} @tab @code{ActualPC} @tab @code{CHAR(10)} +@item @code{tt} @tab @code{AssignedPC} @tab @code{CHAR(10)} +@item @code{tt} @tab @code{ClientID} @tab @code{CHAR(10)} +@item @code{et} @tab @code{EMPLOYID} @tab @code{CHAR(15)} +@item @code{do} @tab @code{CUSTNMBR} @tab @code{CHAR(15)} +@end multitable -@node InnoDB physical structure, InnoDB Insert buffering, Table and index, Table and index -@subsubsection Physical structure of an index +@item +The tables have the indexes shown below: -All indexes in InnoDB are B-trees where the index records are -stored in the leaf pages of the tree. The default size of an index -page is 16 kB. When new records are inserted, InnoDB tries to -leave 1 / 16 of the page free for future insertions and updates -of the index records. +@multitable @columnfractions .1 .9 +@item @strong{Table} @tab @strong{Index} +@item @code{tt} @tab @code{ActualPC} +@item @code{tt} @tab @code{AssignedPC} +@item @code{tt} @tab @code{ClientID} +@item @code{et} @tab @code{EMPLOYID} (primary key) +@item @code{do} @tab @code{CUSTNMBR} (primary key) +@end multitable -If index records are inserted in a sequential (ascending or descending) -order, the resulting index pages will be about 15/16 full. -If records are inserted in a random order, then the pages will be -1/2 - 15/16 full. If the fillfactor of an index page drops below 1/2, -InnoDB will try to contract the index tree to free the page. +@item +The @code{tt.ActualPC} values aren't evenly distributed. +@end itemize -@node InnoDB Insert buffering, InnoDB Adaptive hash, InnoDB physical structure, Table and index -@subsubsection Insert buffering +Initially, before any optimizations have been performed, the @code{EXPLAIN} +statement produces the following information: -It is a common situation in a database application that the -primary key is a unique identifier and new rows are inserted in the -ascending order of the primary key. Thus the insertions to the -clustered index do not require random reads from a disk. +@example +table type possible_keys key key_len ref rows Extra +et ALL PRIMARY NULL NULL NULL 74 +do ALL PRIMARY NULL NULL NULL 2135 +et_1 ALL PRIMARY NULL NULL NULL 74 +tt ALL AssignedPC,ClientID,ActualPC NULL NULL NULL 3872 + range checked for each record (key map: 35) +@end example -On the other hand, secondary indexes are usually non-unique and -insertions happen in a relatively random order into secondary indexes. -This would cause a lot of random disk i/o's without a special mechanism -used in InnoDB. +Because @code{type} is @code{ALL} for each table, this output indicates that +@strong{MySQL} is doing a full join for all tables! This will take quite a +long time, as the product of the number of rows in each table must be +examined! For the case at hand, this is @code{74 * 2135 * 74 * 3872 = +45,268,558,720} rows. If the tables were bigger, you can only imagine how +long it would take. -If an index record should be inserted to a non-unique secondary index, -InnoDB checks if the secondary index page is already in the buffer -pool. If that is the case, InnoDB will do the insertion directly to -the index page. But, if the index page is not found from the buffer -pool, InnoDB inserts the record to a special insert buffer structure. -The insert buffer is kept so small that it entirely fits in the buffer -pool, and insertions can be made to it very fast. +One problem here is that @strong{MySQL} can't (yet) use indexes on columns +efficiently if they are declared differently. In this context, +@code{VARCHAR} and @code{CHAR} are the same unless they are declared as +different lengths. Because @code{tt.ActualPC} is declared as @code{CHAR(10)} +and @code{et.EMPLOYID} is declared as @code{CHAR(15)}, there is a length +mismatch. -The insert buffer is periodically merged to the secondary index -trees in the database. Often we can merge several insertions on the -same page in of the index tree, and hence save disk i/o's. -It has been measured that the insert buffer can speed up insertions -to a table up to 15 times. +To fix this disparity between column lengths, use @code{ALTER TABLE} to +lengthen @code{ActualPC} from 10 characters to 15 characters: -@node InnoDB Adaptive hash, InnoDB Physical record, InnoDB Insert buffering, Table and index -@subsubsection Adaptive hash indexes +@example +mysql> ALTER TABLE tt MODIFY ActualPC VARCHAR(15); +@end example -If a database fits almost entirely in main memory, then the fastest way -to perform queries on it is to use hash indexes. InnoDB has an -automatic mechanism which monitors index searches made to the indexes -defined for a table, and if InnoDB notices that queries could -benefit from building of a hash index, such an index is automatically -built. +Now @code{tt.ActualPC} and @code{et.EMPLOYID} are both @code{VARCHAR(15)}. +Executing the @code{EXPLAIN} statement again produces this result: -But note that the hash index is always built based on an existing -B-tree index on the table. InnoDB can build a hash index on a prefix -of any length of the key defined for the B-tree, depending on -what search pattern InnoDB observes on the B-tree index. -A hash index can be partial: it is not required that the whole -B-tree index is cached in the buffer pool. InnoDB will build -hash indexes on demand to those pages of the index which are -often accessed. +@example +table type possible_keys key key_len ref rows Extra +tt ALL AssignedPC,ClientID,ActualPC NULL NULL NULL 3872 where used +do ALL PRIMARY NULL NULL NULL 2135 + range checked for each record (key map: 1) +et_1 ALL PRIMARY NULL NULL NULL 74 + range checked for each record (key map: 1) +et eq_ref PRIMARY PRIMARY 15 tt.ActualPC 1 +@end example -In a sense, through the adaptive hash index mechanism InnoDB adapts itself -to ample main memory, coming closer to the architecture of main memory -databases. +This is not perfect, but is much better (the product of the @code{rows} +values is now less by a factor of 74). This version is executed in a couple +of seconds. -@node InnoDB Physical record, , InnoDB Adaptive hash, Table and index -@subsubsection Physical record structure +A second alteration can be made to eliminate the column length mismatches +for the @code{tt.AssignedPC = et_1.EMPLOYID} and @code{tt.ClientID = +do.CUSTNMBR} comparisons: -@itemize @bullet -@item -Each index record in InnoDB contains a header of 6 bytes. The header -is used to link consecutive records together, and also in the row level -locking. -@item -Records in the clustered index contain fields for all user-defined -columns. In addition, there is a 6-byte field for the transaction id -and a 7-byte field for the roll pointer. -@item -If the user has not defined a primary key for a table, then each clustered -index record contains also a 6-byte row id field. -@item -Each secondary index record contains also all the fields defined -for the clustered index key. -@item -A record contains also a pointer to each field of the record. -If the total length of the fields in a record is < 128 bytes, then -the pointer is 1 byte, else 2 bytes. -@end itemize +@example +mysql> ALTER TABLE tt MODIFY AssignedPC VARCHAR(15), + MODIFY ClientID VARCHAR(15); +@end example -@subsubsection How an auto-increment column works in InnoDB +Now @code{EXPLAIN} produces the output shown below: -After a database startup, when a user first does an insert to a -table @code{T} -where an auto-increment column has been defined, and the user does not provide -an explicit value for the column, then InnoDB executes @code{SELECT -MAX(auto-inc-column) FROM T}, and assigns that value incremented -by one to the the column and the auto-increment counter of the table. -We say that -the auto-increment counter for table @code{T} has been initialized. +@example +table type possible_keys key key_len ref rows Extra +et ALL PRIMARY NULL NULL NULL 74 +tt ref AssignedPC,ClientID,ActualPC ActualPC 15 et.EMPLOYID 52 where used +et_1 eq_ref PRIMARY PRIMARY 15 tt.AssignedPC 1 +do eq_ref PRIMARY PRIMARY 15 tt.ClientID 1 +@end example -InnoDB follows the same procedure in initializing the auto-increment counter -for a freshly created table. +This is almost as good as it can get. -Note that if the user specifies in an insert the value 0 to the auto-increment -column, then InnoDB treats the row like the value would not have been -specified. +The remaining problem is that, by default, @strong{MySQL} assumes that values +in the @code{tt.ActualPC} column are evenly distributed, and that isn't the +case for the @code{tt} table. Fortunately, it is easy to tell @strong{MySQL} +about this: -After the auto-increment counter has been initialized, if a user inserts -a row where he explicitly specifies the column value, and the value is bigger -than the current counter value, then the counter is set to the specified -column value. If the user does not explicitly specify a value, then InnoDB -increments the counter by one and assigns its new value to the column. +@example +shell> myisamchk --analyze PATH_TO_MYSQL_DATABASE/tt +shell> mysqladmin refresh +@end example -The auto-increment mechanism, when assigning values from the counter, -bypasses locking and transaction handling. Therefore you may also get -gaps in the number sequence if you roll back transactions which have -got numbers from the counter. +Now the join is perfect, and @code{EXPLAIN} produces this result: -The behavior of auto-increment is not defined if a user gives a negative -value to the column or if the value becomes bigger than the maximum -integer that can be stored in the specified integer type. +@example +table type possible_keys key key_len ref rows Extra +tt ALL AssignedPC,ClientID,ActualPC NULL NULL NULL 3872 where used +et eq_ref PRIMARY PRIMARY 15 tt.ActualPC 1 +et_1 eq_ref PRIMARY PRIMARY 15 tt.AssignedPC 1 +do eq_ref PRIMARY PRIMARY 15 tt.ClientID 1 +@end example -@node File space management, Error handling, Table and index, InnoDB -@subsection File space management and disk i/o +Note that the @code{rows} column in the output from @code{EXPLAIN} is an +educated guess from the @strong{MySQL} join optimizer. To optimize a +query, you should check if the numbers are even close to the truth. If not, +you may get better performance by using @code{STRAIGHT_JOIN} in your +@code{SELECT} statement and trying to list the tables in a different order in +the @code{FROM} clause. -@menu -* InnoDB Disk i/o:: -* InnoDB File space:: -* InnoDB File Defragmenting:: -@end menu +@findex DESC +@findex DESCRIBE +@node DESCRIBE, COMMIT, EXPLAIN, Reference +@section @code{DESCRIBE} Syntax (Get Information About Columns) -@node InnoDB Disk i/o, InnoDB File space, File space management, File space management -@subsubsection Disk i/o +@example +@{DESCRIBE | DESC@} tbl_name @{col_name | wild@} +@end example -In disk i/o InnoDB uses asynchronous i/o. On Windows NT -it uses the native asynchronous i/o provided by the operating system. -On Unix, InnoDB uses simulated asynchronous i/o built -into InnoDB: InnoDB creates a number of i/o threads to take care -of i/o operations, such as read-ahead. In a future version we will -add support for simulated aio on Windows NT and native aio on those -versions of Unix which have one. +@code{DESCRIBE} is a shortcut for @code{SHOW COLUMNS FROM}. +@xref{SHOW DATABASE INFO}. -On Windows NT InnoDB uses non-buffered i/o. That means that the disk -pages InnoDB reads or writes are not buffered in the operating system -file cache. This saves some memory bandwidth. +@code{DESCRIBE} provides information about a table's columns. @code{col_name} +may be a column name or a string containing the SQL @samp{%} and @samp{_} +wild-card characters. -You can also use a raw disk in InnoDB, though this has not been tested yet: -just define the raw disk in place of a data file in @file{my.cnf}. -You must give the exact size in bytes of the raw disk in @file{my.cnf}, -because at startup InnoDB checks that the size of the file -is the same as specified in the configuration file. Using a raw disk -you can on some versions of Unix perform non-buffered i/o. +If the column types are different than you expect them to be based on a +@code{CREATE TABLE} statement, note that @strong{MySQL} sometimes +changes column types. @xref{Silent column changes}. -There are two read-ahead heuristics in InnoDB: sequential read-ahead -and random read-ahead. In sequential read-ahead InnoDB notices that -the access pattern to a segment in the tablespace is sequential. -Then InnoDB will post in advance a batch of reads of database pages to the -i/o system. In random read-ahead InnoDB notices that some area -in a tablespace seems to be in the process of being -fully read into the buffer pool. Then InnoDB posts the remaining -reads to the i/o system. +@cindex Oracle compatibility +@cindex compatibility, with Oracle +This statement is provided for Oracle compatibility. -@node InnoDB File space, InnoDB File Defragmenting, InnoDB Disk i/o, File space management -@subsubsection File space management +The @code{SHOW} statement provides similar information. +@xref{SHOW, , @code{SHOW}}. -The data files you define in the configuration file form the tablespace -of InnoDB. The files are simply catenated to form the tablespace, -there is no striping in use. -Currently you cannot directly instruct where the space is allocated -for your tables, except by using the following fact: from a newly created -tablespace InnoDB will allocate space starting from the low end. +@findex BEGIN +@findex COMMIT +@findex ROLLBACK +@node COMMIT, LOCK TABLES, DESCRIBE, Reference +@section @code{BEGIN/COMMIT/ROLLBACK} Syntax -The tablespace consists of database pages whose default size is 16 kB. -The pages are grouped into extents of 64 consecutive pages. The 'files' inside -a tablespace are called segments in InnoDB. The name of the rollback -segment is somewhat misleading because it actually contains many -segments in the tablespace. +By default, @strong{MySQL} runs in @code{autocommit} mode. This means that +as soon as you execute an update, @strong{MySQL} will store the update on +disk. -For each index in InnoDB we allocate two segments: one is for non-leaf -nodes of the B-tree, the other is for the leaf nodes. The idea here is -to achieve better sequentiality for the leaf nodes, which contain the -data. +If you are using transactions safe tables (like @code{BDB}, +@code{InnoDB}, you can put @strong{MySQL} into +non-@code{autocommit} mode with the following command: -When a segment grows inside the tablespace, InnoDB allocates the -first 32 pages to it individually. After that InnoDB starts -to allocate whole extents to the segment. -InnoDB can add to a large segment up to 4 extents at a time to ensure -good sequentiality of data. +@example +SET AUTOCOMMIT=0 +@end example -Some pages in the tablespace contain bitmaps of other pages, and -therefore a few extents in an InnoDB tablespace cannot be -allocated to segments as a whole, but only as individual pages. +After this you must use @code{COMMIT} to store your changes to disk or +@code{ROLLBACK} if you want to ignore the changes you have made since +the beginning of your transaction. -When you issue a query @code{SHOW TABLE STATUS FROM ... LIKE ...} -to ask for available free space in the tablespace, InnoDB will -report you the space which is certainly usable in totally free extents -of the tablespace. InnoDB always reserves some extents for -clean-up and other internal purposes; these reserved extents are not -included in the free space. +If you want to switch from @code{AUTOCOMMIT} mode for one series of +statements, you can use the @code{BEGIN} or @code{BEGIN WORK} statement: -When you delete data from a table, InnoDB will contract the corresponding -B-tree indexes. It depends on the pattern of deletes if that frees -individual pages or extents to the tablespace, so that the freed -space is available for other users. Dropping a table or deleting -all rows from it is guaranteed to release the space to other users, -but remember that deleted rows can be physically removed only in a -purge operation after they are no longer needed in transaction rollback or -consistent read. +@example +BEGIN; +SELECT @@A:=SUM(salary) FROM table1 WHERE type=1; +UPDATE table2 SET summmary=@@A WHERE type=1; +COMMIT; +@end example -@node InnoDB File Defragmenting, , InnoDB File space, File space management -@subsubsection Defragmenting a table +Note that if you are using non-transaction-safe tables, the changes will be +stored at once, independent of the status of the @code{autocommit} mode. -If there are random insertions or deletions -in the indexes of a table, the indexes -may become fragmented. By fragmentation we mean that the physical ordering -of the index pages on the disk is not close to the alphabetical ordering -of the records on the pages, or that there are many unused pages in the -64-page blocks which were allocated to the index. +If you do a @code{ROLLBACK} when you have updated a non-transactional +table you will get an error (@code{ER_WARNING_NOT_COMPLETE_ROLLBACK}) as +a warning. All transactional safe tables will be restored but any +non-transactional table will not change. -It can speed up index scans if you -periodically use @code{mysqldump} to dump the table to -a text file, drop the table, and reload it from the dump. -Another way to do the defragmenting is to @code{ALTER} the table type to -@code{MyISAM} and back to @code{InnoDB} again. -Note that a @code{MyISAM} table must fit in a single file -on your operating system. +If you are using @code{BEGIN} or @code{SET AUTOCOMMIT=0}, you +should use the @strong{MySQL} binary log for backups instead of the +older update log. Transactions are stored in the binary log +in one chunk, upon @code{COMMIT}, to ensure that transactions which are +rolled back are not stored. @xref{Binary log}. -If the insertions to and index are always ascending and -records are deleted only from the end, then the the file space management -algorithm of InnoDB guarantees that fragmentation in the index will -not occur. +The following commands automatically end a transaction (as if you had done +a @code{COMMIT} before executing the command): -@node Error handling, InnoDB restrictions, File space management, InnoDB -@subsection Error handling +@multitable @columnfractions .33 .33 .33 +@item @code{ALTER TABLE} @tab @code{BEGIN} @tab @code{CREATE INDEX} +@item @code{DROP DATABASE} @tab @code{DROP TABLE} @tab @code{RENAME TABLE} +@item @code{TRUNCATE} +@end multitable -The error handling in InnoDB is not always the same as -specified in the ANSI SQL standards. According to the ANSI -standard, any error during an SQL statement should cause the -rollback of that statement. InnoDB sometimes rolls back only -part of the statement. -The following list specifies the error handling of InnoDB. +You can change the isolation level for transactions with +@code{SET TRANSACTION ISOLATION LEVEL ...}. @xref{SET TRANSACTION}. + +@findex LOCK TABLES +@findex UNLOCK TABLES +@node LOCK TABLES, SET OPTION, COMMIT, Reference +@section @code{LOCK TABLES/UNLOCK TABLES} Syntax + +@example +LOCK TABLES tbl_name [AS alias] @{READ | [READ LOCAL] | [LOW_PRIORITY] WRITE@} + [, tbl_name @{READ | [LOW_PRIORITY] WRITE@} ...] +... +UNLOCK TABLES +@end example + +@code{LOCK TABLES} locks tables for the current thread. @code{UNLOCK +TABLES} releases any locks held by the current thread. All tables that +are locked by the current thread are automatically unlocked when the +thread issues another @code{LOCK TABLES}, or when the connection to the +server is closed. + +The main reasons to use @code{LOCK TABLES} are: @itemize @bullet @item -If you run out of file space in the tablespace, -you will get the @strong{MySQL} @code{'Table is full'} error -and InnoDB rolls back the SQL statement. -@item -A transaction deadlock or a timeout in a lock wait will give -@code{'Table handler error 1000000'} and InnoDB rolls back -the SQL statement. +Emulate transactions with not transaction safe tables. @item -A duplicate key error only rolls back the insert of that particular row, -even in a statement like @code{INSERT INTO ... SELECT ...}. -This will probably change so that the SQL statement will be rolled -back if you have not specified the @code{IGNORE} option in your -statement. -@item -A 'row too long' error rolls back the SQL statement. -@item -Other errors are mostly detected by the @strong{MySQL} layer of code, and -they roll back the corresponding SQL statement. +To get more speed with @code{MyISAM} tables when inserting/updating data +over many statements. The main reason this gives more speed is that +@strong{MySQL} will not flush the key cache for the locked tables until +@code{UNLOCK TABLES} is called. @end itemize -@node InnoDB restrictions, InnoDB contact information, Error handling, InnoDB -@subsection Some restrictions on InnoDB tables +If a thread obtains a @code{READ} lock on a table, that thread (and all other +threads) can only read from the table. If a thread obtains a @code{WRITE} +lock on a table, then only the thread holding the lock can @code{READ} from +or @code{WRITE} to the table. Other threads are blocked. -@itemize @bullet +The difference between @code{READ LOCAL} and @code{READ} is that +@code{READ LOCAL} allows non-conflicting @code{INSERT} statements to +execute while the lock is held. This can't however be used if you are +going to manipulate the database files outside @strong{MySQL} while you +hold the lock. -@item @code{SHOW TABLE STATUS} does not give accurate statistics -on InnoDB tables, except for the physical size reserved by the table. -The row count is only a rough estimate used in SQL optimization. +Each thread waits (without timing out) until it obtains all the locks it has +requested. + +@code{WRITE} locks normally have higher priority than @code{READ} locks, to +ensure that updates are processed as soon as possible. This means that if one +thread obtains a @code{READ} lock and then another thread requests a +@code{WRITE} lock, subsequent @code{READ} lock requests will wait until the +@code{WRITE} thread has gotten the lock and released it. You can use +@code{LOW_PRIORITY WRITE} locks to allow other threads to obtain @code{READ} +locks while the thread is waiting for the @code{WRITE} lock. You should only +use @code{LOW_PRIORITY WRITE} locks if you are sure that there will +eventually be a time when no threads will have a @code{READ} lock. + +@code{LOCK TABLES} and @code{UNLOCK TABLES} both commits any active +transactions. + +When you use @code{LOCK TABLES}, you must lock all tables that you are +going to use and you must use the same alias that you are going to use +in your queries! If you are using a table multiple times in a query +(with aliases), you must get a lock for each alias! This policy ensures +that table locking is deadlock free and makes the locking code smaller, +simpler and much faster. + +Note that you should @strong{NOT} lock any tables that you are using with +@code{INSERT DELAYED}. This is because that in this case the @code{INSERT} +is done by a separate thread. + +Normally, you don't have to lock tables, as all single @code{UPDATE} statements +are atomic; no other thread can interfere with any other currently executing +SQL statement. There are a few cases when you would like to lock tables +anyway: +@itemize @bullet @item -If you try to create an unique index on a prefix of a column you will get an -error: +If you are going to run many operations on a bunch of tables, it's much +faster to lock the tables you are going to use. The downside is, of course, +that no other thread can update a @code{READ}-locked table and no other +thread can read a @code{WRITE}-locked table. +@item +If you are using a table handler in @strong{MySQL} that doesn't support +transactions, you must use @code{LOCK TABLES} if you want to ensure that +no other thread comes between a @code{SELECT} and an @code{UPDATE}. The +example shown below requires @code{LOCK TABLES} in order to execute safely: @example -CREATE TABLE T (A CHAR(20), B INT, UNIQUE (A(5))) TYPE = InnoDB; +mysql> LOCK TABLES trans READ, customer WRITE; +mysql> select sum(value) from trans where customer_id= some_id; +mysql> update customer set total_value=sum_from_previous_statement + where customer_id=some_id; +mysql> UNLOCK TABLES; @end example -If you create a non unique index on a prefix of a column, InnoDB will -create an index over the whole column. -@item -@code{INSERT DELAYED} is not supported for InnoDB tables. -@item -The @strong{MySQL} @code{LOCK TABLES} operation does not know of InnoDB -row level locks set in already completed SQL statements: this means that -you can get a table lock on a table even if there still exist transactions -of other users which have row level locks on the same table. Thus -your operations on the table may have to wait if they collide with -these locks of other users. Also a deadlock is possible. However, -this does not endanger transaction integrity, because the row level -locks set by InnoDB will always take care of the integrity. -Also, a table lock prevents other transactions from acquiring more -row level locks (in a conflicting lock mode) on the table. -@item -You cannot have a key on a @code{BLOB} or @code{TEXT} column. -@item -A table cannot contain more than 1000 columns. -@item -@code{DELETE FROM TABLE} does not regenerate the table but instead -deletes all rows, one by one, which is not that fast. In future versions -of @strong{MySQL} you can use @code{TRUNCATE} which is fast. -@item -Before dropping a database with InnoDB tables one has to drop -the individual InnoDB tables first. -@item -The default database page size in InnoDB is 16 kB. By recompiling the -code one can set it from 8 kB to 64 kB. -The maximun row length is slightly less than a half of a database page, -the row length also includes @code{BLOB} and @code{TEXT} type -columns. The restriction on the size of @code{BLOB} and -@code{TEXT} columns will be removed by June 2001 in a future version of -InnoDB. -@item -The maximum data or log file size is 2 GB or 4 GB depending on how large -files your operating system supports. Support for > 4 GB files will -be added to InnoDB in a future version. -@item -The maximum tablespace size is 4 billion database pages. This is also -the maximum size for a table. The minimum tablespace size is 10 MB. +Without @code{LOCK TABLES}, there is a chance that another thread might +insert a new row in the @code{trans} table between execution of the +@code{SELECT} and @code{UPDATE} statements. @end itemize -@node InnoDB contact information, , InnoDB restrictions, InnoDB -@subsection InnoDB contact information - -Contact information of Innobase Oy, producer of the InnoDB engine. -Website: @uref{http://www.innodb.com}. Email: -@email{Heikki.Tuuri@@innodb.com} +By using incremental updates (@code{UPDATE customer SET +value=value+new_value}) or the @code{LAST_INSERT_ID()} function, you can +avoid using @code{LOCK TABLES} in many cases. -@example -phone: 358-9-6969 3250 (office) 358-40-5617367 (mobile) -InnoDB Oy Inc. -World Trade Center Helsinki -Aleksanterinkatu 17 -P.O.Box 800 -00101 Helsinki -Finland -@end example +You can also solve some cases by using the user-level lock functions +@code{GET_LOCK()} and @code{RELEASE_LOCK()}. These locks are saved in a hash +table in the server and implemented with @code{pthread_mutex_lock()} and +@code{pthread_mutex_unlock()} for high speed. +@xref{Miscellaneous functions}. -@cindex tables, @code{BDB} -@cindex tables, @code{Berkeley DB} -@node BDB, , InnoDB, Table types -@section BDB or Berkeley_DB Tables +See @ref{Internal locking}, for more information on locking policy. -@menu -* BDB overview:: Overview of BDB Tables -* BDB install:: Installing BDB -* BDB start:: BDB startup options -* BDB characteristic:: Some characteristic of @code{BDB} tables: -* BDB TODO:: Some things we need to fix for BDB in the near future: -* BDB portability:: Operating systems supported by @strong{BDB} -* BDB errors:: Errors You May Get When Using BDB Tables -@end menu +You can also lock all tables in all databases with read locks with the +@code{FLUSH TABLES WITH READ LOCK} command. @xref{FLUSH}. This is very +convenient way to get backups if you have a file system, like Veritas, +that can take snapshots in time. -@node BDB overview, BDB install, BDB, BDB -@subsection Overview of BDB Tables +@strong{NOTE}: @code{LOCK TABLES} is not transaction-safe and will +automatically commit any active transactions before attempting to lock the +tables. -Support for BDB tables is included in the @strong{MySQL} source distribution -starting from Version 3.23.34 and is activated in the @strong{MySQL}-Max -binary. +@findex SET OPTION +@node SET OPTION, SET TRANSACTION, LOCK TABLES, Reference +@section @code{SET} Syntax -BerkeleyDB, available at @uref{http://www.sleepycat.com/} has provided -@strong{MySQL} with a transactional table handler. By using BerkeleyDB -tables, your tables may have a greater chance of surviving crashes, and also -provides @code{COMMIT} and @code{ROLLBACK} on transactions. The -@strong{MySQL} source distribution comes with a BDB distribution that has a -couple of small patches to make it work more smoothly with @strong{MySQL}. -You can't use a non-patched @code{BDB} version with @strong{MySQL}. +@example +SET [OPTION] SQL_VALUE_OPTION= value, ... +@end example -We at @strong{MySQL AB} are working in close cooperation with Sleepycat to -keep the quality of the @strong{MySQL}/BDB interface high. +@code{SET OPTION} sets various options that affect the operation of the +server or your client. Any option you set remains in effect until the +current session ends, or until you set the option to a different value. -When it comes to supporting BDB tables, we are committed to help our -users to locate the problem and help creating a reproducable test case -for any problems involving BDB tables. Any such test case will be -forwarded to Sleepycat who in turn will help us find and fix the -problem. As this is a two stage operation, any problems with BDB tables -may take a little longer for us to fix than for other table handlers. -However, as the BerkeleyDB code itself has been used by many other -applications than @strong{MySQL}, we don't envision any big problems with -this. @xref{Table handler support}. +@table @code +@item CHARACTER SET character_set_name | DEFAULT +This maps all strings from and to the client with the given mapping. +Currently the only option for @code{character_set_name} is +@code{cp1251_koi8}, but you can easily add new mappings by editing the +@file{sql/convert.cc} file in the @strong{MySQL} source distribution. The +default mapping can be restored by using a @code{character_set_name} value of +@code{DEFAULT}. -@node BDB install, BDB start, BDB overview, BDB -@subsection Installing BDB +Note that the syntax for setting the @code{CHARACTER SET} option differs +from the syntax for setting the other options. -If you have downloaded a binary version of @strong{MySQL} that includes -support for BerkeleyDB, simply follow the instructions for installing a -binary version of @strong{MySQL}. -@xref{Installing binary}. @xref{mysqld-max, , @code{mysqld-max}}. +@item PASSWORD = PASSWORD('some password') +@cindex passwords, setting +Set the password for the current user. Any non-anonymous user can change his +own password! -To compile @strong{MySQL} with Berkeley DB support, download @strong{MySQL} -Version 3.23.34 or newer and configure @code{MySQL} with the -@code{--with-berkeley-db} option. @xref{Installing source}. +@item PASSWORD FOR user = PASSWORD('some password') +Set the password for a specific user on the current server host. Only a user +with access to the @code{mysql} database can do this. The user should be +given in @code{user@@hostname} format, where @code{user} and @code{hostname} +are exactly as they are listed in the @code{User} and @code{Host} columns of +the @code{mysql.user} table entry. For example, if you had an entry with +@code{User} and @code{Host} fields of @code{'bob'} and @code{'%.loc.gov'}, +you would write: @example -cd /path/to/source/of/mysql-3.23.34 -./configure --with-berkeley-db -@end example - -Please refer to the manual provided with the @code{BDB} distribution for -more updated information. +mysql> SET PASSWORD FOR bob@@"%.loc.gov" = PASSWORD("newpass"); -Even though Berkeley DB is in itself very tested and reliable, -the @strong{MySQL} interface is still considered beta quality. -We are actively improving and optimizing it to get it stable very -soon. +or -@node BDB start, BDB characteristic, BDB install, BDB -@subsection BDB startup options +mysql> UPDATE mysql.user SET password=PASSWORD("newpass") where user="bob' and host="%.loc.gov"; +@end example -If you are running with @code{AUTOCOMMIT=0} then your changes in @code{BDB} -tables will not be updated until you execute @code{COMMIT}. Instead of commit -you can execute @code{ROLLBACK} to forget your changes. @xref{COMMIT}. +@item SQL_AUTO_IS_NULL = 0 | 1 +If set to @code{1} (default) then one can find the last inserted row +for a table with an auto_increment row with the following construct: +@code{WHERE auto_increment_column IS NULL}. This is used by some +ODBC programs like Access. -If you are running with @code{AUTOCOMMIT=1} (the default), your changes -will be committed immediately. You can start an extended transaction with -the @code{BEGIN WORK} SQL command, after which your changes will not be -committed until you execute @code{COMMIT} (or decide to @code{ROLLBACK} -the changes). +@item AUTOCOMMIT= 0 | 1 +If set to @code{1} all changes to a table will be done at once. To start +a multi-command transaction, you have to use the @code{BEGIN} +statement. @xref{COMMIT}. If set to @code{0} you have to use @code{COMMIT} / +@code{ROLLBACK} to accept/revoke that transaction. @xref{COMMIT}. Note +that when you change from not @code{AUTOCOMMIT} mode to +@code{AUTOCOMMIT} mode, @strong{MySQL} will do an automatic +@code{COMMIT} on any open transactions. -The following options to @code{mysqld} can be used to change the behavior of -BDB tables: +@item SQL_BIG_TABLES = 0 | 1 +@cindex table is full +If set to @code{1}, all temporary tables are stored on disk rather than in +memory. This will be a little slower, but you will not get the error +@code{The table tbl_name is full} for big @code{SELECT} operations that +require a large temporary table. The default value for a new connection is +@code{0} (that is, use in-memory temporary tables). -@multitable @columnfractions .30 .70 -@item @strong{Option} @tab @strong{Meaning} -@item @code{--bdb-home=directory} @tab Base directory for BDB tables. This should be the same directory you use for --datadir. -@item @code{--bdb-lock-detect=#} @tab Berkeley lock detect. One of (DEFAULT, OLDEST, RANDOM, or YOUNGEST). -@item @code{--bdb-logdir=directory} @tab Berkeley DB log file directory. -@item @code{--bdb-no-sync} @tab Don't synchronously flush logs. -@item @code{--bdb-no-recover} @tab Don't start Berkeley DB in recover mode. -@item @code{--bdb-shared-data} @tab Start Berkeley DB in multi-process mode (Don't use @code{DB_PRIVATE} when initializing Berkeley DB) -@item @code{--bdb-tmpdir=directory} @tab Berkeley DB tempfile name. -@item @code{--skip-bdb} @tab Don't use berkeley db. -@item @code{-O bdb_max_lock=1000} @tab Set the maximum number of locks possible. @xref{SHOW VARIABLES}. -@end multitable +@item SQL_BIG_SELECTS = 0 | 1 +If set to @code{0}, @strong{MySQL} will abort if a @code{SELECT} is attempted +that probably will take a very long time. This is useful when an inadvisable +@code{WHERE} statement has been issued. A big query is defined as a +@code{SELECT} that probably will have to examine more than +@code{max_join_size} rows. The default value for a new connection is +@code{1} (which will allow all @code{SELECT} statements). -If you use @code{--skip-bdb}, @strong{MySQL} will not initialize the -Berkeley DB library and this will save a lot of memory. Of course, -you cannot use @code{BDB} tables if you are using this option. +@item SQL_BUFFER_RESULT = 0 | 1 +@code{SQL_BUFFER_RESULT} will force the result from @code{SELECT}'s +to be put into a temporary table. This will help @strong{MySQL} free the +table locks early and will help in cases where it takes a long time to +send the result set to the client. -Normally you should start @code{mysqld} without @code{--bdb-no-recover} if you -intend to use BDB tables. This may, however, give you problems when you -try to start @code{mysqld} if the BDB log files are corrupted. @xref{Starting -server}. +@item SQL_LOW_PRIORITY_UPDATES = 0 | 1 +If set to @code{1}, all @code{INSERT}, @code{UPDATE}, @code{DELETE}, and +and @code{LOCK TABLE WRITE} statements wait until there is no pending +@code{SELECT} or @code{LOCK TABLE READ} on the affected table. -With @code{bdb_max_lock} you can specify the maximum number of locks -(10000 by default) you can have active on a BDB table. You should -increase this if you get errors of type @code{bdb: Lock table is out of -available locks} or @code{Got error 12 from ...} when you have do long -transactions or when @code{mysqld} has to examine a lot of rows to -calculate the query. +@item SQL_MAX_JOIN_SIZE = value | DEFAULT +Don't allow @code{SELECT}s that will probably need to examine more than +@code{value} row combinations. By setting this value, you can catch +@code{SELECT}s where keys are not used properly and that would probably +take a long time. Setting this to a value other than @code{DEFAULT} will reset +the @code{SQL_BIG_SELECTS} flag. If you set the @code{SQL_BIG_SELECTS} +flag again, the @code{SQL_MAX_JOIN_SIZE} variable will be ignored. +You can set a default value for this variable by starting @code{mysqld} with +@code{-O max_join_size=#}. -You may also want to change @code{binlog_cache_size} and -@code{max_binlog_cache_size} if you are using big multi-line transactions. -@xref{COMMIT}. +@item SQL_SAFE_MODE = 0 | 1 +If set to @code{1}, @strong{MySQL} will abort if an @code{UPDATE} or +@code{DELETE} is attempted that doesn't use a key or @code{LIMIT} in the +@code{WHERE} clause. This makes it possible to catch wrong updates +when creating SQL commands by hand. -@node BDB characteristic, BDB TODO, BDB start, BDB -@subsection Some characteristic of @code{BDB} tables: +@item SQL_SELECT_LIMIT = value | DEFAULT +The maximum number of records to return from @code{SELECT} statements. If +a @code{SELECT} has a @code{LIMIT} clause, the @code{LIMIT} takes precedence +over the value of @code{SQL_SELECT_LIMIT}. The default value for a new +connection is ``unlimited.'' If you have changed the limit, the default value +can be restored by using a @code{SQL_SELECT_LIMIT} value of @code{DEFAULT}. -@itemize @bullet -@item -To be able to rollback transactions BDB maintain log files. For maximum -performance you should place these on another disk than your databases -by using the @code{--bdb_log_dir} options. -@item -@strong{MySQL} performs a checkpoint each time a new BDB log -file is started, and removes any log files that are not needed for -current transactions. One can also run @code{FLUSH LOGS} at any time -to checkpoint the Berkeley DB tables. +@item SQL_LOG_OFF = 0 | 1 +If set to @code{1}, no logging will be done to the standard log for this +client, if the client has the @strong{process} privilege. This does not +affect the update log! -For disaster recovery, one should use table backups plus -@strong{MySQL}'s binary log. @xref{Backup}. +@item SQL_LOG_UPDATE = 0 | 1 +If set to @code{0}, no logging will be done to the update log for the client, +if the client has the @strong{process} privilege. This does not affect the +standard log! -@strong{Warning}: If you delete old log files that are in use, BDB will -not be able to do recovery at all and you may loose data if something -goes wrong. -@item -@strong{MySQL} requires a @code{PRIMARY KEY} in each BDB table to be -able to refer to previously read rows. If you don't create one, -@strong{MySQL} will create an maintain a hidden @code{PRIMARY KEY} for -you. The hidden key has a length of 5 bytes and is incremented for each -insert attempt. -@item -If all columns you access in a @code{BDB} table are part of the same index or -part of the primary key, then @strong{MySQL} can execute the query -without having to access the actual row. In a @code{MyISAM} table the -above holds only if the columns are part of the same index. -@item -The @code{PRIMARY KEY} will be faster than any other key, as the -@code{PRIMARY KEY} is stored together with the row data. As the other keys are -stored as the key data + the @code{PRIMARY KEY}, it's important to keep the -@code{PRIMARY KEY} as short as possible to save disk and get better speed. -@item -@code{LOCK TABLES} works on @code{BDB} tables as with other tables. If -you don't use @code{LOCK TABLE}, @strong{MYSQL} will issue an internal -multiple-write lock on the table to ensure that the table will be -properly locked if another thread issues a table lock. -@item -Internal locking in @code{BDB} tables is done on page level. -@item -@code{SELECT COUNT(*) FROM table_name} is slow as @code{BDB} tables doesn't -maintain a count of the number of rows in the table. -@item -Scanning is slower than with @code{MyISAM} tables as one has data in BDB -tables stored in B-trees and not in a separate data file. -@item -The application must always be prepared to handle cases where -any change of a @code{BDB} table may make an automatic rollback and any -read may fail with a deadlock error. -@item -Keys are not compressed to previous keys as with ISAM or MyISAM -tables. In other words, the key information will take a little more -space in @code{BDB} tables compared to MyISAM tables which don't use -@code{PACK_KEYS=0}. -@item -There is often holes in the BDB table to allow you to insert new rows in -the middle of the key tree. This makes BDB tables somewhat larger than -MyISAM tables. -@item -The optimizer needs to know an approximation of the number of rows in -the table. @strong{MySQL} solves this by counting inserts and -maintaining this in a separate segment in each BDB table. If you don't -do a lot of @code{DELETE} or @code{ROLLBACK}:s this number should be -accurate enough for the @strong{MySQL} optimizer, but as @strong{MySQL} -only store the number on close, it may be wrong if @strong{MySQL} dies -unexpectedly. It should not be fatal even if this number is not 100 % -correct. One can update the number of rows by executing @code{ANALYZE -TABLE} or @code{OPTIMIZE TABLE}. @xref{ANALYZE TABLE} . @xref{OPTIMIZE -TABLE}. -@item -If you get full disk with a @code{BDB} table, you will get an error -(probably error 28) and the transaction should roll back. This is in -contrast with @code{MyISAM} and @code{ISAM} tables where @code{mysqld} will -wait for enough free disk before continuing. -@end itemize +@item SQL_QUOTE_SHOW_CREATE = 0 | 1 +If set to @code{1}, @code{SHOW CREATE TABLE} will quote +table and column names. This is @strong{on} by default, +for replication of tables with fancy column names to work. +@ref{SHOW CREATE TABLE, , @code{SHOW CREATE TABLE}}. -@node BDB TODO, BDB portability, BDB characteristic, BDB -@subsection Some things we need to fix for BDB in the near future: +@item TIMESTAMP = timestamp_value | DEFAULT +Set the time for this client. This is used to get the original timestamp if +you use the update log to restore rows. @code{timestamp_value} should be a +UNIX Epoch timestamp, not a @strong{MySQL} timestamp. -@itemize @bullet -@item -It's very slow to open many BDB tables at the same time. If you are -going to use BDB tables, you should not have a very big table cache (> -256 ?) and you should use @code{--no-auto-rehash} with the @code{mysql} -client. We plan to partly fix this in 4.0. -@item -@code{SHOW TABLE STATUS} doesn't yet provide that much information for BDB -tables. -@item -Optimize performance. -@item -Change to not use page locks at all when we are scanning tables. -@end itemize +@item LAST_INSERT_ID = # +Set the value to be returned from @code{LAST_INSERT_ID()}. This is stored in +the update log when you use @code{LAST_INSERT_ID()} in a command that updates +a table. -@node BDB portability, BDB errors, BDB TODO, BDB -@subsection Operating systems supported by @strong{BDB} +@item INSERT_ID = # +Set the value to be used by the following @code{INSERT} or @code{ALTER TABLE} +command when inserting an @code{AUTO_INCREMENT} value. This is mainly used +with the update log. +@end table -If you after having built @strong{MySQL} with support for BDB tables get -the following error in the log file when you start @code{mysqld}: +@findex ISOLATION LEVEL +@node SET TRANSACTION, GRANT, SET OPTION, Reference +@section @code{SET TRANSACTION} Syntax @example -bdb: architecture lacks fast mutexes: applications cannot be threaded -Can't init dtabases +SET [GLOBAL | SESSION] TRANSACTION ISOLATION LEVEL +[READ UNCOMMITTED | READ COMMITTED | REPEATABLE READ | SERIALIZABLE] @end example -This means that @code{BDB} tables are not supported for your architecture. -In this case you have to rebuild @strong{MySQL} without BDB table support. +Sets the transaction isolation level for the global, whole session or +the next transaction. -NOTE: The following list is not complete; We will update this as we get -more information about this. +The default behavior is to set the isolation level for the next (not started) +transaction. -Currently we know that BDB tables works with the following operating -system. +If you set the @code{GLOBAL} privilege it will affect all new created threads. +You will need the @code{PROCESS} privilege to do do this. -@itemize @bullet -@item -Linux 2.x intel -@item -Solaris sparc -@item -SCO OpenServer -@item -SCO UnixWare 7.0.1 -@end itemize +Setting the @code{SESSION} privilege will affect the following and all +future transactions. -It doesn't work with the following operating systems: +You can set the default isolation level for @code{mysqld} with +@code{--transaction-isolation=...}. @xref{Command-line options}. -@itemize @bullet -@item -Linux 2.x Alpha -@item -Max OS X -@end itemize +@cindex privileges, granting +@cindex privileges, revoking +@cindex global privileges +@cindex revoking, privleges +@cindex granting, privleges -@node BDB errors, , BDB portability, BDB -@subsection Errors You May Get When Using BDB Tables - -@itemize @bullet -@item -If you get the following error in the @code{hostname.err log} when -starting @code{mysqld}: +@findex GRANT +@findex REVOKE +@node GRANT, CREATE INDEX, SET TRANSACTION, Reference +@section @code{GRANT} and @code{REVOKE} Syntax @example -bdb: Ignoring log file: .../log.XXXXXXXXXX: unsupported log version # -@end example -it means that the new @code{BDB} version doesn't support the old log -file format. In this case you have to delete all @code{BDB} log BDB -from your database directory (the files that has the format -@code{log.XXXXXXXXXX} ) and restart @code{mysqld}. We would also -recommend you to do a @code{mysqldump --opt} of your old @code{BDB} -tables, delete the old table and restore the dump. -@item -If you are running in not @code{auto_commit} mode and delete a table you -are using by another thread you may get the following error messages in -the @strong{MySQL} error file: +GRANT priv_type [(column_list)] [, priv_type [(column_list)] ...] + ON @{tbl_name | * | *.* | db_name.*@} + TO user_name [IDENTIFIED BY 'password'] + [, user_name [IDENTIFIED BY 'password'] ...] + [WITH GRANT OPTION] -@example -001119 23:43:56 bdb: Missing log fileid entry -001119 23:43:56 bdb: txn_abort: Log undo failed for LSN: 1 3644744: Invalid +REVOKE priv_type [(column_list)] [, priv_type [(column_list)] ...] + ON @{tbl_name | * | *.* | db_name.*@} + FROM user_name [, user_name ...] @end example -This is not fatal but we don't recommend that you delete tables if you are -not in @code{auto_commit} mode, until this problem is fixed (the fix is -not trivial). -@end itemize +@code{GRANT} is implemented in @strong{MySQL} Version 3.22.11 or later. For +earlier @strong{MySQL} versions, the @code{GRANT} statement does nothing. +The @code{GRANT} and @code{REVOKE} commands allow system administrators +to create users and grant and revoke rights to @strong{MySQL} users at +four privilege levels: -@cindex tutorial -@cindex terminal monitor, defined -@cindex monitor, terminal -@cindex options, provided by MySQL -@node Tutorial, Server, Table types, Top -@chapter MySQL Tutorial +@table @strong +@item Global level +Global privileges apply to all databases on a given server. These privileges +are stored in the @code{mysql.user} table. -@menu -* Connecting-disconnecting:: Connecting to and disconnecting from the server -* Entering queries:: Entering queries -* Database use:: Creating and using a database -* Getting information:: Getting information about databases and tables -* Examples:: Examples -* Batch mode:: Using @code{mysql} in batch mode -* Twin:: Queries from twin project -@end menu +@item Database level +Database privileges apply to all tables in a given database. These privileges +are stored in the @code{mysql.db} and @code{mysql.host} tables. -This chapter provides a tutorial introduction to @strong{MySQL} by showing -how to use the @code{mysql} client program to create and use a simple -database. @code{mysql} (sometimes referred to as the ``terminal monitor'' or -just ``monitor'') is an interactive program that allows you to connect to a -@strong{MySQL} server, run queries, and view the results. @code{mysql} may -also be used in batch mode: you place your queries in a file beforehand, then -tell @code{mysql} to execute the contents of the file. Both ways of using -@code{mysql} are covered here. +@item Table level +Table privileges apply to all columns in a given table. These privileges are +stored in the @code{mysql.tables_priv} table. -To see a list of options provided by @code{mysql}, invoke it with -the @code{--help} option: +@item Column level +Column privileges apply to single columns in a given table. These privileges are +stored in the @code{mysql.columns_priv} table. +@end table + +If you give a grant for a users that doesn't exists, that user is created. +For examples of how @code{GRANT} works, see @ref{Adding users}. + +For the @code{GRANT} and @code{REVOKE} statements, @code{priv_type} may be +specified as any of the following: @example -shell> mysql --help +ALL PRIVILEGES FILE RELOAD +ALTER INDEX SELECT +CREATE INSERT SHUTDOWN +DELETE PROCESS UPDATE +DROP REFERENCES USAGE @end example -This chapter assumes that @code{mysql} is installed on your machine and that -a @strong{MySQL} server is available to which you can connect. If this is -not true, contact your @strong{MySQL} administrator. (If @emph{you} are the -administrator, you will need to consult other sections of this manual.) +@code{ALL} is a synonym for @code{ALL PRIVILEGES}. @code{REFERENCES} is not +yet implemented. @code{USAGE} is currently a synonym for ``no privileges.'' +It can be used when you want to create a user that has no privileges. -This chapter describes the entire process of setting up and using a -database. If you are interested only in accessing an already-existing -database, you may want to skip over the sections that describe how to -create the database and the tables it contains. +To revoke the @strong{grant} privilege from a user, use a @code{priv_type} +value of @code{GRANT OPTION}: -Because this chapter is tutorial in nature, many details are necessarily left -out. Consult the relevant sections of the manual for more -information on the topics covered here. +@example +REVOKE GRANT OPTION ON ... FROM ...; +@end example -@cindex connecting, to the server -@cindex disconnecting, from the server -@cindex server, connecting -@cindex server, disconnecting -@node Connecting-disconnecting, Entering queries, Tutorial, Tutorial -@section Connecting to and Disconnecting from the Server +The only @code{priv_type} values you can specify for a table are @code{SELECT}, +@code{INSERT}, @code{UPDATE}, @code{DELETE}, @code{CREATE}, @code{DROP}, +@code{GRANT}, @code{INDEX}, and @code{ALTER}. -To connect to the server, you'll usually need to provide a @strong{MySQL} -user name when you invoke @code{mysql} and, most likely, a password. If the -server runs on a machine other than the one where you log in, you'll also -need to specify a hostname. Contact your administrator to find out what -connection parameters you should use to connect (that is, what host, user name, -and password to use). Once you know the proper parameters, you should be -able to connect like this: +The only @code{priv_type} values you can specify for a column (that is, when +you use a @code{column_list} clause) are @code{SELECT}, @code{INSERT}, and +@code{UPDATE}. + +You can set global privileges by using @code{ON *.*} syntax. You can set +database privileges by using @code{ON db_name.*} syntax. If you specify +@code{ON *} and you have a current database, you will set the privileges for +that database. (@strong{WARNING:} If you specify @code{ON *} and you +@emph{don't} have a current database, you will affect the global privileges!) + +In order to accommodate granting rights to users from arbitrary hosts, +@strong{MySQL} supports specifying the @code{user_name} value in the form +@code{user@@host}. If you want to specify a @code{user} string +containing special characters (such as @samp{-}), or a @code{host} string +containing special characters or wild-card characters (such as @samp{%}), you +can quote the user or host name (for example, @code{'test-user'@@'test-hostname'}). + +You can specify wild cards in the hostname. For example, +@code{user@@"%.loc.gov"} applies to @code{user} for any host in the +@code{loc.gov} domain, and @code{user@@"144.155.166.%"} applies to @code{user} +for any host in the @code{144.155.166} class C subnet. + +The simple form @code{user} is a synonym for @code{user@@"%"}. +@strong{NOTE:} If you allow anonymous users to connect to the @strong{MySQL} +server (which is the default), you should also add all local users as +@code{user@@localhost} because otherwise the anonymous user entry for the +local host in the @code{mysql.user} table will be used when the user tries to +log into the @strong{MySQL} server from the local machine! Anonymous users +are defined by inserting entries with @code{User=''} into the +@code{mysql.user} table. You can verify if this applies to you by executing +this query: @example -shell> mysql -h host -u user -p -Enter password: ******** +mysql> SELECT Host,User FROM mysql.user WHERE User=''; @end example -The @code{********} represents your password; enter it when @code{mysql} -displays the @code{Enter password:} prompt. +For the moment, @code{GRANT} only supports host, table, database, and +column names up to 60 characters long. A user name can be up to 16 +characters. -If that works, you should see some introductory information followed by a -@code{mysql>} prompt: +The privileges for a table or column are formed from the +logical OR of the privileges at each of the four privilege +levels. For example, if the @code{mysql.user} table specifies that a +user has a global @strong{select} privilege, this can't be denied by an +entry at the database, table, or column level. +The privileges for a column can be calculated as follows: @example -shell> mysql -h host -u user -p -Enter password: ******** -Welcome to the MySQL monitor. Commands end with ; or \g. -Your MySQL connection id is 459 to server version: 3.22.20a-log +global privileges +OR (database privileges AND host privileges) +OR table privileges +OR column privileges +@end example -Type 'help' for help. +In most cases, you grant rights to a user at only one of the privilege +levels, so life isn't normally as complicated as above. The details of the +privilege-checking procedure are presented in +@ref{Privilege system}. -mysql> -@end example +If you grant privileges for a user/hostname combination that does not exist +in the @code{mysql.user} table, an entry is added and remains there until +deleted with a @code{DELETE} command. In other words, @code{GRANT} may +create @code{user} table entries, but @code{REVOKE} will not remove them; +you must do that explicitly using @code{DELETE}. -The prompt tells you that @code{mysql} is ready for you to enter commands. +@cindex passwords, setting +In @strong{MySQL} Version 3.22.12 or later, +if a new user is created or if you have global grant privileges, the user's +password will be set to the password specified by the @code{IDENTIFIED BY} +clause, if one is given. If the user already had a password, it is replaced +by the new one. -Some @strong{MySQL} installations allow users to connect as the anonymous -(unnamed) user to the server running on the local host. If this is the case -on your machine, you should be able to connect to that server by invoking -@code{mysql} without any options: +@strong{WARNING:} If you create a new user but do not specify an +@code{IDENTIFIED BY} clause, the user has no password. This is insecure. -@example -shell> mysql -@end example +Passwords can also be set with the @code{SET PASSWORD} command. +@xref{SET OPTION, , @code{SET OPTION}}. -After you have connected successfully, you can disconnect any time by typing -@code{QUIT} at the @code{mysql>} prompt: +If you grant privileges for a database, an entry in the @code{mysql.db} +table is created if needed. When all privileges for the database have been +removed with @code{REVOKE}, this entry is deleted. -@example -mysql> QUIT -Bye -@end example +If a user doesn't have any privileges on a table, the table is not displayed +when the user requests a list of tables (for example, with a @code{SHOW TABLES} +statement). -You can also disconnect by pressing Control-D. +The @code{WITH GRANT OPTION} clause gives the user the ability to give +to other users any privileges the user has at the specified privilege level. +You should be careful to whom you give the @strong{grant} privilege, as two +users with different privileges may be able to join privileges! -Most examples in the following sections assume you are connected to the -server. They indicate this by the @code{mysql>} prompt. +You cannot grant another user a privilege you don't have yourself; +the @strong{grant} privilege allows you to give away only those privileges +you possess. -@cindex running, queries -@cindex queries, entering -@cindex entering, queries -@node Entering queries, Database use, Connecting-disconnecting, Tutorial -@section Entering Queries +Be aware that when you grant a user the @strong{grant} privilege at a +particular privilege level, any privileges the user already possesses (or +is given in the future!) at that level are also grantable by that user. +Suppose you grant a user the @strong{insert} privilege on a database. If +you then grant the @strong{select} privilege on the database and specify +@code{WITH GRANT OPTION}, the user can give away not only the @strong{select} +privilege, but also @strong{insert}. If you then grant the @strong{update} +privilege to the user on the database, the user can give away the +@strong{insert}, @strong{select} and @strong{update}. -Make sure you are connected to the server, as discussed in the previous -section. Doing so will not in itself select any database to work with, but -that's okay. At this point, it's more important to find out a little about -how to issue queries than to jump right in creating tables, loading data -into them, and retrieving data from them. This section describes the basic -principles of entering commands, using several queries you can try out to -familiarize yourself with how @code{mysql} works. +You should not grant @strong{alter} privileges to a normal user. If you +do that, the user can try to subvert the privilege system by renaming +tables! -Here's a simple command that asks the server to tell you its version number -and the current date. Type it in as shown below following the @code{mysql>} -prompt and hit the RETURN key: +Note that if you are using table or column privileges for even one user, the +server examines table and column privileges for all users and this will slow +down @strong{MySQL} a bit. -@example -mysql> SELECT VERSION(), CURRENT_DATE; -+--------------+--------------+ -| version() | CURRENT_DATE | -+--------------+--------------+ -| 3.22.20a-log | 1999-03-19 | -+--------------+--------------+ -1 row in set (0.01 sec) -mysql> -@end example +When @code{mysqld} starts, all privileges are read into memory. +Database, table, and column privileges take effect at once, and +user-level privileges take effect the next time the user connects. +Modifications to the grant tables that you perform using @code{GRANT} or +@code{REVOKE} are noticed by the server immediately. +If you modify the grant tables manually (using @code{INSERT}, @code{UPDATE}, +etc.), you should execute a @code{FLUSH PRIVILEGES} statement or run +@code{mysqladmin flush-privileges} to tell the server to reload the grant +tables. +@xref{Privilege changes}. -This query illustrates several things about @code{mysql}: +@cindex ANSI SQL, differences from +The biggest differences between the ANSI SQL and @strong{MySQL} versions of +@code{GRANT} are: @itemize @bullet @item -A command normally consists of a SQL statement followed by a semicolon. -(There are some exceptions where a semicolon is not needed. @code{QUIT}, -mentioned earlier, is one of them. We'll get to others later.) +In @strong{MySQL} privileges are given for an username + hostname combination +and not only for an username. @item -When you issue a command, @code{mysql} sends it to the server for execution -and displays the results, then prints another @code{mysql>} to indicate -that it is ready for another command. +ANSI SQL doesn't have global or database-level privileges, and ANSI SQL +doesn't support all privilege types that @strong{MySQL} supports. +@strong{MySQL} doesn't support the ANSI SQL @code{TRIGGER}, @code{EXECUTE} or +@code{UNDER} privileges. @item -@code{mysql} displays query output as a table (rows and columns). The first -row contains labels for the columns. The rows following are the query -results. Normally, column labels are the names of the columns you fetch from -database tables. If you're retrieving the value of an expression rather than -a table column (as in the example just shown), @code{mysql} labels the column -using the expression itself. +ANSI SQL privileges are structured in a hierarchal manner. If you remove +an user, all privileges the user has granted are revoked. In +@strong{MySQL} the granted privileges are not automatically revoked, but +you have to revoke these yourself if needed. @item -@code{mysql} shows how many rows were returned and how long the query took -to execute, which gives you a rough idea of server performance. These values -are imprecise because they represent wall clock time (not CPU or machine -time), and because they are affected by factors such as server load and -network latency. (For brevity, the ``rows in set'' line is not shown in -the remaining examples in this chapter.) +If you in @strong{MySQL} have the @code{INSERT} grant on only part of the +columns in a table, you can execute @code{INSERT} statements on the +table; The columns for which you don't have the @code{INSERT} privilege +will set to their default values. ANSI SQL requires you to have the +@code{INSERT} privilege on all columns. + +@item +When you drop a table in ANSI SQL, all privileges for the table are revoked. +If you revoke a privilege in ANSI SQL, all privileges that were granted based +on this privilege are also revoked. In @strong{MySQL}, privileges can be +dropped only with explicit @code{REVOKE} commands or by manipulating the +@strong{MySQL} grant tables. @end itemize -Keywords may be entered in any lettercase. The following queries are -equivalent: +@cindex indexes +@cindex indexes, multi-part +@cindex multi-part index +@findex CREATE INDEX +@node CREATE INDEX, DROP INDEX, GRANT, Reference +@section @code{CREATE INDEX} Syntax @example -mysql> SELECT VERSION(), CURRENT_DATE; -mysql> select version(), current_date; -mysql> SeLeCt vErSiOn(), current_DATE; +CREATE [UNIQUE|FULLTEXT] INDEX index_name ON tbl_name (col_name[(length)],... ) @end example -Here's another query. It demonstrates that you can use @code{mysql} as a -simple calculator: +The @code{CREATE INDEX} statement doesn't do anything in @strong{MySQL} prior +to Version 3.22. In Version 3.22 or later, @code{CREATE INDEX} is mapped to an +@code{ALTER TABLE} statement to create indexes. +@xref{ALTER TABLE, , @code{ALTER TABLE}}. + +Normally, you create all indexes on a table at the time the table itself +is created with @code{CREATE TABLE}. +@xref{CREATE TABLE, , @code{CREATE TABLE}}. +@code{CREATE INDEX} allows you to add indexes to existing tables. + +A column list of the form @code{(col1,col2,...)} creates a multiple-column +index. Index values are formed by concatenating the values of the given +columns. + +For @code{CHAR} and @code{VARCHAR} columns, indexes can be created that +use only part of a column, using @code{col_name(length)} syntax. (On +@code{BLOB} and @code{TEXT} columns the length is required). The +statement shown below creates an index using the first 10 characters of +the @code{name} column: @example -mysql> SELECT SIN(PI()/4), (4+1)*5; -+-------------+---------+ -| SIN(PI()/4) | (4+1)*5 | -+-------------+---------+ -| 0.707107 | 25 | -+-------------+---------+ +mysql> CREATE INDEX part_of_name ON customer (name(10)); @end example -The commands shown thus far have been relatively short, single-line -statements. You can even enter multiple statements on a single line. -Just end each one with a semicolon: +Because most names usually differ in the first 10 characters, this index should +not be much slower than an index created from the entire @code{name} column. +Also, using partial columns for indexes can make the index file much smaller, +which could save a lot of disk space and might also speed up @code{INSERT} +operations! -@example -mysql> SELECT VERSION(); SELECT NOW(); -+--------------+ -| version() | -+--------------+ -| 3.22.20a-log | -+--------------+ +Note that you can only add an index on a column that can have @code{NULL} +values or on a @code{BLOB}/@code{TEXT} column if you are using +@strong{MySQL} Version 3.23.2 or newer and are using the @code{MyISAM} +table type. -+---------------------+ -| NOW() | -+---------------------+ -| 1999-03-19 00:15:33 | -+---------------------+ -@end example +For more information about how @strong{MySQL} uses indexes, see +@ref{MySQL indexes, , @strong{MySQL} indexes}. -A command need not be given all on a single line, so lengthy commands that -require several lines are not a problem. @code{mysql} determines where your -statement ends by looking for the terminating semicolon, not by looking for -the end of the input line. (In other words, @code{mysql} -accepts free-format input: it collects input lines but does not execute them -until it sees the semicolon.) +@code{FULLTEXT} indexes can index only @code{VARCHAR} and +@code{TEXT} columns, and only in @code{MyISAM} tables. @code{FULLTEXT} indexes +are available in @strong{MySQL} Version 3.23.23 and later. +@ref{Fulltext Search}. -Here's a simple multiple-line statement: +@findex DROP INDEX +@node DROP INDEX, Comments, CREATE INDEX, Reference +@section @code{DROP INDEX} Syntax @example -mysql> SELECT - -> USER() - -> , - -> CURRENT_DATE; -+--------------------+--------------+ -| USER() | CURRENT_DATE | -+--------------------+--------------+ -| joesmith@@localhost | 1999-03-18 | -+--------------------+--------------+ +DROP INDEX index_name ON tbl_name @end example -In this example, notice how the prompt changes from @code{mysql>} to -@code{->} after you enter the first line of a multiple-line query. This is -how @code{mysql} indicates that it hasn't seen a complete statement and is -waiting for the rest. The prompt is your friend, because it provides -valuable feedback. If you use that feedback, you will always be aware of -what @code{mysql} is waiting for. +@code{DROP INDEX} drops the index named @code{index_name} from the table +@code{tbl_name}. @code{DROP INDEX} doesn't do anything in @strong{MySQL} +prior to Version 3.22. In Version 3.22 or later, @code{DROP INDEX} is mapped to an +@code{ALTER TABLE} statement to drop the index. +@xref{ALTER TABLE, , @code{ALTER TABLE}}. -If you decide you don't want to execute a command that you are in the -process of entering, cancel it by typing @code{\c}: +@findex Comment syntax +@cindex comments, adding +@node Comments, CREATE FUNCTION, DROP INDEX, Reference +@section Comment Syntax + +The @strong{MySQL} server supports the @code{# to end of line}, @code{-- +to end of line} and @code{/* in-line or multiple-line */} comment +styles: @example -mysql> SELECT - -> USER() - -> \c -mysql> +mysql> select 1+1; # This comment continues to the end of line +mysql> select 1+1; -- This comment continues to the end of line +mysql> select 1 /* this is an in-line comment */ + 1; +mysql> select 1+ +/* +this is a +multiple-line comment +*/ +1; @end example -Here, too, notice the prompt. It switches back to @code{mysql>} after you -type @code{\c}, providing feedback to indicate that @code{mysql} is ready -for a new command. +Note that the @code{--} comment style requires you to have at least one space +after the @code{--}! -The following table shows each of the prompts you may see and summarizes what -they mean about the state that @code{mysql} is in: +Although the server understands the comment syntax just described, +there are some limitations on the way that the @code{mysql} client +parses @code{/* ... */} comments: -@cindex prompts, meanings -@multitable @columnfractions .10 .9 -@item @strong{Prompt} @tab @strong{Meaning} -@item @code{mysql>} @tab Ready for new command. -@item @code{@ @ @ @ ->} @tab Waiting for next line of multiple-line command. -@item @code{@ @ @ @ '>} @tab Waiting for next line, collecting a string that begins -with a single quote (@samp{'}). -@item @code{@ @ @ @ ">} @tab Waiting for next line, collecting a string that begins -with a double quote (@samp{"}). -@end multitable +@itemize @bullet +@item +Single-quote and double-quote characters are taken to indicate the beginning +of a quoted string, even within a comment. If the quote is not matched by a +second quote within the comment, the parser doesn't realize the comment has +ended. If you are running @code{mysql} interactively, you can tell that it +has gotten confused like this because the prompt changes from @code{mysql>} +to @code{'>} or @code{">}. -Multiple-line statements commonly occur by accident when you intend to -issue a command on a single line, but forget the terminating semicolon. In -this case, @code{mysql} waits for more input: +@item +A semicolon is taken to indicate the end of the current SQL statement +and anything following it to indicate the beginning of the next statement. +@end itemize + +These limitations apply both when you run @code{mysql} interactively +and when you put commands in a file and tell @code{mysql} to read its +input from that file with @code{mysql < some-file}. + +@strong{MySQL} doesn't support the @samp{--} ANSI SQL comment style. +@xref{Missing comments}. + +@findex CREATE FUNCTION +@findex DROP FUNCTION +@findex UDF functions +@findex User-defined functions +@findex Functions, user-defined +@node CREATE FUNCTION, Reserved words, Comments, Reference +@section @code{CREATE FUNCTION/DROP FUNCTION} Syntax @example -mysql> SELECT USER() - -> +CREATE [AGGREGATE] FUNCTION function_name RETURNS @{STRING|REAL|INTEGER@} + SONAME shared_library_name + +DROP FUNCTION function_name @end example -If this happens to you (you think you've entered a statement but the only -response is a @code{->} prompt), most likely @code{mysql} is waiting for the -semicolon. If you don't notice what the prompt is telling you, you might sit -there for a while before realizing what you need to do. Enter a semicolon to -complete the statement, and @code{mysql} will execute it: +A user-definable function (UDF) is a way to extend @strong{MySQL} with a new +function that works like native (built in) @strong{MySQL} functions such as +@code{ABS()} and @code{CONCAT()}. -@example -mysql> SELECT USER() - -> ; -+--------------------+ -| USER() | -+--------------------+ -| joesmith@@localhost | -+--------------------+ -@end example +@code{AGGREGATE} is a new option for @strong{MySQL} Version 3.23. An +@code{AGGREGATE} function works exactly like a native @strong{MySQL} +@code{GROUP} function like @code{SUM} or @code{COUNT()}. -The @code{'>} and @code{">} prompts occur during string collection. -In @strong{MySQL}, you can write strings surrounded by either @samp{'} -or @samp{"} characters (for example, @code{'hello'} or @code{"goodbye"}), -and @code{mysql} lets you enter strings that span multiple lines. When you -see a @code{'>} or @code{">} prompt, it means that you've entered a line -containing a string that begins with a @samp{'} or @samp{"} quote character, -but have not yet entered the matching quote that terminates the string. -That's fine if you really are entering a multiple-line string, but how likely -is that? Not very. More often, the @code{'>} and @code{">} prompts indicate -that you've inadvertantly left out a quote character. For example: +@code{CREATE FUNCTION} saves the function's name, type, and shared library +name in the @code{mysql.func} system table. You must have the +@strong{insert} and @strong{delete} privileges for the @code{mysql} database +to create and drop functions. -@example -mysql> SELECT * FROM my_table WHERE name = "Smith AND age < 30; - "> -@end example +All active functions are reloaded each time the server starts, unless +you start @code{mysqld} with the @code{--skip-grant-tables} option. In +this case, UDF initialization is skipped and UDFs are unavailable. +(An active function is one that has been loaded with @code{CREATE FUNCTION} +and not removed with @code{DROP FUNCTION}.) -If you enter this @code{SELECT} statement, then hit RETURN and wait for the -result, nothing will happen. Instead of wondering why this -query takes so long, notice the clue provided by the @code{">} prompt. It -tells you that @code{mysql} expects to see the rest of an unterminated -string. (Do you see the error in the statement? The string @code{"Smith} is -missing the second quote.) +For instructions on writing user-definable functions, see @ref{Adding +functions}. For the UDF mechanism to work, functions must be written in C or +C++, your operating system must support dynamic loading and you must have +compiled @code{mysqld} dynamically (not statically). -At this point, what do you do? The simplest thing is to cancel the command. -However, you cannot just type @code{\c} in this case, because @code{mysql} -interprets it as part of the string that it is collecting! Instead, enter -the closing quote character (so @code{mysql} knows you've finished the -string), then type @code{\c}: +@cindex keywords +@cindex reserved words, exceptions +@node Reserved words, , CREATE FUNCTION, Reference +@section Is MySQL Picky About Reserved Words? -@example -mysql> SELECT * FROM my_table WHERE name = "Smith AND age < 30; - "> "\c -mysql> -@end example +A common problem stems from trying to create a table with column names that +use the names of datatypes or functions built into @strong{MySQL}, such as +@code{TIMESTAMP} or @code{GROUP}. You're allowed to do it (for example, +@code{ABS} is an allowed column name), but whitespace is not allowed between +a function name and the @samp{(} when using functions whose names are also +column names. -The prompt changes back to @code{mysql>}, indicating that @code{mysql} -is ready for a new command. +The following words are explicitly reserved in @strong{MySQL}. Most of +them are forbidden by ANSI SQL92 as column and/or table names +(for example, @code{group}). +A few are reserved because @strong{MySQL} needs them and is +(currently) using a @code{yacc} parser: -It's important to know what the @code{'>} and @code{">} prompts signify, -because if you mistakenly enter an unterminated string, any further lines you -type will appear to be ignored by @code{mysql} --- including a line -containing @code{QUIT}! This can be quite confusing, especially if you -don't know that you need to supply the terminating quote before you can -cancel the current command. +@c This is fixed by including the symbols table from lex.h here and then running +@c fix-mysql-reserved-words in emacs (or let David do it): +@c (defun fix-mysql-reserved-words () +@c (interactive) +@c (let ((cnt 0)) +@c (insert "\n@item ") +@c (while (looking-at "[ \t]*{ +\"\\([^\"]+\\)\"[ \t]*,.*\n") +@c (replace-match "@code{\\1}") +@c (incf cnt) +@c (if (> cnt 3) +@c (progn +@c (setf cnt 0) +@c (insert "\n@item ")) +@c (insert " @tab "))))) +@c But remove the non alphanumeric entries by hand first. +@c Updated after 3.23.4 990928 by David -@cindex databases, creating -@cindex databases, using -@cindex creating, databases -@node Database use, Getting information, Entering queries, Tutorial -@section Creating and Using a Database +@multitable @columnfractions .25 .25 .25 .25 +@item @code{action} @tab @code{add} @tab @code{aggregate} @tab @code{all} +@item @code{alter} @tab @code{after} @tab @code{and} @tab @code{as} +@item @code{asc} @tab @code{avg} @tab @code{avg_row_length} @tab @code{auto_increment} +@item @code{between} @tab @code{bigint} @tab @code{bit} @tab @code{binary} +@item @code{blob} @tab @code{bool} @tab @code{both} @tab @code{by} +@item @code{cascade} @tab @code{case} @tab @code{char} @tab @code{character} +@item @code{change} @tab @code{check} @tab @code{checksum} @tab @code{column} +@item @code{columns} @tab @code{comment} @tab @code{constraint} @tab @code{create} +@item @code{cross} @tab @code{current_date} @tab @code{current_time} @tab @code{current_timestamp} +@item @code{data} @tab @code{database} @tab @code{databases} @tab @code{date} +@item @code{datetime} @tab @code{day} @tab @code{day_hour} @tab @code{day_minute} +@item @code{day_second} @tab @code{dayofmonth} @tab @code{dayofweek} @tab @code{dayofyear} +@item @code{dec} @tab @code{decimal} @tab @code{default} @tab @code{delayed} +@item @code{delay_key_write} @tab @code{delete} @tab @code{desc} @tab @code{describe} +@item @code{distinct} @tab @code{distinctrow} @tab @code{double} @tab @code{drop} +@item @code{end} @tab @code{else} @tab @code{escape} @tab @code{escaped} +@item @code{enclosed} @tab @code{enum} @tab @code{explain} @tab @code{exists} +@item @code{fields} @tab @code{file} @tab @code{first} @tab @code{float} +@item @code{float4} @tab @code{float8} @tab @code{flush} @tab @code{foreign} +@item @code{from} @tab @code{for} @tab @code{full} @tab @code{function} +@item @code{global} @tab @code{grant} @tab @code{grants} @tab @code{group} +@item @code{having} @tab @code{heap} @tab @code{high_priority} @tab @code{hour} +@item @code{hour_minute} @tab @code{hour_second} @tab @code{hosts} @tab @code{identified} +@item @code{ignore} @tab @code{in} @tab @code{index} @tab @code{infile} +@item @code{inner} @tab @code{insert} @tab @code{insert_id} @tab @code{int} +@item @code{integer} @tab @code{interval} @tab @code{int1} @tab @code{int2} +@item @code{int3} @tab @code{int4} @tab @code{int8} @tab @code{into} +@item @code{if} @tab @code{is} @tab @code{isam} @tab @code{join} +@item @code{key} @tab @code{keys} @tab @code{kill} @tab @code{last_insert_id} +@item @code{leading} @tab @code{left} @tab @code{length} @tab @code{like} +@item @code{lines} @tab @code{limit} @tab @code{load} @tab @code{local} +@item @code{lock} @tab @code{logs} @tab @code{long} @tab @code{longblob} +@item @code{longtext} @tab @code{low_priority} @tab @code{max} @tab @code{max_rows} +@item @code{match} @tab @code{mediumblob} @tab @code{mediumtext} @tab @code{mediumint} +@item @code{middleint} @tab @code{min_rows} @tab @code{minute} @tab @code{minute_second} +@item @code{modify} @tab @code{month} @tab @code{monthname} @tab @code{myisam} +@item @code{natural} @tab @code{numeric} @tab @code{no} @tab @code{not} +@item @code{null} @tab @code{on} @tab @code{optimize} @tab @code{option} +@item @code{optionally} @tab @code{or} @tab @code{order} @tab @code{outer} +@item @code{outfile} @tab @code{pack_keys} @tab @code{partial} @tab @code{password} +@item @code{precision} @tab @code{primary} @tab @code{procedure} @tab @code{process} +@item @code{processlist} @tab @code{privileges} @tab @code{read} @tab @code{real} +@item @code{references} @tab @code{reload} @tab @code{regexp} @tab @code{rename} +@item @code{replace} @tab @code{restrict} @tab @code{returns} @tab @code{revoke} +@item @code{rlike} @tab @code{row} @tab @code{rows} @tab @code{second} +@item @code{select} @tab @code{set} @tab @code{show} @tab @code{shutdown} +@item @code{smallint} @tab @code{soname} @tab @code{sql_big_tables} @tab @code{sql_big_selects} +@item @code{sql_low_priority_updates} @tab @code{sql_log_off} @tab @code{sql_log_update} @tab @code{sql_select_limit} +@item @code{sql_small_result} @tab @code{sql_big_result} @tab @code{sql_warnings} @tab @code{straight_join} +@item @code{starting} @tab @code{status} @tab @code{string} @tab @code{table} +@item @code{tables} @tab @code{temporary} @tab @code{terminated} @tab @code{text} +@item @code{then} @tab @code{time} @tab @code{timestamp} @tab @code{tinyblob} +@item @code{tinytext} @tab @code{tinyint} @tab @code{trailing} @tab @code{to} +@item @code{type} @tab @code{use} @tab @code{using} @tab @code{unique} +@item @code{unlock} @tab @code{unsigned} @tab @code{update} @tab @code{usage} +@item @code{values} @tab @code{varchar} @tab @code{variables} @tab @code{varying} +@item @code{varbinary} @tab @code{with} @tab @code{write} @tab @code{when} +@item @code{where} @tab @code{year} @tab @code{year_month} @tab @code{zerofill} +@end multitable -@menu -* Creating database:: Creating a database -* Creating tables:: Creating a table -* Loading tables:: Loading data into a table -* Retrieving data:: Retrieving information from a table -@end menu +The following symbols (from the table above) are disallowed by ANSI SQL +but allowed by @strong{MySQL} as column/table names. This is because some +of these names are very natural names and a lot of people have already +used them. -Now that you know how to enter commands, it's time to access a database. +@itemize @bullet +@item @code{ACTION} +@item @code{BIT} +@item @code{DATE} +@item @code{ENUM} +@item @code{NO} +@item @code{TEXT} +@item @code{TIME} +@item @code{TIMESTAMP} +@end itemize -Suppose you have several pets in your home (your menagerie) and you'd -like to keep track of various types of information about them. You can do so -by creating tables to hold your data and loading them with the desired -information. Then you can answer different sorts of questions about your -animals by retrieving data from the tables. This section shows you how to: +@cindex table types, choosing +@cindex @code{BDB} table type +@cindex @code{Berkeley_db} table type +@cindex @code{HEAP} table type +@cindex @code{ISAM} table type +@cindex @code{InnoDB} table type +@cindex @code{MERGE} table type +@cindex MySQL table types +@cindex @code{MyISAM} table type +@cindex types, of tables +@node Table types, Tutorial, Reference, Top +@chapter MySQL Table Types + +As of @strong{MySQL} Version 3.23.6, you can choose between three basic +table formats (@code{ISAM}, @code{HEAP} and @code{MyISAM}. Newer +@strong{MySQL} may support additional table type (@code{BDB}, +or @code{InnoDB}), depending on how you compile it. + +When you create a new table, you can tell @strong{MySQL} which table +type it should use for the table. @strong{MySQL} will always create a +@code{.frm} file to hold the table and column definitions. Depending on +the table type, the index and data will be stored in other files. + +Note that to use @code{InnoDB} tables you have to use at least +the @code{innodb_data_file_path} startup option. @xref{InnoDB start}. + +The default table type in @strong{MySQL} is @code{MyISAM}. If you are +trying to use a table type that is not compiled-in or activated, +@strong{MySQL} will instead create a table of type @code{MyISAM}. This +is a very useful feature when you want to copy tables between different +SQL servers that supports different table types (like copying tables to +a slave that is optimized for speed by not having transactional tables). +This automatic table changing can however also be very confusing for new +@strong{MySQL} users. We plan to fix this by introducing warnings in +@strong{MySQL} 4.0 and giving a warning when a table type is automatically +changed. + +You can convert tables between different types with the @code{ALTER +TABLE} statement. @xref{ALTER TABLE, , @code{ALTER TABLE}}. + +Note that @strong{MySQL} supports two different kinds of +tables. Transaction-safe tables (@code{BDB}, @code{InnoDB} +and not transaction-safe tables (@code{HEAP}, @code{ISAM}, +@code{MERGE}, and @code{MyISAM}). + +Advantages of transaction-safe tables (TST): @itemize @bullet @item -Create a database +Safer. Even if @strong{MySQL} crashes or you get hardware problems, you +can get your data back, either by automatic recovery or from a backup ++ the transaction log. @item -Create a table +You can combine many statements and accept these all in one go with +the @code{COMMIT} command. @item -Load data into the table +You can execute @code{ROLLBACK} to ignore your changes (if you are not +running in auto commit mode). @item -Retrieve data from the table in various ways +If an update fails, all your changes will be restored. (With NTST tables all +changes that have taken place are permanent) +@end itemize + +Advantages of not transaction-safe tables (NTST): + +@itemize @bullet @item -Use multiple tables +Much faster as there is no transaction overhead. +@item +Will use less disk space as there is no overhead of transactions. +@item +Will use less memory to do updates. @end itemize -The menagerie database will be simple (deliberately), but it is not difficult -to think of real-world situations in which a similar type of database might -be used. For example, a database like this could be used by a farmer to keep -track of livestock, or by a veterinarian to keep track of patient records. -A menagerie distribution containing some of the queries and sample data used -in the following sections can be obtained from the @strong{MySQL} Web site. -It's available in either -@uref{http://www.mysql.com/Downloads/Contrib/Examples/menagerie.tar.gz,compressed @code{tar} format} -or -@uref{http://www.mysql.com/Downloads/Contrib/Examples/menagerie.zip,Zip format}. +You can combine TST and NTST tables in the same statements to get the best +of both worlds. -Use the @code{SHOW} statement to find out what databases currently exist -on the server: +@menu +* MyISAM:: MyISAM tables +* MERGE:: MERGE tables +* ISAM:: ISAM tables +* HEAP:: HEAP tables +* InnoDB:: InnoDB tables +* BDB:: BDB or Berkeley_db tables +@end menu -@example -mysql> SHOW DATABASES; -+----------+ -| Database | -+----------+ -| mysql | -| test | -| tmp | -+----------+ -@end example +@node MyISAM, MERGE, Table types, Table types +@section MyISAM Tables -The list of databases is probably different on your machine, but the -@code{mysql} and @code{test} databases are likely to be among them. The -@code{mysql} database is required because it describes user access -privileges. The @code{test} database is often provided as a workspace for -users to try things out. +@code{MyISAM} is the default table type in @strong{MySQL} Version 3.23. It's +based on the @code{ISAM} code and has a lot of useful extensions. -If the @code{test} database exists, try to access it: +The index is stored in a file with the @code{.MYI} (MYIndex) extension, +and the data is stored in a file with the @code{.MYD} (MYData) extension. +You can check/repair @code{MyISAM} tables with the @code{myisamchk} +utility. @xref{Crash recovery}. You can compress @code{MyISAM} tables with +@code{myisampack} to take up much less space. @xref{myisampack}. -@example -mysql> USE test -Database changed -@end example +The following is new in @code{MyISAM}: -Note that @code{USE}, like @code{QUIT}, does not require a semicolon. (You -can terminate such statements with a semicolon if you like; it does no harm.) -The @code{USE} statement is special in another way, too: it must be given on -a single line. +@itemize @bullet +@item +There is a flag in the @code{MyISAM} file that indicates whether or not +the table was closed correctly. If @code{mysqld} is started with +@code{--myisam-recover}, @code{MyISAM} tables will automatically be +checked and/or repaired on open if the table wasn't closed properly. +@item +You can @code{INSERT} new rows in a table that doesn't have free blocks +in the middle of the data file, at the same time other threads are +reading from the table (concurrent insert). An free block can come from +an update of a dynamic length row with much data to a row with less data +or when deleting rows. When all free blocks are used up, all future +inserts will be concurrent again. +@item +Support for big files (63-bit) on filesystems/operating systems that +support big files. +@item +All data is stored with the low byte first. This makes the data machine +and OS independent. The only requirement is that the machine uses +two's-complement signed integers (as every machine for the last 20 years +has) and IEEE floating-point format (also totally dominant among +mainstream machines). The only area of machines that may not support +binary compatibility are embedded systems (because they sometimes have +peculiar processors). -You can use the @code{test} database (if you have access to it) for the -examples that follow, but anything you create in that database can be -removed by anyone else with access to it. For this reason, you should -probably ask your @strong{MySQL} administrator for permission to use a -database of your own. Suppose you want to call yours @code{menagerie}. The -administrator needs to execute a command like this: +There is no big speed penalty in storing data low byte first; The bytes +in a table row is normally unaligned and it doesn't take that much more +power to read an unaligned byte in order than in reverse order. The +actual fetch-column-value code is also not time critical compared to +other code. +@item +All number keys are stored with high byte first to give better index +compression. +@item +Internal handling of one @code{AUTO_INCREMENT} column. @code{MyISAM} +will automatically update this on @code{INSERT/UPDATE}. The +@code{AUTO_INCREMENT} value can be reset with @code{myisamchk}. This +will make @code{AUTO_INCREMENT} columns faster (at least 10 %) and old +numbers will not be reused as with the old @code{ISAM}. Note that when an +@code{AUTO_INCREMENT} is defined on the end of a multi-part-key the old +behavior is still present. +@item +When inserted in sorted order (as when you are using an @code{AUTO_INCREMENT} +column) the key tree will be split so that the high node only contains one +key. This will improve the space utilization in the key tree. +@item +@code{BLOB} and @code{TEXT} columns can be indexed. +@item +@code{NULL} values are allowed in indexed columns. This takes 0-1 +bytes/key. +@item +Maximum key length is 500 bytes by default (can be changed by +recompiling). In cases of keys longer than 250 bytes, a bigger key +block size than the default of 1024 bytes is used for this key. +@item +Maximum number of keys/table is 32 as default. This can be enlarged to 64 +without having to recompile @code{myisamchk}. +@item +@code{myisamchk} will mark tables as checked if one runs it with +@code{--update-state}. @code{myisamchk --fast} will only check those +tables that don't have this mark. +@item +@code{myisamchk -a} stores statistics for key parts (and not only for +whole keys as in @code{ISAM}). +@item +Dynamic size rows will now be much less fragmented when mixing deletes +with updates and inserts. This is done by automatically combining adjacent +deleted blocks and by extending blocks if the next block is deleted. +@item +@code{myisampack} can pack @code{BLOB} and @code{VARCHAR} columns. +@item +You can use put the datafile and index file on different directories +to get more speed (with the @code{DATA/INDEX DIRECTORY="path"} option to +@code{CREATE TABLE}). @xref{CREATE TABLE}. +@end itemize -@example -mysql> GRANT ALL ON menagerie.* TO your_mysql_name; -@end example +@code{MyISAM} also supports the following things, which @strong{MySQL} +will be able to use in the near future: -where @code{your_mysql_name} is the @strong{MySQL} user name assigned to -you. +@itemize @bullet +@item +Support for a true @code{VARCHAR} type; A @code{VARCHAR} column starts +with a length stored in 2 bytes. +@item +Tables with @code{VARCHAR} may have fixed or dynamic record length. +@item +@code{VARCHAR} and @code{CHAR} may be up to 64K. +All key segments have their own language definition. This will enable +@strong{MySQL} to have different language definitions per column. +@item +A hashed computed index can be used for @code{UNIQUE}. This will allow +you to have @code{UNIQUE} on any combination of columns in a table. (You +can't search on a @code{UNIQUE} computed index, however.) +@end itemize -@cindex selecting, databases -@cindex databases, selecting -@node Creating database, Creating tables, Database use, Database use -@subsection Creating and Selecting a Database +Note that index files are usually much smaller with @code{MyISAM} than with +@code{ISAM}. This means that @code{MyISAM} will normally use less +system resources than @code{ISAM}, but will need more CPU when inserting +data into a compressed index. -If the administrator creates your database for you when setting up your -permissions, you can begin using it. Otherwise, you need to create it -yourself: +The following options to @code{mysqld} can be used to change the behavior of +@code{MyISAM} tables. @xref{SHOW VARIABLES}. -@example -mysql> CREATE DATABASE menagerie; -@end example +@multitable @columnfractions .40 .60 +@item @strong{Option} @tab @strong{Meaning} +@item @code{--myisam-recover=#} @tab Automatic recover of crashed tables. +@item @code{-O myisam_sort_buffer_size=#} @tab Buffer used when recovering tables. +@item @code{--delay-key-write-for-all-tables} @tab Don't flush key buffers between writes for any MyISAM table +@item @code{-O myisam_max_extra_sort_file_size=#} @tab Used to help @strong{MySQL} to decide when to use the slow but safe key cache index create method. @strong{NOTE} that this parameter is given in megabytes! +@item @code{-O myisam_max_sort_file_size=#} @tab Don't use the fast sort index method to created index if the temporary file would get bigger than this. +@strong{NOTE} that this paramter is given in megabytes! +@end multitable -Under Unix, database names are case sensitive (unlike SQL keywords), so you -must always refer to your database as @code{menagerie}, not as -@code{Menagerie}, @code{MENAGERIE}, or some other variant. This is also true -for table names. (Under Windows, this restriction does not apply, although -you must refer to databases and tables using the same lettercase throughout a -given query.) +The automatic recovery is activated if you start @code{mysqld} with +@code{--myisam-recover=#}. @xref{Command-line options}. +On open, the table is checked if it's marked as crashed or if the open +count variable for the table is not 0 and you are running with +@code{--skip-locking}. If either of the above is true the following happens. -Creating a database does not select it for use; you must do that explicitly. -To make @code{menagerie} the current database, use this command: +@itemize @bullet +@item +The table is checked for errors. +@item +If we found an error, try to do a fast repair (with sorting and without +re-creating the data file) of the table. +@item +If the repair fails because of an error in the data file (for example a +duplicate key error), we try again, but this time we re-create the data file. +@item +If the repair fails, retry once more with the old repair option method +(write row by row without sorting) which should be able to repair any +type of error with little disk requirements.. +@end itemize + +If the recover wouldn't be able to recover all rows from a previous +completed statement and you didn't specify @code{FORCE} as an option to +@code{myisam-recover}, then the automatic repair will abort with an error +message in the error file: @example -mysql> USE menagerie -Database changed +Error: Couldn't repair table: test.g00pages @end example -Your database needs to be created only once, but you must select it for use -each time you begin a @code{mysql} session. You can do this by issuing a -@code{USE} statement as shown above. Alternatively, you can select the -database on the command line when you invoke @code{mysql}. Just specify its -name after any connection parameters that you might need to provide. For -example: +If you in this case had used the @code{FORCE} option you would instead have got +a warning in the error file: @example -shell> mysql -h host -u user -p menagerie -Enter password: ******** +Warning: Found 344 of 354 rows when repairing ./test/g00pages @end example -Note that @code{menagerie} is not your password on the command just shown. -If you want to supply your password on the command line after the @code{-p} -option, you must do so with no intervening space (for example, as -@code{-pmypassword}, not as @code{-p mypassword}). However, putting your -password on the command line is not recommended, because doing so exposes it -to snooping by other users logged in on your machine. +Note that if you run automatic recover with the @code{BACKUP} option, +you should have a cron script that automatically moves file with names +like @file{tablename-datetime.BAK} from the database directories to a +backup media. -@cindex tables, creating -@cindex creating, tables -@node Creating tables, Loading tables, Creating database, Database use -@subsection Creating a Table +@xref{Command-line options}. -Creating the database is the easy part, but at this point it's empty, as -@code{SHOW TABLES} will tell you: +@menu +* Key space:: Space needed for keys +* MyISAM table formats:: MyISAM table formats +* MyISAM table problems:: +@end menu -@example -mysql> SHOW TABLES; -Empty set (0.00 sec) -@end example +@cindex key space, MyISAM +@node Key space, MyISAM table formats, MyISAM, MyISAM +@subsection Space Needed for Keys -The harder part is deciding what the structure of your database should be: -what tables you will need and what columns will be in each of them. +@strong{MySQL} can support different index types, but the normal type is +ISAM or MyISAM. These use a B-tree index, and you can roughly calculate +the size for the index file as @code{(key_length+4)/0.67}, summed over +all keys. (This is for the worst case when all keys are inserted in +sorted order and we don't have any compressed keys.) -You'll want a table that contains a record for each of your pets. This can -be called the @code{pet} table, and it should contain, as a bare minimum, -each animal's name. Because the name by itself is not very interesting, the -table should contain other information. For example, if more than one person -in your family keeps pets, you might want to list each animal's owner. You -might also want to record some basic descriptive information such as species -and sex. +String indexes are space compressed. If the first index part is a +string, it will also be prefix compressed. Space compression makes the +index file smaller than the above figures if the string column has a lot +of trailing space or is a @code{VARCHAR} column that is not always used +to the full length. Prefix compression is used on keys that start +with a string. Prefix compression helps if there are many strings +with an identical prefix. -How about age? That might be of interest, but it's not a good thing to store -in a database. Age changes as time passes, which means you'd have to update -your records often. Instead, it's better to store a fixed value such as -date of birth. Then, whenever you need age, you can calculate it as the -difference between the current date and the birth date. @strong{MySQL} -provides functions for doing date arithmetic, so this is not difficult. -Storing birth date rather than age has other advantages, too: +In @code{MyISAM} tables, you can also prefix compress numbers by specifying +@code{PACK_KEYS=1} when you create the table. This helps when you have +many integer keys that have an identical prefix when the numbers are stored +high-byte first. -@itemize @bullet -@item -You can use the database for tasks such as generating reminders for upcoming -pet birthdays. (If you think this type of query is somewhat silly, note that -it is the same question you might ask in the context of a business database -to identify clients to whom you'll soon need to send out birthday greetings, -for that computer-assisted personal touch.) +@node MyISAM table formats, MyISAM table problems, Key space, MyISAM +@subsection MyISAM Table Formats -@item -You can calculate age in relation to dates other than the current date. For -example, if you store death date in the database, you can easily calculate -how old a pet was when it died. -@end itemize +@strong{MyISAM} supports 3 different table types. Two of them are chosen +automatically depending on the type of columns you are using. The third, +compressed tables, can only be created with the @code{myisampack} tool. -You can probably think of other types of information that would be useful in -the @code{pet} table, but the ones identified so far are sufficient for now: -name, owner, species, sex, birth, and death. +@menu +* Static format:: Static (Fixed-length) table characteristics +* Dynamic format:: Dynamic table characteristics +* Compressed format:: Compressed table characteristics +@end menu -Use a @code{CREATE TABLE} statement to specify the layout of your table: +@node Static format, Dynamic format, MyISAM table formats, MyISAM table formats +@subsubsection Static (Fixed-length) Table Characteristics -@example -mysql> CREATE TABLE pet (name VARCHAR(20), owner VARCHAR(20), - -> species VARCHAR(20), sex CHAR(1), birth DATE, death DATE); -@end example +This is the default format. It's used when the table contains no +@code{VARCHAR}, @code{BLOB}, or @code{TEXT} columns. -@code{VARCHAR} is a good choice for the @code{name}, @code{owner}, and -@code{species} columns because the column values will vary in length. The -lengths of those columns need not all be the same, and need not be -@code{20}. You can pick any length from @code{1} to @code{255}, whatever -seems most reasonable to you. (If you make a poor choice and it turns -out later that you need a longer field, @strong{MySQL} provides an -@code{ALTER TABLE} statement.) +This format is the simplest and most secure format. It is also the +fastest of the on-disk formats. The speed comes from the easy way data +can be found on disk. When looking up something with an index and static +format it is very simple. Just multiply the row number by the row length. -Animal sex can be represented in a variety of ways, for example, @code{"m"} -and @code{"f"}, or perhaps @code{"male"} and @code{"female"}. It's simplest -to use the single characters @code{"m"} and @code{"f"}. +Also, when scanning a table it is very easy to read a constant number of +records with each disk read. -The use of the @code{DATE} data type for the @code{birth} and @code{death} -columns is a fairly obvious choice. +The security is evidenced if your computer crashes when writing to a +fixed-size MyISAM file, in which case @code{myisamchk} can easily figure out where each +row starts and ends. So it can usually reclaim all records except the +partially written one. Note that in @strong{MySQL} all indexes can always be +reconstructed: -Now that you have created a table, @code{SHOW TABLES} should produce some -output: +@itemize @bullet +@item +All @code{CHAR}, @code{NUMERIC}, and @code{DECIMAL} columns are space-padded +to the column width. +@item +Very quick. +@item +Easy to cache. +@item +Easy to reconstruct after a crash, because records are located in fixed +positions. +@item +Doesn't have to be reorganized (with @code{myisamchk}) unless a huge number of +records are deleted and you want to return free disk space to the operating +system. +@item +Usually requires more disk space than dynamic tables. +@end itemize -@example -mysql> SHOW TABLES; -+---------------------+ -| Tables in menagerie | -+---------------------+ -| pet | -+---------------------+ -@end example +@cindex dynamic table characteristics +@cindex tables, dynamic +@node Dynamic format, Compressed format, Static format, MyISAM table formats +@subsubsection Dynamic Table Characteristics -To verify that your table was created the way you expected, use -a @code{DESCRIBE} statement: +This format is used if the table contains any @code{VARCHAR}, @code{BLOB}, +or @code{TEXT} columns or if the table was created with +@code{ROW_FORMAT=dynamic}. -@example -mysql> DESCRIBE pet; -+---------+-------------+------+-----+---------+-------+ -| Field | Type | Null | Key | Default | Extra | -+---------+-------------+------+-----+---------+-------+ -| name | varchar(20) | YES | | NULL | | -| owner | varchar(20) | YES | | NULL | | -| species | varchar(20) | YES | | NULL | | -| sex | char(1) | YES | | NULL | | -| birth | date | YES | | NULL | | -| death | date | YES | | NULL | | -+---------+-------------+------+-----+---------+-------+ -@end example +This format is a little more complex because each row has to have a header +that says how long it is. One record can also end up at more than one +location when it is made longer at an update. -You can use @code{DESCRIBE} any time, for example, if you forget the names of -the columns in your table or what types they are. +@cindex tables, defragment +You can use @code{OPTIMIZE table} or @code{myisamchk} to defragment a +table. If you have static data that you access/change a lot in the same +table as some @code{VARCHAR} or @code{BLOB} columns, it might be a good +idea to move the dynamic columns to other tables just to avoid +fragmentation: -@cindex loading, tables -@cindex tables, loading data -@cindex data, loading into tables -@node Loading tables, Retrieving data, Creating tables, Database use -@subsection Loading Data into a Table +@itemize @bullet +@item +All string columns are dynamic (except those with a length less than 4). +@item +Each record is preceded by a bitmap indicating which columns are empty +(@code{''}) for string columns, or zero for numeric columns. (This isn't +the same as columns containing @code{NULL} values.) If a string column +has a length of zero after removal of trailing spaces, or a numeric +column has a value of zero, it is marked in the bit map and not saved to +disk. Non-empty strings are saved as a length byte plus the string +contents. +@item +Usually takes much less disk space than fixed-length tables. +@item +Each record uses only as much space as is required. If a record becomes +larger, it is split into as many pieces as are required. This results in record +fragmentation. +@item +If you update a row with information that extends the row length, the +row will be fragmented. In this case, you may have to run @code{myisamchk +-r} from time to time to get better performance. Use @code{myisamchk -ei +tbl_name} for some statistics. +@item +Not as easy to reconstruct after a crash, because a record may be fragmented +into many pieces and a link (fragment) may be missing. +@item +The expected row length for dynamic sized records is: -After creating your table, you need to populate it. The @code{LOAD DATA} and -@code{INSERT} statements are useful for this. +@example +3 ++ (number of columns + 7) / 8 ++ (number of char columns) ++ packed size of numeric columns ++ length of strings ++ (number of NULL columns + 7) / 8 +@end example -Suppose your pet records can be described as shown below. -(Observe that @strong{MySQL} expects dates in @code{YYYY-MM-DD} format; -this may be different than what you are used to.) +There is a penalty of 6 bytes for each link. A dynamic record is linked +whenever an update causes an enlargement of the record. Each new link will be +at least 20 bytes, so the next enlargement will probably go in the same link. +If not, there will be another link. You may check how many links there are +with @code{myisamchk -ed}. All links may be removed with @code{myisamchk -r}. +@end itemize -@multitable @columnfractions .16 .16 .16 .16 .16 .16 -@item @strong{name} @tab @strong{owner} @tab @strong{species} @tab @strong{sex} @tab @strong{birth} @tab @strong{death} -@item Fluffy @tab Harold @tab cat @tab f @tab 1993-02-04 @tab -@item Claws @tab Gwen @tab cat @tab m @tab 1994-03-17 @tab -@item Buffy @tab Harold @tab dog @tab f @tab 1989-05-13 @tab -@item Fang @tab Benny @tab dog @tab m @tab 1990-08-27 @tab -@item Bowser @tab Diane @tab dog @tab m @tab 1989-08-31 @tab 1995-07-29 -@item Chirpy @tab Gwen @tab bird @tab f @tab 1998-09-11 @tab -@item Whistler @tab Gwen @tab bird @tab @tab 1997-12-09 @tab -@item Slim @tab Benny @tab snake @tab m @tab 1996-04-29 @tab -@end multitable +@cindex tables, compressed format +@node Compressed format, , Dynamic format, MyISAM table formats +@subsubsection Compressed Table Characteristics -Because you are beginning with an empty table, an easy way to populate it is to -create a text file containing a row for each of your animals, then load the -contents of the file into the table with a single statement. +This is a read-only type that is generated with the optional +@code{myisampack} tool (@code{pack_isam} for @code{ISAM} tables): -You could create a text file @file{pet.txt} containing one record per line, -with values separated by tabs, and given in the order in which the columns -were listed in the @code{CREATE TABLE} statement. For missing values (such -as unknown sexes or death dates for animals that are still living), you can -use @code{NULL} values. To represent these in your text file, use -@code{\N}. For example, the record for Whistler the bird would look like -this (where the whitespace between values is a single tab character): +@itemize @bullet +@item +All @strong{MySQL} distributions, even those that existed before @strong{MySQL} +went GPL, can read tables that were compressed with @code{myisampack}. +@item +Compressed tables take very little disk space. This minimizes disk usage, which +is very nice when using slow disks (like CD-ROMs). +@item +Each record is compressed separately (very little access overhead). The +header for a record is fixed (1-3 bytes) depending on the biggest record in the +table. Each column is compressed differently. Some of the compression types +are: +@itemize @minus +@item +There is usually a different Huffman table for each column. +@item +Suffix space compression. +@item +Prefix space compression. +@item +Numbers with value @code{0} are stored using 1 bit. +@item +If values in an integer column have a small range, the column is stored using +the smallest possible type. For example, a @code{BIGINT} column (8 bytes) may +be stored as a @code{TINYINT} column (1 byte) if all values are in the range +@code{0} to @code{255}. +@item +If a column has only a small set of possible values, the column type is +converted to @code{ENUM}. +@item +A column may use a combination of the above compressions. +@end itemize +@item +Can handle fixed- or dynamic-length records, but not @code{BLOB} or @code{TEXT} +columns. +@item +Can be uncompressed with @code{myisamchk}. +@end itemize -@multitable @columnfractions .15 .15 .15 .15 .25 .15 -@item @code{Whistler} @tab @code{Gwen} @tab @code{bird} @tab @code{\N} @tab @code{1997-12-09} @tab @code{\N} -@end multitable +@node MyISAM table problems, , MyISAM table formats, MyISAM +@subsection MyISAM table problems. -To load the text file @file{pet.txt} into the @code{pet} table, use this -command: +The file format that @strong{MySQL} uses to store data has been extensively +tested, but there are always circumstances that may cause database tables +to become corrupted. -@example -mysql> LOAD DATA LOCAL INFILE "pet.txt" INTO TABLE pet; -@end example +@menu +* Corrupted MyISAM tables:: +* MyISAM table close:: +@end menu -You can specify the column value separator and end of line marker explicitly -in the @code{LOAD DATA} statement if you wish, but the defaults are tab and -linefeed. These are sufficient for the statement to read the file -@file{pet.txt} properly. +@node Corrupted MyISAM tables, MyISAM table close, MyISAM table problems, MyISAM table problems +@subsubsection Corrupted MyISAM tables. -When you want to add new records one at a time, the @code{INSERT} statement -is useful. In its simplest form, you supply values for each column, in the -order in which the columns were listed in the @code{CREATE TABLE} statement. -Suppose Diane gets a new hamster named Puffball. You could add a new record -using an @code{INSERT} statement like this: +Even if the MyISAM table format is very reliable (all changes to a table +is written before the SQL statements returns) , you can still get +corrupted tables if some of the following things happens: -@example -mysql> INSERT INTO pet - -> VALUES ('Puffball','Diane','hamster','f','1999-03-30',NULL); -@end example +@itemize @bullet +@item +The @code{mysqld} process being killed in the middle of a write. +@item +Unexpected shutdown of the computer (for example, if the computer is turned +off). +@item +A hardware error. +@item +You are using an external program (like myisamchk) on a live table. +@item +A software bug in the MySQL or MyISAM code. +@end itemize -Note that string and date values are specified as quoted strings here. Also, -with @code{INSERT}, you can insert @code{NULL} directly to represent a -missing value. You do not use @code{\N} like you do with @code{LOAD DATA}. +Typial typical symptoms for a corrupt table is: -From this example, you should be able to see that there would be a lot more -typing involved to load -your records initially using several @code{INSERT} statements rather -than a single @code{LOAD DATA} statement. +@itemize @bullet +@item +You get the error @code{Incorrect key file for table: '...'. Try to repair it} +while selecting data from the table. +@item +Queries doesn't find rows in the table or returns incomplete data. +@end itemize -@cindex data, retrieving -@cindex tables, retrieving data -@cindex retrieving, data from tables -@cindex unloading, tables -@node Retrieving data, , Loading tables, Database use -@subsection Retrieving Information from a Table +You can check if a table is ok with the command @code{CHECK +TABLE}. @xref{CHECK TABLE}. -@menu -* Selecting all:: Selecting all data -* Selecting rows:: Selecting particular rows -* Selecting columns:: Selecting particular columns -* Sorting rows:: Sorting rows -* Date calculations:: Date calculations -* Working with NULL:: Working with @code{NULL} values -* Pattern matching:: Pattern matching -* Counting rows:: Counting rows -* Multiple tables:: -@end menu +You can repair a corrupted table with @code{REPAIR TABLE}. @xref{REPAIR TABLE}. +You can also repair a table, when @code{mysqld} is not running with +the @code{myisamchk} command. @code{myisamchk syntax}. -The @code{SELECT} statement is used to pull information from a table. -The general form of the statement is: +If your tables get corrupted a lot you should try to find the reason for +this! @xref{Crashing}. -@example -SELECT what_to_select -FROM which_table -WHERE conditions_to_satisfy -@end example +In this case the most important thing to know is if the table got +corrupted if the @code{mysqld} died (one can easily verify this by +checking if there is a recent row @code{restarted mysqld} in the mysqld +error file). If this isn't the case, then you should try to make a test +case of this. @xref{Reproduceable test case}. -@code{what_to_select} indicates what you want to see. This can be a list of -columns, or @code{*} to indicate ``all columns.'' @code{which_table} -indicates the table from which you want to retrieve data. The @code{WHERE} -clause is optional. If it's present, @code{conditions_to_satisfy} specifies -conditions that rows must satisfy to qualify for retrieval. +@node MyISAM table close, , Corrupted MyISAM tables, MyISAM table problems +@subsubsection Clients is using or hasn't closed the table properly -@node Selecting all, Selecting rows, Retrieving data, Retrieving data -@subsubsection Selecting All Data +Each @code{MyISAM} @code{.MYI} file has in the header a counter that can +be used to check if a table has been closed properly. -The simplest form of @code{SELECT} retrieves everything from a table: +If you get the following warning from @code{CHECK TABLE} or @code{myisamchk}: @example -mysql> SELECT * FROM pet; -+----------+--------+---------+------+------------+------------+ -| name | owner | species | sex | birth | death | -+----------+--------+---------+------+------------+------------+ -| Fluffy | Harold | cat | f | 1993-02-04 | NULL | -| Claws | Gwen | cat | m | 1994-03-17 | NULL | -| Buffy | Harold | dog | f | 1989-05-13 | NULL | -| Fang | Benny | dog | m | 1990-08-27 | NULL | -| Bowser | Diane | dog | m | 1998-08-31 | 1995-07-29 | -| Chirpy | Gwen | bird | f | 1998-09-11 | NULL | -| Whistler | Gwen | bird | NULL | 1997-12-09 | NULL | -| Slim | Benny | snake | m | 1996-04-29 | NULL | -| Puffball | Diane | hamster | f | 1999-03-30 | NULL | -+----------+--------+---------+------+------------+------------+ +# clients is using or hasn't closed the table properly @end example -This form of @code{SELECT} is useful if you want to review your entire table, -for instance, after you've just loaded it with your initial dataset. As it -happens, the output just shown reveals an error in your data file: Bowser -appears to have been born after he died! Consulting your original pedigree -papers, you find that the correct birth year is 1989, not 1998. +this means that this counter has come out of sync. This doesn't mean +that the table is corrupted, but means that you should at least do a +check on the table to verify that it's ok. -There are are least a couple of ways to fix this: +The counter works as follows: @itemize @bullet @item -Edit the file @file{pet.txt} to correct the error, then empty the table -and reload it using @code{DELETE} and @code{LOAD DATA}: - -@example -mysql> SET AUTOCOMMIT=1; # Used for quick re-create of the table -mysql> DELETE FROM pet; -mysql> LOAD DATA LOCAL INFILE "pet.txt" INTO TABLE pet; -@end example +The first time a table is updated in @strong{MySQL}, a counter in the +header of the index files is incremented. +@item +The counter is not changed during further updates. +@item +When the last instance of a table is closed (because of a @code{FLUSH} or +because there isn't room in the table cache) the counter is +decremented if the table has been updated at any point. +@item +When you repair the table or check the table and it was ok, +the counter is reset to 0. +@item +To avoid problems with interaction with other processes that may do a +check on the table, the counter is not decremented on close if it was 0. +@end itemize -However, if you do this, you must also re-enter the record for Puffball. +In other words, the only ways this can go out of sync are: +@itemize @bullet @item -Fix only the erroneous record with an @code{UPDATE} statement: - -@example -mysql> UPDATE pet SET birth = "1989-08-31" WHERE name = "Bowser"; -@end example +The @code{MyISAM} tables are copied without a @code{LOCK} and +@code{FLUSH TABLES}. +@item +@strong{MySQL} has crashed between an update and the final close +(Note that the table may still be ok, as @strong{MySQL} always issues writes +for everything between each statement). +@item +Someone has done a @code{myisamchk --repair} or @code{myisamchk +--update-state}on a table that was in use by @code{mysqld}. +@item +Many @code{mysqld} servers are using the table and one has done a +@code{REPAIR} or @code{CHECK} of the table while it was in use by +another server. In this setup the @code{CHECK} is safe to do (even if +you will get the warning from other servers), but @code{REPAIR} should +be avoided as it currently replaces the data file with a new one, which +is not signaled to the other servers. @end itemize -As shown above, it is easy to retrieve an entire table. But typically you -don't want to do that, particularly when the table becomes large. Instead, -you're usually more interested in answering a particular question, in which -case you specify some constraints on the information you want. Let's look at -some selection queries in terms of questions about your pets that they -answer. +@cindex tables, merging +@cindex MERGE tables, defined +@node MERGE, ISAM, MyISAM, Table types +@section MERGE Tables -@cindex rows, selecting -@cindex tables, selecting rows -@node Selecting rows, Selecting columns, Selecting all, Retrieving data -@subsubsection Selecting Particular Rows +@code{MERGE} tables are new in @strong{MySQL} Version 3.23.25. The code +is still in gamma, but should be resonable stable. -You can select only particular rows from your table. For example, if you want -to verify the change that you made to Bowser's birth date, select Bowser's -record like this: +A @code{MERGE} table is a collection of identical @code{MyISAM} tables +that can be used as one. You can only @code{SELECT}, @code{DELETE}, and +@code{UPDATE} from the collection of tables. If you @code{DROP} the +@code{MERGE} table, you are only dropping the @code{MERGE} +specification. -@example -mysql> SELECT * FROM pet WHERE name = "Bowser"; -+--------+-------+---------+------+------------+------------+ -| name | owner | species | sex | birth | death | -+--------+-------+---------+------+------------+------------+ -| Bowser | Diane | dog | m | 1989-08-31 | 1995-07-29 | -+--------+-------+---------+------+------------+------------+ -@end example +Note that @code{DELETE FROM merge_table} used without a @code{WHERE} +will only clear the mapping for the table, not delete everything in the +mapped tables. (We plan to fix this in 4.0). -The output confirms that the year is correctly recorded now as 1989, not 1998. +With identical tables we mean that all tables are created with identical +column and key information. You can't put a MERGE over tables where the +columns are packed differently or doesn't have exactly the same columns. +Some of the tables can however be compressed with @code{myisampack}. +@xref{myisampack}. -String comparisons are normally case insensitive, so you can specify the -name as @code{"bowser"}, @code{"BOWSER"}, etc. The query result will be -the same. +When you create a @code{MERGE} table, you will get a @code{.frm} table +definition file and a @code{.MRG} table list file. The @code{.MRG} just +contains a list of the index files (@code{.MYI} files) that should +be used as one. -You can specify conditions on any column, not just @code{name}. For example, -if you want to know which animals were born after 1998, test the @code{birth} -column: +For the moment you need to have @code{SELECT}, @code{UPDATE}, and +@code{DELETE} privileges on the tables you map to a @code{MERGE} table. -@example -mysql> SELECT * FROM pet WHERE birth >= "1998-1-1"; -+----------+-------+---------+------+------------+-------+ -| name | owner | species | sex | birth | death | -+----------+-------+---------+------+------------+-------+ -| Chirpy | Gwen | bird | f | 1998-09-11 | NULL | -| Puffball | Diane | hamster | f | 1999-03-30 | NULL | -+----------+-------+---------+------+------------+-------+ -@end example +@code{MERGE} tables can help you solve the following problems: -You can combine conditions, for example, to locate female dogs: +@itemize @bullet +@item +Easily manage a set of log tables. For example, you can put data from +different months into separate files, compress some of them with +@code{myisampack}, and then create a @code{MERGE} to use these as one. +@item +Give you more speed. You can split a big read-only table based on some +criteria and then put the different table part on different disks. +A @code{MERGE} table on this could be much faster than using +the big table. (You can, of course, also use a RAID to get the same +kind of benefits.) +@item +Do more efficient searches. If you know exactly what you are looking +after, you can search in just one of the split tables for some queries +and use @strong{MERGE} table for others. You can even have many +different @code{MERGE} tables active, with possible overlapping files. +@item +More efficient repairs. It's easier to repair the individual files that +are mapped to a @code{MERGE} file than trying to repair a real big file. +@item +Instant mapping of many files as one. A @code{MERGE} table uses the +index of the individual tables. It doesn't need to maintain an index of +its one. This makes @code{MERGE} table collections VERY fast to make or +remap. Note that you must specify the key definitions when you create +a @code{MERGE} table!. +@item +If you have a set of tables that you join to a big table on demand or +batch, you should instead create a @code{MERGE} table on them on demand. +This is much faster and will save a lot of disk space. +@item +Go around the file size limit for the operating system. +@item +You can create an alias/synonym for a table by just using MERGE over one +table. There shouldn't be any really notable performance impacts of doing this +(only a couple of indirect calls and memcpy's for each read). +@end itemize + +The disadvantages with @code{MERGE} tables are: + +@itemize @bullet +@item +You can't use @code{INSERT} on @code{MERGE} tables, as @strong{MySQL} +can't know in which of the tables we should insert the row. +@item +You can only use identical @code{MyISAM} tables for a @code{MERGE} table. +@item +@code{MERGE} tables uses more file descriptors. If you are using a +@strong{MERGE} that maps over 10 tables and 10 users are using this, you +are using 10*10 + 10 file descriptors. (10 data files for 10 users +and 10 shared index files.) +@item +Key reads are slower. When you do a read on a key, the @code{MERGE} +handler will need to issue a read on all underlying tables to check +which one most closely matches the given key. If you then do a 'read-next' +then the merge table handler will need to search the read buffers +to find the next key. Only when one key buffer is used up, the handler +will need to read the next key block. This makes @code{MERGE} keys much slower +on @code{eq_ref} searches, but not much slower on @code{ref} searches. +@xref{EXPLAIN}. +@item +You can't do @code{DROP TABLE}, @code{ALTER TABLE} or @code{DELETE FROM +table_name} without a @code{WHERE} clause on any of the table that is +mapped by a @code{MERGE} table that is 'open'. If you do this, the +@code{MERGE} table may still refer to the original table and you will +get unexpected results. +@end itemize + +The following example shows you how to use @code{MERGE} tables: @example -mysql> SELECT * FROM pet WHERE species = "dog" AND sex = "f"; -+-------+--------+---------+------+------------+-------+ -| name | owner | species | sex | birth | death | -+-------+--------+---------+------+------------+-------+ -| Buffy | Harold | dog | f | 1989-05-13 | NULL | -+-------+--------+---------+------+------------+-------+ +CREATE TABLE t1 (a INT AUTO_INCREMENT PRIMARY KEY, message CHAR(20)); +CREATE TABLE t2 (a INT AUTO_INCREMENT PRIMARY KEY, message CHAR(20)); +INSERT INTO t1 (message) VALUES ("Testing"),("table"),("t1"); +INSERT INTO t2 (message) VALUES ("Testing"),("table"),("t2"); +CREATE TABLE total (a INT NOT NULL, message CHAR(20), KEY(a)) TYPE=MERGE UNION=(t1,t2); @end example -The preceding query uses the @code{AND} logical operator. There is also an -@code{OR} operator: +Note that we didn't create a @code{UNIQUE} or @code{PRIMARY KEY} in the +@code{total} table as the key isn't going to be unique in the @code{total} +table. + +Note that you can also manipulate the @code{.MRG} file directly from +the outside of the @strong{MySQL} server: @example -mysql> SELECT * FROM pet WHERE species = "snake" OR species = "bird"; -+----------+-------+---------+------+------------+-------+ -| name | owner | species | sex | birth | death | -+----------+-------+---------+------+------------+-------+ -| Chirpy | Gwen | bird | f | 1998-09-11 | NULL | -| Whistler | Gwen | bird | NULL | 1997-12-09 | NULL | -| Slim | Benny | snake | m | 1996-04-29 | NULL | -+----------+-------+---------+------+------------+-------+ +shell> cd /mysql-data-directory/current-database +shell> ls -1 t1.MYI t2.MYI > total.MRG +shell> mysqladmin flush-tables @end example -@code{AND} and @code{OR} may be intermixed. If you do that, it's a good idea -to use parentheses to indicate how conditions should be grouped: +Now you can do things like: @example -mysql> SELECT * FROM pet WHERE (species = "cat" AND sex = "m") - -> OR (species = "dog" AND sex = "f"); -+-------+--------+---------+------+------------+-------+ -| name | owner | species | sex | birth | death | -+-------+--------+---------+------+------------+-------+ -| Claws | Gwen | cat | m | 1994-03-17 | NULL | -| Buffy | Harold | dog | f | 1989-05-13 | NULL | -+-------+--------+---------+------+------------+-------+ +mysql> select * from total; ++---+---------+ +| a | message | ++---+---------+ +| 1 | Testing | +| 2 | table | +| 3 | t1 | +| 1 | Testing | +| 2 | table | +| 3 | t2 | ++---+---------+ @end example -@cindex columns, selecting -@cindex tables, selecting columns -@node Selecting columns, Sorting rows, Selecting rows, Retrieving data -@subsubsection Selecting Particular Columns +To remap a @code{MERGE} table you can do one of the following: +@itemize @bullet +@item +@code{DROP} the table and re-create it +@item +Use @code{ALTER TABLE table_name UNION(...)} +@item +Change the @code{.MRG} file and issue a @code{FLUSH TABLE} on the +@code{MERGE} table and all underlying tables to force the handler to +read the new definition file. +@end itemize -If you don't want to see entire rows from your table, just name the columns -in which you're interested, separated by commas. For example, if you want to -know when your animals were born, select the @code{name} and @code{birth} -columns: +@cindex tables, ISAM +@node ISAM, HEAP, MERGE, Table types +@section ISAM Tables -@example -mysql> SELECT name, birth FROM pet; -+----------+------------+ -| name | birth | -+----------+------------+ -| Fluffy | 1993-02-04 | -| Claws | 1994-03-17 | -| Buffy | 1989-05-13 | -| Fang | 1990-08-27 | -| Bowser | 1989-08-31 | -| Chirpy | 1998-09-11 | -| Whistler | 1997-12-09 | -| Slim | 1996-04-29 | -| Puffball | 1999-03-30 | -+----------+------------+ -@end example +You can also use the deprecated ISAM table type. This will disappear +rather soon because @code{MyISAM} is a better implementation of the same +thing. ISAM uses a @code{B-tree} index. The index is stored in a file +with the @code{.ISM} extension, and the data is stored in a file with the +@code{.ISD} extension. You can check/repair ISAM tables with the +@code{isamchk} utility. @xref{Crash recovery}. -To find out who owns pets, use this query: +@code{ISAM} has the following features/properties: -@example -mysql> SELECT owner FROM pet; -+--------+ -| owner | -+--------+ -| Harold | -| Gwen | -| Harold | -| Benny | -| Diane | -| Gwen | -| Gwen | -| Benny | -| Diane | -+--------+ -@end example +@itemize @bullet +@item Compressed and fixed-length keys +@item Fixed and dynamic record length +@item 16 keys with 16 key parts/key +@item Max key length 256 (default) +@item Data is stored in machine format; this is fast, but is machine/OS dependent. +@end itemize -@findex DISTINCT -However, notice that the query simply retrieves the @code{owner} field from -each record, and some of them appear more than once. To minimize the output, -retrieve each unique output record just once by adding the keyword -@code{DISTINCT}: +Most of the things true for @code{MyISAM} tables are also true for @code{ISAM} +tables. @xref{MyISAM}. The major differences compared to @code{MyISAM} +tables are: -@example -mysql> SELECT DISTINCT owner FROM pet; -+--------+ -| owner | -+--------+ -| Benny | -| Diane | -| Gwen | -| Harold | -+--------+ -@end example +@itemize @bullet +@item ISAM tables are not binary portable across OS/Platforms. +@item Can't handle tables > 4G. +@item Only support prefix compression on strings. +@item Smaller key limits. +@item Dynamic tables get more fragmented. +@item Tables are compressed with @code{pack_isam} rather than with @code{myisampack}. +@end itemize -You can use a @code{WHERE} clause to combine row selection with column -selection. For example, to get birth dates for dogs and cats only, -use this query: +If you want to convert an @code{ISAM} table to a @code{MyISAM} table so +that you can use utilities such as @code{mysqlcheck}, use an @code{ALTER +TABLE} statement: @example -mysql> SELECT name, species, birth FROM pet - -> WHERE species = "dog" OR species = "cat"; -+--------+---------+------------+ -| name | species | birth | -+--------+---------+------------+ -| Fluffy | cat | 1993-02-04 | -| Claws | cat | 1994-03-17 | -| Buffy | dog | 1989-05-13 | -| Fang | dog | 1990-08-27 | -| Bowser | dog | 1989-08-31 | -+--------+---------+------------+ +mysql> ALTER TABLE tbl_name TYPE = MYISAM; @end example -@cindex rows, sorting -@cindex sorting, table rows -@cindex sorting, data -@cindex tables, sorting rows -@node Sorting rows, Date calculations, Selecting columns, Retrieving data -@subsubsection Sorting Rows +@cindex tables, @code{HEAP} +@node HEAP, InnoDB, ISAM, Table types +@section HEAP Tables -You may have noticed in the preceding examples that the result rows are -displayed in no particular order. However, it's often easier to examine -query output when the rows are sorted in some meaningful way. To sort a -result, use an @code{ORDER BY} clause. +@code{HEAP} tables use a hashed index and are stored in memory. This +makes them very fast, but if @strong{MySQL} crashes you will lose all +data stored in them. @code{HEAP} is very useful for temporary tables! -Here are animal birthdays, sorted by date: +The @strong{MySQL} internal HEAP tables use 100% dynamic hashing +without overflow areas. There is no extra space needed for free lists. +@code{HEAP} tables also don't have problems with delete + inserts, which +normally is common with hashed tables: @example -mysql> SELECT name, birth FROM pet ORDER BY birth; -+----------+------------+ -| name | birth | -+----------+------------+ -| Buffy | 1989-05-13 | -| Bowser | 1989-08-31 | -| Fang | 1990-08-27 | -| Fluffy | 1993-02-04 | -| Claws | 1994-03-17 | -| Slim | 1996-04-29 | -| Whistler | 1997-12-09 | -| Chirpy | 1998-09-11 | -| Puffball | 1999-03-30 | -+----------+------------+ +mysql> CREATE TABLE test TYPE=HEAP SELECT ip,SUM(downloads) as down + FROM log_table GROUP BY ip; +mysql> SELECT COUNT(ip),AVG(down) FROM test; +mysql> DROP TABLE test; @end example -To sort in reverse order, add the @code{DESC} (descending) keyword to the -name of the column you are sorting by: +Here are some things you should consider when you use @code{HEAP} tables: -@example -mysql> SELECT name, birth FROM pet ORDER BY birth DESC; -+----------+------------+ -| name | birth | -+----------+------------+ -| Puffball | 1999-03-30 | -| Chirpy | 1998-09-11 | -| Whistler | 1997-12-09 | -| Slim | 1996-04-29 | -| Claws | 1994-03-17 | -| Fluffy | 1993-02-04 | -| Fang | 1990-08-27 | -| Bowser | 1989-08-31 | -| Buffy | 1989-05-13 | -+----------+------------+ -@end example +@itemize @bullet +@item +You should always use specify @code{MAX_ROWS} in the @code{CREATE} statement +to ensure that you accidentally do not use all memory. +@item +Indexes will only be used with @code{=} and @code{<=>} (but are VERY fast). +@item +@code{HEAP} tables can only use whole keys to search for a row; compare this +to @code{MyISAM} tables where any prefix of the key can be used to find rows. +@item +@code{HEAP} tables use a fixed record length format. +@item +@code{HEAP} doesn't support @code{BLOB}/@code{TEXT} columns. +@item +@code{HEAP} doesn't support @code{AUTO_INCREMENT} columns. +@item +@code{HEAP} doesn't support an index on a @code{NULL} column. +@item +You can have non-unique keys in a @code{HEAP} table (this isn't common for +hashed tables). +@item +@code{HEAP} tables are shared between all clients (just like any other +table). +@item +You can't search for the next entry in order (that is, to use the index +to do an @code{ORDER BY}). +@item +Data for @code{HEAP} tables are allocated in small blocks. The tables +are 100% dynamic (on inserting). No overflow areas and no extra key +space are needed. Deleted rows are put in a linked list and are +reused when you insert new data into the table. +@item +You need enough extra memory for all HEAP tables that you want to use at +the same time. +@item +To free memory, you should execute @code{DELETE FROM heap_table}, +@code{TRUNCATE heap_table} or @code{DROP TABLE heap_table}. +@item +@strong{MySQL} cannot find out approximately how many rows there +are between two values (this is used by the range optimizer to decide which +index to use). This may affect some queries if you change a @code{MyISAM} +table to a @code{HEAP} table. +@item +To ensure that you accidentally don't do anything foolish, you can't create +@code{HEAP} tables bigger than @code{max_heap_table_size}. +@end itemize -You can sort on multiple columns. For example, to sort by type of -animal, then by birth date within animal type with youngest animals first, -use the following query: +The memory needed for one row in a @code{HEAP} table is: @example -mysql> SELECT name, species, birth FROM pet ORDER BY species, birth DESC; -+----------+---------+------------+ -| name | species | birth | -+----------+---------+------------+ -| Chirpy | bird | 1998-09-11 | -| Whistler | bird | 1997-12-09 | -| Claws | cat | 1994-03-17 | -| Fluffy | cat | 1993-02-04 | -| Fang | dog | 1990-08-27 | -| Bowser | dog | 1989-08-31 | -| Buffy | dog | 1989-05-13 | -| Puffball | hamster | 1999-03-30 | -| Slim | snake | 1996-04-29 | -+----------+---------+------------+ +SUM_OVER_ALL_KEYS(max_length_of_key + sizeof(char*) * 2) ++ ALIGN(length_of_row+1, sizeof(char*)) @end example -Note that the @code{DESC} keyword applies only to the column name immediately -preceding it (@code{birth}); @code{species} values are still sorted in -ascending order. +@code{sizeof(char*)} is 4 on 32-bit machines and 8 on 64-bit machines. -@cindex date calculations -@cindex calculating, dates -@cindex extracting, dates -@cindex age, calculating -@node Date calculations, Working with NULL, Sorting rows, Retrieving data -@subsubsection Date Calculations +@node InnoDB, BDB, HEAP, Table types +@section InnoDB Tables -@strong{MySQL} provides several functions that you can use to perform -calculations on dates, for example, to calculate ages or extract -parts of dates. +@menu +* InnoDB overview:: InnoDB tables overview +* InnoDB start:: InnoDB startup options +* InnoDB init:: Creating InnoDB table space. +* Using InnoDB tables:: Creating InnoDB tables +* Adding and removing:: Adding and removing InnoDB data and log files +* Backing up:: Backing up and recovering an InnoDB database +* Moving:: Moving an InnoDB database to another machine +* InnoDB transaction model:: InnoDB transaction model. +* Implementation:: Implementation of multiversioning +* Table and index:: Table and index structures +* File space management:: File space management and disk i/o +* Error handling:: Error handling +* InnoDB restrictions:: Some restrictions on InnoDB tables +* InnoDB contact information:: InnoDB contact information. +@end menu -To determine how many years old each of your pets is, compute age as the -difference between the birth date and the current date. Do this by -converting the two dates to days, take the difference, and divide by 365 (the -number of days in a year): +@node InnoDB overview, InnoDB start, InnoDB, InnoDB +@subsection InnoDB tables overview -@example -mysql> SELECT name, (TO_DAYS(NOW())-TO_DAYS(birth))/365 FROM pet; -+----------+-------------------------------------+ -| name | (TO_DAYS(NOW())-TO_DAYS(birth))/365 | -+----------+-------------------------------------+ -| Fluffy | 6.15 | -| Claws | 5.04 | -| Buffy | 9.88 | -| Fang | 8.59 | -| Bowser | 9.58 | -| Chirpy | 0.55 | -| Whistler | 1.30 | -| Slim | 2.92 | -| Puffball | 0.00 | -+----------+-------------------------------------+ -@end example +InnoDB tables are included in the @strong{MySQL} source distribution +starting from 3.23.34a and are activated in the @strong{MySQL -max} +binary. -Although the query works, there are some things about it that could be -improved. First, the result could be scanned more easily if the rows were -presented in some order. Second, the heading for the age column isn't very -meaningful. +If you have downloaded a binary version of @strong{MySQL} that includes +support for InnoDB (mysqld-max), simply follow the instructions for +installing a binary version of @strong{MySQL}. @xref{Installing binary}. +@xref{mysqld-max, , @code{mysqld-max}}. -The first problem can be handled by adding an @code{ORDER BY name} clause to -sort the output by name. To deal with the column heading, provide a name for -the column so that a different label appears in the output (this is called a -column alias): +To compile @strong{MySQL} with InnoDB support, download MySQL-3.23.37 or newer +and configure @code{MySQL} with the @code{--with-innodb} option. +@xref{Installing source}. @example -mysql> SELECT name, (TO_DAYS(NOW())-TO_DAYS(birth))/365 AS age - -> FROM pet ORDER BY name; -+----------+------+ -| name | age | -+----------+------+ -| Bowser | 9.58 | -| Buffy | 9.88 | -| Chirpy | 0.55 | -| Claws | 5.04 | -| Fang | 8.59 | -| Fluffy | 6.15 | -| Puffball | 0.00 | -| Slim | 2.92 | -| Whistler | 1.30 | -+----------+------+ +cd /path/to/source/of/mysql-3.23.37 +./configure --with-innodb @end example -To sort the output by @code{age} rather than @code{name}, just use a -different @code{ORDER BY} clause: +InnoDB provides @strong{MySQL} with a transaction-safe table handler with +commit, rollback, and crash recovery capabilities. InnoDB does +locking on row level, and also provides an Oracle-style consistent +non-locking read in @code{SELECTS}, which increases transaction +concurrency. There is not need for lock escalation in InnoDB, +because row level locks in InnoDB fit in very small space. -@example -mysql> SELECT name, (TO_DAYS(NOW())-TO_DAYS(birth))/365 AS age - -> FROM pet ORDER BY age; -+----------+------+ -| name | age | -+----------+------+ -| Puffball | 0.00 | -| Chirpy | 0.55 | -| Whistler | 1.30 | -| Slim | 2.92 | -| Claws | 5.04 | -| Fluffy | 6.15 | -| Fang | 8.59 | -| Bowser | 9.58 | -| Buffy | 9.88 | -+----------+------+ -@end example +InnoDB has been designed for maximum performance when processing +large data volumes. Its CPU efficiency is probably not +matched by any other disk-based relational database engine. -A similar query can be used to determine age at death for animals that have -died. You determine which animals these are by checking whether or not the -@code{death} value is @code{NULL}. Then, for those with non-@code{NULL} -values, compute the difference between the @code{death} and @code{birth} -values: +You can find the latest information about InnoDB at +@uref{http://www.innodb.com}. The most up-to-date version of the +InnoDB manual is always placed there, and you can also order commercial +support for InnoDB. -@example -mysql> SELECT name, birth, death, (TO_DAYS(death)-TO_DAYS(birth))/365 AS age - -> FROM pet WHERE death IS NOT NULL ORDER BY age; -+--------+------------+------------+------+ -| name | birth | death | age | -+--------+------------+------------+------+ -| Bowser | 1989-08-31 | 1995-07-29 | 5.91 | -+--------+------------+------------+------+ -@end example +Technically, InnoDB is a database backend placed under @strong{MySQL}. InnoDB +has its own buffer pool for caching data and indexes in main +memory. InnoDB stores its tables and indexes in a tablespace, which +may consist of several files. This is different from, for example, +@code{MyISAM} tables where each table is stored as a separate file. -The query uses @code{death IS NOT NULL} rather than @code{death != NULL} -because @code{NULL} is a special value. This is explained later. -@xref{Working with NULL, , Working with @code{NULL}}. +InnoDB is distributed under the GNU GPL License Version 2 (of June 1991). +In the source distribution of @strong{MySQL}, InnoDB appears as a subdirectory. -What if you want to know which animals have birthdays next month? For this -type of calculation, year and day are irrelevant; you simply want to extract -the month part of the @code{birth} column. @strong{MySQL} provides several -date-part extraction functions, such as @code{YEAR()}, @code{MONTH()}, and -@code{DAYOFMONTH()}. @code{MONTH()} is the appropriate function here. To -see how it works, run a simple query that displays the value of both -@code{birth} and @code{MONTH(birth)}: +@node InnoDB start, InnoDB init, InnoDB overview, InnoDB +@subsection InnoDB startup options -@example -mysql> SELECT name, birth, MONTH(birth) FROM pet; -+----------+------------+--------------+ -| name | birth | MONTH(birth) | -+----------+------------+--------------+ -| Fluffy | 1993-02-04 | 2 | -| Claws | 1994-03-17 | 3 | -| Buffy | 1989-05-13 | 5 | -| Fang | 1990-08-27 | 8 | -| Bowser | 1989-08-31 | 8 | -| Chirpy | 1998-09-11 | 9 | -| Whistler | 1997-12-09 | 12 | -| Slim | 1996-04-29 | 4 | -| Puffball | 1999-03-30 | 3 | -+----------+------------+--------------+ -@end example - -Finding animals with birthdays in the upcoming month is easy, too. Suppose -the current month is April. Then the month value is @code{4} and you look -for animals born in May (month 5) like this: +Beginning from @strong{MySQL}-3.23.37 the prefix of the options is changed +from @code{innobase_...} to @code{innodb_...}. -@example -mysql> SELECT name, birth FROM pet WHERE MONTH(birth) = 5; -+-------+------------+ -| name | birth | -+-------+------------+ -| Buffy | 1989-05-13 | -+-------+------------+ -@end example +To use InnoDB tables you @strong{MUST} specify configuration parameters +in the @strong{MySQL} configuration file in the @code{[mysqld]} section of +the configuration file @file{my.cnf}. @xref{Option files}. -There is a small complication if the current month is December, of course. -You don't just add one to the month number (@code{12}) and look for animals -born in month 13, because there is no such month. Instead, you look for -animals born in January (month 1). +The only required parameter to use InnoDB is @code{innodb_data_file_path}, +but you should set others if you want to get a better performance. -You can even write the query so that it works no matter what the current -month is. That way you don't have to use a particular month number -in the query. @code{DATE_ADD()} allows you to add a time interval to a -given date. If you add a month to the value of @code{NOW()}, then extract -the month part with @code{MONTH()}, the result produces the month in which to -look for birthdays: +Suppose you have a Windows NT machine with 128 MB RAM and a single 10 GB +hard disk. Below is an example of possible configuration parameters in +@file{my.cnf} for InnoDB: @example -mysql> SELECT name, birth FROM pet - -> WHERE MONTH(birth) = MONTH(DATE_ADD(NOW(), INTERVAL 1 MONTH)); +innodb_data_file_path = ibdata1:2000M;ibdata2:2000M +innodb_data_home_dir = c:\ibdata +set-variable = innodb_mirrored_log_groups=1 +innodb_log_group_home_dir = c:\iblogs +set-variable = innodb_log_files_in_group=3 +set-variable = innodb_log_file_size=30M +set-variable = innodb_log_buffer_size=8M +innodb_flush_log_at_trx_commit=1 +innodb_log_arch_dir = c:\iblogs +innodb_log_archive=0 +set-variable = innodb_buffer_pool_size=80M +set-variable = innodb_additional_mem_pool_size=10M +set-variable = innodb_file_io_threads=4 +set-variable = innodb_lock_wait_timeout=50 @end example -A different way to accomplish the same task is to add @code{1} to get the -next month after the current one (after using the modulo function (@code{MOD}) -to wrap around the month value to @code{0} if it is currently -@code{12}): +Suppose you have a Linux machine with 512 MB RAM and +three 20 GB hard disks (at directory paths @file{/}, +@file{/dr2} and @file{/dr3}). +Below is an example of possible configuration parameters in @file{my.cnf} for +InnoDB: @example -mysql> SELECT name, birth FROM pet - -> WHERE MONTH(birth) = MOD(MONTH(NOW()), 12) + 1; +innodb_data_file_path = ibdata/ibdata1:2000M;dr2/ibdata/ibdata2:2000M +innodb_data_home_dir = / +set-variable = innodb_mirrored_log_groups=1 +innodb_log_group_home_dir = /dr3 +set-variable = innodb_log_files_in_group=3 +set-variable = innodb_log_file_size=50M +set-variable = innodb_log_buffer_size=8M +innodb_flush_log_at_trx_commit=1 +innodb_log_arch_dir = /dr3/iblogs +innodb_log_archive=0 +set-variable = innodb_buffer_pool_size=400M +set-variable = innodb_additional_mem_pool_size=20M +set-variable = innodb_file_io_threads=4 +set-variable = innodb_lock_wait_timeout=50 @end example -Note that @code{MONTH} returns a number between 1 and 12. And -@code{MOD(something,12)} returns a number between 0 and 11. So the -addition has to be after the @code{MOD()} otherwise we would go from -November (11) to January (1). +Note that we have placed the two data files on different disks. +The reason for the name @code{innodb_data_file_path} is that +you can also specify paths to your data files, and +@code{innodb_data_home_dir} is just textually catenated +before your data file paths, adding a possible slash or +backslash in between. InnoDB will fill the tablespace +formed by the data files from bottom up. In some cases it will +improve the performance of the database if all data is not placed +on the same physical disk. Putting log files on a different disk from +data is very often beneficial for performance. -@findex NULL -@cindex NULL value -@node Working with NULL, Pattern matching, Date calculations, Retrieving data -@subsubsection Working with @code{NULL} Values +The meanings of the configuration parameters are the following: -The @code{NULL} value can be surprising until you get used to it. -Conceptually, @code{NULL} means missing value or unknown value and it -is treated somewhat differently than other values. To test for @code{NULL}, -you cannot use the arithmetic comparison operators such as @code{=}, @code{<}, -or @code{!=}. To demonstrate this for yourself, try the following query: +@multitable @columnfractions .30 .70 +@item @code{innodb_data_home_dir} @tab +The common part of the directory path for all InnoDB data files. +@item @code{innodb_data_file_path} @tab +Paths to individual data files and their sizes. The full directory path +to each data file is acquired by concatenating innodb_data_home_dir to +the paths specified here. The file sizes are specified in megabytes, +hence the 'M' after the size specification above. Do not set a file size +bigger than 4000M, and on most operating systems not bigger than 2000M. +InnoDB also understands the abbreviation 'G', 1G meaning 1024M. +The sum of the sizes of the files must be at least 10 MB. +@item @code{innodb_mirrored_log_groups} @tab +Number of identical copies of log groups we +keep for the database. Currently this should be set to 1. +@item @code{innodb_log_group_home_dir} @tab +Directory path to InnoDB log files. +@item @code{innodb_log_files_in_group} @tab +Number of log files in the log group. InnoDB writes to the files in a +circular fashion. Value 3 is recommended here. +@item @code{innodb_log_file_size} @tab +Size of each log file in a log group in megabytes. Sensible values range +from 1M to the size of the buffer pool specified below. The bigger the +value, the less checkpoint flush activity is needed in the buffer pool, +saving disk i/o. But bigger log files also mean that recovery will be +slower in case of a crash. File size restriction as for a data file. +@item @code{innodb_log_buffer_size} @tab +The size of the buffer which InnoDB uses to write log to the log files +on disk. Sensible values range from 1M to half the combined size of log +files. A big log buffer allows large transactions to run without a need +to write the log to disk until the transaction commit. Thus, if you have +big transactions, making the log buffer big will save disk i/o. +@item @code{innodb_flush_log_at_trx_commit} @tab +Normally this is set to 1, meaning that at a transaction commit the log +is flushed to disk, and the modifications made by the transaction become +permanent, and survive a database crash. If you are willing to +compromise this safety, and you are running small transactions, you may +set this to 0 to reduce disk i/o to the logs. +@item @code{innodb_log_arch_dir} @tab +The directory where fully written log files would be archived if we used +log archiving. The value of this parameter should currently be set the +same as @code{innodb_log_group_home_dir}. +@item @code{innodb_log_archive} @tab +This value should currently be set to 0. As recovery from a backup is +done by @strong{MySQL} using its own log files, there is currently no need to +archive InnoDB log files. +@item @code{innodb_buffer_pool_size} @tab +The size of the memory buffer InnoDB uses to cache data and indexes of +its tables. The bigger you set this the less disk i/o is needed to +access data in tables. On a dedicated database server you may set this +parameter up to 90 % of the machine physical memory size. Do not set it +too large, though, because competition of the physical memory may cause +paging in the operating system. +@item @code{innodb_additional_mem_pool_size} @tab +Size of a memory pool InnoDB uses to store data dictionary information +and other internal data structures. A sensible value for this might be +2M, but the more tables you have in your application the more you will +need to allocate here. If InnoDB runs out of memory in this pool, it +will start to allocate memory from the operating system, and write +warning messages to the @strong{MySQL} error log. +@item @code{innodb_file_io_threads} @tab +Number of file i/o threads in InnoDB. Normally, this should be 4, but +on Windows NT disk i/o may benefit from a larger number. +@item @code{innodb_lock_wait_timeout} @tab +Timeout in seconds an InnoDB transaction may wait for a lock before +being rolled back. InnoDB automatically detects transaction deadlocks +in its own lock table and rolls back the transaction. If you use +@code{LOCK TABLES} command, or other transaction-safe table handlers +than InnoDB in the same transaction, then a deadlock may arise which +InnoDB cannot notice. In cases like this the timeout is useful to +resolve the situation. +@item @code{innodb_flush_method} @tab +(Available from 3.23.40 up.) +The default value for this is @code{fdatasync}. +Another option is @code{O_DSYNC}. +Options @code{littlesync} and @code{nosync} have the +risk that in an operating system crash or a power outage you may easily +end up with a half-written database page, and you have to do a recovery +from a backup. See the section "InnoDB performance tuning", item 6, below +for tips on how to set this parameter. If you are happy with your database +performance it is wisest not to specify this parameter at all, in which +case it will get the default value. +@end multitable + +@node InnoDB init, Using InnoDB tables, InnoDB start, InnoDB +@subsection Creating InnoDB table space + +Suppose you have installed @strong{MySQL} and have edited @file{my.cnf} so that +it contains the necessary InnoDB configuration parameters. +Before starting @strong{MySQL} you should check that the directories you have +specified for InnoDB data files and log files exist and that you have +access rights to those directories. InnoDB +cannot create directories, only files. Check also you have enough disk space +for the data and log files. + +When you now start @strong{MySQL}, InnoDB will start creating your data files +and log files. InnoDB will print something like the following: @example -mysql> SELECT 1 = NULL, 1 != NULL, 1 < NULL, 1 > NULL; -+----------+-----------+----------+----------+ -| 1 = NULL | 1 != NULL | 1 < NULL | 1 > NULL | -+----------+-----------+----------+----------+ -| NULL | NULL | NULL | NULL | -+----------+-----------+----------+----------+ +~/mysqlm/sql > mysqld +InnoDB: The first specified data file /home/heikki/data/ibdata1 did not exist: +InnoDB: a new database to be created! +InnoDB: Setting file /home/heikki/data/ibdata1 size to 134217728 +InnoDB: Database physically writes the file full: wait... +InnoDB: Data file /home/heikki/data/ibdata2 did not exist: new to be created +InnoDB: Setting file /home/heikki/data/ibdata2 size to 262144000 +InnoDB: Database physically writes the file full: wait... +InnoDB: Log file /home/heikki/data/logs/ib_logfile0 did not exist: new to be c +reated +InnoDB: Setting log file /home/heikki/data/logs/ib_logfile0 size to 5242880 +InnoDB: Log file /home/heikki/data/logs/ib_logfile1 did not exist: new to be c +reated +InnoDB: Setting log file /home/heikki/data/logs/ib_logfile1 size to 5242880 +InnoDB: Log file /home/heikki/data/logs/ib_logfile2 did not exist: new to be c +reated +InnoDB: Setting log file /home/heikki/data/logs/ib_logfile2 size to 5242880 +InnoDB: Started +mysqld: ready for connections @end example -Clearly you get no meaningful results from these comparisons. Use -the @code{IS NULL} and @code{IS NOT NULL} operators instead: +A new InnoDB database has now been created. You can connect to the @strong{MySQL} +server with the usual @strong{MySQL} client programs like @code{mysql}. +When you shut down the @strong{MySQL} server with @file{mysqladmin shutdown}, +InnoDB output will be like the following: @example -mysql> SELECT 1 IS NULL, 1 IS NOT NULL; -+-----------+---------------+ -| 1 IS NULL | 1 IS NOT NULL | -+-----------+---------------+ -| 0 | 1 | -+-----------+---------------+ +010321 18:33:34 mysqld: Normal shutdown +010321 18:33:34 mysqld: Shutdown Complete +InnoDB: Starting shutdown... +InnoDB: Shutdown completed @end example -In @strong{MySQL}, 0 or @code{NULL} means false and anything else means true. -The default truth value from a boolean operation is 1. +You can now look at the data files and logs directories and you +will see the files created. The log directory will also contain +a small file named @file{ib_arch_log_0000000000}. That file +resulted from the database creation, after which InnoDB switched off +log archiving. +When @strong{MySQL} is again started, the output will be like the following: -This special treatment of @code{NULL} is why, in the previous section, it -was necessary to determine which animals are no longer alive using -@code{death IS NOT NULL} instead of @code{death != NULL}. +@example +~/mysqlm/sql > mysqld +InnoDB: Started +mysqld: ready for connections +@end example -@cindex pattern matching -@cindex matching, patterns -@cindex expressions, extended -@node Pattern matching, Counting rows, Working with NULL, Retrieving data -@subsubsection Pattern Matching +@menu +* Error creating InnoDB:: +@end menu -@strong{MySQL} provides standard SQL pattern matching as well as a form of -pattern matching based on extended regular expressions similar to those used -by Unix utilities such as @code{vi}, @code{grep}, and @code{sed}. +@node Error creating InnoDB, , InnoDB init, InnoDB init +@subsubsection If something goes wrong in database creation -SQL pattern matching allows you to use @samp{_} to match any single -character and @samp{%} to match an arbitrary number of characters (including -zero characters). In @strong{MySQL}, SQL patterns are case insensitive by -default. Some examples are shown below. Note that you do not use @code{=} -or @code{!=} when you use SQL patterns; use the @code{LIKE} or @code{NOT -LIKE} comparison operators instead. +If something goes wrong in an InnoDB database creation, you should +delete all files created by InnoDB. This means all data files, all log +files, the small archived log file, and in the case you already did +create some InnoDB tables, delete also the corresponding @file{.frm} +files for these tables from the @strong{MySQL} database +directories. Then you can try the InnoDB database creation again. -To find names beginning with @samp{b}: +@node Using InnoDB tables, Adding and removing, InnoDB init, InnoDB +@subsection Creating InnoDB tables + +Suppose you have started the @strong{MySQL} client with the command +@code{mysql test}. +To create a table in the InnoDB format you must specify +@code{TYPE = InnoDB} in the table creation SQL command: @example -mysql> SELECT * FROM pet WHERE name LIKE "b%"; -+--------+--------+---------+------+------------+------------+ -| name | owner | species | sex | birth | death | -+--------+--------+---------+------+------------+------------+ -| Buffy | Harold | dog | f | 1989-05-13 | NULL | -| Bowser | Diane | dog | m | 1989-08-31 | 1995-07-29 | -+--------+--------+---------+------+------------+------------+ +CREATE TABLE CUSTOMER (A INT, B CHAR (20), INDEX (A)) TYPE = InnoDB; @end example -To find names ending with @samp{fy}: +This SQL command will create a table and an index on column @code{A} +into the InnoDB tablespace consisting of the data files you specified +in @file{my.cnf}. In addition @strong{MySQL} will create a file +@file{CUSTOMER.frm} to the @strong{MySQL} database directory @file{test}. +Internally, InnoDB will add to its own data dictionary an entry +for table @code{'test/CUSTOMER'}. Thus you can create a table +of the same name @code{CUSTOMER} in another database of @strong{MySQL}, and +the table names will not collide inside InnoDB. + +You can query the amount of free space in the InnoDB tablespace +by issuing the table status command of @strong{MySQL} for any table you have +created with @code{TYPE = InnoDB}. Then the amount of free +space in the tablespace appears in the table comment section in the +output of @code{SHOW}. An example: @example -mysql> SELECT * FROM pet WHERE name LIKE "%fy"; -+--------+--------+---------+------+------------+-------+ -| name | owner | species | sex | birth | death | -+--------+--------+---------+------+------------+-------+ -| Fluffy | Harold | cat | f | 1993-02-04 | NULL | -| Buffy | Harold | dog | f | 1989-05-13 | NULL | -+--------+--------+---------+------+------------+-------+ +SHOW TABLE STATUS FROM test LIKE 'CUSTOMER' @end example -To find names containing a @samp{w}: +Note that the statistics @code{SHOW} gives about InnoDB tables +are only approximate: they are used in SQL optimization. Table and +index reserved sizes in bytes are accurate, though. -@example -mysql> SELECT * FROM pet WHERE name LIKE "%w%"; -+----------+-------+---------+------+------------+------------+ -| name | owner | species | sex | birth | death | -+----------+-------+---------+------+------------+------------+ -| Claws | Gwen | cat | m | 1994-03-17 | NULL | -| Bowser | Diane | dog | m | 1989-08-31 | 1995-07-29 | -| Whistler | Gwen | bird | NULL | 1997-12-09 | NULL | -+----------+-------+---------+------+------------+------------+ -@end example +NOTE: @code{DROP DATABASE} does not currently work for InnoDB tables! +You must drop the tables individually. Also take care not to delete or +add @file{.frm} files to your InnoDB database manually: use +@code{CREATE TABLE} and @code{DROP TABLE} commands. +InnoDB has its own internal data dictionary, and you will get problems +if the @strong{MySQL} @file{.frm} files are out of 'sync' with the InnoDB +internal data dictionary. -To find names containing exactly five characters, use the @samp{_} pattern -character: +@subsubsection Converting MyISAM tables to InnoDB + +InnoDB does not have a special optimization for separate index creation. +Therefore it does not pay to export and import the table and create indexes +afterwards. +The fastest way to alter a table to InnoDB is to do the inserts +directly to an InnoDB table, that is, use @code{ALTER TABLE ... TYPE=INNODB}, +or create an empty InnoDB table with identical definitions and insert +the rows with @code{INSERT INTO ... SELECT * FROM ...}. + +To get better control over the insertion process, it may be good to insert +big tables in pieces: @example -mysql> SELECT * FROM pet WHERE name LIKE "_____"; -+-------+--------+---------+------+------------+-------+ -| name | owner | species | sex | birth | death | -+-------+--------+---------+------+------------+-------+ -| Claws | Gwen | cat | m | 1994-03-17 | NULL | -| Buffy | Harold | dog | f | 1989-05-13 | NULL | -+-------+--------+---------+------+------------+-------+ +INSERT INTO newtable SELECT * FROM oldtable WHERE yourkey > something + AND yourkey <= somethingelse; @end example -The other type of pattern matching provided by @strong{MySQL} uses extended -regular expressions. When you test for a match for this type of pattern, use -the @code{REGEXP} and @code{NOT REGEXP} operators (or @code{RLIKE} and -@code{NOT RLIKE}, which are synonyms). +After all data has been inserted you can rename the tables. -Some characteristics of extended regular expressions are: +During the conversion of big tables you should set the InnoDB +buffer pool size big +to reduce disk i/o. Not bigger than 80 % of the physical memory, though. +You should set InnoDB log files big, and also the log buffer large. -@itemize @bullet -@item -@samp{.} matches any single character. +Make sure you do not run out of tablespace: InnoDB tables take a lot +more space than MyISAM tables. If an @code{ALTER TABLE} runs out +of space, it will start a rollback, and that can take hours if it is +disk-bound. +In inserts InnoDB uses the insert buffer to merge secondary index records +to indexes in batches. That saves a lot of disk i/o. In rollback no such +mechanism is used, and the rollback can take 30 times longer than the +insertion. -@item -A character class @samp{[...]} matches any character within the brackets. -For example, @samp{[abc]} matches @samp{a}, @samp{b}, or @samp{c}. To name a -range of characters, use a dash. @samp{[a-z]} matches any lowercase letter, -whereas @samp{[0-9]} matches any digit. +In the case of a runaway rollback, if you do not have valuable data in your +database, +it is better that you kill the database process and delete all InnoDB data +and log files and all InnoDB table @file{.frm} files, and start +your job again, rather than wait for millions of disk i/os to complete. -@item -@samp{*} matches zero or more instances of the thing preceding it. For -example, @samp{x*} matches any number of @samp{x} characters, -@samp{[0-9]*} matches any number of digits, and @samp{.*} matches any -number of anything. +@node Adding and removing, Backing up, Using InnoDB tables, InnoDB +@subsection Adding and removing InnoDB data and log files -@item -Regular expressions are case sensitive, but you can use a character class to -match both lettercases if you wish. For example, @samp{[aA]} matches -lowercase or uppercase @samp{a} and @samp{[a-zA-Z]} matches any letter in -either case. +You cannot increase the size of an InnoDB data file. To add more into +your tablespace you have to add a new data file. To do this you have to +shut down your @strong{MySQL} database, edit the @file{my.cnf} file, adding a +new file to @code{innodb_data_file_path}, and then start @strong{MySQL} +again. -@item -The pattern matches if it occurs anywhere in the value being tested. -(SQL patterns match only if they match the entire value.) +Currently you cannot remove a data file from InnoDB. To decrease the +size of your database you have to use @code{mysqldump} to dump +all your tables, create a new database, and import your tables to the +new database. -@item -To anchor a pattern so that it must match the beginning or end of the value -being tested, use @samp{^} at the beginning or @samp{$} at the end of the -pattern. -@end itemize +If you want to change the number or the size of your InnoDB log files, +you have to shut down @strong{MySQL} and make sure that it shuts down without errors. +Then copy the old log files into a safe place just in case something +went wrong in the shutdown and you will need them to recover the +database. Delete then the old log files from the log file directory, +edit @file{my.cnf}, and start @strong{MySQL} again. InnoDB will tell +you at the startup that it is creating new log files. -To demonstrate how extended regular expressions work, the @code{LIKE} queries -shown above are rewritten below to use @code{REGEXP}. +@node Backing up, Moving, Adding and removing, InnoDB +@subsection Backing up and recovering an InnoDB database -To find names beginning with @samp{b}, use @samp{^} to match the beginning of -the name: +The key to safe database management is taking regular backups. +To take a 'binary' backup of your database you have to do the following: -@example -mysql> SELECT * FROM pet WHERE name REGEXP "^b"; -+--------+--------+---------+------+------------+------------+ -| name | owner | species | sex | birth | death | -+--------+--------+---------+------+------------+------------+ -| Buffy | Harold | dog | f | 1989-05-13 | NULL | -| Bowser | Diane | dog | m | 1989-08-31 | 1995-07-29 | -+--------+--------+---------+------+------------+------------+ -@end example +@itemize @bullet +@item +Shut down your @strong{MySQL} database and make sure it shuts down without errors. +@item +Copy all your data files into a safe place. +@item +Copy all your InnoDB log files to a safe place. +@item +Copy your @file{my.cnf} configuration file(s) to a safe place. +@item +Copy all the @file{.frm} files for your InnoDB tables into a +safe place. +@end itemize -Prior to @strong{MySQL} Version 3.23.4, @code{REGEXP} is case sensitive, -and the previous query will return no rows. To match either lowercase or -uppercase @samp{b}, use this query instead: +There is currently no on-line or incremental backup tool available for +InnoDB, though they are in the TODO list. -@example -mysql> SELECT * FROM pet WHERE name REGEXP "^[bB]"; -@end example +In addition to taking the binary backups described above, +you should also regularly take dumps of your tables with +@file{mysqldump}. The reason to this is that a binary file +may be corrupted without you noticing it. Dumped tables are stored +into text files which are human-readable and much simpler than +database binary files. Seeing table corruption from dumped files +is easier, and since their format is simpler, the chance for +serious data corruption in them is smaller. -From @strong{MySQL} 3.23.4 on, to force a @code{REGEXP} comparison to -be case sensitive, use the @code{BINARY} keyword to make one of the -strings a binary string. This query will match only lowercase @samp{b} -at the beginning of a name: +A good idea is to take the dumps at the same time you take a binary +backup of your database. You have to shut out all clients from your +database to get a consistent snapshot of all your tables into your +dumps. Then you can take the binary backup, and you will then have +a consistent snapshot of your database in two formats. -@example -mysql> SELECT * FROM pet WHERE name REGEXP BINARY "^b"; -@end example +To be able to recover your InnoDB database to the present from the +binary backup described above, you have to run your @strong{MySQL} database +with the general logging and log archiving of @strong{MySQL} switched on. Here +by the general logging we mean the logging mechanism of the @strong{MySQL} server +which is independent of InnoDB logs. -To find names ending with @samp{fy}, use @samp{$} to match the end of the -name: +To recover from a crash of your @strong{MySQL} server process, the only thing +you have to do is to restart it. InnoDB will automatically check the +logs and perform a roll-forward of the database to the present. +InnoDB will automatically roll back uncommitted transactions which were +present at the time of the crash. During recovery, InnoDB will print +out something like the following: @example -mysql> SELECT * FROM pet WHERE name REGEXP "fy$"; -+--------+--------+---------+------+------------+-------+ -| name | owner | species | sex | birth | death | -+--------+--------+---------+------+------------+-------+ -| Fluffy | Harold | cat | f | 1993-02-04 | NULL | -| Buffy | Harold | dog | f | 1989-05-13 | NULL | -+--------+--------+---------+------+------------+-------+ +~/mysqlm/sql > mysqld +InnoDB: Database was not shut down normally. +InnoDB: Starting recovery from log files... +InnoDB: Starting log scan based on checkpoint at +InnoDB: log sequence number 0 13674004 +InnoDB: Doing recovery: scanned up to log sequence number 0 13739520 +InnoDB: Doing recovery: scanned up to log sequence number 0 13805056 +InnoDB: Doing recovery: scanned up to log sequence number 0 13870592 +InnoDB: Doing recovery: scanned up to log sequence number 0 13936128 +... +InnoDB: Doing recovery: scanned up to log sequence number 0 20555264 +InnoDB: Doing recovery: scanned up to log sequence number 0 20620800 +InnoDB: Doing recovery: scanned up to log sequence number 0 20664692 +InnoDB: 1 uncommitted transaction(s) which must be rolled back +InnoDB: Starting rollback of uncommitted transactions +InnoDB: Rolling back trx no 16745 +InnoDB: Rolling back of trx no 16745 completed +InnoDB: Rollback of uncommitted transactions completed +InnoDB: Starting an apply batch of log records to the database... +InnoDB: Apply batch completed +InnoDB: Started +mysqld: ready for connections @end example -To find names containing a lowercase or uppercase @samp{w}, use this query: +If your database gets corrupted or your disk fails, you have +to do the recovery from a backup. In the case of corruption, you should +first find a backup which is not corrupted. From a backup do the recovery +from the general log files of @strong{MySQL} according to instructions in the +MySQL manual. -@example -mysql> SELECT * FROM pet WHERE name REGEXP "w"; -+----------+-------+---------+------+------------+------------+ -| name | owner | species | sex | birth | death | -+----------+-------+---------+------+------------+------------+ -| Claws | Gwen | cat | m | 1994-03-17 | NULL | -| Bowser | Diane | dog | m | 1989-08-31 | 1995-07-29 | -| Whistler | Gwen | bird | NULL | 1997-12-09 | NULL | -+----------+-------+---------+------+------------+------------+ -@end example +@menu +* InnoDB checkpoints:: +@end menu -Because a regular expression pattern matches if it occurs anywhere in the -value, it is not necessary in the previous query to put a wild card on either -side of the pattern to get it to match the entire value like it would be if -you used a SQL pattern. +@node InnoDB checkpoints, , Backing up, Backing up +@subsubsection Checkpoints -To find names containing exactly five characters, use @samp{^} and @samp{$} -to match the beginning and end of the name, and five instances of @samp{.} -in between: +InnoDB implements a checkpoint mechanism called a fuzzy +checkpoint. InnoDB will flush modified database pages from the buffer +pool in small batches, there is no need to flush the buffer pool +in one single batch, which would in practice stop processing +of user SQL statements for a while. -@example -mysql> SELECT * FROM pet WHERE name REGEXP "^.....$"; -+-------+--------+---------+------+------------+-------+ -| name | owner | species | sex | birth | death | -+-------+--------+---------+------+------------+-------+ -| Claws | Gwen | cat | m | 1994-03-17 | NULL | -| Buffy | Harold | dog | f | 1989-05-13 | NULL | -+-------+--------+---------+------+------------+-------+ -@end example +In crash recovery InnoDB looks for a checkpoint label written +to the log files. It knows that all modifications to the database +before the label are already present on the disk image of the database. +Then InnoDB scans the log files forward from the place of the checkpoint +applying the logged modifications to the database. -You could also write the previous query using the @samp{@{n@}} -``repeat-@code{n}-times'' operator: +InnoDB writes to the log files in a circular fashion. +All committed modifications which make the database pages in the buffer +pool different from the images on disk must be available in the log files +in case InnoDB has to do a recovery. This means that when InnoDB starts +to reuse a log file in the circular fashion, it has to make sure that the +database page images on disk already contain the modifications +logged in the log file InnoDB is going to reuse. In other words, InnoDB +has to make a checkpoint and often this involves flushing of +modified database pages to disk. -@example -mysql> SELECT * FROM pet WHERE name REGEXP "^.@{5@}$"; -+-------+--------+---------+------+------------+-------+ -| name | owner | species | sex | birth | death | -+-------+--------+---------+------+------------+-------+ -| Claws | Gwen | cat | m | 1994-03-17 | NULL | -| Buffy | Harold | dog | f | 1989-05-13 | NULL | -+-------+--------+---------+------+------------+-------+ -@end example +The above explains why making your log files very big may save +disk i/o in checkpointing. It can make sense to set +the total size of the log files as big as the buffer pool or even bigger. +The drawback in big log files is that crash recovery can last longer +because there will be more log to apply to the database. -@cindex rows, counting -@cindex tables, counting rows -@cindex counting, table rows -@node Counting rows, Multiple tables, Pattern matching, Retrieving data -@subsubsection Counting Rows +@node Moving, InnoDB transaction model, Backing up, InnoDB +@subsection Moving an InnoDB database to another machine -Databases are often used to answer the question, ``How often does a certain -type of data occur in a table?'' For example, you might want to know how -many pets you have, or how many pets each owner has, or you might want to -perform various kinds of censuses on your animals. +InnoDB data and log files are binary-compatible on all platforms +if the floating point number format on the machines is the same. +You can move an InnoDB database simply by copying all the relevant +files, which we already listed in the previous section on backing up +a database. If the floating point formats on the machines are +different but you have not used @code{FLOAT} or @code{DOUBLE} +data types in your tables then the procedure is the same: just copy +the relevant files. If the formats are different and your tables +contain floating point data, you have to use @file{mysqldump} +and @file{mysqlimport} to move those tables. -Counting the total number of animals you have is the same question as ``How -many rows are in the @code{pet} table?'' because there is one record per pet. -The @code{COUNT()} function counts the number of non-@code{NULL} results, so -the query to count your animals looks like this: +A performance tip is to switch off the auto commit when you import +data into your database, assuming your tablespace has enough space for +the big rollback segment the big import transaction will generate. +Do the commit only after importing a whole table or a segment of +a table. -@example -mysql> SELECT COUNT(*) FROM pet; -+----------+ -| COUNT(*) | -+----------+ -| 9 | -+----------+ -@end example +@node InnoDB transaction model, Implementation, Moving, InnoDB +@subsection InnoDB transaction model -Earlier, you retrieved the names of the people who owned pets. You can -use @code{COUNT()} if you want to find out how many pets each owner has: +In the InnoDB transaction model the goal has been to combine the best +sides of a multiversioning database to traditional two-phase locking. +InnoDB does locking on row level and runs queries by default +as non-locking consistent reads, in the style of Oracle. +The lock table in InnoDB is stored so space-efficiently that lock +escalation is not needed: typically several users are allowed +to lock every row in the database, or any random subset of the rows, +without InnoDB running out of memory. -@example -mysql> SELECT owner, COUNT(*) FROM pet GROUP BY owner; -+--------+----------+ -| owner | COUNT(*) | -+--------+----------+ -| Benny | 2 | -| Diane | 2 | -| Gwen | 3 | -| Harold | 2 | -+--------+----------+ -@end example +In InnoDB all user activity happens inside transactions. If the +auto commit mode is used in @strong{MySQL}, then each SQL statement +will form a single transaction. If the auto commit mode is +switched off, then we can think that a user always has a transaction +open. If he issues +the SQL @code{COMMIT} or @code{ROLLBACK} statement, that +ends the current transaction, and a new starts. Both statements +will release all InnoDB locks that were set during the +current transaction. A @code{COMMIT} means that the +changes made in the current transaction are made permanent +and become visible to other users. A @code{ROLLBACK} +on the other hand cancels all modifications made by the current +transaction. -Note the use of @code{GROUP BY} to group together all records for each -@code{owner}. Without it, all you get is an error message: +@menu +* InnoDB consistent read:: +* InnoDB locking reads:: +* InnoDB Next-key locking:: +* InnoDB Locks set:: +* InnoDB Deadlock detection:: +@end menu -@example -mysql> SELECT owner, COUNT(owner) FROM pet; -ERROR 1140 at line 1: Mixing of GROUP columns (MIN(),MAX(),COUNT()...) -with no GROUP columns is illegal if there is no GROUP BY clause -@end example +@node InnoDB consistent read, InnoDB locking reads, InnoDB transaction model, InnoDB transaction model +@subsubsection Consistent read -@code{COUNT()} and @code{GROUP BY} are useful for characterizing your -data in various ways. The following examples show different ways to -perform animal census operations. +A consistent read means that InnoDB uses its multiversioning to +present to a query a snapshot of the database at a point in time. +The query will see the changes made by exactly those transactions that +committed before that point of time, and no changes made by later +or uncommitted transactions. The exception to this rule is that the +query will see the changes made by the transaction itself which issues +the query. -Number of animals per species: +When a transaction issues its first consistent read, InnoDB assigns +the snapshot, or the point of time, which all consistent reads in the +same transaction will use. In the snapshot are all transactions that +committed before assigning the snapshot. Thus the consistent reads +within the same transaction will also be consistent with respect to each +other. You can get a fresher snapshot for your queries by committing +the current transaction and after that issuing new queries. -@example -mysql> SELECT species, COUNT(*) FROM pet GROUP BY species; -+---------+----------+ -| species | COUNT(*) | -+---------+----------+ -| bird | 2 | -| cat | 2 | -| dog | 3 | -| hamster | 1 | -| snake | 1 | -+---------+----------+ -@end example +Consistent read is the default mode in which InnoDB processes +@code{SELECT} statements. A consistent read does not set any locks +on the tables it accesses, and therefore other users are free to +modify those tables at the same time a consistent read is being performed +on the table. -Number of animals per sex: +@node InnoDB locking reads, InnoDB Next-key locking, InnoDB consistent read, InnoDB transaction model +@subsubsection Locking reads -@example -mysql> SELECT sex, COUNT(*) FROM pet GROUP BY sex; -+------+----------+ -| sex | COUNT(*) | -+------+----------+ -| NULL | 1 | -| f | 4 | -| m | 4 | -+------+----------+ -@end example +A consistent read is not convenient in some circumstances. +Suppose you want to add a new row into your table @code{CHILD}, +and make sure that the child already has a parent in table +@code{PARENT}. -(In this output, @code{NULL} indicates sex unknown.) +Suppose you use a consistent read to read the table @code{PARENT} +and indeed see the parent of the child in the table. Can you now safely +add the child row to table @code{CHILD}? No, because it may +happen that meanwhile some other user has deleted the parent row +from the table @code{PARENT}, and you are not aware of that. -Number of animals per combination of species and sex: +The solution is to perform the @code{SELECT} in a locking +mode, @code{LOCK IN SHARE MODE}. @example -mysql> SELECT species, sex, COUNT(*) FROM pet GROUP BY species, sex; -+---------+------+----------+ -| species | sex | COUNT(*) | -+---------+------+----------+ -| bird | NULL | 1 | -| bird | f | 1 | -| cat | f | 1 | -| cat | m | 1 | -| dog | f | 1 | -| dog | m | 2 | -| hamster | f | 1 | -| snake | m | 1 | -+---------+------+----------+ +SELECT * FROM PARENT WHERE NAME = 'Jones' LOCK IN SHARE MODE; @end example -You need not retrieve an entire table when you use @code{COUNT()}. For -example, the previous query, when performed just on dogs and cats, looks like -this: +Performing a read in share mode means that we read the latest +available data, and set a shared mode lock on the rows we read. +If the latest data belongs to a yet uncommitted transaction of another +user, we will wait until that transaction commits. +A shared mode lock prevents others from updating or deleting +the row we have read. After we see that the above query returns +the parent @code{'Jones'}, we can safely add his child +to table @code{CHILD}, and commit our transaction. +This example shows how to implement referential +integrity in your application code. -@example -mysql> SELECT species, sex, COUNT(*) FROM pet - -> WHERE species = "dog" OR species = "cat" - -> GROUP BY species, sex; -+---------+------+----------+ -| species | sex | COUNT(*) | -+---------+------+----------+ -| cat | f | 1 | -| cat | m | 1 | -| dog | f | 1 | -| dog | m | 2 | -+---------+------+----------+ -@end example +Let us look at another example: we have an integer counter field in +a table @code{CHILD_CODES} which we use to assign +a unique identifier to each child we add to table @code{CHILD}. +Obviously, using a consistent read or a shared mode read +to read the present value of the counter is not a good idea, since +then two users of the database may see the same value for the +counter, and we will get a duplicate key error when we add +the two children with the same identifier to the table. -Or, if you wanted the number of animals per sex only for known-sex animals: +In this case there are two good ways to implement the +reading and incrementing of the counter: (1) update the counter +first by incrementing it by 1 and only after that read it, +or (2) read the counter first with +a lock mode @code{FOR UPDATE}, and increment after that: @example -mysql> SELECT species, sex, COUNT(*) FROM pet - -> WHERE sex IS NOT NULL - -> GROUP BY species, sex; -+---------+------+----------+ -| species | sex | COUNT(*) | -+---------+------+----------+ -| bird | f | 1 | -| cat | f | 1 | -| cat | m | 1 | -| dog | f | 1 | -| dog | m | 2 | -| hamster | f | 1 | -| snake | m | 1 | -+---------+------+----------+ +SELECT COUNTER_FIELD FROM CHILD_CODES FOR UPDATE; +UPDATE CHILD_CODES SET COUNTER_FIELD = COUNTER_FIELD + 1; @end example -@cindex tables, multiple -@node Multiple tables, , Counting rows, Retrieving data -@subsubsection Using More Than one Table - -The @code{pet} table keeps track of which pets you have. If you want to -record other information about them, such as events in their lives like -visits to the vet or when litters are born, you need another table. What -should this table look like? It needs: - -@itemize @bullet -@item -To contain the pet name so you know which animal each event pertains -to. - -@item -A date so you know when the event occurred. +A @code{SELECT ... FOR UPDATE} will read the latest +available data setting exclusive locks on each row it reads. +Thus it sets the same locks a searched SQL @code{UPDATE} would set +on the rows. -@item -A field to describe the event. +@node InnoDB Next-key locking, InnoDB Locks set, InnoDB locking reads, InnoDB transaction model +@subsubsection Next-key locking: avoiding the phantom problem -@item -An event type field, if you want to be able to categorize events. -@end itemize +In row level locking InnoDB uses an algorithm called next-key locking. +InnoDB does the row level locking so that when it searches or +scans an index of a table, it sets shared or exclusive locks +on the index records in encounters. Thus the row level locks are +more precisely called index record locks. -Given these considerations, the @code{CREATE TABLE} statement for the -@code{event} table might look like this: +The locks InnoDB sets on index records also affect the 'gap' +before that index record. If a user has a shared or exclusive +lock on record R in an index, then another user cannot insert +a new index record immediately before R in the index order. +This locking of gaps is done to prevent the so-called phantom +problem. Suppose I want to read and lock all children with identifier +bigger than 100 from table @code{CHILD}, +and update some field in the selected rows. @example -mysql> CREATE TABLE event (name VARCHAR(20), date DATE, - -> type VARCHAR(15), remark VARCHAR(255)); +SELECT * FROM CHILD WHERE ID > 100 FOR UPDATE; @end example -As with the @code{pet} table, it's easiest to load the initial records -by creating a tab-delimited text file containing the information: - -@multitable @columnfractions .15 .15 .15 .55 -@item Fluffy @tab 1995-05-15 @tab litter @tab 4 kittens, 3 female, 1 male -@item Buffy @tab 1993-06-23 @tab litter @tab 5 puppies, 2 female, 3 male -@item Buffy @tab 1994-06-19 @tab litter @tab 3 puppies, 3 female -@item Chirpy @tab 1999-03-21 @tab vet @tab needed beak straightened -@item Slim @tab 1997-08-03 @tab vet @tab broken rib -@item Bowser @tab 1991-10-12 @tab kennel -@item Fang @tab 1991-10-12 @tab kennel -@item Fang @tab 1998-08-28 @tab birthday @tab Gave him a new chew toy -@item Claws @tab 1998-03-17 @tab birthday @tab Gave him a new flea collar -@item Whistler @tab 1998-12-09 @tab birthday @tab First birthday -@end multitable - -Load the records like this: +Suppose there is an index on table @code{CHILD} on column +@code{ID}. Our query will scan that index starting from +the first record where @code{ID} is bigger than 100. +Now, if the locks set on the index records would not lock out +inserts made in the gaps, a new child might meanwhile be +inserted to the table. If now I in my transaction execute @example -mysql> LOAD DATA LOCAL INFILE "event.txt" INTO TABLE event; +SELECT * FROM CHILD WHERE ID > 100 FOR UPDATE; @end example -Based on what you've learned from the queries you've run on the @code{pet} -table, you should be able to perform retrievals on the records in the -@code{event} table; the principles are the same. But when is the -@code{event} table by itself insufficient to answer questions you might ask? +again, I will see a new child in the result set the query returns. +This is against the isolation principle of transactions: +a transaction should be able to run so that the data +it has read does not change during the transaction. If we regard +a set of rows as a data item, then the new 'phantom' child would break +this isolation principle. -Suppose you want to find out the ages of each pet when they had their -litters. The @code{event} table indicates when this occurred, but to -calculate the age of the mother, you need her birth date. Because that is -stored in the @code{pet} table, you need both tables for the query: +When InnoDB scans an index it can also lock the gap +after the last record in the index. Just that happens in the previous +example: the locks set by InnoDB will prevent any insert to +the table where @code{ID} would be bigger than 100. -@example -mysql> SELECT pet.name, (TO_DAYS(date) - TO_DAYS(birth))/365 AS age, remark - -> FROM pet, event - -> WHERE pet.name = event.name AND type = "litter"; -+--------+------+-----------------------------+ -| name | age | remark | -+--------+------+-----------------------------+ -| Fluffy | 2.27 | 4 kittens, 3 female, 1 male | -| Buffy | 4.12 | 5 puppies, 2 female, 3 male | -| Buffy | 5.10 | 3 puppies, 3 female | -+--------+------+-----------------------------+ -@end example +You can use the next-key locking to implement a uniqueness +check in your application: if you read your data in share mode +and do not see a duplicate for a row you are going to insert, +then you can safely insert your row and know that the next-key +lock set on the successor of your row during the read will prevent +anyone meanwhile inserting a duplicate for your row. Thus the next-key +locking allows you to 'lock' the non-existence of something in your +table. -There are several things to note about this query: +@node InnoDB Locks set, InnoDB Deadlock detection, InnoDB Next-key locking, InnoDB transaction model +@subsubsection Locks set by different SQL statements in InnoDB @itemize @bullet @item -The @code{FROM} clause lists two tables because the query needs to pull -information from both of them. - +@code{SELECT ... FROM ...} : this is a consistent read, reading a +snapshot of the database and setting no locks. @item -When combining (joining) information from multiple tables, you need to -specify how records in one table can be matched to records in the other. -This is easy because they both have a @code{name} column. The query uses -@code{WHERE} clause to match up records in the two tables based on the -@code{name} values. - +@code{SELECT ... FROM ... LOCK IN SHARE MODE} : sets shared next-key locks +on all index records the read encounters. @item -Because the @code{name} column occurs in both tables, you must be specific -about which table you mean when referring to the column. This is done -by prepending the table name to the column name. +@code{SELECT ... FROM ... FOR UPDATE} : sets exclusive next-key locks +on all index records the read encounters. +@item +@code{INSERT INTO ... VALUES (...)} : sets an exclusive lock +on the inserted row; note that this lock is not a next-key lock +and does not prevent other users from inserting to the gap before the +inserted row. If a duplicate key error occurs, sets a shared lock +on the duplicate index record. +@item +@code{INSERT INTO T SELECT ... FROM S WHERE ...} sets an exclusive +(non-next-key) lock on each row inserted into @code{T}. Does +the search on @code{S} as a consistent read, but sets shared next-key +locks on @code{S} if the @strong{MySQL} logging is on. InnoDB has to set +locks in the latter case because in roll-forward recovery from a +backup every SQL statement has to be executed in exactly the same +way as it was done originally. +@item +@code{CREATE TABLE ... SELECT ...} performs the @code{SELECT} +as a consistent read or with shared locks, like in the previous +item. +@item +@code{REPLACE} is done like an insert if there is no collision +on a unique key. Otherwise, an exclusive next-key lock is placed +on the row which has to be updated. +@item +@code{UPDATE ... SET ... WHERE ...} : sets an exclusive next-key +lock on every record the search encounters. +@item +@code{DELETE FROM ... WHERE ...} : sets an exclusive next-key +lock on every record the search encounters. +@item +@code{LOCK TABLES ... } : sets table locks. In the implementation +the @strong{MySQL} layer of code sets these locks. The automatic deadlock detection +of InnoDB cannot detect deadlocks where such table locks are involved: +see the next section below. See also section 13 'InnoDB restrictions' +about the following: since @strong{MySQL} does know about row level locks, +it is possible that you +get a table lock on a table where another user currently has row level +locks. But that does not put transaction integerity into danger. @end itemize -You need not have two different tables to perform a join. Sometimes it is -useful to join a table to itself, if you want to compare records in a table -to other records in that same table. For example, to find breeding pairs -among your pets, you can join the @code{pet} table with itself to pair up -males and females of like species: - -@example -mysql> SELECT p1.name, p1.sex, p2.name, p2.sex, p1.species - -> FROM pet AS p1, pet AS p2 - -> WHERE p1.species = p2.species AND p1.sex = "f" AND p2.sex = "m"; -+--------+------+--------+------+---------+ -| name | sex | name | sex | species | -+--------+------+--------+------+---------+ -| Fluffy | f | Claws | m | cat | -| Buffy | f | Fang | m | dog | -| Buffy | f | Bowser | m | dog | -+--------+------+--------+------+---------+ -@end example - -In this query, we specify aliases for the table name in order -to refer to the columns and keep straight which instance of the table -each column reference is associated with. +@node InnoDB Deadlock detection, , InnoDB Locks set, InnoDB transaction model +@subsubsection Deadlock detection and rollback -@cindex databases, information about -@cindex tables, information about -@findex DESCRIBE -@node Getting information, Examples, Database use, Tutorial -@section Getting Information About Databases and Tables +InnoDB automatically detects a deadlock of transactions and rolls +back the transaction whose lock request was the last one to build +a deadlock, that is, a cycle in the waits-for graph of transactions. +InnoDB cannot detect deadlocks where a lock set by a @strong{MySQL} +@code{LOCK TABLES} statement is involved, or if a lock set +in another table handler than InnoDB is involved. You have to resolve +these situations using @code{innodb_lock_wait_timeout} set in +@file{my.cnf}. -What if you forget the name of a database or table, or what the structure of -a given table is (for example, what its columns are called)? @strong{MySQL} -addresses this problem through several statements that provide information -about the databases and tables it supports. +When InnoDB performs a complete rollback of a transaction, all the +locks of the transaction are released. However, if just a single SQL +statement is rolled back as a result of an error, some of the locks +set by the SQL statement may be preserved. This is because InnoDB +stores row locks in a format where it cannot afterwards know which was +set by which SQL statement. -You have already seen @code{SHOW DATABASES}, which lists the databases -managed by the server. To find out which database is currently selected, -use the @code{DATABASE()} function: +@subsection Performance tuning tips -@example -mysql> SELECT DATABASE(); -+------------+ -| DATABASE() | -+------------+ -| menagerie | -+------------+ -@end example +@strong{1.} +If the Unix @file{top} or the Windows @file{Task Manager} shows that +the CPU usage percentage with your workload is less than 70 %, +your workload is probably +disk-bound. Maybe you are making too many transaction commits, or the +buffer pool is too small. +Making the buffer pool bigger can help, but do not set +it bigger than 80 % of physical memory. -If you haven't selected any database yet, the result is blank. +@strong{2.} +Wrap several modifications into one transaction. InnoDB must +flush the log to disk at each transaction commit, if that transaction +made modifications to the database. Since the rotation speed of a disk +is typically +at most 167 revolutions/second, that constrains the number of commits +to the same 167/second if the disk does not fool the operating system. -To find out what tables the current database contains (for example, when -you're not sure about the name of a table), use this command: +@strong{3.} +If you can afford the loss of some latest committed transactions, you can +set the @file{my.cnf} parameter @code{innodb_flush_log_at_trx_commit} +to zero. InnoDB tries to flush the log anyway once in a second, +though the flush is not guaranteed. -@example -mysql> SHOW TABLES; -+---------------------+ -| Tables in menagerie | -+---------------------+ -| event | -| pet | -+---------------------+ -@end example +@strong{4.} +Make your log files big, even as big as the buffer pool. When InnoDB +has written the log files full, it has to write the modified contents +of the buffer pool to disk in a checkpoint. Small log files will cause many +unnecessary disk writes. The drawback in big log files is that recovery +time will be longer. -If you want to find out about the structure of a table, the @code{DESCRIBE} -command is useful; it displays information about each of a table's columns: +@strong{5.} +Also the log buffer should be quite big, say 8 MB. + +@strong{6.} (Relevant from 3.23.39 up.) +In some versions of Linux and Unix, flushing files to disk with the Unix +@code{fdatasync} and other similar methods is surprisingly slow. +The default method InnoDB uses is the @code{fdatasync} function. +If you are not satisfied with the database write performance, you may +try setting @code{innodb_flush_method} in @file{my.cnf} +to @code{O_DSYNC}, though O_DSYNC seems to be slower on most systems. +You can also try setting it to @code{littlesync}, which means that +InnoDB does not call the file flush for every write it does to a +file, but only +in log flush at transaction commits and data file flush at a checkpoint. +The drawback in @code{littlesync} is that if the operating system +crashes, you can easily end up with a half-written database page, +and you have to +do a recovery from a backup. With @code{nosync} you have even less safety: +InnoDB will only flush the database files to disk at database shutdown + +@strong{7.} In importing data to InnoDB, make sure that MySQL does not have +@code{autocommit=1} on. Then every insert requires a log flush to disk. +Put before your plain SQL import file line @example -mysql> DESCRIBE pet; -+---------+-------------+------+-----+---------+-------+ -| Field | Type | Null | Key | Default | Extra | -+---------+-------------+------+-----+---------+-------+ -| name | varchar(20) | YES | | NULL | | -| owner | varchar(20) | YES | | NULL | | -| species | varchar(20) | YES | | NULL | | -| sex | char(1) | YES | | NULL | | -| birth | date | YES | | NULL | | -| death | date | YES | | NULL | | -+---------+-------------+------+-----+---------+-------+ +set autocommit=0; @end example -@code{Field} indicates the column name, @code{Type} is the data type for -the column, @code{Null} indicates whether or not the column can contain -@code{NULL} values, @code{Key} indicates whether or not the column is -indexed, and @code{Default} specifies the column's default value. +and after it -If you have indexes on a table, -@code{SHOW INDEX FROM tbl_name} produces information about them. +@example +commit; +@end example -@cindex queries, examples -@cindex examples, queries -@node Examples, Batch mode, Getting information, Tutorial -@section Examples of Common Queries +If you use the @file{mysqldump} option @code{--opt}, you will get dump +files which are fast to import also to an InnoDB table, even without wrapping +them to the above @code{set autocommit=0; ... commit;} wrappers. -Here are examples of how to solve some common problems with -@strong{MySQL}. +@strong{8.} +Beware of big rollbacks of mass inserts: InnoDB uses the insert buffer +to save disk i/o in inserts, but in a corresponding rollback no such +mechanism is used. A disk-bound rollback can take 30 times the time +of the corresponding insert. Killing the database process will not +help because the rollback will start again at the database startup. The +only way to get rid of a runaway rollback is to increase the buffer pool +so that the rollback becomes CPU-bound and runs fast, or delete the whole +InnoDB database. -Some of the examples use the table @code{shop} to hold the price of each -article (item number) for certain traders (dealers). Supposing that each -trader has a single fixed price per article, then (@code{item}, -@code{trader}) is a primary key for the records. +@strong{9.} +Beware also of other big disk-bound operations. +Use @code{DROP TABLE} +or @code{TRUNCATE} (from MySQL-4.0 up) to empty a table, not +@code{DELETE FROM yourtable}. -Start the command line tool @code{mysql} and select a database: +@strong{10.} +Use the multi-line @code{INSERT} to reduce +communication overhead between the client and the server if you need +to insert many rows: @example -mysql your-database-name +INSERT INTO yourtable VALUES (1, 2), (5, 5); @end example -(In most @strong{MySQL} installations, you can use the database-name 'test'). - -You can create the example table as: - -@example -CREATE TABLE shop ( - article INT(4) UNSIGNED ZEROFILL DEFAULT '0000' NOT NULL, - dealer CHAR(20) DEFAULT '' NOT NULL, - price DOUBLE(16,2) DEFAULT '0.00' NOT NULL, - PRIMARY KEY(article, dealer)); +This tip is of course valid for inserts into any table type, not just InnoDB. -INSERT INTO shop VALUES -(1,'A',3.45),(1,'B',3.99),(2,'A',10.99),(3,'B',1.45),(3,'C',1.69), -(3,'D',1.25),(4,'D',19.95); -@end example +@node Implementation, Table and index, InnoDB transaction model, InnoDB +@subsection Implementation of multiversioning -Okay, so the example data is: +Since InnoDB is a multiversioned database, it must keep information +of old versions of rows in the tablespace. This information is stored +in a data structure we call a rollback segment after an analogous +data structure in Oracle. -@example -mysql> SELECT * FROM shop; +InnoDB internally adds two fields to each row stored in the database. +A 6-byte field tells the transaction identifier for the last +transaction which inserted or updated the row. Also a deletion +is internally treated as an update where a special bit in the row +is set to mark it as deleted. Each row also contains a 7-byte +field called the roll pointer. The roll pointer points to an +undo log record written to the rollback segment. If the row was +updated, then the undo log record contains the information necessary +to rebuild the content of the row before it was updated. -+---------+--------+-------+ -| article | dealer | price | -+---------+--------+-------+ -| 0001 | A | 3.45 | -| 0001 | B | 3.99 | -| 0002 | A | 10.99 | -| 0003 | B | 1.45 | -| 0003 | C | 1.69 | -| 0003 | D | 1.25 | -| 0004 | D | 19.95 | -+---------+--------+-------+ -@end example +InnoDB uses the information in the rollback segment to perform the +undo operations needed in a transaction rollback. It also uses the +information to build earlier versions of a row for a consistent +read. -@menu -* example-Maximum-column:: The maximum value for a column -* example-Maximum-row:: The row holding the maximum of a certain column -* example-Maximum-column-group:: Maximum of column per group -* example-Maximum-column-group-row:: The rows holding the group-wise maximum of a certain field -* example-user-variables:: Using user variables -* example-Foreign keys:: Using foreign keys -* Searching on two keys:: -* Calculating days:: -@end menu +Undo logs in the rollback segment are divided into insert and update +undo logs. Insert undo logs are only needed in transaction rollback +and can be discarded as soon as the transaction commits. Update undo logs +are used also in consistent reads, and they can be discarded only after +there is no transaction present for which InnoDB has assigned +a snapshot that in a consistent read could need the information +in the update undo log to build an earlier version of a database +row. -@node example-Maximum-column, example-Maximum-row, Examples, Examples -@subsection The Maximum Value for a Column +You must remember to commit your transactions regularly. Otherwise +InnoDB cannot discard data from the update undo logs, and the +rollback segment may grow too big, filling up your tablespace. -``What's the highest item number?'' +The physical size of an undo log record in the rollback segment +is typically smaller than the corresponding inserted or updated +row. You can use this information to calculate the space need +for your rollback segment. -@example -SELECT MAX(article) AS article FROM shop +In our multiversioning scheme a row is not physically removed from +the database immediately when you delete it with an SQL statement. +Only when InnoDB can discard the update undo log record written for +the deletion, it can also physically remove the corresponding row and +its index records from the database. This removal operation is +called a purge, and it is quite fast, usually taking the same order of +time as the SQL statement which did the deletion. -+---------+ -| article | -+---------+ -| 4 | -+---------+ -@end example +@node Table and index, File space management, Implementation, InnoDB +@subsection Table and index structures -@node example-Maximum-row, example-Maximum-column-group, example-Maximum-column, Examples -@subsection The Row Holding the Maximum of a Certain Column +Every InnoDB table has a special index called the clustered index +where the data of the rows is stored. If you define a +@code{PRIMARY KEY} on your table, then the index of the primary key +will be the clustered index. -``Find number, dealer, and price of the most expensive article.'' +If you do not define a primary key for +your table, InnoDB will internally generate a clustered index +where the rows are ordered by the row id InnoDB assigns +to the rows in such a table. The row id is a 6-byte field which +monotonically increases as new rows are inserted. Thus the rows +ordered by the row id will be physically in the insertion order. -In ANSI SQL this is easily done with a sub-query: +Accessing a row through the clustered index is fast, because +the row data will be on the same page where the index search +leads us. In many databases the data is traditionally stored on a different +page from the index record. If a table is large, the clustered +index architecture often saves a disk i/o when compared to the +traditional solution. -@example -SELECT article, dealer, price -FROM shop -WHERE price=(SELECT MAX(price) FROM shop) -@end example +The records in non-clustered indexes (we also call them secondary indexes), +in InnoDB contain the primary key value for the row. InnoDB +uses this primary key value to search for the row from the clustered +index. Note that if the primary key is long, the secondary indexes +will use more space. -In @strong{MySQL} (which does not yet have sub-selects), just do it in -two steps: +@menu +* InnoDB physical structure:: +* InnoDB Insert buffering:: +* InnoDB Adaptive hash:: +* InnoDB Physical record:: +@end menu -@enumerate -@item -Get the maximum price value from the table with a @code{SELECT} statement. -@item -Using this value compile the actual query: -@example -SELECT article, dealer, price -FROM shop -WHERE price=19.95 -@end example -@end enumerate +@node InnoDB physical structure, InnoDB Insert buffering, Table and index, Table and index +@subsubsection Physical structure of an index -Another solution is to sort all rows descending by price and only -get the first row using the @strong{MySQL} specific @code{LIMIT} clause: +All indexes in InnoDB are B-trees where the index records are +stored in the leaf pages of the tree. The default size of an index +page is 16 kB. When new records are inserted, InnoDB tries to +leave 1 / 16 of the page free for future insertions and updates +of the index records. -@example -SELECT article, dealer, price -FROM shop -ORDER BY price DESC -LIMIT 1 -@end example +If index records are inserted in a sequential (ascending or descending) +order, the resulting index pages will be about 15/16 full. +If records are inserted in a random order, then the pages will be +1/2 - 15/16 full. If the fillfactor of an index page drops below 1/2, +InnoDB will try to contract the index tree to free the page. -@strong{NOTE}: If there are several most expensive articles (for example, each 19.95) -the @code{LIMIT} solution shows only one of them! +@node InnoDB Insert buffering, InnoDB Adaptive hash, InnoDB physical structure, Table and index +@subsubsection Insert buffering -@node example-Maximum-column-group, example-Maximum-column-group-row, example-Maximum-row, Examples -@subsection Maximum of Column per Group +It is a common situation in a database application that the +primary key is a unique identifier and new rows are inserted in the +ascending order of the primary key. Thus the insertions to the +clustered index do not require random reads from a disk. -``What's the highest price per article?'' +On the other hand, secondary indexes are usually non-unique and +insertions happen in a relatively random order into secondary indexes. +This would cause a lot of random disk i/o's without a special mechanism +used in InnoDB. -@example -SELECT article, MAX(price) AS price -FROM shop -GROUP BY article +If an index record should be inserted to a non-unique secondary index, +InnoDB checks if the secondary index page is already in the buffer +pool. If that is the case, InnoDB will do the insertion directly to +the index page. But, if the index page is not found from the buffer +pool, InnoDB inserts the record to a special insert buffer structure. +The insert buffer is kept so small that it entirely fits in the buffer +pool, and insertions can be made to it very fast. -+---------+-------+ -| article | price | -+---------+-------+ -| 0001 | 3.99 | -| 0002 | 10.99 | -| 0003 | 1.69 | -| 0004 | 19.95 | -+---------+-------+ -@end example +The insert buffer is periodically merged to the secondary index +trees in the database. Often we can merge several insertions on the +same page in of the index tree, and hence save disk i/o's. +It has been measured that the insert buffer can speed up insertions +to a table up to 15 times. -@node example-Maximum-column-group-row, example-user-variables, example-Maximum-column-group, Examples -@subsection The Rows Holding the Group-wise Maximum of a Certain Field +@node InnoDB Adaptive hash, InnoDB Physical record, InnoDB Insert buffering, Table and index +@subsubsection Adaptive hash indexes -``For each article, find the dealer(s) with the most expensive price.'' +If a database fits almost entirely in main memory, then the fastest way +to perform queries on it is to use hash indexes. InnoDB has an +automatic mechanism which monitors index searches made to the indexes +defined for a table, and if InnoDB notices that queries could +benefit from building of a hash index, such an index is automatically +built. -In ANSI SQL, I'd do it with a sub-query like this: +But note that the hash index is always built based on an existing +B-tree index on the table. InnoDB can build a hash index on a prefix +of any length of the key defined for the B-tree, depending on +what search pattern InnoDB observes on the B-tree index. +A hash index can be partial: it is not required that the whole +B-tree index is cached in the buffer pool. InnoDB will build +hash indexes on demand to those pages of the index which are +often accessed. -@example -SELECT article, dealer, price -FROM shop s1 -WHERE price=(SELECT MAX(s2.price) - FROM shop s2 - WHERE s1.article = s2.article); -@end example +In a sense, through the adaptive hash index mechanism InnoDB adapts itself +to ample main memory, coming closer to the architecture of main memory +databases. -In @strong{MySQL} it's best do it in several steps: +@node InnoDB Physical record, , InnoDB Adaptive hash, Table and index +@subsubsection Physical record structure -@enumerate +@itemize @bullet @item -Get the list of (article,maxprice). +Each index record in InnoDB contains a header of 6 bytes. The header +is used to link consecutive records together, and also in the row level +locking. @item -For each article get the corresponding rows that have the stored maximum -price. -@end enumerate +Records in the clustered index contain fields for all user-defined +columns. In addition, there is a 6-byte field for the transaction id +and a 7-byte field for the roll pointer. +@item +If the user has not defined a primary key for a table, then each clustered +index record contains also a 6-byte row id field. +@item +Each secondary index record contains also all the fields defined +for the clustered index key. +@item +A record contains also a pointer to each field of the record. +If the total length of the fields in a record is < 128 bytes, then +the pointer is 1 byte, else 2 bytes. +@end itemize -This can easily be done with a temporary table: +@subsubsection How an auto-increment column works in InnoDB -@example -CREATE TEMPORARY TABLE tmp ( - article INT(4) UNSIGNED ZEROFILL DEFAULT '0000' NOT NULL, - price DOUBLE(16,2) DEFAULT '0.00' NOT NULL); +After a database startup, when a user first does an insert to a +table @code{T} +where an auto-increment column has been defined, and the user does not provide +an explicit value for the column, then InnoDB executes @code{SELECT +MAX(auto-inc-column) FROM T}, and assigns that value incremented +by one to the the column and the auto-increment counter of the table. +We say that +the auto-increment counter for table @code{T} has been initialized. -LOCK TABLES shop read; +InnoDB follows the same procedure in initializing the auto-increment counter +for a freshly created table. -INSERT INTO tmp SELECT article, MAX(price) FROM shop GROUP BY article; +Note that if the user specifies in an insert the value 0 to the auto-increment +column, then InnoDB treats the row like the value would not have been +specified. -SELECT shop.article, dealer, shop.price FROM shop, tmp -WHERE shop.article=tmp.article AND shop.price=tmp.price; +After the auto-increment counter has been initialized, if a user inserts +a row where he explicitly specifies the column value, and the value is bigger +than the current counter value, then the counter is set to the specified +column value. If the user does not explicitly specify a value, then InnoDB +increments the counter by one and assigns its new value to the column. -UNLOCK TABLES; +The auto-increment mechanism, when assigning values from the counter, +bypasses locking and transaction handling. Therefore you may also get +gaps in the number sequence if you roll back transactions which have +got numbers from the counter. -DROP TABLE tmp; -@end example +The behavior of auto-increment is not defined if a user gives a negative +value to the column or if the value becomes bigger than the maximum +integer that can be stored in the specified integer type. -If you don't use a @code{TEMPORARY} table, you must also lock the 'tmp' table. +@node File space management, Error handling, Table and index, InnoDB +@subsection File space management and disk i/o -``Can it be done with a single query?'' +@menu +* InnoDB Disk i/o:: +* InnoDB File space:: +* InnoDB File Defragmenting:: +@end menu -Yes, but only by using a quite inefficient trick that I call the -``MAX-CONCAT trick'': +@node InnoDB Disk i/o, InnoDB File space, File space management, File space management +@subsubsection Disk i/o -@example -SELECT article, - SUBSTRING( MAX( CONCAT(LPAD(price,6,'0'),dealer) ), 7) AS dealer, - 0.00+LEFT( MAX( CONCAT(LPAD(price,6,'0'),dealer) ), 6) AS price -FROM shop -GROUP BY article; +In disk i/o InnoDB uses asynchronous i/o. On Windows NT +it uses the native asynchronous i/o provided by the operating system. +On Unix, InnoDB uses simulated asynchronous i/o built +into InnoDB: InnoDB creates a number of i/o threads to take care +of i/o operations, such as read-ahead. In a future version we will +add support for simulated aio on Windows NT and native aio on those +versions of Unix which have one. -+---------+--------+-------+ -| article | dealer | price | -+---------+--------+-------+ -| 0001 | B | 3.99 | -| 0002 | A | 10.99 | -| 0003 | C | 1.69 | -| 0004 | D | 19.95 | -+---------+--------+-------+ -@end example +On Windows NT InnoDB uses non-buffered i/o. That means that the disk +pages InnoDB reads or writes are not buffered in the operating system +file cache. This saves some memory bandwidth. -The last example can, of course, be made a bit more efficient by doing the -splitting of the concatenated column in the client. +You can also use a raw disk in InnoDB, though this has not been tested yet: +just define the raw disk in place of a data file in @file{my.cnf}. +You must give the exact size in bytes of the raw disk in @file{my.cnf}, +because at startup InnoDB checks that the size of the file +is the same as specified in the configuration file. Using a raw disk +you can on some versions of Unix perform non-buffered i/o. -@node example-user-variables, example-Foreign keys, example-Maximum-column-group-row, Examples -@subsection Using user variables +There are two read-ahead heuristics in InnoDB: sequential read-ahead +and random read-ahead. In sequential read-ahead InnoDB notices that +the access pattern to a segment in the tablespace is sequential. +Then InnoDB will post in advance a batch of reads of database pages to the +i/o system. In random read-ahead InnoDB notices that some area +in a tablespace seems to be in the process of being +fully read into the buffer pool. Then InnoDB posts the remaining +reads to the i/o system. -You can use @strong{MySQL} user variables to remember results without -having to store them in a temporary variables in the client. -@xref{Variables}. +@node InnoDB File space, InnoDB File Defragmenting, InnoDB Disk i/o, File space management +@subsubsection File space management -For example, to find the articles with the highest and lowest price you -can do: - -@example -select @@min_price:=min(price),@@max_price:=max(price) from shop; -select * from shop where price=@@min_price or price=@@max_price; - -+---------+--------+-------+ -| article | dealer | price | -+---------+--------+-------+ -| 0003 | D | 1.25 | -| 0004 | D | 19.95 | -+---------+--------+-------+ -@end example +The data files you define in the configuration file form the tablespace +of InnoDB. The files are simply catenated to form the tablespace, +there is no striping in use. +Currently you cannot directly instruct where the space is allocated +for your tables, except by using the following fact: from a newly created +tablespace InnoDB will allocate space starting from the low end. -@cindex foreign keys -@cindex keys, foreign -@node example-Foreign keys, Searching on two keys, example-user-variables, Examples -@subsection Using Foreign Keys +The tablespace consists of database pages whose default size is 16 kB. +The pages are grouped into extents of 64 consecutive pages. The 'files' inside +a tablespace are called segments in InnoDB. The name of the rollback +segment is somewhat misleading because it actually contains many +segments in the tablespace. -You don't need foreign keys to join 2 tables. +For each index in InnoDB we allocate two segments: one is for non-leaf +nodes of the B-tree, the other is for the leaf nodes. The idea here is +to achieve better sequentiality for the leaf nodes, which contain the +data. -The only thing @strong{MySQL} doesn't do is @code{CHECK} to make sure that -the keys you use really exist in the table(s) you're referencing and it -doesn't automatically delete rows from table with a foreign key -definition. If you use your keys like normal, it'll work just fine: +When a segment grows inside the tablespace, InnoDB allocates the +first 32 pages to it individually. After that InnoDB starts +to allocate whole extents to the segment. +InnoDB can add to a large segment up to 4 extents at a time to ensure +good sequentiality of data. +Some pages in the tablespace contain bitmaps of other pages, and +therefore a few extents in an InnoDB tablespace cannot be +allocated to segments as a whole, but only as individual pages. -@example -CREATE TABLE persons ( - id SMALLINT UNSIGNED NOT NULL AUTO_INCREMENT, - name CHAR(60) NOT NULL, - PRIMARY KEY (id) -); +When you issue a query @code{SHOW TABLE STATUS FROM ... LIKE ...} +to ask for available free space in the tablespace, InnoDB will +report you the space which is certainly usable in totally free extents +of the tablespace. InnoDB always reserves some extents for +clean-up and other internal purposes; these reserved extents are not +included in the free space. -CREATE TABLE shirts ( - id SMALLINT UNSIGNED NOT NULL AUTO_INCREMENT, - style ENUM('t-shirt', 'polo', 'dress') NOT NULL, - color ENUM('red', 'blue', 'orange', 'white', 'black') NOT NULL, - owner SMALLINT UNSIGNED NOT NULL REFERENCES persons, - PRIMARY KEY (id) -); +When you delete data from a table, InnoDB will contract the corresponding +B-tree indexes. It depends on the pattern of deletes if that frees +individual pages or extents to the tablespace, so that the freed +space is available for other users. Dropping a table or deleting +all rows from it is guaranteed to release the space to other users, +but remember that deleted rows can be physically removed only in a +purge operation after they are no longer needed in transaction rollback or +consistent read. +@node InnoDB File Defragmenting, , InnoDB File space, File space management +@subsubsection Defragmenting a table -INSERT INTO persons VALUES (NULL, 'Antonio Paz'); +If there are random insertions or deletions +in the indexes of a table, the indexes +may become fragmented. By fragmentation we mean that the physical ordering +of the index pages on the disk is not close to the alphabetical ordering +of the records on the pages, or that there are many unused pages in the +64-page blocks which were allocated to the index. -INSERT INTO shirts VALUES -(NULL, 'polo', 'blue', LAST_INSERT_ID()), -(NULL, 'dress', 'white', LAST_INSERT_ID()), -(NULL, 't-shirt', 'blue', LAST_INSERT_ID()); +It can speed up index scans if you +periodically use @code{mysqldump} to dump the table to +a text file, drop the table, and reload it from the dump. +Another way to do the defragmenting is to @code{ALTER} the table type to +@code{MyISAM} and back to @code{InnoDB} again. +Note that a @code{MyISAM} table must fit in a single file +on your operating system. +If the insertions to and index are always ascending and +records are deleted only from the end, then the the file space management +algorithm of InnoDB guarantees that fragmentation in the index will +not occur. -INSERT INTO persons VALUES (NULL, 'Lilliana Angelovska'); +@node Error handling, InnoDB restrictions, File space management, InnoDB +@subsection Error handling -INSERT INTO shirts VALUES -(NULL, 'dress', 'orange', LAST_INSERT_ID()), -(NULL, 'polo', 'red', LAST_INSERT_ID()), -(NULL, 'dress', 'blue', LAST_INSERT_ID()), -(NULL, 't-shirt', 'white', LAST_INSERT_ID()); +The error handling in InnoDB is not always the same as +specified in the ANSI SQL standards. According to the ANSI +standard, any error during an SQL statement should cause the +rollback of that statement. InnoDB sometimes rolls back only +part of the statement. +The following list specifies the error handling of InnoDB. +@itemize @bullet +@item +If you run out of file space in the tablespace, +you will get the @strong{MySQL} @code{'Table is full'} error +and InnoDB rolls back the SQL statement. +@item +A transaction deadlock or a timeout in a lock wait will give +@code{'Table handler error 1000000'} and InnoDB rolls back +the SQL statement. +@item +A duplicate key error only rolls back the insert of that particular row, +even in a statement like @code{INSERT INTO ... SELECT ...}. +This will probably change so that the SQL statement will be rolled +back if you have not specified the @code{IGNORE} option in your +statement. +@item +A 'row too long' error rolls back the SQL statement. +@item +Other errors are mostly detected by the @strong{MySQL} layer of code, and +they roll back the corresponding SQL statement. +@end itemize -SELECT * FROM persons; -+----+---------------------+ -| id | name | -+----+---------------------+ -| 1 | Antonio Paz | -| 2 | Lilliana Angelovska | -+----+---------------------+ +@node InnoDB restrictions, InnoDB contact information, Error handling, InnoDB +@subsection Some restrictions on InnoDB tables -SELECT * FROM shirts; -+----+---------+--------+-------+ -| id | style | color | owner | -+----+---------+--------+-------+ -| 1 | polo | blue | 1 | -| 2 | dress | white | 1 | -| 3 | t-shirt | blue | 1 | -| 4 | dress | orange | 2 | -| 5 | polo | red | 2 | -| 6 | dress | blue | 2 | -| 7 | t-shirt | white | 2 | -+----+---------+--------+-------+ +@itemize @bullet +@item @code{SHOW TABLE STATUS} does not give accurate statistics +on InnoDB tables, except for the physical size reserved by the table. +The row count is only a rough estimate used in SQL optimization. -SELECT s.* FROM persons p, shirts s - WHERE p.name LIKE 'Lilliana%' - AND s.owner = p.id - AND s.color <> 'white'; +@item +If you try to create an unique index on a prefix of a column you will get an +error: -+----+-------+--------+-------+ -| id | style | color | owner | -+----+-------+--------+-------+ -| 4 | dress | orange | 2 | -| 5 | polo | red | 2 | -| 6 | dress | blue | 2 | -+----+-------+--------+-------+ +@example +CREATE TABLE T (A CHAR(20), B INT, UNIQUE (A(5))) TYPE = InnoDB; @end example -@findex UNION -@cindex searching, two keys -@cindex keys, searching on two -@node Searching on two keys, Calculating days, example-Foreign keys, Examples -@subsection Searching on Two Keys +If you create a non unique index on a prefix of a column, InnoDB will +create an index over the whole column. +@item +@code{INSERT DELAYED} is not supported for InnoDB tables. +@item +The @strong{MySQL} @code{LOCK TABLES} operation does not know of InnoDB +row level locks set in already completed SQL statements: this means that +you can get a table lock on a table even if there still exist transactions +of other users which have row level locks on the same table. Thus +your operations on the table may have to wait if they collide with +these locks of other users. Also a deadlock is possible. However, +this does not endanger transaction integrity, because the row level +locks set by InnoDB will always take care of the integrity. +Also, a table lock prevents other transactions from acquiring more +row level locks (in a conflicting lock mode) on the table. +@item +You cannot have a key on a @code{BLOB} or @code{TEXT} column. +@item +A table cannot contain more than 1000 columns. +@item +@code{DELETE FROM TABLE} does not regenerate the table but instead +deletes all rows, one by one, which is not that fast. In future versions +of @strong{MySQL} you can use @code{TRUNCATE} which is fast. +@item +Before dropping a database with InnoDB tables one has to drop +the individual InnoDB tables first. +@item +The default database page size in InnoDB is 16 kB. By recompiling the +code one can set it from 8 kB to 64 kB. +The maximun row length is slightly less than a half of a database page, +the row length also includes @code{BLOB} and @code{TEXT} type +columns. The restriction on the size of @code{BLOB} and +@code{TEXT} columns will be removed by June 2001 in a future version of +InnoDB. +@item +The maximum data or log file size is 2 GB or 4 GB depending on how large +files your operating system supports. Support for > 4 GB files will +be added to InnoDB in a future version. +@item +The maximum tablespace size is 4 billion database pages. This is also +the maximum size for a table. The minimum tablespace size is 10 MB. +@end itemize -@strong{MySQL} doesn't yet optimize when you search on two different -keys combined with @code{OR} (Searching on one key with different @code{OR} -parts is optimized quite good): +@node InnoDB contact information, , InnoDB restrictions, InnoDB +@subsection InnoDB contact information + +Contact information of Innobase Oy, producer of the InnoDB engine. +Website: @uref{http://www.innodb.com}. Email: +@email{Heikki.Tuuri@@innodb.com} @example -SELECT field1_index, field2_index FROM test_table WHERE field1_index = '1' -OR field2_index = '1' +phone: 358-9-6969 3250 (office) 358-40-5617367 (mobile) +InnoDB Oy Inc. +World Trade Center Helsinki +Aleksanterinkatu 17 +P.O.Box 800 +00101 Helsinki +Finland @end example -The reason is that we haven't yet had time to come up with an efficient -way to handle this in the general case. (The @code{AND} handling is, -in comparison, now completely general and works very well). +@cindex tables, @code{BDB} +@cindex tables, @code{Berkeley DB} +@node BDB, , InnoDB, Table types +@section BDB or Berkeley_DB Tables -For the moment you can solve this very efficiently by using a -@code{TEMPORARY} table. This type of optimization is also very good if -you are using very complicated queries where the SQL server does the -optimizations in the wrong order. +@menu +* BDB overview:: Overview of BDB Tables +* BDB install:: Installing BDB +* BDB start:: BDB startup options +* BDB characteristic:: Some characteristic of @code{BDB} tables: +* BDB TODO:: Some things we need to fix for BDB in the near future: +* BDB portability:: Operating systems supported by @strong{BDB} +* BDB errors:: Errors You May Get When Using BDB Tables +@end menu -@example -CREATE TEMPORARY TABLE tmp -SELECT field1_index, field2_index FROM test_table WHERE field1_index = '1'; -INSERT INTO tmp -SELECT field1_index, field2_index FROM test_table WHERE field2_index = '1'; -SELECT * from tmp; -DROP TABLE tmp; -@end example +@node BDB overview, BDB install, BDB, BDB +@subsection Overview of BDB Tables -The above way to solve this query is in effect an @code{UNION} of two queries. +Support for BDB tables is included in the @strong{MySQL} source distribution +starting from Version 3.23.34 and is activated in the @strong{MySQL}-Max +binary. -@cindex bit_functions, example -@findex BIT_OR -@findex BIT_COUNT -@findex << -@node Calculating days, , Searching on two keys, Examples -@subsection Calculating visits per day +BerkeleyDB, available at @uref{http://www.sleepycat.com/} has provided +@strong{MySQL} with a transactional table handler. By using BerkeleyDB +tables, your tables may have a greater chance of surviving crashes, and also +provides @code{COMMIT} and @code{ROLLBACK} on transactions. The +@strong{MySQL} source distribution comes with a BDB distribution that has a +couple of small patches to make it work more smoothly with @strong{MySQL}. +You can't use a non-patched @code{BDB} version with @strong{MySQL}. -The following shows an idea of how you can use the bit group functions -to calculate the number of days per month a user has visited a web page. +We at @strong{MySQL AB} are working in close cooperation with Sleepycat to +keep the quality of the @strong{MySQL}/BDB interface high. -@example -CREATE TABLE t1 (year YEAR(4), month INT(2) UNSIGNED ZEROFILL, day INT(2) UNSIGNED ZEROFILL); -INSERT INTO t1 VALUES(2000,1,1),(2000,1,20),(2000,1,30),(2000,2,2),(2000,2,23),(2000,2,23); +When it comes to supporting BDB tables, we are committed to help our +users to locate the problem and help creating a reproducable test case +for any problems involving BDB tables. Any such test case will be +forwarded to Sleepycat who in turn will help us find and fix the +problem. As this is a two stage operation, any problems with BDB tables +may take a little longer for us to fix than for other table handlers. +However, as the BerkeleyDB code itself has been used by many other +applications than @strong{MySQL}, we don't envision any big problems with +this. @xref{Table handler support}. -SELECT year,month,BIT_COUNT(BIT_OR(1<<day)) AS days FROM t1 GROUP BY year,month; +@node BDB install, BDB start, BDB overview, BDB +@subsection Installing BDB -Which returns: +If you have downloaded a binary version of @strong{MySQL} that includes +support for BerkeleyDB, simply follow the instructions for installing a +binary version of @strong{MySQL}. +@xref{Installing binary}. @xref{mysqld-max, , @code{mysqld-max}}. -+------+-------+------+ -| year | month | days | -+------+-------+------+ -| 2000 | 01 | 3 | -| 2000 | 02 | 2 | -+------+-------+------+ +To compile @strong{MySQL} with Berkeley DB support, download @strong{MySQL} +Version 3.23.34 or newer and configure @code{MySQL} with the +@code{--with-berkeley-db} option. @xref{Installing source}. + +@example +cd /path/to/source/of/mysql-3.23.34 +./configure --with-berkeley-db @end example -The above calculates how many different days was used for a given -year/month combination, with automatic removal of duplicate entries. +Please refer to the manual provided with the @code{BDB} distribution for +more updated information. -@cindex modes, batch -@cindex batch mode -@cindex running, batch mode -@cindex script files -@cindex files, script -@node Batch mode, Twin, Examples, Tutorial -@section Using @code{mysql} in Batch Mode +Even though Berkeley DB is in itself very tested and reliable, +the @strong{MySQL} interface is still considered beta quality. +We are actively improving and optimizing it to get it stable very +soon. -In the previous sections, you used @code{mysql} interactively to enter -queries and view the results. You can also run @code{mysql} in batch -mode. To do this, put the commands you want to run in a file, then -tell @code{mysql} to read its input from the file: +@node BDB start, BDB characteristic, BDB install, BDB +@subsection BDB startup options -@example -shell> mysql < batch-file -@end example +If you are running with @code{AUTOCOMMIT=0} then your changes in @code{BDB} +tables will not be updated until you execute @code{COMMIT}. Instead of commit +you can execute @code{ROLLBACK} to forget your changes. @xref{COMMIT}. -If you need to specify connection parameters on the command line, the -command might look like this: +If you are running with @code{AUTOCOMMIT=1} (the default), your changes +will be committed immediately. You can start an extended transaction with +the @code{BEGIN WORK} SQL command, after which your changes will not be +committed until you execute @code{COMMIT} (or decide to @code{ROLLBACK} +the changes). -@example -shell> mysql -h host -u user -p < batch-file -Enter password: ******** -@end example +The following options to @code{mysqld} can be used to change the behavior of +BDB tables: -When you use @code{mysql} this way, you are creating a script file, then -executing the script. +@multitable @columnfractions .30 .70 +@item @strong{Option} @tab @strong{Meaning} +@item @code{--bdb-home=directory} @tab Base directory for BDB tables. This should be the same directory you use for --datadir. +@item @code{--bdb-lock-detect=#} @tab Berkeley lock detect. One of (DEFAULT, OLDEST, RANDOM, or YOUNGEST). +@item @code{--bdb-logdir=directory} @tab Berkeley DB log file directory. +@item @code{--bdb-no-sync} @tab Don't synchronously flush logs. +@item @code{--bdb-no-recover} @tab Don't start Berkeley DB in recover mode. +@item @code{--bdb-shared-data} @tab Start Berkeley DB in multi-process mode (Don't use @code{DB_PRIVATE} when initializing Berkeley DB) +@item @code{--bdb-tmpdir=directory} @tab Berkeley DB tempfile name. +@item @code{--skip-bdb} @tab Don't use berkeley db. +@item @code{-O bdb_max_lock=1000} @tab Set the maximum number of locks possible. @xref{SHOW VARIABLES}. +@end multitable -Why use a script? Here are a few reasons: +If you use @code{--skip-bdb}, @strong{MySQL} will not initialize the +Berkeley DB library and this will save a lot of memory. Of course, +you cannot use @code{BDB} tables if you are using this option. -@itemize @bullet -@item -If you run a query repeatedly (say, every day or every week), making it a -script allows you to avoid retyping it each time you execute it. +Normally you should start @code{mysqld} without @code{--bdb-no-recover} if you +intend to use BDB tables. This may, however, give you problems when you +try to start @code{mysqld} if the BDB log files are corrupted. @xref{Starting +server}. -@item -You can generate new queries from existing ones that are similar by copying -and editing script files. +With @code{bdb_max_lock} you can specify the maximum number of locks +(10000 by default) you can have active on a BDB table. You should +increase this if you get errors of type @code{bdb: Lock table is out of +available locks} or @code{Got error 12 from ...} when you have do long +transactions or when @code{mysqld} has to examine a lot of rows to +calculate the query. -@item -Batch mode can also be useful while you're developing a query, particularly -for multiple-line commands or multiple-statement sequences of commands. If -you make a mistake, you don't have to retype everything. Just edit your -script to correct the error, then tell @code{mysql} to execute it again. +You may also want to change @code{binlog_cache_size} and +@code{max_binlog_cache_size} if you are using big multi-line transactions. +@xref{COMMIT}. +@node BDB characteristic, BDB TODO, BDB start, BDB +@subsection Some characteristic of @code{BDB} tables: + +@itemize @bullet @item -If you have a query that produces a lot of output, you can run the output -through a pager rather than watching it scroll off the top of your screen: +To be able to rollback transactions BDB maintain log files. For maximum +performance you should place these on another disk than your databases +by using the @code{--bdb_log_dir} options. +@item +@strong{MySQL} performs a checkpoint each time a new BDB log +file is started, and removes any log files that are not needed for +current transactions. One can also run @code{FLUSH LOGS} at any time +to checkpoint the Berkeley DB tables. -@example -shell> mysql < batch-file | more -@end example +For disaster recovery, one should use table backups plus +@strong{MySQL}'s binary log. @xref{Backup}. +@strong{Warning}: If you delete old log files that are in use, BDB will +not be able to do recovery at all and you may loose data if something +goes wrong. @item -You can catch the output in a file for further processing: +@strong{MySQL} requires a @code{PRIMARY KEY} in each BDB table to be +able to refer to previously read rows. If you don't create one, +@strong{MySQL} will create an maintain a hidden @code{PRIMARY KEY} for +you. The hidden key has a length of 5 bytes and is incremented for each +insert attempt. +@item +If all columns you access in a @code{BDB} table are part of the same index or +part of the primary key, then @strong{MySQL} can execute the query +without having to access the actual row. In a @code{MyISAM} table the +above holds only if the columns are part of the same index. +@item +The @code{PRIMARY KEY} will be faster than any other key, as the +@code{PRIMARY KEY} is stored together with the row data. As the other keys are +stored as the key data + the @code{PRIMARY KEY}, it's important to keep the +@code{PRIMARY KEY} as short as possible to save disk and get better speed. +@item +@code{LOCK TABLES} works on @code{BDB} tables as with other tables. If +you don't use @code{LOCK TABLE}, @strong{MYSQL} will issue an internal +multiple-write lock on the table to ensure that the table will be +properly locked if another thread issues a table lock. +@item +Internal locking in @code{BDB} tables is done on page level. +@item +@code{SELECT COUNT(*) FROM table_name} is slow as @code{BDB} tables doesn't +maintain a count of the number of rows in the table. +@item +Scanning is slower than with @code{MyISAM} tables as one has data in BDB +tables stored in B-trees and not in a separate data file. +@item +The application must always be prepared to handle cases where +any change of a @code{BDB} table may make an automatic rollback and any +read may fail with a deadlock error. +@item +Keys are not compressed to previous keys as with ISAM or MyISAM +tables. In other words, the key information will take a little more +space in @code{BDB} tables compared to MyISAM tables which don't use +@code{PACK_KEYS=0}. +@item +There is often holes in the BDB table to allow you to insert new rows in +the middle of the key tree. This makes BDB tables somewhat larger than +MyISAM tables. +@item +The optimizer needs to know an approximation of the number of rows in +the table. @strong{MySQL} solves this by counting inserts and +maintaining this in a separate segment in each BDB table. If you don't +do a lot of @code{DELETE} or @code{ROLLBACK}:s this number should be +accurate enough for the @strong{MySQL} optimizer, but as @strong{MySQL} +only store the number on close, it may be wrong if @strong{MySQL} dies +unexpectedly. It should not be fatal even if this number is not 100 % +correct. One can update the number of rows by executing @code{ANALYZE +TABLE} or @code{OPTIMIZE TABLE}. @xref{ANALYZE TABLE} . @xref{OPTIMIZE +TABLE}. +@item +If you get full disk with a @code{BDB} table, you will get an error +(probably error 28) and the transaction should roll back. This is in +contrast with @code{MyISAM} and @code{ISAM} tables where @code{mysqld} will +wait for enough free disk before continuing. +@end itemize -@example -shell> mysql < batch-file > mysql.out -@end example +@node BDB TODO, BDB portability, BDB characteristic, BDB +@subsection Some things we need to fix for BDB in the near future: +@itemize @bullet @item -You can distribute your script to other people so they can run the commands, -too. - +It's very slow to open many BDB tables at the same time. If you are +going to use BDB tables, you should not have a very big table cache (> +256 ?) and you should use @code{--no-auto-rehash} with the @code{mysql} +client. We plan to partly fix this in 4.0. @item -Some situations do not allow for interactive use, for example, when you run -a query from a @code{cron} job. In this case, you must use batch mode. +@code{SHOW TABLE STATUS} doesn't yet provide that much information for BDB +tables. +@item +Optimize performance. +@item +Change to not use page locks at all when we are scanning tables. @end itemize -The default output format is different (more concise) when you run -@code{mysql} in batch mode than when you use it interactively. For -example, the output of @code{SELECT DISTINCT species FROM pet} looks like -this when run interactively: +@node BDB portability, BDB errors, BDB TODO, BDB +@subsection Operating systems supported by @strong{BDB} + +If you after having built @strong{MySQL} with support for BDB tables get +the following error in the log file when you start @code{mysqld}: @example -+---------+ -| species | -+---------+ -| bird | -| cat | -| dog | -| hamster | -| snake | -+---------+ +bdb: architecture lacks fast mutexes: applications cannot be threaded +Can't init dtabases @end example -But like this when run in batch mode: +This means that @code{BDB} tables are not supported for your architecture. +In this case you have to rebuild @strong{MySQL} without BDB table support. -@example -species -bird -cat -dog -hamster -snake -@end example +NOTE: The following list is not complete; We will update this as we get +more information about this. -If you want to get the interactive output format in batch mode, use -@code{mysql -t}. To echo to the output the commands that are executed, use -@code{mysql -vvv}. +Currently we know that BDB tables works with the following operating +system. -@cindex Twin Studies, queries -@cindex queries, Twin Studeis project -@node Twin, , Batch mode, Tutorial -@section Queries from Twin Project +@itemize @bullet +@item +Linux 2.x intel +@item +Solaris sparc +@item +SCO OpenServer +@item +SCO UnixWare 7.0.1 +@end itemize -At Analytikerna and Lentus, we have been doing the systems and field work -for a big research project. This project is a collaboration between the -Institute of Environmental Medicine at Karolinska Institutet Stockholm -and the Section on Clinical Research in Aging and Psychology at the -University of Southern California. +It doesn't work with the following operating systems: -The project involves a screening part where all twins in Sweden older -than 65 years are interviewed by telephone. Twins who meet certain -criteria are passed on to the next stage. In this latter stage, twins who -want to participate are visited by a doctor/nurse team. Some of the -examinations include physical and neuropsychological examination, -laboratory testing, neuroimaging, psychological status assessment, and family -history collection. In addition, data are collected on medical and -environmental risk factors. +@itemize @bullet +@item +Linux 2.x Alpha +@item +Max OS X +@end itemize -More information about Twin studies can be found at: +@node BDB errors, , BDB portability, BDB +@subsection Errors You May Get When Using BDB Tables + +@itemize @bullet +@item +If you get the following error in the @code{hostname.err log} when +starting @code{mysqld}: @example -@url{http://www.imm.ki.se/TWIN/TWINUKW.HTM} +bdb: Ignoring log file: .../log.XXXXXXXXXX: unsupported log version # @end example +it means that the new @code{BDB} version doesn't support the old log +file format. In this case you have to delete all @code{BDB} log BDB +from your database directory (the files that has the format +@code{log.XXXXXXXXXX} ) and restart @code{mysqld}. We would also +recommend you to do a @code{mysqldump --opt} of your old @code{BDB} +tables, delete the old table and restore the dump. +@item +If you are running in not @code{auto_commit} mode and delete a table you +are using by another thread you may get the following error messages in +the @strong{MySQL} error file: -The latter part of the project is administered with a Web interface -written using Perl and @strong{MySQL}. +@example +001119 23:43:56 bdb: Missing log fileid entry +001119 23:43:56 bdb: txn_abort: Log undo failed for LSN: 1 3644744: Invalid +@end example -Each night all data from the interviews are moved into a @strong{MySQL} -database. +This is not fatal but we don't recommend that you delete tables if you are +not in @code{auto_commit} mode, until this problem is fixed (the fix is +not trivial). +@end itemize + + +@cindex tutorial +@cindex terminal monitor, defined +@cindex monitor, terminal +@cindex options, provided by MySQL +@node Tutorial, Server, Table types, Top +@chapter MySQL Tutorial @menu -* Twin pool:: Find all non-distributed twins -* Twin event:: Show a table on twin pair status +* Connecting-disconnecting:: Connecting to and disconnecting from the server +* Entering queries:: Entering queries +* Database use:: Creating and using a database +* Getting information:: Getting information about databases and tables +* Examples:: Examples +* Batch mode:: Using @code{mysql} in batch mode +* Twin:: Queries from twin project @end menu -@node Twin pool, Twin event, Twin, Twin -@subsection Find all Non-distributed Twins +This chapter provides a tutorial introduction to @strong{MySQL} by showing +how to use the @code{mysql} client program to create and use a simple +database. @code{mysql} (sometimes referred to as the ``terminal monitor'' or +just ``monitor'') is an interactive program that allows you to connect to a +@strong{MySQL} server, run queries, and view the results. @code{mysql} may +also be used in batch mode: you place your queries in a file beforehand, then +tell @code{mysql} to execute the contents of the file. Both ways of using +@code{mysql} are covered here. -The following query is used to determine who goes into the second part of the -project: +To see a list of options provided by @code{mysql}, invoke it with +the @code{--help} option: @example -select - concat(p1.id, p1.tvab) + 0 as tvid, - concat(p1.christian_name, " ", p1.surname) as Name, - p1.postal_code as Code, - p1.city as City, - pg.abrev as Area, - if(td.participation = "Aborted", "A", " ") as A, - p1.dead as dead1, - l.event as event1, - td.suspect as tsuspect1, - id.suspect as isuspect1, - td.severe as tsevere1, - id.severe as isevere1, - p2.dead as dead2, - l2.event as event2, - h2.nurse as nurse2, - h2.doctor as doctor2, - td2.suspect as tsuspect2, - id2.suspect as isuspect2, - td2.severe as tsevere2, - id2.severe as isevere2, - l.finish_date -from - twin_project as tp - /* For Twin 1 */ - left join twin_data as td on tp.id = td.id and tp.tvab = td.tvab - left join informant_data as id on tp.id = id.id and tp.tvab = id.tvab - left join harmony as h on tp.id = h.id and tp.tvab = h.tvab - left join lentus as l on tp.id = l.id and tp.tvab = l.tvab - /* For Twin 2 */ - left join twin_data as td2 on p2.id = td2.id and p2.tvab = td2.tvab - left join informant_data as id2 on p2.id = id2.id and p2.tvab = id2.tvab - left join harmony as h2 on p2.id = h2.id and p2.tvab = h2.tvab - left join lentus as l2 on p2.id = l2.id and p2.tvab = l2.tvab, - person_data as p1, - person_data as p2, - postal_groups as pg -where - /* p1 gets main twin and p2 gets his/her twin. */ - /* ptvab is a field inverted from tvab */ - p1.id = tp.id and p1.tvab = tp.tvab and - p2.id = p1.id and p2.ptvab = p1.tvab and - /* Just the sceening survey */ - tp.survey_no = 5 and - /* Skip if partner died before 65 but allow emigration (dead=9) */ - (p2.dead = 0 or p2.dead = 9 or - (p2.dead = 1 and - (p2.death_date = 0 or - (((to_days(p2.death_date) - to_days(p2.birthday)) / 365) - >= 65)))) - and - ( - /* Twin is suspect */ - (td.future_contact = 'Yes' and td.suspect = 2) or - /* Twin is suspect - Informant is Blessed */ - (td.future_contact = 'Yes' and td.suspect = 1 and id.suspect = 1) or - /* No twin - Informant is Blessed */ - (ISNULL(td.suspect) and id.suspect = 1 and id.future_contact = 'Yes') or - /* Twin broken off - Informant is Blessed */ - (td.participation = 'Aborted' - and id.suspect = 1 and id.future_contact = 'Yes') or - /* Twin broken off - No inform - Have partner */ - (td.participation = 'Aborted' and ISNULL(id.suspect) and p2.dead = 0)) - and - l.event = 'Finished' - /* Get at area code */ - and substring(p1.postal_code, 1, 2) = pg.code - /* Not already distributed */ - and (h.nurse is NULL or h.nurse=00 or h.doctor=00) - /* Has not refused or been aborted */ - and not (h.status = 'Refused' or h.status = 'Aborted' - or h.status = 'Died' or h.status = 'Other') -order by - tvid; +shell> mysql --help @end example -Some explanations: -@table @asis -@item @code{concat(p1.id, p1.tvab) + 0 as tvid} -We want to sort on the concatenated @code{id} and @code{tvab} in -numerical order. Adding @code{0} to the result causes @strong{MySQL} to -treat the result as a number. -@item column @code{id} -This identifies a pair of twins. It is a key in all tables. -@item column @code{tvab} -This identifies a twin in a pair. It has a value of @code{1} or @code{2}. -@item column @code{ptvab} -This is an inverse of @code{tvab}. When @code{tvab} is @code{1} this is -@code{2}, and vice versa. It exists to save typing and to make it easier for -@strong{MySQL} to optimize the query. -@end table - -This query demonstrates, among other things, how to do lookups on a -table from the same table with a join (@code{p1} and @code{p2}). In the example, this -is used to check whether a twin's partner died before the age of 65. If so, -the row is not returned. - -All of the above exist in all tables with twin-related information. We -have a key on both @code{id,tvab} (all tables), and @code{id,ptvab} -(@code{person_data}) to make queries faster. +This chapter assumes that @code{mysql} is installed on your machine and that +a @strong{MySQL} server is available to which you can connect. If this is +not true, contact your @strong{MySQL} administrator. (If @emph{you} are the +administrator, you will need to consult other sections of this manual.) -On our production machine (A 200MHz UltraSPARC), this query returns -about 150-200 rows and takes less than one second. +This chapter describes the entire process of setting up and using a +database. If you are interested only in accessing an already-existing +database, you may want to skip over the sections that describe how to +create the database and the tables it contains. -The current number of records in the tables used above: -@multitable @columnfractions .3 .5 -@item @strong{Table} @tab @strong{Rows} -@item @code{person_data} @tab 71074 -@item @code{lentus} @tab 5291 -@item @code{twin_project} @tab 5286 -@item @code{twin_data} @tab 2012 -@item @code{informant_data} @tab 663 -@item @code{harmony} @tab 381 -@item @code{postal_groups} @tab 100 -@end multitable +Because this chapter is tutorial in nature, many details are necessarily left +out. Consult the relevant sections of the manual for more +information on the topics covered here. -@node Twin event, , Twin pool, Twin -@subsection Show a Table on Twin Pair Status +@cindex connecting, to the server +@cindex disconnecting, from the server +@cindex server, connecting +@cindex server, disconnecting +@node Connecting-disconnecting, Entering queries, Tutorial, Tutorial +@section Connecting to and Disconnecting from the Server -Each interview ends with a status code called @code{event}. The query -shown below is used to display a table over all twin pairs combined by -event. This indicates in how many pairs both twins are finished, in how many -pairs one twin is finished and the other refused, and so on. +To connect to the server, you'll usually need to provide a @strong{MySQL} +user name when you invoke @code{mysql} and, most likely, a password. If the +server runs on a machine other than the one where you log in, you'll also +need to specify a hostname. Contact your administrator to find out what +connection parameters you should use to connect (that is, what host, user name, +and password to use). Once you know the proper parameters, you should be +able to connect like this: @example -select - t1.event, - t2.event, - count(*) -from - lentus as t1, - lentus as t2, - twin_project as tp -where - /* We are looking at one pair at a time */ - t1.id = tp.id - and t1.tvab=tp.tvab - and t1.id = t2.id - /* Just the sceening survey */ - and tp.survey_no = 5 - /* This makes each pair only appear once */ - and t1.tvab='1' and t2.tvab='2' -group by - t1.event, t2.event; - +shell> mysql -h host -u user -p +Enter password: ******** @end example -@cindex functions, server -@cindex server functions -@node Server, Replication, Tutorial, Top -@chapter MySQL Server Functions - -@menu -* Languages:: What languages are supported by @strong{MySQL}? -* Table size:: How big @strong{MySQL} tables can be -@end menu +The @code{********} represents your password; enter it when @code{mysql} +displays the @code{Enter password:} prompt. -This chapter describes the languages @strong{MySQL} supports, how sorting -works in @strong{MySQL}, and how to add new character sets to @strong{MySQL}. -You will also find information about maximum table sizes in this chapter. +If that works, you should see some introductory information followed by a +@code{mysql>} prompt: -@cindex error messages, languages -@cindex messages, languages -@cindex files, error messages -@cindex language support -@node Languages, Table size, Server, Server -@section What Languages Are Supported by MySQL? -@code{mysqld} can issue error messages in the following languages: -Czech, Danish, Dutch, English (the default), Estonian, French, German, Greek, -Hungarian, Italian, Japanese, Korean, Norwegian, Norwegian-ny, Polish, -Portuguese, Romanian, Russian, Slovak, Spanish, and Swedish. +@example +shell> mysql -h host -u user -p +Enter password: ******** +Welcome to the MySQL monitor. Commands end with ; or \g. +Your MySQL connection id is 459 to server version: 3.22.20a-log -To start @code{mysqld} with a particular language, use either the -@code{--language=lang} or @code{-L lang} options. For example: +Type 'help' for help. -@example -shell> mysqld --language=swedish +mysql> @end example -or: +The prompt tells you that @code{mysql} is ready for you to enter commands. + +Some @strong{MySQL} installations allow users to connect as the anonymous +(unnamed) user to the server running on the local host. If this is the case +on your machine, you should be able to connect to that server by invoking +@code{mysql} without any options: @example -shell> mysqld --language=/usr/local/share/swedish +shell> mysql @end example -Note that all language names are specified in lowercase. - -The language files are located (by default) in -@file{@var{mysql_base_dir}/share/@var{LANGUAGE}/}. - -To update the error message file, you should edit the @file{errmsg.txt} file -and execute the following command to generate the @file{errmsg.sys} file: +After you have connected successfully, you can disconnect any time by typing +@code{QUIT} at the @code{mysql>} prompt: @example -shell> comp_err errmsg.txt errmsg.sys +mysql> QUIT +Bye @end example -If you upgrade to a newer version of @strong{MySQL}, remember to repeat -your changes with the new @file{errmsg.txt} file. +You can also disconnect by pressing Control-D. -@menu -* Character sets:: The character set used for data and sorting -* Adding character set:: Adding a new character set -* Character arrays:: The character definition arrays -* String collating:: String collating support -* Multi-byte characters:: Multi-byte character support -@end menu +Most examples in the following sections assume you are connected to the +server. They indicate this by the @code{mysql>} prompt. -@cindex character sets -@cindex data, character sets -@cindex sorting, character sets -@node Character sets, Adding character set, Languages, Languages -@subsection The Character Set Used for Data and Sorting +@cindex running, queries +@cindex queries, entering +@cindex entering, queries +@node Entering queries, Database use, Connecting-disconnecting, Tutorial +@section Entering Queries -By default, @strong{MySQL} uses the ISO-8859-1 (Latin1) character set -with sorting according to Swedish/Finnish. This is the character set suitable -in the USA and western Europe. +Make sure you are connected to the server, as discussed in the previous +section. Doing so will not in itself select any database to work with, but +that's okay. At this point, it's more important to find out a little about +how to issue queries than to jump right in creating tables, loading data +into them, and retrieving data from them. This section describes the basic +principles of entering commands, using several queries you can try out to +familiarize yourself with how @code{mysql} works. -All standard @strong{MySQL} binaries are compiled with -@code{--with-extra-charsets=complex}. This will add code to all -standard programs to be able to handle @code{latin1} and all multi-byte -character sets within the binary. Other character sets will be -loaded from a character-set definition file when needed. +Here's a simple command that asks the server to tell you its version number +and the current date. Type it in as shown below following the @code{mysql>} +prompt and hit the RETURN key: -The character set determines what characters are allowed in names and how -things are sorted by the @code{ORDER BY} and @code{GROUP BY} clauses of -the @code{SELECT} statement. +@example +mysql> SELECT VERSION(), CURRENT_DATE; ++--------------+--------------+ +| version() | CURRENT_DATE | ++--------------+--------------+ +| 3.22.20a-log | 1999-03-19 | ++--------------+--------------+ +1 row in set (0.01 sec) +mysql> +@end example -You can change the character set with the @code{--default-character-set} -option when you start the server. The character sets available depend -on the @code{--with-charset=charset} and @code{--with-extra-charset= -list-of-charset | complex | all} options to @code{configure}, and the -character set configuration files listed in -@file{SHAREDIR/charsets/Index}. @xref{configure options}. +This query illustrates several things about @code{mysql}: -If you change the character set when running @strong{MySQL} (which may -also change the sort order), you must run myisamchk -r -q on all -tables. Otherwise your indexes may not be ordered correctly. +@itemize @bullet +@item +A command normally consists of a SQL statement followed by a semicolon. +(There are some exceptions where a semicolon is not needed. @code{QUIT}, +mentioned earlier, is one of them. We'll get to others later.) -When a client connects to a @strong{MySQL} server, the server sends the -default character set in use to the client. The client will switch to -use this character set for this connection. +@item +When you issue a command, @code{mysql} sends it to the server for execution +and displays the results, then prints another @code{mysql>} to indicate +that it is ready for another command. -One should use @code{mysql_real_escape_string()} when escaping strings -for a SQL query. @code{mysql_real_escape_string()} is identical to the -old @code{mysql_escape_string()} function, except that it takes the MYSQL -connection handle as the first parameter. +@item +@code{mysql} displays query output as a table (rows and columns). The first +row contains labels for the columns. The rows following are the query +results. Normally, column labels are the names of the columns you fetch from +database tables. If you're retrieving the value of an expression rather than +a table column (as in the example just shown), @code{mysql} labels the column +using the expression itself. -If the client is compiled with different paths than where the server is -installed and the user who configured @strong{MySQL} didn't included all -character sets in the @strong{MySQL} binary, one must specify for -the client where it can find the additional character sets it will need -if the server runs with a different character set than the client. +@item +@code{mysql} shows how many rows were returned and how long the query took +to execute, which gives you a rough idea of server performance. These values +are imprecise because they represent wall clock time (not CPU or machine +time), and because they are affected by factors such as server load and +network latency. (For brevity, the ``rows in set'' line is not shown in +the remaining examples in this chapter.) +@end itemize -One can specify this by putting in a @strong{MySQL} option file: +Keywords may be entered in any lettercase. The following queries are +equivalent: @example -[client] -character-sets-dir=/usr/local/mysql/share/mysql/charsets +mysql> SELECT VERSION(), CURRENT_DATE; +mysql> select version(), current_date; +mysql> SeLeCt vErSiOn(), current_DATE; @end example -where the path points to where the dynamic @strong{MySQL} character sets -are stored. - -One can force the client to use specific character set by specifying: +Here's another query. It demonstrates that you can use @code{mysql} as a +simple calculator: @example -[client] -default-character-set=character-set-name +mysql> SELECT SIN(PI()/4), (4+1)*5; ++-------------+---------+ +| SIN(PI()/4) | (4+1)*5 | ++-------------+---------+ +| 0.707107 | 25 | ++-------------+---------+ @end example -but normally this is never needed. - -@cindex character sets, adding -@cindex adding, character sets -@node Adding character set, Character arrays, Character sets, Languages -@subsection Adding a New Character Set - -To add another character set to @strong{MySQL}, use the following procedure. +The commands shown thus far have been relatively short, single-line +statements. You can even enter multiple statements on a single line. +Just end each one with a semicolon: -Decide if the set is simple or complex. If the character set -does not need to use special string collating routines for -sorting and does not need multi-byte character support, it is -simple. If it needs either of those features, it is complex. +@example +mysql> SELECT VERSION(); SELECT NOW(); ++--------------+ +| version() | ++--------------+ +| 3.22.20a-log | ++--------------+ -For example, @code{latin1} and @code{danish} are simple charactersets while -@code{big5} or @code{czech} are complex character sets. ++---------------------+ +| NOW() | ++---------------------+ +| 1999-03-19 00:15:33 | ++---------------------+ +@end example -In the following section, we have assumed that you name your character -set @code{MYSET}. +A command need not be given all on a single line, so lengthy commands that +require several lines are not a problem. @code{mysql} determines where your +statement ends by looking for the terminating semicolon, not by looking for +the end of the input line. (In other words, @code{mysql} +accepts free-format input: it collects input lines but does not execute them +until it sees the semicolon.) -For a simple character set do the following: +Here's a simple multiple-line statement: -@enumerate -@item -Add MYSET to the end of the @file{sql/share/charsets/Index} file -Assign an unique number to it. +@example +mysql> SELECT + -> USER() + -> , + -> CURRENT_DATE; ++--------------------+--------------+ +| USER() | CURRENT_DATE | ++--------------------+--------------+ +| joesmith@@localhost | 1999-03-18 | ++--------------------+--------------+ +@end example -@item -Create the file @file{sql/share/charsets/MYSET.conf}. -(You can use @file{sql/share/charsets/latin1.conf} as a base for this). +In this example, notice how the prompt changes from @code{mysql>} to +@code{->} after you enter the first line of a multiple-line query. This is +how @code{mysql} indicates that it hasn't seen a complete statement and is +waiting for the rest. The prompt is your friend, because it provides +valuable feedback. If you use that feedback, you will always be aware of +what @code{mysql} is waiting for. -The syntax for the file very simple: +If you decide you don't want to execute a command that you are in the +process of entering, cancel it by typing @code{\c}: -@itemize @bullet -@item -Comments start with a '#' character and proceed to the end of the line. -@item -Words are separated by arbitrary amounts of whitespace. -@item -When defining the character set, every word must be a number in hexadecimal -format -@item -The @code{ctype} array takes up the first 257 words. The -@code{to_lower}, @code{to_upper} and @code{sort_order} arrays take up -256 words each after that. -@end itemize +@example +mysql> SELECT + -> USER() + -> \c +mysql> +@end example -@xref{Character arrays}. +Here, too, notice the prompt. It switches back to @code{mysql>} after you +type @code{\c}, providing feedback to indicate that @code{mysql} is ready +for a new command. -@item -Add the character set name to the @code{CHARSETS_AVAILABLE} and -@code{COMPILED_CHARSETS} lists in @code{configure.in}. +The following table shows each of the prompts you may see and summarizes what +they mean about the state that @code{mysql} is in: -@item -Reconfigure, recompile, and test. +@cindex prompts, meanings +@multitable @columnfractions .10 .9 +@item @strong{Prompt} @tab @strong{Meaning} +@item @code{mysql>} @tab Ready for new command. +@item @code{@ @ @ @ ->} @tab Waiting for next line of multiple-line command. +@item @code{@ @ @ @ '>} @tab Waiting for next line, collecting a string that begins +with a single quote (@samp{'}). +@item @code{@ @ @ @ ">} @tab Waiting for next line, collecting a string that begins +with a double quote (@samp{"}). +@end multitable -@end enumerate +Multiple-line statements commonly occur by accident when you intend to +issue a command on a single line, but forget the terminating semicolon. In +this case, @code{mysql} waits for more input: -For a complex character set do the following: +@example +mysql> SELECT USER() + -> +@end example -@enumerate -@item -Create the file @file{strings/ctype-MYSET.c} in the @strong{MySQL} source -distribution. +If this happens to you (you think you've entered a statement but the only +response is a @code{->} prompt), most likely @code{mysql} is waiting for the +semicolon. If you don't notice what the prompt is telling you, you might sit +there for a while before realizing what you need to do. Enter a semicolon to +complete the statement, and @code{mysql} will execute it: -@item -Add MYSET to the end of the @file{sql/share/charsets/Index} file. -Assign an unique number to it. +@example +mysql> SELECT USER() + -> ; ++--------------------+ +| USER() | ++--------------------+ +| joesmith@@localhost | ++--------------------+ +@end example -@item -Look at one of the existing @file{ctype-*.c} files to see what needs to -be defined, for example @file{strings/ctype-big5.c}. Note that the -arrays in your file must have names like @code{ctype_MYSET}, -@code{to_lower_MYSET}, and so on. This corresponds to the arrays -in the simple character set. @xref{Character arrays}. For a complex -character set +The @code{'>} and @code{">} prompts occur during string collection. +In @strong{MySQL}, you can write strings surrounded by either @samp{'} +or @samp{"} characters (for example, @code{'hello'} or @code{"goodbye"}), +and @code{mysql} lets you enter strings that span multiple lines. When you +see a @code{'>} or @code{">} prompt, it means that you've entered a line +containing a string that begins with a @samp{'} or @samp{"} quote character, +but have not yet entered the matching quote that terminates the string. +That's fine if you really are entering a multiple-line string, but how likely +is that? Not very. More often, the @code{'>} and @code{">} prompts indicate +that you've inadvertantly left out a quote character. For example: -@item -Near the top of the file, place a special comment like this: +@example +mysql> SELECT * FROM my_table WHERE name = "Smith AND age < 30; + "> +@end example + +If you enter this @code{SELECT} statement, then hit RETURN and wait for the +result, nothing will happen. Instead of wondering why this +query takes so long, notice the clue provided by the @code{">} prompt. It +tells you that @code{mysql} expects to see the rest of an unterminated +string. (Do you see the error in the statement? The string @code{"Smith} is +missing the second quote.) + +At this point, what do you do? The simplest thing is to cancel the command. +However, you cannot just type @code{\c} in this case, because @code{mysql} +interprets it as part of the string that it is collecting! Instead, enter +the closing quote character (so @code{mysql} knows you've finished the +string), then type @code{\c}: @example -/* - * This comment is parsed by configure to create ctype.c, - * so don't change it unless you know what you are doing. - * - * .configure. number_MYSET=MYNUMBER - * .configure. strxfrm_multiply_MYSET=N - * .configure. mbmaxlen_MYSET=N - */ +mysql> SELECT * FROM my_table WHERE name = "Smith AND age < 30; + "> "\c +mysql> @end example -The @code{configure} program uses this comment to include -the character set into the @strong{MySQL} library automatically. +The prompt changes back to @code{mysql>}, indicating that @code{mysql} +is ready for a new command. -The strxfrm_multiply and mbmaxlen lines will be explained in -the following sections. Only include them if you the string -collating functions or the multi-byte character set functions, -respectively. +It's important to know what the @code{'>} and @code{">} prompts signify, +because if you mistakenly enter an unterminated string, any further lines you +type will appear to be ignored by @code{mysql} --- including a line +containing @code{QUIT}! This can be quite confusing, especially if you +don't know that you need to supply the terminating quote before you can +cancel the current command. -@item -You should then create some of the following functions: +@cindex databases, creating +@cindex databases, using +@cindex creating, databases +@node Database use, Getting information, Entering queries, Tutorial +@section Creating and Using a Database + +@menu +* Creating database:: Creating a database +* Creating tables:: Creating a table +* Loading tables:: Loading data into a table +* Retrieving data:: Retrieving information from a table +@end menu + +Now that you know how to enter commands, it's time to access a database. + +Suppose you have several pets in your home (your menagerie) and you'd +like to keep track of various types of information about them. You can do so +by creating tables to hold your data and loading them with the desired +information. Then you can answer different sorts of questions about your +animals by retrieving data from the tables. This section shows you how to: @itemize @bullet -@item @code{my_strncoll_MYSET()} -@item @code{my_strcoll_MYSET()} -@item @code{my_strxfrm_MYSET()} -@item @code{my_like_range_MYSET()} +@item +Create a database +@item +Create a table +@item +Load data into the table +@item +Retrieve data from the table in various ways +@item +Use multiple tables @end itemize -@xref{String collating}. +The menagerie database will be simple (deliberately), but it is not difficult +to think of real-world situations in which a similar type of database might +be used. For example, a database like this could be used by a farmer to keep +track of livestock, or by a veterinarian to keep track of patient records. +A menagerie distribution containing some of the queries and sample data used +in the following sections can be obtained from the @strong{MySQL} Web site. +It's available in either +@uref{http://www.mysql.com/Downloads/Contrib/Examples/menagerie.tar.gz,compressed @code{tar} format} +or +@uref{http://www.mysql.com/Downloads/Contrib/Examples/menagerie.zip,Zip format}. -@item -Add the character set name to the @code{CHARSETS_AVAILABLE} and -@code{COMPILED_CHARSETS} lists in @code{configure.in}. +Use the @code{SHOW} statement to find out what databases currently exist +on the server: -@item -Reconfigure, recompile, and test. -@end enumerate +@example +mysql> SHOW DATABASES; ++----------+ +| Database | ++----------+ +| mysql | +| test | +| tmp | ++----------+ +@end example -The file @file{sql/share/charsets/README} includes some more instructions. +The list of databases is probably different on your machine, but the +@code{mysql} and @code{test} databases are likely to be among them. The +@code{mysql} database is required because it describes user access +privileges. The @code{test} database is often provided as a workspace for +users to try things out. -If you want to have the character set included in the @strong{MySQL} -distribution, mail a patch to @email{internals@@lists.mysql.com}. +If the @code{test} database exists, try to access it: -@node Character arrays, String collating, Adding character set, Languages -@subsection The character definition arrays +@example +mysql> USE test +Database changed +@end example -@code{to_lower[]} and @code{to_upper[]} are simple arrays that hold the -lowercase and uppercase characters corresponding to each member of the -character set. For example: +Note that @code{USE}, like @code{QUIT}, does not require a semicolon. (You +can terminate such statements with a semicolon if you like; it does no harm.) +The @code{USE} statement is special in another way, too: it must be given on +a single line. + +You can use the @code{test} database (if you have access to it) for the +examples that follow, but anything you create in that database can be +removed by anyone else with access to it. For this reason, you should +probably ask your @strong{MySQL} administrator for permission to use a +database of your own. Suppose you want to call yours @code{menagerie}. The +administrator needs to execute a command like this: @example -to_lower['A'] should contain 'a' -to_upper['a'] should contain 'A' +mysql> GRANT ALL ON menagerie.* TO your_mysql_name; @end example -@code{sort_order[]} is a map indicating how characters should be ordered for -comparison and sorting purposes. For many character sets, this is the same as -@code{to_upper[]} (which means sorting will be case insensitive). -@strong{MySQL} will sort characters based on the value of -@code{sort_order[character]}. For more complicated sorting rules, see -the discussion of string collating below. @xref{String collating}. +where @code{your_mysql_name} is the @strong{MySQL} user name assigned to +you. -@code{ctype[]} is an array of bit values, with one element for one character. -(Note that @code{to_lower[]}, @code{to_upper[]}, and @code{sort_order[]} -are indexed by character value, but @code{ctype[]} is indexed by character -value + 1. This is an old legacy to be able to handle EOF.) +@cindex selecting, databases +@cindex databases, selecting +@node Creating database, Creating tables, Database use, Database use +@subsection Creating and Selecting a Database -You can find the following bitmask definitions in @file{m_ctype.h}: +If the administrator creates your database for you when setting up your +permissions, you can begin using it. Otherwise, you need to create it +yourself: @example -#define _U 01 /* Uppercase */ -#define _L 02 /* Lowercase */ -#define _N 04 /* Numeral (digit) */ -#define _S 010 /* Spacing character */ -#define _P 020 /* Punctuation */ -#define _C 040 /* Control character */ -#define _B 0100 /* Blank */ -#define _X 0200 /* heXadecimal digit */ +mysql> CREATE DATABASE menagerie; @end example -The @code{ctype[]} entry for each character should be the union of the -applicable bitmask values that describe the character. For example, -@code{'A'} is an uppercase character (@code{_U}) as well as a -hexadecimal digit (@code{_X}), so @code{ctype['A'+1]} should contain the -value: +Under Unix, database names are case sensitive (unlike SQL keywords), so you +must always refer to your database as @code{menagerie}, not as +@code{Menagerie}, @code{MENAGERIE}, or some other variant. This is also true +for table names. (Under Windows, this restriction does not apply, although +you must refer to databases and tables using the same lettercase throughout a +given query.) + +Creating a database does not select it for use; you must do that explicitly. +To make @code{menagerie} the current database, use this command: @example -_U + _X = 01 + 0200 = 0201 +mysql> USE menagerie +Database changed @end example -@cindex collating, strings -@cindex string collating -@node String collating, Multi-byte characters, Character arrays, Languages -@subsection String Collating Support +Your database needs to be created only once, but you must select it for use +each time you begin a @code{mysql} session. You can do this by issuing a +@code{USE} statement as shown above. Alternatively, you can select the +database on the command line when you invoke @code{mysql}. Just specify its +name after any connection parameters that you might need to provide. For +example: -If the sorting rules for your language are too complex to be handled -with the simple @code{sort_order[]} table, you need to use the string -collating functions. +@example +shell> mysql -h host -u user -p menagerie +Enter password: ******** +@end example -Right now the best documentation on this is the character sets that are -already implemented. Look at the big5, czech, gbk, sjis, and tis160 -character sets for examples. +Note that @code{menagerie} is not your password on the command just shown. +If you want to supply your password on the command line after the @code{-p} +option, you must do so with no intervening space (for example, as +@code{-pmypassword}, not as @code{-p mypassword}). However, putting your +password on the command line is not recommended, because doing so exposes it +to snooping by other users logged in on your machine. -You must specify the @code{strxfrm_multiply_MYSET=N} value in the -special comment at the top of the file. @code{N} should be set to -the maximum ratio the strings may grow during @code{my_strxfrm_MYSET} (it -must be a positive integer). +@cindex tables, creating +@cindex creating, tables +@node Creating tables, Loading tables, Creating database, Database use +@subsection Creating a Table -@cindex characters, multi-byte -@cindex multi-byte characters -@node Multi-byte characters, , String collating, Languages -@subsection Multi-byte Character Support +Creating the database is the easy part, but at this point it's empty, as +@code{SHOW TABLES} will tell you: -If your want to add support for a new character set that includes -multi-byte characters, you need to use the multi-byte character -functions. +@example +mysql> SHOW TABLES; +Empty set (0.00 sec) +@end example -Right now the best documentation on this is the character sets that are -already implemented. Look at the euc_kr, gb2312, gbk, sjis and ujis -character sets for examples. These are implemented in the -@code{ctype-'charset'.c} files in the @file{strings} directory. +The harder part is deciding what the structure of your database should be: +what tables you will need and what columns will be in each of them. -You must specify the @code{mbmaxlen_MYSET=N} value in the special -comment at the top of the source file. @code{N} should be set to the -size in bytes of the largest character in the set. +You'll want a table that contains a record for each of your pets. This can +be called the @code{pet} table, and it should contain, as a bare minimum, +each animal's name. Because the name by itself is not very interesting, the +table should contain other information. For example, if more than one person +in your family keeps pets, you might want to list each animal's owner. You +might also want to record some basic descriptive information such as species +and sex. -@cindex tables, maximum size -@cindex size of tables -@cindex operating systems, file size limits -@cindex limits, file size -@cindex files, size limits -@node Table size, , Languages, Server -@section How Big MySQL Tables Can Be +How about age? That might be of interest, but it's not a good thing to store +in a database. Age changes as time passes, which means you'd have to update +your records often. Instead, it's better to store a fixed value such as +date of birth. Then, whenever you need age, you can calculate it as the +difference between the current date and the birth date. @strong{MySQL} +provides functions for doing date arithmetic, so this is not difficult. +Storing birth date rather than age has other advantages, too: -@strong{MySQL} Version 3.22 has a 4G limit on table size. With the new -@code{MyISAM} in @strong{MySQL} Version 3.23 the maximum table size is -pushed up to 8 million terabytes (2 ^ 63 bytes). +@itemize @bullet +@item +You can use the database for tasks such as generating reminders for upcoming +pet birthdays. (If you think this type of query is somewhat silly, note that +it is the same question you might ask in the context of a business database +to identify clients to whom you'll soon need to send out birthday greetings, +for that computer-assisted personal touch.) -Note, however, that operating systems have their own file size -limits. Here are some examples: +@item +You can calculate age in relation to dates other than the current date. For +example, if you store death date in the database, you can easily calculate +how old a pet was when it died. +@end itemize -@multitable @columnfractions .5 .5 -@item @strong{Operating System} @tab @strong{File Size Limit} -@item Linux-Intel 32 bit @tab 2G, 4G or more, depends on Linux version -@item Linux-Alpha @tab 8T (?) -@item Solaris 2.5.1 @tab 2G (possible 4G with patch) -@item Solaris 2.6 @tab 4G -@item Solaris 2.7 Intel @tab 4G -@item Solaris 2.7 ULTRA-SPARC @tab 8T (?) -@end multitable +You can probably think of other types of information that would be useful in +the @code{pet} table, but the ones identified so far are sufficient for now: +name, owner, species, sex, birth, and death. -On Linux 2.2 you can get bigger tables than 2G by using the LFS patch for -the ext2 file system. On Linux 2.4 there exists also patches for ReiserFS -to get support for big files. +Use a @code{CREATE TABLE} statement to specify the layout of your table: -This means that the table size for @strong{MySQL} is normally limited by -the operating system. +@example +mysql> CREATE TABLE pet (name VARCHAR(20), owner VARCHAR(20), + -> species VARCHAR(20), sex CHAR(1), birth DATE, death DATE); +@end example -By default, @strong{MySQL} tables have a maximum size of about 4G. You can -check the maximum table size for a table with the @code{SHOW TABLE STATUS} -command or with the @code{myisamchk -dv table_name}. -@xref{SHOW}. +@code{VARCHAR} is a good choice for the @code{name}, @code{owner}, and +@code{species} columns because the column values will vary in length. The +lengths of those columns need not all be the same, and need not be +@code{20}. You can pick any length from @code{1} to @code{255}, whatever +seems most reasonable to you. (If you make a poor choice and it turns +out later that you need a longer field, @strong{MySQL} provides an +@code{ALTER TABLE} statement.) -If you need bigger tables than 4G (and your operating system supports -this), you should set the @code{AVG_ROW_LENGTH} and @code{MAX_ROWS} -parameter when you create your table. @xref{CREATE TABLE}. You can -also set these later with @code{ALTER TABLE}. @xref{ALTER TABLE}. +Animal sex can be represented in a variety of ways, for example, @code{"m"} +and @code{"f"}, or perhaps @code{"male"} and @code{"female"}. It's simplest +to use the single characters @code{"m"} and @code{"f"}. -If your big table is going to be read-only, you could use -@code{myisampack} to merge and compress many tables to one. -@code{myisampack} usually compresses a table by at least 50%, so you can -have, in effect, much bigger tables. @xref{myisampack, , -@code{myisampack}}. +The use of the @code{DATE} data type for the @code{birth} and @code{death} +columns is a fairly obvious choice. -You can go around the operating system file limit for @code{MyISAM} data -files by using the @code{RAID} option. @xref{CREATE TABLE}. +Now that you have created a table, @code{SHOW TABLES} should produce some +output: -Another solution can be the included MERGE library, which allows you to -handle a collection of identical tables as one. @xref{MERGE, MERGE -tables}. +@example +mysql> SHOW TABLES; ++---------------------+ +| Tables in menagerie | ++---------------------+ +| pet | ++---------------------+ +@end example -@cindex replication -@cindex increasing, speed -@cindex speed, increasing -@cindex databases, replicating -@node Replication, Fulltext Search, Server, Top -@chapter Replication in MySQL +To verify that your table was created the way you expected, use +a @code{DESCRIBE} statement: -@menu -* Replication Intro:: Introduction -* Replication Implementation:: Replication Implementation Overview -* Replication HOWTO:: HOWTO -* Replication Features:: Replication Features -* Replication Options:: Replication Options in my.cnf -* Replication SQL:: SQL Commands related to replication -* Replication FAQ:: Frequently Asked Questions about replication -* Replication Problems:: Troubleshooting Replication. -@end menu +@example +mysql> DESCRIBE pet; ++---------+-------------+------+-----+---------+-------+ +| Field | Type | Null | Key | Default | Extra | ++---------+-------------+------+-----+---------+-------+ +| name | varchar(20) | YES | | NULL | | +| owner | varchar(20) | YES | | NULL | | +| species | varchar(20) | YES | | NULL | | +| sex | char(1) | YES | | NULL | | +| birth | date | YES | | NULL | | +| death | date | YES | | NULL | | ++---------+-------------+------+-----+---------+-------+ +@end example -This chapter describes the various replication features in @strong{MySQL}. -It serves as a reference to the options available with replication. -You will be introduced to replication and learn how to implement it. -Towards the end, there are some frequently asked questions and descriptions -of problems and how to solve them. +You can use @code{DESCRIBE} any time, for example, if you forget the names of +the columns in your table or what types they are. -@node Replication Intro, Replication Implementation, Replication, Replication -@section Introduction +@cindex loading, tables +@cindex tables, loading data +@cindex data, loading into tables +@node Loading tables, Retrieving data, Creating tables, Database use +@subsection Loading Data into a Table -One way replication can be used is to increase both robustness and -speed. For robustness you can have two systems and can switch to the backup if -you have problems with the master. The extra speed is achieved by -sending a part of the non-updating queries to the replica server. Of -course this only works if non-updating queries dominate, but that is the -normal case. +After creating your table, you need to populate it. The @code{LOAD DATA} and +@code{INSERT} statements are useful for this. -Starting in Version 3.23.15, @strong{MySQL} supports one-way replication -internally. One server acts as the master, while the other acts as the -slave. Note that one server could play the roles of master in one pair -and slave in the other. The master server keeps a binary log of updates -(@xref{Binary log}.) and an index file to binary logs to keep track of -log rotation. The slave, upon connecting, informs the master where it -left off since the last successfully propagated update, catches up on -the updates, and then blocks and waits for the master to notify it of -the new updates. - -Note that if you are replicating a database, all updates to this -database should be done through the master! +Suppose your pet records can be described as shown below. +(Observe that @strong{MySQL} expects dates in @code{YYYY-MM-DD} format; +this may be different than what you are used to.) -On older servers one can use the update log to do simple replication. -@xref{Log Replication}. +@multitable @columnfractions .16 .16 .16 .16 .16 .16 +@item @strong{name} @tab @strong{owner} @tab @strong{species} @tab @strong{sex} @tab @strong{birth} @tab @strong{death} +@item Fluffy @tab Harold @tab cat @tab f @tab 1993-02-04 @tab +@item Claws @tab Gwen @tab cat @tab m @tab 1994-03-17 @tab +@item Buffy @tab Harold @tab dog @tab f @tab 1989-05-13 @tab +@item Fang @tab Benny @tab dog @tab m @tab 1990-08-27 @tab +@item Bowser @tab Diane @tab dog @tab m @tab 1989-08-31 @tab 1995-07-29 +@item Chirpy @tab Gwen @tab bird @tab f @tab 1998-09-11 @tab +@item Whistler @tab Gwen @tab bird @tab @tab 1997-12-09 @tab +@item Slim @tab Benny @tab snake @tab m @tab 1996-04-29 @tab +@end multitable -Another benefit of using replication is that one can get live backups of -the system by doing a backup on a slave instead of doing it on the -master. @xref{Backup}. +Because you are beginning with an empty table, an easy way to populate it is to +create a text file containing a row for each of your animals, then load the +contents of the file into the table with a single statement. -@cindex master-slave setup -@node Replication Implementation, Replication HOWTO, Replication Intro, Replication -@section Replication Implementation Overview +You could create a text file @file{pet.txt} containing one record per line, +with values separated by tabs, and given in the order in which the columns +were listed in the @code{CREATE TABLE} statement. For missing values (such +as unknown sexes or death dates for animals that are still living), you can +use @code{NULL} values. To represent these in your text file, use +@code{\N}. For example, the record for Whistler the bird would look like +this (where the whitespace between values is a single tab character): -@strong{MySQL} replication is based on the server keeping track of all -changes to your database (updates, deletes, etc) in the binary -log. (@xref{Binary log}.) and the slave server(s) reading the saved -queries from the master server's binary log so that the slave can -execute the same queries on its copy of the data. +@multitable @columnfractions .15 .15 .15 .15 .25 .15 +@item @code{Whistler} @tab @code{Gwen} @tab @code{bird} @tab @code{\N} @tab @code{1997-12-09} @tab @code{\N} +@end multitable -It is @strong{very important} to realize that the binary log is simply a -record starting from a fixed point in time (the moment you enable binary -logging). Any slaves which you set up will need copies of all the data -from your master as it existed the moment that you enabled binary -logging on the master. If you start your slaves with data that doesn't -agree with what was on the master @strong{when the binary log was -started}, your slaves may fail. +To load the text file @file{pet.txt} into the @code{pet} table, use this +command: -A future version (4.0) of @strong{MySQL} will remove the need to keep a -(possibly large) snapshot of data for new slaves that you might wish to -set up through the live backup functionality with no locking required. -However, at this time, it is necessary to block all writes either with a -global read lock or by shutting down the master while taking a snapshot. +@example +mysql> LOAD DATA LOCAL INFILE "pet.txt" INTO TABLE pet; +@end example -Once a slave is properly configured and running, it will simply connect -to the master and wait for updates to process. If the master goes away -or the slave loses connectivity with your master, it will keep trying to -connect every @code{master-connect-retry} seconds until it is able to -reconnect and resume listening for updates. +You can specify the column value separator and end of line marker explicitly +in the @code{LOAD DATA} statement if you wish, but the defaults are tab and +linefeed. These are sufficient for the statement to read the file +@file{pet.txt} properly. -Each slave keeps track of where it left off. The master server has no -knowledge of how many slaves there are or which ones are up-to-date at -any given time. +When you want to add new records one at a time, the @code{INSERT} statement +is useful. In its simplest form, you supply values for each column, in the +order in which the columns were listed in the @code{CREATE TABLE} statement. +Suppose Diane gets a new hamster named Puffball. You could add a new record +using an @code{INSERT} statement like this: -The next section explains the master/slave setup process in more detail. +@example +mysql> INSERT INTO pet + -> VALUES ('Puffball','Diane','hamster','f','1999-03-30',NULL); +@end example -@node Replication HOWTO, Replication Features, Replication Implementation, Replication -@section HOWTO +Note that string and date values are specified as quoted strings here. Also, +with @code{INSERT}, you can insert @code{NULL} directly to represent a +missing value. You do not use @code{\N} like you do with @code{LOAD DATA}. -Below is a quick description of how to set up complete replication on -your current @strong{MySQL} server. It assumes you want to replicate all -your databases and have not configured replication before. You will need -to shutdown your master server briefly to complete the steps outlined -below. +From this example, you should be able to see that there would be a lot more +typing involved to load +your records initially using several @code{INSERT} statements rather +than a single @code{LOAD DATA} statement. -@enumerate -@item -Make sure you have a recent version of @strong{MySQL} installed on the master -and slave(s). +@cindex data, retrieving +@cindex tables, retrieving data +@cindex retrieving, data from tables +@cindex unloading, tables +@node Retrieving data, , Loading tables, Database use +@subsection Retrieving Information from a Table -Use Version 3.23.29 or higher. Previous releases used a different binary -log format and had bugs which have been fixed in newer releases. Please, -do not report bugs until you have verified that the problem is present -in the latest release. -@item -Set up special a replication user on the master with the @code{FILE} -privilege and permission to connect from all the slaves. If the user is -only doing replication (which is recommended), you don't need to grant any -additional privileges. +@menu +* Selecting all:: Selecting all data +* Selecting rows:: Selecting particular rows +* Selecting columns:: Selecting particular columns +* Sorting rows:: Sorting rows +* Date calculations:: Date calculations +* Working with NULL:: Working with @code{NULL} values +* Pattern matching:: Pattern matching +* Counting rows:: Counting rows +* Multiple tables:: +@end menu -For example, to create a user named @code{repl} which can access your -master from any host, you might use this command: +The @code{SELECT} statement is used to pull information from a table. +The general form of the statement is: @example -GRANT FILE ON *.* TO repl@@"%" IDENTIFIED BY '<password>'; +SELECT what_to_select +FROM which_table +WHERE conditions_to_satisfy @end example -@item -Shut down @strong{MySQL} on the master. - -@example -mysqladmin -u root -p<password> shutdown -@end example +@code{what_to_select} indicates what you want to see. This can be a list of +columns, or @code{*} to indicate ``all columns.'' @code{which_table} +indicates the table from which you want to retrieve data. The @code{WHERE} +clause is optional. If it's present, @code{conditions_to_satisfy} specifies +conditions that rows must satisfy to qualify for retrieval. -@item -Snapshot all the data on your master server. +@node Selecting all, Selecting rows, Retrieving data, Retrieving data +@subsubsection Selecting All Data -The easiest way to do this (on Unix) is to simply use @strong{tar} to -produce an archive of your entire data directory. The exact data -directory location depends on your installation. +The simplest form of @code{SELECT} retrieves everything from a table: @example -tar -cvf /tmp/mysql-snapshot.tar /path/to/data-dir +mysql> SELECT * FROM pet; ++----------+--------+---------+------+------------+------------+ +| name | owner | species | sex | birth | death | ++----------+--------+---------+------+------------+------------+ +| Fluffy | Harold | cat | f | 1993-02-04 | NULL | +| Claws | Gwen | cat | m | 1994-03-17 | NULL | +| Buffy | Harold | dog | f | 1989-05-13 | NULL | +| Fang | Benny | dog | m | 1990-08-27 | NULL | +| Bowser | Diane | dog | m | 1998-08-31 | 1995-07-29 | +| Chirpy | Gwen | bird | f | 1998-09-11 | NULL | +| Whistler | Gwen | bird | NULL | 1997-12-09 | NULL | +| Slim | Benny | snake | m | 1996-04-29 | NULL | +| Puffball | Diane | hamster | f | 1999-03-30 | NULL | ++----------+--------+---------+------+------------+------------+ @end example -Windows users can use WinZip or similar software to create an archive of -the data directory. +This form of @code{SELECT} is useful if you want to review your entire table, +for instance, after you've just loaded it with your initial dataset. As it +happens, the output just shown reveals an error in your data file: Bowser +appears to have been born after he died! Consulting your original pedigree +papers, you find that the correct birth year is 1989, not 1998. + +There are are least a couple of ways to fix this: +@itemize @bullet @item -In @code{my.cnf} on the master add @code{log-bin} and -@code{server-id=unique number} to the @code{[mysqld]} section and -restart it. It is very important that the id of the slave is different from -the id of the master. Think of @code{server-id} as something similar -to the IP address - it uniquely identifies the server instance in the -community of replication partners. +Edit the file @file{pet.txt} to correct the error, then empty the table +and reload it using @code{DELETE} and @code{LOAD DATA}: @example -[mysqld] -log-bin -server-id=1 +mysql> SET AUTOCOMMIT=1; # Used for quick re-create of the table +mysql> DELETE FROM pet; +mysql> LOAD DATA LOCAL INFILE "pet.txt" INTO TABLE pet; @end example -@item -Restart @strong{MySQL} on the master. +However, if you do this, you must also re-enter the record for Puffball. @item -Add the following to @code{my.cnf} on the slave(s): +Fix only the erroneous record with an @code{UPDATE} statement: @example -master-host=<hostname of the master> -master-user=<replication user name> -master-password=<replication user password> -master-port=<TCP/IP port for master> -server-id=<some unique number between 2 and 2^32-1> +mysql> UPDATE pet SET birth = "1989-08-31" WHERE name = "Bowser"; @end example +@end itemize -replacing the values in <> with what is relevant to your system. - -@code{server-id} must be different for each server participating in -replication. If you don't specify a server-id, it will be set to 1 if -you have not defined @code{master-host}, else it will be set to 2. Note -that in the case of @code{server-id} omission the master will refuse -connections from all slaves, and the slave will refuse to connect to a -master. Thus, omitting @code{server-id} is only good for backup with a -binary log. +As shown above, it is easy to retrieve an entire table. But typically you +don't want to do that, particularly when the table becomes large. Instead, +you're usually more interested in answering a particular question, in which +case you specify some constraints on the information you want. Let's look at +some selection queries in terms of questions about your pets that they +answer. +@cindex rows, selecting +@cindex tables, selecting rows +@node Selecting rows, Selecting columns, Selecting all, Retrieving data +@subsubsection Selecting Particular Rows -@item -Copy the snapshot data into your data directory on your slave(s). Make -sure that the privileges on the files and directories are correct. The -user which @strong{MySQL} runs as needs to be able to read and write to -them, just as on the master. +You can select only particular rows from your table. For example, if you want +to verify the change that you made to Bowser's birth date, select Bowser's +record like this: -@item Restart the slave(s). +@example +mysql> SELECT * FROM pet WHERE name = "Bowser"; ++--------+-------+---------+------+------------+------------+ +| name | owner | species | sex | birth | death | ++--------+-------+---------+------+------------+------------+ +| Bowser | Diane | dog | m | 1989-08-31 | 1995-07-29 | ++--------+-------+---------+------+------------+------------+ +@end example -@end enumerate +The output confirms that the year is correctly recorded now as 1989, not 1998. -After you have done the above, the slave(s) should connect to the master -and catch up on any updates which happened since the snapshot was taken. +String comparisons are normally case insensitive, so you can specify the +name as @code{"bowser"}, @code{"BOWSER"}, etc. The query result will be +the same. -If you have forgotten to set @code{server-id} for the slave you will get -the following error in the error log file: +You can specify conditions on any column, not just @code{name}. For example, +if you want to know which animals were born after 1998, test the @code{birth} +column: @example -Warning: one should set server_id to a non-0 value if master_host is set. -The server will not act as a slave. +mysql> SELECT * FROM pet WHERE birth >= "1998-1-1"; ++----------+-------+---------+------+------------+-------+ +| name | owner | species | sex | birth | death | ++----------+-------+---------+------+------------+-------+ +| Chirpy | Gwen | bird | f | 1998-09-11 | NULL | +| Puffball | Diane | hamster | f | 1999-03-30 | NULL | ++----------+-------+---------+------+------------+-------+ @end example -If you have forgot to do this for the master, the slaves will not be -able to connect to the master. - -If a slave is not able to replicate for any reason, you will find error -messages in the error log on the slave. - -Once a slave is replicating, you will find a file called -@code{master.info} in the same directory as your error log. The -@code{master.info} file is used by the slave to keep track of how much -of the master's binary log is has processed. @strong{Do not} remove or -edit the file, unless you really know what you are doing. Even in that case, -it is preferred that you use @code{CHANGE MASTER TO} command. +You can combine conditions, for example, to locate female dogs: -@cindex options, replication -@cindex @code{my.cnf} file -@cindex files,@code{my.cnf} -@node Replication Features, Replication Options, Replication HOWTO, Replication -@section Replication Features and known problems +@example +mysql> SELECT * FROM pet WHERE species = "dog" AND sex = "f"; ++-------+--------+---------+------+------------+-------+ +| name | owner | species | sex | birth | death | ++-------+--------+---------+------+------------+-------+ +| Buffy | Harold | dog | f | 1989-05-13 | NULL | ++-------+--------+---------+------+------------+-------+ +@end example -Below is an explanation of what is supported and what is not: +The preceding query uses the @code{AND} logical operator. There is also an +@code{OR} operator: -@itemize @bullet -@item -Replication will be done correctly with @code{AUTO_INCREMENT}, -@code{LAST_INSERT_ID}, and @code{TIMESTAMP} values. -@item -@code{RAND()} in updates does not replicate properly. Use -@code{RAND(some_non_rand_expr)} if you are replicating updates with -@code{RAND()}. You can, for example, use @code{UNIX_TIMESTAMP()} for the -argument to @code{RAND()}. -@item -You have to use the same character set (@code{--default-character-set}) -on the master and the slave. If not, you may get duplicate key errors on -the slave, because a key that is regarded as unique on the master may -not be that in the other character set. -@item -@code{LOAD DATA INFILE} will be handled properly as long as the file -still resides on the master server at the time of update -propagation. @code{LOAD LOCAL DATA INFILE} will be skipped. -@item -Update queries that use user variables are not replication-safe (yet). -@item -@code{FLUSH} commands are not stored in the binary log and are because -of this not replicated to the slaves. This is not normally a problem as -@code{FLUSH} doesn't change anything. This does however mean that if you -update the @code{MySQL} privilege tables directly without using -@code{GRANT} statement and you replicate the @code{MySQL} privilege -database, you must do a @code{FLUSH PRIVILEGES} on your slaves to put -the new privileges into effect. -@item -Temporary tables starting in 3.23.29 are replicated properly with the -exception of the case when you shut down slave server ( not just slave thread), -you have some temporary tables open, and the are used in subsequent updates. -To deal with this problem, to shut down the slave, do @code{SLAVE STOP}, then -check @code{Slave_open_temp_tables} variable to see if it is 0, then issue -@code{mysqladmin shutdown}. If the number is not 0, restart the slave thread -with @code{SLAVE START} and see -if you have better luck next time. There will be a cleaner solution, but it -has to wait until version 4.0. -In earlier versions temporary tables are not being replicated properly - we -recommend that you either upgrade, or execute @code{SET SQL_LOG_BIN=0} on -your clients before all queries with temp tables. -@item -@strong{MySQL} only supports one master and many slaves. We will in 4.x -add a voting algorithm to automatically change master if something goes -wrong with the current master. We will also introduce 'agent' processes -to help doing load balancing by sending select queries to different -slaves. -@item -Starting in Version 3.23.26, it is safe to connect servers in a circular -master-slave relationship with @code{log-slave-updates} enabled. -Note, however, that many queries will not work right in this kind of -setup unless your client code is written to take care of the potential -problems that can happen from updates that occur in different sequence -on different servers. +@example +mysql> SELECT * FROM pet WHERE species = "snake" OR species = "bird"; ++----------+-------+---------+------+------------+-------+ +| name | owner | species | sex | birth | death | ++----------+-------+---------+------+------------+-------+ +| Chirpy | Gwen | bird | f | 1998-09-11 | NULL | +| Whistler | Gwen | bird | NULL | 1997-12-09 | NULL | +| Slim | Benny | snake | m | 1996-04-29 | NULL | ++----------+-------+---------+------+------------+-------+ +@end example -This means that you can do a setup like the following: +@code{AND} and @code{OR} may be intermixed. If you do that, it's a good idea +to use parentheses to indicate how conditions should be grouped: @example -A -> B -> C -> A +mysql> SELECT * FROM pet WHERE (species = "cat" AND sex = "m") + -> OR (species = "dog" AND sex = "f"); ++-------+--------+---------+------+------------+-------+ +| name | owner | species | sex | birth | death | ++-------+--------+---------+------+------------+-------+ +| Claws | Gwen | cat | m | 1994-03-17 | NULL | +| Buffy | Harold | dog | f | 1989-05-13 | NULL | ++-------+--------+---------+------+------------+-------+ @end example -This setup will only works if you only do non conflicting updates -between the tables. In other words, if you insert data in A and C, you -should never insert a row in A that may have a conflicting key with a -row insert in C. You should also not update the sam rows on two servers -if the order in which the updates are applied matters. +@cindex columns, selecting +@cindex tables, selecting columns +@node Selecting columns, Sorting rows, Selecting rows, Retrieving data +@subsubsection Selecting Particular Columns -Note that the log format has changed in Version 3.23.26 so that -pre-3.23.26 slaves will not be able to read it. -@item -If the query on the slave gets an error, the slave thread will -terminate, and a message will appear in the @code{.err} file. You should -then connect to the slave manually, fix the cause of the error (for -example, non-existent table), and then run @code{SLAVE START} sql -command (available starting in Version 3.23.16). In Version 3.23.15, you -will have to restart the server. -@item -If connection to the master is lost, the slave will retry immediately, -and then in case of failure every @code{master-connect-retry} (default -60) seconds. Because of this, it is safe to shut down the master, and -then restart it after a while. The slave will also be able to deal with -network connectivity outages. -@item -Shutting down the slave (cleanly) is also safe, as it keeps track of -where it left off. Unclean shutdowns might produce problems, especially -if disk cache was not synced before the system died. Your system fault -tolerance will be greatly increased if you have a good UPS. -@item -If the master is listening on a non-standard port, you will also need to -specify this with @code{master-port} parameter in @code{my.cnf} . -@item -In Version 3.23.15, all of the tables and databases will be -replicated. Starting in Version 3.23.16, you can restrict replication to -a set of databases with @code{replicate-do-db} directives in -@code{my.cnf} or just exclude a set of databases with -@code{replicate-ignore-db}. Note that up until Version 3.23.23, there was a bug -that did not properly deal with @code{LOAD DATA INFILE} if you did it in -a database that was excluded from replication. -@item -Starting in Version 3.23.16, @code{SET SQL_LOG_BIN = 0} will turn off -replication (binary) logging on the master, and @code{SET SQL_LOG_BIN = -1} will turn in back on - you must have the process privilege to do -this. -@item -Starting in Version 3.23.19, you can clean up stale replication leftovers when -something goes wrong and you want a clean start with @code{FLUSH MASTER} -and @code{FLUSH SLAVE} commands. In Version 3.23.26 we have renamed them to -@code{RESET MASTER} and @code{RESET SLAVE} respectively to clarify -what they do. The old @code{FLUSH} variants still work, though, for -compatibility. -@item -Starting in Version 3.23.21, you can use @code{LOAD TABLE FROM MASTER} for -network backup and to set up replication initially. We have recently -received a number of bug reports concerning it that we are investigating, so -we recommend that you use it only in testing until we make it more stable. -@item -Starting in Version 3.23.23, you can change masters and adjust log position -with @code{CHANGE MASTER TO}. -@item -Starting in Version 3.23.23, you tell the master that updates in certain -databases should not be logged to the binary log with @code{binlog-ignore-db}. -@item -Starting in Version 3.23.26, you can use @code{replicate-rewrite-db} to tell -the slave to apply updates from one database on the master to the one -with a different name on the slave. -@item -Starting in Version 3.23.28, you can use @code{PURGE MASTER LOGS TO 'log-name'} -to get rid of old logs while the slave is running. -@end itemize +If you don't want to see entire rows from your table, just name the columns +in which you're interested, separated by commas. For example, if you want to +know when your animals were born, select the @code{name} and @code{birth} +columns: -@node Replication Options, Replication SQL, Replication Features, Replication -@section Replication Options in my.cnf +@example +mysql> SELECT name, birth FROM pet; ++----------+------------+ +| name | birth | ++----------+------------+ +| Fluffy | 1993-02-04 | +| Claws | 1994-03-17 | +| Buffy | 1989-05-13 | +| Fang | 1990-08-27 | +| Bowser | 1989-08-31 | +| Chirpy | 1998-09-11 | +| Whistler | 1997-12-09 | +| Slim | 1996-04-29 | +| Puffball | 1999-03-30 | ++----------+------------+ +@end example -If you are using replication, we recommend you to use @strong{MySQL} Version -3.23.30 or later. Older versions work, but they do have some bugs and are -missing some features. +To find out who owns pets, use this query: -On both master and slave you need to use the @code{server-id} option. -This sets an unique replication id. You should pick a unique value in the -range between 1 to 2^32-1 for each master and slave. -Example: @code{server-id=3} +@example +mysql> SELECT owner FROM pet; ++--------+ +| owner | ++--------+ +| Harold | +| Gwen | +| Harold | +| Benny | +| Diane | +| Gwen | +| Gwen | +| Benny | +| Diane | ++--------+ +@end example -The following table has the options you can use for the @strong{MASTER}: +@findex DISTINCT +However, notice that the query simply retrieves the @code{owner} field from +each record, and some of them appear more than once. To minimize the output, +retrieve each unique output record just once by adding the keyword +@code{DISTINCT}: -@multitable @columnfractions .3 .7 +@example +mysql> SELECT DISTINCT owner FROM pet; ++--------+ +| owner | ++--------+ +| Benny | +| Diane | +| Gwen | +| Harold | ++--------+ +@end example -@item @strong{Option} @tab @strong{Description} -@item @code{log-bin=filename} @tab -Write to a binary update log to the specified location. Note that if you -give it a parameter with an extension (for example, -@code{log-bin=/mysql/logs/replication.log} ) versions up to 3.23.24 will -not work right during replication if you do @code{FLUSH LOGS} . The -problem is fixed in Version 3.23.25. If you are using this kind of log -name, @code{FLUSH LOGS} will be ignored on binlog. To clear the log, run -@code{FLUSH MASTER}, and do not forget to run @code{FLUSH SLAVE} on all -slaves. In Version 3.23.26 and in later versions you should use -@code{RESET MASTER} and @code{RESET SLAVE} +You can use a @code{WHERE} clause to combine row selection with column +selection. For example, to get birth dates for dogs and cats only, +use this query: -@item @code{log-bin-index=filename} @tab -Because the user could issue the @code{FLUSH LOGS} command, we need to -know which log is currently active and which ones have been rotated out -and in what sequence. This information is stored in the binary log index file. -The default is `hostname`.index. You can use this option if you want to -be a rebel. (Example: @code{log-bin-index=db.index}) +@example +mysql> SELECT name, species, birth FROM pet + -> WHERE species = "dog" OR species = "cat"; ++--------+---------+------------+ +| name | species | birth | ++--------+---------+------------+ +| Fluffy | cat | 1993-02-04 | +| Claws | cat | 1994-03-17 | +| Buffy | dog | 1989-05-13 | +| Fang | dog | 1990-08-27 | +| Bowser | dog | 1989-08-31 | ++--------+---------+------------+ +@end example -@item @code{sql-bin-update-same} @tab -If set, setting @code{SQL_LOG_BIN} to a value will automatically set -@code{SQL_LOG_UPDATE} to the same value and vice versa. +@cindex rows, sorting +@cindex sorting, table rows +@cindex sorting, data +@cindex tables, sorting rows +@node Sorting rows, Date calculations, Selecting columns, Retrieving data +@subsubsection Sorting Rows -@item @code{binlog-do-db=database_name} @tab -Tells the master it should log updates for the specified database, and -exclude all others not explicitly mentioned. -(Example: @code{binlog-do-db=some_database}) +You may have noticed in the preceding examples that the result rows are +displayed in no particular order. However, it's often easier to examine +query output when the rows are sorted in some meaningful way. To sort a +result, use an @code{ORDER BY} clause. -@item @code{binlog-ignore-db=database_name} @tab -Tells the master that updates to the given database should not be logged -to the binary log (Example: @code{binlog-ignore-db=some_database}) -@end multitable +Here are animal birthdays, sorted by date: -The following table has the options you can use for the @strong{SLAVE}: +@example +mysql> SELECT name, birth FROM pet ORDER BY birth; ++----------+------------+ +| name | birth | ++----------+------------+ +| Buffy | 1989-05-13 | +| Bowser | 1989-08-31 | +| Fang | 1990-08-27 | +| Fluffy | 1993-02-04 | +| Claws | 1994-03-17 | +| Slim | 1996-04-29 | +| Whistler | 1997-12-09 | +| Chirpy | 1998-09-11 | +| Puffball | 1999-03-30 | ++----------+------------+ +@end example -@multitable @columnfractions .3 .7 +To sort in reverse order, add the @code{DESC} (descending) keyword to the +name of the column you are sorting by: -@item @strong{Option} @tab @strong{Description} -@item @code{master-host=host} @tab -Master hostname or IP address for replication. If not set, the slave -thread will not be started. -(Example: @code{master-host=db-master.mycompany.com}) +@example +mysql> SELECT name, birth FROM pet ORDER BY birth DESC; ++----------+------------+ +| name | birth | ++----------+------------+ +| Puffball | 1999-03-30 | +| Chirpy | 1998-09-11 | +| Whistler | 1997-12-09 | +| Slim | 1996-04-29 | +| Claws | 1994-03-17 | +| Fluffy | 1993-02-04 | +| Fang | 1990-08-27 | +| Bowser | 1989-08-31 | +| Buffy | 1989-05-13 | ++----------+------------+ +@end example -@item @code{master-user=username} @tab -The user the slave thread will us for authentication when connecting to -the master. The user must have @code{FILE} privilege. If the master user -is not set, user @code{test} is assumed. (Example: -@code{master-user=scott}) +You can sort on multiple columns. For example, to sort by type of +animal, then by birth date within animal type with youngest animals first, +use the following query: -@item @code{master-password=password} @tab -The password the slave thread will authenticate with when connecting to -the master. If not set, an empty password is assumed. (Example: -@code{master-password=tiger}) +@example +mysql> SELECT name, species, birth FROM pet ORDER BY species, birth DESC; ++----------+---------+------------+ +| name | species | birth | ++----------+---------+------------+ +| Chirpy | bird | 1998-09-11 | +| Whistler | bird | 1997-12-09 | +| Claws | cat | 1994-03-17 | +| Fluffy | cat | 1993-02-04 | +| Fang | dog | 1990-08-27 | +| Bowser | dog | 1989-08-31 | +| Buffy | dog | 1989-05-13 | +| Puffball | hamster | 1999-03-30 | +| Slim | snake | 1996-04-29 | ++----------+---------+------------+ +@end example -@item @code{master-port=portnumber} @tab -The port the master is listening on. If not set, the compiled setting of -@code{MYSQL_PORT} is assumed. If you have not tinkered with -@code{configure} options, this should be 3306. (Example: -@code{master-port=3306}) +Note that the @code{DESC} keyword applies only to the column name immediately +preceding it (@code{birth}); @code{species} values are still sorted in +ascending order. -@item @code{master-connect-retry=seconds} @tab -The number of seconds the slave thread will sleep before retrying to -connect to the master in case the master goes down or the connection is -lost. Default is 60. (Example: @code{master-connect-retry=60}) +@cindex date calculations +@cindex calculating, dates +@cindex extracting, dates +@cindex age, calculating +@node Date calculations, Working with NULL, Sorting rows, Retrieving data +@subsubsection Date Calculations -@item @code{master-info-file=filename} @tab -The location of the file that remembers where we left off on the master -during the replication process. The default is master.info in the data -directory. Sasha: The only reason I see for ever changing the default -is the desire to be rebelious. (Example: -@code{master-info-file=master.info}) +@strong{MySQL} provides several functions that you can use to perform +calculations on dates, for example, to calculate ages or extract +parts of dates. -@item @code{replicate-do-table=db_name.table_name} @tab -Tells the slave thread to restrict replication to the specified database. -To specify more than one table, use the directive multiple times, -once for each table. . -(Example: @code{replicate-do-table=some_db.some_table}) +To determine how many years old each of your pets is, compute age as the +difference between the birth date and the current date. Do this by +converting the two dates to days, take the difference, and divide by 365 (the +number of days in a year): -@item @code{replicate-ignore-table=db_name.table_name} @tab -Tells the slave thread to not replicate to the specified table. To -specify more than one table to ignore, use the directive multiple -times, once for each table.(Example: -@code{replicate-ignore-table=db_name.some_table}) +@example +mysql> SELECT name, (TO_DAYS(NOW())-TO_DAYS(birth))/365 FROM pet; ++----------+-------------------------------------+ +| name | (TO_DAYS(NOW())-TO_DAYS(birth))/365 | ++----------+-------------------------------------+ +| Fluffy | 6.15 | +| Claws | 5.04 | +| Buffy | 9.88 | +| Fang | 8.59 | +| Bowser | 9.58 | +| Chirpy | 0.55 | +| Whistler | 1.30 | +| Slim | 2.92 | +| Puffball | 0.00 | ++----------+-------------------------------------+ +@end example -@item @code{replicate-wild-do-table=db_name.table_name} @tab -Tells the slave thread to restrict replication to the tables that match the -specified wildcard pattern. . -To specify more than one table, use the directive multiple times, -once for each table. . -(Example: @code{replicate-do-table=foo%.bar%} will replicate only updates -to tables in all databases that start with foo and whose table names -start with bar) +Although the query works, there are some things about it that could be +improved. First, the result could be scanned more easily if the rows were +presented in some order. Second, the heading for the age column isn't very +meaningful. -@item @code{replicate-wild-ignore-table=db_name.table_name} @tab -Tells the slave thread to not replicate to the tables that match the given -wild card pattern. To -specify more than one table to ignore, use the directive multiple -times, once for each table.(Example: -@code{replicate-ignore-table=foo%.bar%} - will not upates to tables in all databases that start with foo and whose table names -start with bar) +The first problem can be handled by adding an @code{ORDER BY name} clause to +sort the output by name. To deal with the column heading, provide a name for +the column so that a different label appears in the output (this is called a +column alias): -@item @code{replicate-ignore-db=database_name} @tab -Tells the slave thread to not replicate to the specified database. To -specify more than one database to ignore, use the directive multiple -times, once for each database. This option will not work if you use cross -database updates. If you need cross database updates to work, make sure -you have 3.23.28 or later, and use -@code{replicate-wild-ignore-table=db_name.%}(Example: -@code{replicate-ignore-db=some_db}) +@example +mysql> SELECT name, (TO_DAYS(NOW())-TO_DAYS(birth))/365 AS age + -> FROM pet ORDER BY name; ++----------+------+ +| name | age | ++----------+------+ +| Bowser | 9.58 | +| Buffy | 9.88 | +| Chirpy | 0.55 | +| Claws | 5.04 | +| Fang | 8.59 | +| Fluffy | 6.15 | +| Puffball | 0.00 | +| Slim | 2.92 | +| Whistler | 1.30 | ++----------+------+ +@end example -@item @code{replicate-do-db=database_name} @tab -Tells the slave thread to restrict replication to the specified database. -To specify more than one database, use the directive multiple times, -once for each database. Note that this will only work if you do not use -cross-database queries such as @code{UPDATE some_db.some_table SET -foo='bar'} while having selected a different or no database. If you need -cross database updates to work, make sure -you have 3.23.28 or later, and use -@code{replicate-wild-do-table=db_name.%} -(Example: @code{replicate-do-db=some_db}) +To sort the output by @code{age} rather than @code{name}, just use a +different @code{ORDER BY} clause: -@item @code{log-slave-updates} @tab -Tells the slave to log the updates from the slave thread to the binary -log. Off by default. You will need to turn it on if you plan to -daisy-chain the slaves. +@example +mysql> SELECT name, (TO_DAYS(NOW())-TO_DAYS(birth))/365 AS age + -> FROM pet ORDER BY age; ++----------+------+ +| name | age | ++----------+------+ +| Puffball | 0.00 | +| Chirpy | 0.55 | +| Whistler | 1.30 | +| Slim | 2.92 | +| Claws | 5.04 | +| Fluffy | 6.15 | +| Fang | 8.59 | +| Bowser | 9.58 | +| Buffy | 9.88 | ++----------+------+ +@end example -@item @code{replicate-rewrite-db=from_name->to_name} @tab -Updates to a database with a different name than the original (Example: -@code{replicate-rewrite-db=master_db_name->slave_db_name} +A similar query can be used to determine age at death for animals that have +died. You determine which animals these are by checking whether or not the +@code{death} value is @code{NULL}. Then, for those with non-@code{NULL} +values, compute the difference between the @code{death} and @code{birth} +values: -@item @code{skip-slave-start} @tab -Tells the slave server not to start the slave on the startup. The user -can start it later with @code{SLAVE START}. +@example +mysql> SELECT name, birth, death, (TO_DAYS(death)-TO_DAYS(birth))/365 AS age + -> FROM pet WHERE death IS NOT NULL ORDER BY age; ++--------+------------+------------+------+ +| name | birth | death | age | ++--------+------------+------------+------+ +| Bowser | 1989-08-31 | 1995-07-29 | 5.91 | ++--------+------------+------------+------+ +@end example -@item @code{slave_read_timeout=#} -Number of seconds to wait for more data from the master before aborting -the read. -@end multitable +The query uses @code{death IS NOT NULL} rather than @code{death != NULL} +because @code{NULL} is a special value. This is explained later. +@xref{Working with NULL, , Working with @code{NULL}}. -@cindex SQL commands, replication -@cindex commands, replication -@cindex replication, commands -@node Replication SQL, Replication FAQ, Replication Options, Replication -@section SQL Commands Related to Replication +What if you want to know which animals have birthdays next month? For this +type of calculation, year and day are irrelevant; you simply want to extract +the month part of the @code{birth} column. @strong{MySQL} provides several +date-part extraction functions, such as @code{YEAR()}, @code{MONTH()}, and +@code{DAYOFMONTH()}. @code{MONTH()} is the appropriate function here. To +see how it works, run a simple query that displays the value of both +@code{birth} and @code{MONTH(birth)}: -Replication can be controlled through the SQL interface. Below is the -summary of commands: +@example +mysql> SELECT name, birth, MONTH(birth) FROM pet; ++----------+------------+--------------+ +| name | birth | MONTH(birth) | ++----------+------------+--------------+ +| Fluffy | 1993-02-04 | 2 | +| Claws | 1994-03-17 | 3 | +| Buffy | 1989-05-13 | 5 | +| Fang | 1990-08-27 | 8 | +| Bowser | 1989-08-31 | 8 | +| Chirpy | 1998-09-11 | 9 | +| Whistler | 1997-12-09 | 12 | +| Slim | 1996-04-29 | 4 | +| Puffball | 1999-03-30 | 3 | ++----------+------------+--------------+ +@end example -@multitable @columnfractions .30 .70 -@item @strong{Command} @tab @strong{Description} +Finding animals with birthdays in the upcoming month is easy, too. Suppose +the current month is April. Then the month value is @code{4} and you look +for animals born in May (month 5) like this: -@item @code{SLAVE START} - @tab Starts the slave thread. (Slave) +@example +mysql> SELECT name, birth FROM pet WHERE MONTH(birth) = 5; ++-------+------------+ +| name | birth | ++-------+------------+ +| Buffy | 1989-05-13 | ++-------+------------+ +@end example -@item @code{SLAVE STOP} - @tab Stops the slave thread. (Slave) +There is a small complication if the current month is December, of course. +You don't just add one to the month number (@code{12}) and look for animals +born in month 13, because there is no such month. Instead, you look for +animals born in January (month 1). -@item @code{SET SQL_LOG_BIN=0} - @tab Disables update logging if the user has process privilege. - Ignored otherwise. (Master) +You can even write the query so that it works no matter what the current +month is. That way you don't have to use a particular month number +in the query. @code{DATE_ADD()} allows you to add a time interval to a +given date. If you add a month to the value of @code{NOW()}, then extract +the month part with @code{MONTH()}, the result produces the month in which to +look for birthdays: -@item @code{SET SQL_LOG_BIN=1} - @tab Re-enables update logging if the user has process privilege. - Ignored otherwise. (Master) +@example +mysql> SELECT name, birth FROM pet + -> WHERE MONTH(birth) = MONTH(DATE_ADD(NOW(), INTERVAL 1 MONTH)); +@end example -@item @code{SET SQL_SLAVE_SKIP_COUNTER=n} - @tab Skip the next @code{n} events from the master. Only valid when -the slave thread is not running, otherwise, gives an error. Useful for -recovering from replication glitches. +A different way to accomplish the same task is to add @code{1} to get the +next month after the current one (after using the modulo function (@code{MOD}) +to wrap around the month value to @code{0} if it is currently +@code{12}): -@item @code{RESET MASTER} - @tab Deletes all binary logs listed in the index file, resetting the binlog -index file to be empty. In pre-3.23.26 versions, @code{FLUSH MASTER} (Master) +@example +mysql> SELECT name, birth FROM pet + -> WHERE MONTH(birth) = MOD(MONTH(NOW()), 12) + 1; +@end example -@item @code{RESET SLAVE} - @tab Makes the slave forget its replication position in the master -logs. In pre 3.23.26 versions the command was called -@code{FLUSH SLAVE}(Slave) +Note that @code{MONTH} returns a number between 1 and 12. And +@code{MOD(something,12)} returns a number between 0 and 11. So the +addition has to be after the @code{MOD()} otherwise we would go from +November (11) to January (1). -@item @code{LOAD TABLE tblname FROM MASTER} - @tab Downloads a copy of the table from master to the slave. (Slave) +@findex NULL +@cindex NULL value +@node Working with NULL, Pattern matching, Date calculations, Retrieving data +@subsubsection Working with @code{NULL} Values -@item @code{CHANGE MASTER TO master_def_list} - @tab Changes the master parameters to the values specified in -@code{master_def_list} and restarts the slave thread. @code{master_def_list} -is a comma-separated list of @code{master_def} where @code{master_def} is -one of the following: @code{MASTER_HOST}, @code{MASTER_USER}, -@code{MASTER_PASSWORD}, @code{MASTER_PORT}, @code{MASTER_CONNECT_RETRY}, -@code{MASTER_LOG_FILE}, @code{MASTER_LOG_POS}. Example: +The @code{NULL} value can be surprising until you get used to it. +Conceptually, @code{NULL} means missing value or unknown value and it +is treated somewhat differently than other values. To test for @code{NULL}, +you cannot use the arithmetic comparison operators such as @code{=}, @code{<}, +or @code{!=}. To demonstrate this for yourself, try the following query: @example +mysql> SELECT 1 = NULL, 1 != NULL, 1 < NULL, 1 > NULL; ++----------+-----------+----------+----------+ +| 1 = NULL | 1 != NULL | 1 < NULL | 1 > NULL | ++----------+-----------+----------+----------+ +| NULL | NULL | NULL | NULL | ++----------+-----------+----------+----------+ +@end example -CHANGE MASTER TO - MASTER_HOST='master2.mycompany.com', - MASTER_USER='replication', - MASTER_PASSWORD='bigs3cret', - MASTER_PORT=3306, - MASTER_LOG_FILE='master2-bin.001', - MASTER_LOG_POS=4; +Clearly you get no meaningful results from these comparisons. Use +the @code{IS NULL} and @code{IS NOT NULL} operators instead: +@example +mysql> SELECT 1 IS NULL, 1 IS NOT NULL; ++-----------+---------------+ +| 1 IS NULL | 1 IS NOT NULL | ++-----------+---------------+ +| 0 | 1 | ++-----------+---------------+ @end example -You only need to specify the values that need to be changed. The values that -you omit will stay the same with the exception of when you change the host or -the port. In that case, the slave will assume that since you are connecting to -a different host or a different port, the master is different. Therefore, the -old values of log and position are not applicable anymore, and will -automatically be reset to an empty string and 0, respectively (the start -values). Note that if you restart the slave, it will remember its last master. -If this is not desirable, you should delete the @file{master.info} file before -restarting, and the slave will read its master from @code{my.cnf} or the -command line. (Slave) +In @strong{MySQL}, 0 or @code{NULL} means false and anything else means true. +The default truth value from a boolean operation is 1. -@item @code{SHOW MASTER STATUS} @tab Provides status information on the binlog of the master. (Master) +This special treatment of @code{NULL} is why, in the previous section, it +was necessary to determine which animals are no longer alive using +@code{death IS NOT NULL} instead of @code{death != NULL}. -@item @code{SHOW SLAVE STATUS} @tab Provides status information on essential parameters of the slave thread. (Slave) -@item @code{SHOW MASTER LOGS} @tab Only available starting in Version 3.23.28. Lists the binary logs on the master. You should use this command prior to @code{PURGE MASTER LOGS TO} to find out how far you should go. +@cindex pattern matching +@cindex matching, patterns +@cindex expressions, extended +@node Pattern matching, Counting rows, Working with NULL, Retrieving data +@subsubsection Pattern Matching -@item @code{PURGE MASTER LOGS TO 'logname'} - @tab Available starting in Version 3.23.28. Deletes all the -replication logs that are listed in the log -index as being prior to the specified log, and removed them from the -log index, so that the given log now becomes first. Example: +@strong{MySQL} provides standard SQL pattern matching as well as a form of +pattern matching based on extended regular expressions similar to those used +by Unix utilities such as @code{vi}, @code{grep}, and @code{sed}. + +SQL pattern matching allows you to use @samp{_} to match any single +character and @samp{%} to match an arbitrary number of characters (including +zero characters). In @strong{MySQL}, SQL patterns are case insensitive by +default. Some examples are shown below. Note that you do not use @code{=} +or @code{!=} when you use SQL patterns; use the @code{LIKE} or @code{NOT +LIKE} comparison operators instead. + +To find names beginning with @samp{b}: @example -PURGE MASTER LOGS TO 'mysql-bin.010' +mysql> SELECT * FROM pet WHERE name LIKE "b%"; ++--------+--------+---------+------+------------+------------+ +| name | owner | species | sex | birth | death | ++--------+--------+---------+------+------------+------------+ +| Buffy | Harold | dog | f | 1989-05-13 | NULL | +| Bowser | Diane | dog | m | 1989-08-31 | 1995-07-29 | ++--------+--------+---------+------+------------+------------+ @end example -This command will do nothing and fail with an error if you have an -active slave that is currently reading one of the logs you are trying to -delete. However, if you have a dormant slave, and happen to purge one of -the logs it wants to read, the slave will be unable to replicate once it -comes up. The command is safe to run while slaves are replicating - you -do not need to stop them. +To find names ending with @samp{fy}: -You must first check all the slaves with @code{SHOW SLAVE STATUS} to -see which log they are on, then do a listing of the logs on the -master with @code{SHOW MASTER LOGS}, find the earliest log among all -the slaves (if all the slaves are up to date, this will be the -last log on the list), backup all the logs you are about to delete -(optional) and purge up to the target log. +@example +mysql> SELECT * FROM pet WHERE name LIKE "%fy"; ++--------+--------+---------+------+------------+-------+ +| name | owner | species | sex | birth | death | ++--------+--------+---------+------+------------+-------+ +| Fluffy | Harold | cat | f | 1993-02-04 | NULL | +| Buffy | Harold | dog | f | 1989-05-13 | NULL | ++--------+--------+---------+------+------------+-------+ +@end example -@end multitable +To find names containing a @samp{w}: -@node Replication FAQ, Replication Problems, Replication SQL, Replication -@section Replication FAQ +@example +mysql> SELECT * FROM pet WHERE name LIKE "%w%"; ++----------+-------+---------+------+------------+------------+ +| name | owner | species | sex | birth | death | ++----------+-------+---------+------+------------+------------+ +| Claws | Gwen | cat | m | 1994-03-17 | NULL | +| Bowser | Diane | dog | m | 1989-08-31 | 1995-07-29 | +| Whistler | Gwen | bird | NULL | 1997-12-09 | NULL | ++----------+-------+---------+------+------------+------------+ +@end example -@cindex @code{Binlog_Dump} -@strong{Q}: Why do I sometimes see more than one @code{Binlog_Dump} thread on -the master after I have restarted the slave? +To find names containing exactly five characters, use the @samp{_} pattern +character: -@strong{A}: @code{Binlog_Dump} is a continuous process that is handled by the -server in the following way: +@example +mysql> SELECT * FROM pet WHERE name LIKE "_____"; ++-------+--------+---------+------+------------+-------+ +| name | owner | species | sex | birth | death | ++-------+--------+---------+------+------------+-------+ +| Claws | Gwen | cat | m | 1994-03-17 | NULL | +| Buffy | Harold | dog | f | 1989-05-13 | NULL | ++-------+--------+---------+------+------------+-------+ +@end example + +The other type of pattern matching provided by @strong{MySQL} uses extended +regular expressions. When you test for a match for this type of pattern, use +the @code{REGEXP} and @code{NOT REGEXP} operators (or @code{RLIKE} and +@code{NOT RLIKE}, which are synonyms). + +Some characteristics of extended regular expressions are: @itemize @bullet @item -Catch up on the updates. +@samp{.} matches any single character. + @item -Once there are no more updates left, go into @code{pthread_cond_wait()}, -from which we can be awakened either by an update or a kill. +A character class @samp{[...]} matches any character within the brackets. +For example, @samp{[abc]} matches @samp{a}, @samp{b}, or @samp{c}. To name a +range of characters, use a dash. @samp{[a-z]} matches any lowercase letter, +whereas @samp{[0-9]} matches any digit. + @item -On wake up, check the reason. If we are not supposed to die, continue -the @code{Binlog_dump} loop. +@samp{*} matches zero or more instances of the thing preceding it. For +example, @samp{x*} matches any number of @samp{x} characters, +@samp{[0-9]*} matches any number of digits, and @samp{.*} matches any +number of anything. + @item -If there is some fatal error, such as detecting a dead client, -terminate the loop. +Regular expressions are case sensitive, but you can use a character class to +match both lettercases if you wish. For example, @samp{[aA]} matches +lowercase or uppercase @samp{a} and @samp{[a-zA-Z]} matches any letter in +either case. + +@item +The pattern matches if it occurs anywhere in the value being tested. +(SQL patterns match only if they match the entire value.) + +@item +To anchor a pattern so that it must match the beginning or end of the value +being tested, use @samp{^} at the beginning or @samp{$} at the end of the +pattern. @end itemize -So if the slave thread stops on the slave, the corresponding -@code{Binlog_Dump} thread on the master will not notice it until after -at least one update to the master (or a kill), which is needed to wake -it up from @code{pthread_cond_wait()}. In the meantime, the slave -could have opened another connection, which resulted in another -@code{Binlog_Dump} thread. +To demonstrate how extended regular expressions work, the @code{LIKE} queries +shown above are rewritten below to use @code{REGEXP}. -The above problem should not be present in Version 3.23.26 and later -versions. In Version 3.23.26 we added @code{server-id} to each -replication server, and now all the old zombie threads are killed on the -master when a new replication thread connects from the same slave +To find names beginning with @samp{b}, use @samp{^} to match the beginning of +the name: -@strong{Q}: How do I rotate replication logs? +@example +mysql> SELECT * FROM pet WHERE name REGEXP "^b"; ++--------+--------+---------+------+------------+------------+ +| name | owner | species | sex | birth | death | ++--------+--------+---------+------+------------+------------+ +| Buffy | Harold | dog | f | 1989-05-13 | NULL | +| Bowser | Diane | dog | m | 1989-08-31 | 1995-07-29 | ++--------+--------+---------+------+------------+------------+ +@end example -@strong{A}: In Version 3.23.28 you should use @code{PURGE MASTER LOGS -TO} command after determining which logs can be deleted, and optionally -backing them up first. In earlier versions the process is much more -painful, and cannot be safely done without stopping all the slaves in -the case that you plan to re-use log names. You will need to stop the -slave threads, edit the binary log index file, delete all the old logs, -restart the master, start slave threads, and then remove the old log files. +Prior to @strong{MySQL} Version 3.23.4, @code{REGEXP} is case sensitive, +and the previous query will return no rows. To match either lowercase or +uppercase @samp{b}, use this query instead: +@example +mysql> SELECT * FROM pet WHERE name REGEXP "^[bB]"; +@end example -@strong{Q}: How do I upgrade on a hot replication setup? +From @strong{MySQL} 3.23.4 on, to force a @code{REGEXP} comparison to +be case sensitive, use the @code{BINARY} keyword to make one of the +strings a binary string. This query will match only lowercase @samp{b} +at the beginning of a name: -@strong{A}: If you are upgrading pre-3.23.26 versions, you should just -lock the master tables, let the slave catch up, then run @code{FLUSH -MASTER} on the master, and @code{FLUSH SLAVE} on the slave to reset the -logs, then restart new versions of the master and the slave. Note that -the slave can stay down for some time - since the master is logging -all the updates, the slave will be able to catch up once it is up and -can connect. +@example +mysql> SELECT * FROM pet WHERE name REGEXP BINARY "^b"; +@end example -After 3.23.26, we have locked the replication protocol for modifications, so -you can upgrade masters and slave on the fly to a newer 3.23 version and you -can have different versions of @strong{MySQL} running on the slave and the -master, as long as they are both newer than 3.23.26. +To find names ending with @samp{fy}, use @samp{$} to match the end of the +name: -@cindex replication, two-way -@strong{Q}: What issues should I be aware of when setting up two-way -replication? +@example +mysql> SELECT * FROM pet WHERE name REGEXP "fy$"; ++--------+--------+---------+------+------------+-------+ +| name | owner | species | sex | birth | death | ++--------+--------+---------+------+------------+-------+ +| Fluffy | Harold | cat | f | 1993-02-04 | NULL | +| Buffy | Harold | dog | f | 1989-05-13 | NULL | ++--------+--------+---------+------+------------+-------+ +@end example -@strong{A}: @strong{MySQL} replication currently does not support any -locking protocol between master and slave to guarantee the atomicity of -a distributed (cross-server) update. In in other words, it is possible -for client A to make an update to co-master 1, and in the meantime, -before it propagates to co-master 2, client B could make an update to -co-master 2 that will make the update of client A work differently than -it did on co-master 1. Thus when the update of client A will make it -to co-master 2, it will produce tables that will be different than -what you have on co-master 1, even after all the updates from co-master -2 have also propagated. So you should not co-chain two servers in a -two-way replication relationship, unless you are sure that you updates -can safely happen in any order, or unless you take care of mis-ordered -updates somehow in the client code. +To find names containing a lowercase or uppercase @samp{w}, use this query: +@example +mysql> SELECT * FROM pet WHERE name REGEXP "w"; ++----------+-------+---------+------+------------+------------+ +| name | owner | species | sex | birth | death | ++----------+-------+---------+------+------------+------------+ +| Claws | Gwen | cat | m | 1994-03-17 | NULL | +| Bowser | Diane | dog | m | 1989-08-31 | 1995-07-29 | +| Whistler | Gwen | bird | NULL | 1997-12-09 | NULL | ++----------+-------+---------+------+------------+------------+ +@end example -You must also realize that two-way replication actually does not improve -performance very much, if at all, as far as updates are concerned. Both -servers need to do the same amount of updates each, as you would have -one server do. The only difference is that there will be a little less -lock contention, because the updates originating on another server will -be serialized in one slave thread. This benefit, though, might be -offset by network delays. +Because a regular expression pattern matches if it occurs anywhere in the +value, it is not necessary in the previous query to put a wild card on either +side of the pattern to get it to match the entire value like it would be if +you used a SQL pattern. -@cindex performance, improving -@cindex increasing, performance -@strong{Q}: How can I use replication to improve performance of my system? +To find names containing exactly five characters, use @samp{^} and @samp{$} +to match the beginning and end of the name, and five instances of @samp{.} +in between: -@strong{A}: You should set up one server as the master, and direct all -writes to it, and configure as many slaves as you have the money and -rackspace for, distributing the reads among the master and the slaves. -You can also start the slaves with @code{--skip-bdb}, -@code{--low-priority-updates} and @code{--delay-key-write-for-all-tables} -to get speed improvements for the slave. In this case the slave will -use non-transactional @code{MyISAM} tables instead of @code{BDB} tables -to get more speed. +@example +mysql> SELECT * FROM pet WHERE name REGEXP "^.....$"; ++-------+--------+---------+------+------------+-------+ +| name | owner | species | sex | birth | death | ++-------+--------+---------+------+------------+-------+ +| Claws | Gwen | cat | m | 1994-03-17 | NULL | +| Buffy | Harold | dog | f | 1989-05-13 | NULL | ++-------+--------+---------+------+------------+-------+ +@end example -@strong{Q}: What should I do to prepare my client code to use -performance-enhancing replication? +You could also write the previous query using the @samp{@{n@}} +``repeat-@code{n}-times'' operator: -@strong{A}: -If the part of your code that is responsible for database access has -been properly abstracted/modularized, converting it to run with the -replicated setup should be very smooth and easy - just change the -implementation of your database access to read from some slave or the -master, and to always write to the master. If your code does not have -this level of abstraction, -setting up a replicated system will give you an opportunity/motivation -to it clean up. - You should start by creating a wrapper library -/module with the following functions: +@example +mysql> SELECT * FROM pet WHERE name REGEXP "^.@{5@}$"; ++-------+--------+---------+------+------------+-------+ +| name | owner | species | sex | birth | death | ++-------+--------+---------+------+------------+-------+ +| Claws | Gwen | cat | m | 1994-03-17 | NULL | +| Buffy | Harold | dog | f | 1989-05-13 | NULL | ++-------+--------+---------+------+------------+-------+ +@end example -@itemize -@item -@code{safe_writer_connect()} -@item -@code{safe_reader_connect()} -@item -@code{safe_reader_query()} -@item -@code{safe_writer_query()} -@end itemize +@cindex rows, counting +@cindex tables, counting rows +@cindex counting, table rows +@node Counting rows, Multiple tables, Pattern matching, Retrieving data +@subsubsection Counting Rows -@code{safe_} means that the function will take care of handling all -the error conditions. +Databases are often used to answer the question, ``How often does a certain +type of data occur in a table?'' For example, you might want to know how +many pets you have, or how many pets each owner has, or you might want to +perform various kinds of censuses on your animals. -You should then convert your client code to use the wrapper library. -It may be a painful and scary process at first, but it will pay off in -the long run. All applications that follow the above pattern will be -able to take advantage of one-master/many slaves solution. The -code will be a lot easier to maintain, and adding troubleshooting -options will be trivial. You will just need to modify one or two -functions, for example, to log how long each query took, or which -query, among your many thousands, gave you an error. If you have written a lot of code already, -you may want to automate the conversion task by using Monty's -@code{replace} utility, which comes with the standard distribution of -@strong{MySQL}, or just write your own Perl script. Hopefully, your -code follows some recognizable pattern. If not, then you are probably -better off re-writing it anyway, or at least going through and manually -beating it into a pattern. +Counting the total number of animals you have is the same question as ``How +many rows are in the @code{pet} table?'' because there is one record per pet. +The @code{COUNT()} function counts the number of non-@code{NULL} results, so +the query to count your animals looks like this: -Note that, of course, you can use different names for the -functions. What is important is having unified interface for connecting -for reads, connecting for writes, doing a read, and doing a write. +@example +mysql> SELECT COUNT(*) FROM pet; ++----------+ +| COUNT(*) | ++----------+ +| 9 | ++----------+ +@end example +Earlier, you retrieved the names of the people who owned pets. You can +use @code{COUNT()} if you want to find out how many pets each owner has: -@strong{Q}: When and how much can @strong{MySQL} replication improve the performance -of my system? +@example +mysql> SELECT owner, COUNT(*) FROM pet GROUP BY owner; ++--------+----------+ +| owner | COUNT(*) | ++--------+----------+ +| Benny | 2 | +| Diane | 2 | +| Gwen | 3 | +| Harold | 2 | ++--------+----------+ +@end example -@strong{A}: @strong{MySQL} replication is most beneficial for a system -with frequent reads and not so frequent writes. In theory, by using a -one master/many slaves setup you can scale by adding more slaves until -you either run out of network bandwidth, or your update -load grows to the point -that the master cannot handle it. +Note the use of @code{GROUP BY} to group together all records for each +@code{owner}. Without it, all you get is an error message: -In order to determine how many slaves you can get before the added -benefits begin to level out, and how much you can improve performance -of your site, you need to know your query patterns, and empirically - (by benchmarking) determine the relationship between the throughput -on reads (reads per second, or @code{max_reads}) and on writes -@code{max_writes}) on a typical master and a typical slave. The -example below will show you a rather simplified calculation of what you -can get with replication for our imagined system. +@example +mysql> SELECT owner, COUNT(owner) FROM pet; +ERROR 1140 at line 1: Mixing of GROUP columns (MIN(),MAX(),COUNT()...) +with no GROUP columns is illegal if there is no GROUP BY clause +@end example -Let's say our system load consists of 10% writes and 90% reads, and we -have determined that @code{max_reads} = 1200 - 2 * @code{max_writes}, -or in other words, our system can do 1200 reads per second with no -writes, our average write is twice as slow as average read, -and the relationship is -linear. Let us suppose that our master and slave are of the same -capacity, and we have N slaves and 1 master. Then we have for each -server (master or slave): +@code{COUNT()} and @code{GROUP BY} are useful for characterizing your +data in various ways. The following examples show different ways to +perform animal census operations. -@code{reads = 1200 - 2 * writes} (from bencmarks) +Number of animals per species: -@code{reads = 9* writes / (N + 1) } (reads split, but writes go -to all servers) +@example +mysql> SELECT species, COUNT(*) FROM pet GROUP BY species; ++---------+----------+ +| species | COUNT(*) | ++---------+----------+ +| bird | 2 | +| cat | 2 | +| dog | 3 | +| hamster | 1 | +| snake | 1 | ++---------+----------+ +@end example -@code{9*writes/(N+1) + 2 * writes = 1200} +Number of animals per sex: -@code{writes = 1200/(2 + 9/(N+1)} +@example +mysql> SELECT sex, COUNT(*) FROM pet GROUP BY sex; ++------+----------+ +| sex | COUNT(*) | ++------+----------+ +| NULL | 1 | +| f | 4 | +| m | 4 | ++------+----------+ +@end example -So if N = 0, which means we have no replication, our system can handle -1200/11, about 109 writes per second (which means we will have 9 times -as many reads due to the nature of our application). +(In this output, @code{NULL} indicates sex unknown.) -If N = 1, we can get up to 184 writes per second. +Number of animals per combination of species and sex: -If N = 8, we get up to 400. +@example +mysql> SELECT species, sex, COUNT(*) FROM pet GROUP BY species, sex; ++---------+------+----------+ +| species | sex | COUNT(*) | ++---------+------+----------+ +| bird | NULL | 1 | +| bird | f | 1 | +| cat | f | 1 | +| cat | m | 1 | +| dog | f | 1 | +| dog | m | 2 | +| hamster | f | 1 | +| snake | m | 1 | ++---------+------+----------+ +@end example -If N = 17, 480 writes. +You need not retrieve an entire table when you use @code{COUNT()}. For +example, the previous query, when performed just on dogs and cats, looks like +this: -Eventually as N approaches infinity (and our budget negative infinity), -we can get very close to 600 writes per second, increasing system -throughput about 5.5 times. However, with only 8 servers, we increased -it almost 4 times already. +@example +mysql> SELECT species, sex, COUNT(*) FROM pet + -> WHERE species = "dog" OR species = "cat" + -> GROUP BY species, sex; ++---------+------+----------+ +| species | sex | COUNT(*) | ++---------+------+----------+ +| cat | f | 1 | +| cat | m | 1 | +| dog | f | 1 | +| dog | m | 2 | ++---------+------+----------+ +@end example -Note that our computations assumed infinite network bandwidth, and -neglected several other factors that could turn out to be significant on -your system. In many cases, you may not be able to make a computation -similar to the one above that will accurately predict what will happen -on your system if you add N replication slaves. However, answering the -following questions should help you decided whether and how much, if at -all, the replication will improve the performance of your system: +Or, if you wanted the number of animals per sex only for known-sex animals: -@itemize @bullet -@item -What is the read/write ratio on your system? -@item -How much more write load can one server handle if you reduce the reads? -@item -How many slaves do you have bandwidth for on your network? -@end itemize +@example +mysql> SELECT species, sex, COUNT(*) FROM pet + -> WHERE sex IS NOT NULL + -> GROUP BY species, sex; ++---------+------+----------+ +| species | sex | COUNT(*) | ++---------+------+----------+ +| bird | f | 1 | +| cat | f | 1 | +| cat | m | 1 | +| dog | f | 1 | +| dog | m | 2 | +| hamster | f | 1 | +| snake | m | 1 | ++---------+------+----------+ +@end example -@strong{Q}: How can I use replication to provide redundancy/high -availability? +@cindex tables, multiple +@node Multiple tables, , Counting rows, Retrieving data +@subsubsection Using More Than one Table -@strong{A}: With the currently available features, you would have to -set up a master and a slave (or several slaves), and write a script -that will monitor the -master to see if it is up, and instruct your applications and -the slaves of the master change in case of failure. Some suggestions: +The @code{pet} table keeps track of which pets you have. If you want to +record other information about them, such as events in their lives like +visits to the vet or when litters are born, you need another table. What +should this table look like? It needs: @itemize @bullet @item -To tell a slave to change the master use the @code{CHANGE MASTER TO} command. +To contain the pet name so you know which animal each event pertains +to. + @item -A good way to keep your applications informed where the master is by -having a dynamic DNS entry for the master. With @strong{bind} you can -use @code{nsupdate} to dynamically update your DNS. +A date so you know when the event occurred. + @item -You should run your slaves with the @code{log-bin} option and without -@code{log-slave-updates}. This way the slave will be ready to become a -master as soon as you issue @code{STOP SLAVE}; @code{RESET MASTER}, and -@code{CHANGE MASTER TO} on the other slaves. It will also help you catch -spurious updates that may happen because of misconfiguration of the -slave (ideally, you want to configure access rights so that no client -can update the slave, except for the slave thread) combined with the -bugs in your client programs (they should never update the slave -directly). +A field to describe the event. +@item +An event type field, if you want to be able to categorize events. @end itemize -We are currently working on integrating an automatic master election -system into @strong{MySQL}, but until it is ready, you will have to -create your own monitoring tools. +Given these considerations, the @code{CREATE TABLE} statement for the +@code{event} table might look like this: -@node Replication Problems, , Replication FAQ, Replication -@section Troubleshooting Replication +@example +mysql> CREATE TABLE event (name VARCHAR(20), date DATE, + -> type VARCHAR(15), remark VARCHAR(255)); +@end example -If you have followed the instructions, and your replication setup is not -working, first eliminate the user error factor by checking the following: +As with the @code{pet} table, it's easiest to load the initial records +by creating a tab-delimited text file containing the information: + +@multitable @columnfractions .15 .15 .15 .55 +@item Fluffy @tab 1995-05-15 @tab litter @tab 4 kittens, 3 female, 1 male +@item Buffy @tab 1993-06-23 @tab litter @tab 5 puppies, 2 female, 3 male +@item Buffy @tab 1994-06-19 @tab litter @tab 3 puppies, 3 female +@item Chirpy @tab 1999-03-21 @tab vet @tab needed beak straightened +@item Slim @tab 1997-08-03 @tab vet @tab broken rib +@item Bowser @tab 1991-10-12 @tab kennel +@item Fang @tab 1991-10-12 @tab kennel +@item Fang @tab 1998-08-28 @tab birthday @tab Gave him a new chew toy +@item Claws @tab 1998-03-17 @tab birthday @tab Gave him a new flea collar +@item Whistler @tab 1998-12-09 @tab birthday @tab First birthday +@end multitable + +Load the records like this: + +@example +mysql> LOAD DATA LOCAL INFILE "event.txt" INTO TABLE event; +@end example + +Based on what you've learned from the queries you've run on the @code{pet} +table, you should be able to perform retrievals on the records in the +@code{event} table; the principles are the same. But when is the +@code{event} table by itself insufficient to answer questions you might ask? + +Suppose you want to find out the ages of each pet when they had their +litters. The @code{event} table indicates when this occurred, but to +calculate the age of the mother, you need her birth date. Because that is +stored in the @code{pet} table, you need both tables for the query: + +@example +mysql> SELECT pet.name, (TO_DAYS(date) - TO_DAYS(birth))/365 AS age, remark + -> FROM pet, event + -> WHERE pet.name = event.name AND type = "litter"; ++--------+------+-----------------------------+ +| name | age | remark | ++--------+------+-----------------------------+ +| Fluffy | 2.27 | 4 kittens, 3 female, 1 male | +| Buffy | 4.12 | 5 puppies, 2 female, 3 male | +| Buffy | 5.10 | 3 puppies, 3 female | ++--------+------+-----------------------------+ +@end example + +There are several things to note about this query: @itemize @bullet @item -Is the master logging to the binary log? Check with @code{SHOW MASTER STATUS}. -If it is, @code{Position} will be non-zero. If not, verify that you have -given the master @code{log-bin} option and have set @code{server-id}. -@item -Is the slave running? Check with @code{SHOW SLAVE STATUS}. The answer is found -in @code{Slave_running} column. If not, verify slave options and check the -error log for messages. +The @code{FROM} clause lists two tables because the query needs to pull +information from both of them. + @item -If the slave is running, did it establish connection with the master? Do -@code{SHOW PROCESSLIST}, find the thread with @code{system user} value in -@code{User} column and @code{none} in the @code{Host} column, and check the -@code{State} column. If it says @code{connecting to master}, verify the -privileges for the replication user on the master, master host name, your -DNS setup, whether the master is actually running, whether it is reachable -from the slave, and if all that seems ok, read the error logs. +When combining (joining) information from multiple tables, you need to +specify how records in one table can be matched to records in the other. +This is easy because they both have a @code{name} column. The query uses +@code{WHERE} clause to match up records in the two tables based on the +@code{name} values. + @item -If the slave was running, but then stopped, look at SHOW SLAVE STATUS -output and check the error logs. It usually -happens when some query that succeeded on the master fails on the slave. This -should never happen if you have taken a proper snapshot of the master, and -never modify the data on the slave outside of the slave thread. If it does, -it is a bug, read below on how to report it. -@item -If a query on that succeeded on the master refuses to run on the slave, and -a full database resync ( the proper thing to do ) does not seem feasible, -try the following: -@itemize bullet -@item -First see if there is some stray record in the way. Understand how it got -there, then delete it and run @code{SLAVE START} -@item -If the above does not work or does not apply, try to understand if it would -be safe to make the update manually ( if needed) and then ignore the next -query from the master. -@item -If you have decided you can skip the next query, do -@code{SET SQL_SLAVE_SKIP_COUNTER=1; SLAVE START;} to skip a query that -does not use auto_increment, or last_insert_id or -@code{SET SQL_SLAVE_SKIP_COUNTER=2; SLAVE START;} otherwise. The reason -auto_increment/last_insert_id queries are different is that they take -two events in the binary log of the master. - -@item -If you are sure the slave started out perfectly in sync with the master, -and no one has updated the tables involved outside of slave thread, -report the bug, so -you will not have to do the above tricks again. -@end itemize -@item -Make sure you are not running into an old bug by upgrading to the most recent -version. -@item -If all else fails, read the error logs. If they are big, -@code{grep -i slave /path/to/your-log.err} on the slave. There is no -generic pattern to search for on the master, as the only errors it logs -are general system errors - if it can, it will send the error to the slave -when things go wrong. +Because the @code{name} column occurs in both tables, you must be specific +about which table you mean when referring to the column. This is done +by prepending the table name to the column name. @end itemize -When you have determined that there is no user error involved, and replication -still either does not work at all or is unstable, it is time to start working -on a bug report. We need to get as much info as possible from you to be able -to track down the bug. Please do spend some time and effort preparing a good -bug report. Ideally, we would like to have a test case in the format found in -@code{mysql-test/t/rpl*} directory of the source tree. If you submit a test -case like that, you can expect a patch within a day or two in most cases, -although, of course, you mileage may vary depending on a number of factors. - -Second best option is a just program with easily configurable connection -arguments for the master and the slave that will demonstrate the problem on our -systems. You can write one in Perl or in C, depending on which language you -know better. - -If you have one of the above ways to demonstrate the bug, use -@code{mysqlbug} to prepare a bug report and send it to -@email{bugs@@lists.mysql.com}. If you have a phantom - a problem that -does occur but you cannot duplicate "at will": +You need not have two different tables to perform a join. Sometimes it is +useful to join a table to itself, if you want to compare records in a table +to other records in that same table. For example, to find breeding pairs +among your pets, you can join the @code{pet} table with itself to pair up +males and females of like species: -@itemize @bullet -@item -Verify that there is no user error involved. For example, if you update the -slave outside of the slave thread, the data will be out of sync, and you can -have unique key violations on updates, in which case the slave thread will -stop and wait for you to clean up the tables manually to bring them in sync. -@item -Run slave with @code{log-slave-updates} and @code{log-bin} - this will keep -a log of all updates on the slave. -@item -Save all evidence before resetting the replication. If we have no or only -sketchy information, it would take us a while to track down the problem. The -evidence you should collect is: -@itemize @bullet -@item -All binary logs on the master -@item -All binary log on the slave -@item -The output of @code{SHOW MASTER STATUS} on the master at the time -you have discovered the problem -@item -The output of @code{SHOW SLAVE STATUS} on the master at the time -you have discovered the problem -@item -Error logs on the master and on the slave -@end itemize -@item -Use @code{mysqlbinlog} to examine the binary logs. The following should -be helpful -to find the trouble query, for example: @example -mysqlbinlog -j pos_from_slave_status /path/to/log_from_slave_status | head +mysql> SELECT p1.name, p1.sex, p2.name, p2.sex, p1.species + -> FROM pet AS p1, pet AS p2 + -> WHERE p1.species = p2.species AND p1.sex = "f" AND p2.sex = "m"; ++--------+------+--------+------+---------+ +| name | sex | name | sex | species | ++--------+------+--------+------+---------+ +| Fluffy | f | Claws | m | cat | +| Buffy | f | Fang | m | dog | +| Buffy | f | Bowser | m | dog | ++--------+------+--------+------+---------+ @end example -@end itemize -Once you have collected the evidence on the phantom problem, try hard to -isolate it into a separate test case first. Then report the problem to -@email{bugs@@lists.mysql.com} with as much info as possible. +In this query, we specify aliases for the table name in order +to refer to the columns and keep straight which instance of the table +each column reference is associated with. -@cindex searching, full-text -@cindex full-text search -@cindex FULLTEXT -@node Fulltext Search, Performance, Replication, Top -@chapter MySQL Full-text Search +@cindex databases, information about +@cindex tables, information about +@findex DESCRIBE +@node Getting information, Examples, Database use, Tutorial +@section Getting Information About Databases and Tables -Since Version 3.23.23, @strong{MySQL} has support for full-text indexing -and searching. Full-text indexes in @strong{MySQL} are an index of type -@code{FULLTEXT}. @code{FULLTEXT} indexes can be created from @code{VARCHAR} -and @code{TEXT} columns at @code{CREATE TABLE} time or added later with -@code{ALTER TABLE} or @code{CREATE INDEX}. For large datasets, adding -@code{FULLTEXT} index with @code{ALTER TABLE} (or @code{CREATE INDEX}) would -be much faster than inserting rows into the empty table with a @code{FULLTEXT} -index. +What if you forget the name of a database or table, or what the structure of +a given table is (for example, what its columns are called)? @strong{MySQL} +addresses this problem through several statements that provide information +about the databases and tables it supports. -Full-text search is performed with the @code{MATCH} function. +You have already seen @code{SHOW DATABASES}, which lists the databases +managed by the server. To find out which database is currently selected, +use the @code{DATABASE()} function: @example -mysql> CREATE TABLE articles ( - -> id INT UNSIGNED AUTO_INCREMENT NOT NULL PRIMARY KEY, - -> title VARCHAR(200), - -> body TEXT, - -> FULLTEXT (title,body) - -> ); -Query OK, 0 rows affected (0.00 sec) - -mysql> INSERT INTO articles VALUES - -> (0,'MySQL Tutorial', 'DBMS stands for DataBase Management ...'), - -> (0,'How To Use MySQL Efficiently', 'After you went through a ...'), - -> (0,'Optimizing MySQL','In this tutorial we will show how to ...'), - -> (0,'1001 MySQL Trick','1. Never run mysqld as root. 2. Normalize ...'), - -> (0,'MySQL vs. YourSQL', 'In the following database comparison we ...'), - -> (0,'MySQL Security', 'When configured properly, MySQL could be ...'); -Query OK, 5 rows affected (0.00 sec) -Records: 5 Duplicates: 0 Warnings: 0 - -mysql> SELECT * FROM articles WHERE MATCH (title,body) AGAINST ('database'); -+----+-------------------+---------------------------------------------+ -| id | title | body | -+----+-------------------+---------------------------------------------+ -| 5 | MySQL vs. YourSQL | In the following database comparison we ... | -| 1 | MySQL Tutorial | DBMS stands for DataBase Management ... | -+----+-------------------+---------------------------------------------+ -2 rows in set (0.00 sec) +mysql> SELECT DATABASE(); ++------------+ +| DATABASE() | ++------------+ +| menagerie | ++------------+ @end example -The function @code{MATCH} matches a natural language query @code{AGAINST} -a text collection (which is simply the set of columns covered by a -@code{FULLTEXT} index). For every row in a table it returns relevance - -a similarity measure between the text in that row (in the columns that are -part of the collection) and the query. When it is used in a @code{WHERE} -clause (see example above) the rows returned are automatically sorted with -relevance decreasing. Relevance is a non-negative floating-point number. -Zero relevance means no similarity. Relevance is computed based on the -number of words in the row, the number of unique words in that row, the -total number of words in the collection, and the number of documents (rows) -that contain a particular word. +If you haven't selected any database yet, the result is blank. -The above is a basic example of using @code{MATCH} function. Rows are -returned with relevance decreasing. +To find out what tables the current database contains (for example, when +you're not sure about the name of a table), use this command: @example -mysql> SELECT id,MATCH (title,body) AGAINST ('Tutorial') FROM articles; -+----+-----------------------------------------+ -| id | MATCH (title,body) AGAINST ('Tutorial') | -+----+-----------------------------------------+ -| 1 | 0.64840710366884 | -| 2 | 0 | -| 3 | 0.66266459031789 | -| 4 | 0 | -| 5 | 0 | -| 6 | 0 | -+----+-----------------------------------------+ -5 rows in set (0.00 sec) +mysql> SHOW TABLES; ++---------------------+ +| Tables in menagerie | ++---------------------+ +| event | +| pet | ++---------------------+ @end example -This example shows how to retrieve the relevances. As neither @code{WHERE} -nor @code{ORDER BY} clauses are present, returned rows are not ordered. +If you want to find out about the structure of a table, the @code{DESCRIBE} +command is useful; it displays information about each of a table's columns: @example -mysql> SELECT id, body, MATCH (title,body) AGAINST ( - -> 'Security implications of running MySQL as root') AS score - -> FROM articles WHERE MATCH (title,body) AGAINST - -> ('Security implications of running MySQL as root'); -+----+-----------------------------------------------+-----------------+ -| id | body | score | -+----+-----------------------------------------------+-----------------+ -| 4 | 1. Never run mysqld as root. 2. Normalize ... | 1.5055546709332 | -| 6 | When configured properly, MySQL could be ... | 1.31140957288 | -+----+-----------------------------------------------+-----------------+ -2 rows in set (0.00 sec) +mysql> DESCRIBE pet; ++---------+-------------+------+-----+---------+-------+ +| Field | Type | Null | Key | Default | Extra | ++---------+-------------+------+-----+---------+-------+ +| name | varchar(20) | YES | | NULL | | +| owner | varchar(20) | YES | | NULL | | +| species | varchar(20) | YES | | NULL | | +| sex | char(1) | YES | | NULL | | +| birth | date | YES | | NULL | | +| death | date | YES | | NULL | | ++---------+-------------+------+-----+---------+-------+ @end example -This is more complex example - the query returns the relevance and still -sorts the rows with relevance decreasing. To achieve it one should specify -@code{MATCH} twice. Note, that this will cause no additional overhead, as -@strong{MySQL} optimizer will notice that these two @code{MATCH} calls are -identical and will call full-text search code only once. +@code{Field} indicates the column name, @code{Type} is the data type for +the column, @code{Null} indicates whether or not the column can contain +@code{NULL} values, @code{Key} indicates whether or not the column is +indexed, and @code{Default} specifies the column's default value. -@strong{MySQL} uses a very simple parser to split text into words. A -``word'' is any sequence of letters, numbers, @samp{'}, and @samp{_}. Any -``word'' that is present in the stopword list or just too short (3 -characters or less) is ignored. +If you have indexes on a table, +@code{SHOW INDEX FROM tbl_name} produces information about them. -Every correct word in the collection and in the query is weighted, -according to its significance in the query or collection. This way, a -word that is present in many documents will have lower weight (and may -even have a zero weight), because it has lower semantic value in this -particular collection. Otherwise, if the word is rare, it will receive a -higher weight. The weights of the words are then combined to compute the -relevance of the row. +@cindex queries, examples +@cindex examples, queries +@node Examples, Batch mode, Getting information, Tutorial +@section Examples of Common Queries -Such a technique works best with large collections (in fact, it was -carefully tuned this way). For very small tables, word distribution -does not reflect adequately their semantical value, and this model -may sometimes produce bizarre results. +Here are examples of how to solve some common problems with +@strong{MySQL}. + +Some of the examples use the table @code{shop} to hold the price of each +article (item number) for certain traders (dealers). Supposing that each +trader has a single fixed price per article, then (@code{item}, +@code{trader}) is a primary key for the records. + +Start the command line tool @code{mysql} and select a database: @example -mysql> SELECT * FROM articles WHERE MATCH (title,body) AGAINST ('MySQL'); -Empty set (0.00 sec) +mysql your-database-name @end example -Search for the word @code{MySQL} produces no results in the above example. -Word @code{MySQL} is present in more than half of rows, and as such, is -effectively treated as a stopword (that is, with semantical value zero). -It is, really, the desired behavior - a natural language query should not -return every second row in 1GB table. +(In most @strong{MySQL} installations, you can use the database-name 'test'). -A word that matches half of rows in a table is less likely to locate relevant -documents. In fact, it will most likely find plenty of irrelevant documents. -We all know this happens far too often when we are trying to find something on -the Internet with a search engine. It is with this reasoning that such rows -have been assigned a low semantical value in @strong{this particular dataset}. +You can create the example table as: + +@example +CREATE TABLE shop ( + article INT(4) UNSIGNED ZEROFILL DEFAULT '0000' NOT NULL, + dealer CHAR(20) DEFAULT '' NOT NULL, + price DOUBLE(16,2) DEFAULT '0.00' NOT NULL, + PRIMARY KEY(article, dealer)); + +INSERT INTO shop VALUES +(1,'A',3.45),(1,'B',3.99),(2,'A',10.99),(3,'B',1.45),(3,'C',1.69), +(3,'D',1.25),(4,'D',19.95); +@end example + +Okay, so the example data is: + +@example +mysql> SELECT * FROM shop; + ++---------+--------+-------+ +| article | dealer | price | ++---------+--------+-------+ +| 0001 | A | 3.45 | +| 0001 | B | 3.99 | +| 0002 | A | 10.99 | +| 0003 | B | 1.45 | +| 0003 | C | 1.69 | +| 0003 | D | 1.25 | +| 0004 | D | 19.95 | ++---------+--------+-------+ +@end example @menu -* Fulltext Fine-tuning:: -* Fulltext Features to Appear in MySQL 4.0:: -* Fulltext TODO:: +* example-Maximum-column:: The maximum value for a column +* example-Maximum-row:: The row holding the maximum of a certain column +* example-Maximum-column-group:: Maximum of column per group +* example-Maximum-column-group-row:: The rows holding the group-wise maximum of a certain field +* example-user-variables:: Using user variables +* example-Foreign keys:: Using foreign keys +* Searching on two keys:: +* Calculating days:: @end menu -@node Fulltext Fine-tuning, Fulltext Features to Appear in MySQL 4.0, Fulltext Search, Fulltext Search -@section Fine-tuning MySQL Full-text Search +@node example-Maximum-column, example-Maximum-row, Examples, Examples +@subsection The Maximum Value for a Column -Unfortunately, full-text search has no user-tunable parameters yet, -although adding some is very high on the TODO. However, if you have a -@strong{MySQL} source distribution (@xref{Installing source}.), you can -somewhat alter the full-text search behavior. +``What's the highest item number?'' -Note that full-text search was carefully tuned for the best searching -effectiveness. Modifying the default behavior will, in most cases, -only make the search results worse. Do not alter the @strong{MySQL} sources -unless you know what you are doing! +@example +SELECT MAX(article) AS article FROM shop -@itemize ++---------+ +| article | ++---------+ +| 4 | ++---------+ +@end example + +@node example-Maximum-row, example-Maximum-column-group, example-Maximum-column, Examples +@subsection The Row Holding the Maximum of a Certain Column + +``Find number, dealer, and price of the most expensive article.'' + +In ANSI SQL this is easily done with a sub-query: -@item -Minimal length of word to be indexed is defined in -@code{myisam/ftdefs.h} file by the line @example -#define MIN_WORD_LEN 4 +SELECT article, dealer, price +FROM shop +WHERE price=(SELECT MAX(price) FROM shop) @end example -Change it to the value you prefer, recompile @strong{MySQL}, and rebuild -your @code{FULLTEXT} indexes. -@item -The stopword list is defined in @code{myisam/ft_static.c} -Modify it to your taste, recompile @strong{MySQL} and rebuild -your @code{FULLTEXT} indexes. +In @strong{MySQL} (which does not yet have sub-selects), just do it in +two steps: +@enumerate @item -The 50% threshold is caused by the particular weighting scheme chosen. To -disable it, change the following line in @code{myisam/ftdefs.h}: +Get the maximum price value from the table with a @code{SELECT} statement. +@item +Using this value compile the actual query: @example -#define GWS_IN_USE GWS_PROB +SELECT article, dealer, price +FROM shop +WHERE price=19.95 @end example -to +@end enumerate + +Another solution is to sort all rows descending by price and only +get the first row using the @strong{MySQL} specific @code{LIMIT} clause: + @example -#define GWS_IN_USE GWS_FREQ +SELECT article, dealer, price +FROM shop +ORDER BY price DESC +LIMIT 1 @end example -and recompile @strong{MySQL}. -There is no need to rebuild the indexes in this case. - -@end itemize -@node Fulltext Features to Appear in MySQL 4.0, Fulltext TODO, Fulltext Fine-tuning, Fulltext Search -@section New Features of Full-text Search to Appear in MySQL 4.0 +@strong{NOTE}: If there are several most expensive articles (for example, each 19.95) +the @code{LIMIT} solution shows only one of them! -This section includes a list of the fulltext features that are already -implemented in the 4.0 tree. It explains -@strong{More functions for full-text search} entry of @ref{TODO MySQL 4.0}. +@node example-Maximum-column-group, example-Maximum-column-group-row, example-Maximum-row, Examples +@subsection Maximum of Column per Group -@itemize @bullet -@item @code{REPAIR TABLE} with @code{FULLTEXT} indexes, -@code{ALTER TABLE} with @code{FULLTEXT} indexes, and -@code{OPTIMIZE TABLE} with @code{FULLTEXT} indexes are now -up to 100 times faster. +``What's the highest price per article?'' -@item @code{MATCH ... AGAINST} now supports the following -@strong{boolean operators}: +@example +SELECT article, MAX(price) AS price +FROM shop +GROUP BY article -@itemize @bullet -@item @code{+}word means the that word @strong{must} be present in every -row returned. -@item @code{-}word means the that word @strong{must not} be present in every -row returned. -@item @code{<} and @code{>} can be used to decrease and increase word -weight in the query. -@item @code{~} can be used to assign a @strong{negative} weight to a noise -word. -@item @code{*} is a truncation operator. -@end itemize ++---------+-------+ +| article | price | ++---------+-------+ +| 0001 | 3.99 | +| 0002 | 10.99 | +| 0003 | 1.69 | +| 0004 | 19.95 | ++---------+-------+ +@end example -Boolean search utilizes a more simplistic way of calculating the relevance, -that does not have a 50% threshold. +@node example-Maximum-column-group-row, example-user-variables, example-Maximum-column-group, Examples +@subsection The Rows Holding the Group-wise Maximum of a Certain Field -@item Searches are now up to 2 times faster due to optimized search algorithm. +``For each article, find the dealer(s) with the most expensive price.'' -@item Utility program @code{ft_dump} added for low-level @code{FULLTEXT} -index operations (querying/dumping/statistics). +In ANSI SQL, I'd do it with a sub-query like this: -@end itemize +@example +SELECT article, dealer, price +FROM shop s1 +WHERE price=(SELECT MAX(s2.price) + FROM shop s2 + WHERE s1.article = s2.article); +@end example -@node Fulltext TODO, , Fulltext Features to Appear in MySQL 4.0, Fulltext Search -@section Full-text Search TODO +In @strong{MySQL} it's best do it in several steps: -@itemize @bullet -@item Make all operations with @code{FULLTEXT} index @strong{faster}. -@item Support for braces @code{()} in boolean full-text search. -@item Phrase search, proximity operators -@item Boolean search can work without @code{FULLTEXT} index -(yes, @strong{very} slow). -@item Support for "always-index words". They could be any strings -the user wants to treat as words, examples are "C++", "AS/400", "TCP/IP", etc. -@item Support for full-text search in @code{MERGE} tables. -@item Support for multi-byte charsets. -@item Make stopword list to depend of the language of the data. -@item Stemming (dependent of the language of the data, of course). -@item Generic user-supplyable UDF (?) preparser. -@item Make the model more flexible (by adding some adjustable -parameters to @code{FULLTEXT} in @code{CREATE/ALTER TABLE}). -@end itemize +@enumerate +@item +Get the list of (article,maxprice). +@item +For each article get the corresponding rows that have the stored maximum +price. +@end enumerate -@cindex performance, maximizing -@cindex optimization -@node Performance, MySQL Benchmarks, Fulltext Search, Top -@chapter Getting Maximum Performance from MySQL +This can easily be done with a temporary table: -Optimization is a complicated task because it ultimately requires -understanding of the whole system. While it may be possible to do some -local optimizations with small knowledge of your system/application, the -more optimal you want your system to become the more you will have to -know about it. +@example +CREATE TEMPORARY TABLE tmp ( + article INT(4) UNSIGNED ZEROFILL DEFAULT '0000' NOT NULL, + price DOUBLE(16,2) DEFAULT '0.00' NOT NULL); -So this chapter will try to explain and give some examples of different -ways to optimize @strong{MySQL}. But remember that there are always some -(increasingly harder) additional ways to make the system even faster. +LOCK TABLES shop read; -@menu -* Optimize Basics:: Optimization overview -* System:: System/Compile time and startup parameter tuning -* Data size:: Get your data as small as possible -* MySQL indexes:: How @strong{MySQL} uses indexes -* Query Speed:: Speed of queries that access or update data -* Tips:: Other optimization tips -* Benchmarks:: Using your own benchmarks -* Design:: Design choices -* Design Limitations:: MySQL design limitations/tradeoffs -* Portability:: Portability -* Internal use:: What have we used MySQL for? -@end menu +INSERT INTO tmp SELECT article, MAX(price) FROM shop GROUP BY article; -@node Optimize Basics, System, Performance, Performance -@section Optimization Overview +SELECT shop.article, dealer, shop.price FROM shop, tmp +WHERE shop.article=tmp.article AND shop.price=tmp.price; -The most important part for getting a system fast is of course the basic -design. You also need to know what kinds of things your system will be -doing, and what your bottlenecks are. +UNLOCK TABLES; -The most common bottlenecks are: -@itemize @bullet -@item Disk seeks. -It takes time for the disk to find a piece of data. With modern disks in -1999, the mean time for this is usually lower than 10ms, so we can in -theory do about 1000 seeks a second. This time improves slowly with new -disks and is very hard to optimize for a single table. The way to -optimize this is to spread the data on more than one disk. -@item Disk reading/writing. -When the disk is at the correct position we need to read the data. With -modern disks in 1999, one disk delivers something like 10-20Mb/s. This -is easier to optimize than seeks because you can read in parallel from -multiple disks. -@item CPU cycles. -When we have the data in main memory (or if it already were -there) we need to process it to get to our result. Having small -tables compared to the memory is the most common limiting -factor. But then, with small tables speed is usually not the problem. -@item Memory bandwidth. -When the CPU needs more data than can fit in the CPU cache the main -memory bandwidth becomes a bottleneck. This is an uncommon bottleneck -for most systems, but one should be aware of it. -@end itemize +DROP TABLE tmp; +@end example -@cindex compiling, optimizing -@cindex system optimization -@cindex startup parameters, tuning -@node System, Data size, Optimize Basics, Performance -@section System/Compile Time and Startup Parameter Tuning +If you don't use a @code{TEMPORARY} table, you must also lock the 'tmp' table. -We start with the system level things since some of these decisions have -to be made very early. In other cases a fast look at this part may -suffice because it not that important for the big gains. However, it is always -nice to have a feeling about how much one could gain by changing things -at this level. +``Can it be done with a single query?'' -The default OS to use is really important! To get the most use of -multiple CPU machines one should use Solaris (because the threads works -really nice) or Linux (because the 2.2 kernel has really good SMP -support). Also on 32-bit machines Linux has a 2G file size limit by -default. Hopefully this will be fixed soon when new filesystems are -released (XFS/Reiserfs). If you have a desperate need for files bigger -than 2G on Linux-intel 32 bit, you should get the LFS patch for the ext2 -file system. +Yes, but only by using a quite inefficient trick that I call the +``MAX-CONCAT trick'': -Because we have not run @strong{MySQL} in production on that many platforms, we -advice you to test your intended platform before choosing it, if possible. +@example +SELECT article, + SUBSTRING( MAX( CONCAT(LPAD(price,6,'0'),dealer) ), 7) AS dealer, + 0.00+LEFT( MAX( CONCAT(LPAD(price,6,'0'),dealer) ), 6) AS price +FROM shop +GROUP BY article; -@cindex locking -Other tips: -@itemize @bullet -@item -If you have enough RAM, you could remove all swap devices. Some -operating systems will use a swap device in some contexts even if you -have free memory. -@item -Use the @code{--skip-locking} @strong{MySQL} option to avoid external -locking. Note that this will not impact @strong{MySQL}'s functionality as -long as you only run one server. Just remember to take down the server (or -lock relevant parts) before you run @code{myisamchk}. On some system -this switch is mandatory because the external locking does not work in any -case. ++---------+--------+-------+ +| article | dealer | price | ++---------+--------+-------+ +| 0001 | B | 3.99 | +| 0002 | A | 10.99 | +| 0003 | C | 1.69 | +| 0004 | D | 19.95 | ++---------+--------+-------+ +@end example -The @code{--skip-locking} option is on by default when compiling with -MIT-pthreads, because @code{flock()} isn't fully supported by -MIT-pthreads on all platforms. It's also on default for Linux -as Linux file locking are not yet safe. +The last example can, of course, be made a bit more efficient by doing the +splitting of the concatenated column in the client. -The only case when you can't use @code{--skip-locking} is if you run -multiple @strong{MySQL} @emph{servers} (not clients) on the same data, -or run @code{myisamchk} on the table without first flushing and locking -the @code{mysqld} server tables first. +@node example-user-variables, example-Foreign keys, example-Maximum-column-group-row, Examples +@subsection Using user variables -You can still use @code{LOCK TABLES}/@code{UNLOCK TABLES} even if you -are using @code{--skip-locking} -@end itemize +You can use @strong{MySQL} user variables to remember results without +having to store them in a temporary variables in the client. +@xref{Variables}. -@menu -* Compile and link options:: How compiling and linking affects the speed of MySQL -* Disk issues:: Disk issues -* Symbolic links:: Using Symbolic Links -* Server parameters:: Tuning server parameters -* Table cache:: How MySQL opens and closes tables -* Creating many tables:: Drawbacks of creating large numbers of tables in the same database -* Open tables:: Why so many open tables? -* Memory use:: How MySQL uses memory -* Internal locking:: How MySQL locks tables -* Table locking:: Table locking issues -* DNS:: -@end menu +For example, to find the articles with the highest and lowest price you +can do: -@node Compile and link options, Disk issues, System, System -@subsection How Compiling and Linking Affects the Speed of MySQL +@example +select @@min_price:=min(price),@@max_price:=max(price) from shop; +select * from shop where price=@@min_price or price=@@max_price; -Most of the following tests are done on Linux with the -@strong{MySQL} benchmarks, but they should give some indication for -other operating systems and workloads. ++---------+--------+-------+ +| article | dealer | price | ++---------+--------+-------+ +| 0003 | D | 1.25 | +| 0004 | D | 19.95 | ++---------+--------+-------+ +@end example -@cindex linking, speed -@cindex compiling, speed -@cindex speed, compiling -@cindex speed, linking +@cindex foreign keys +@cindex keys, foreign +@node example-Foreign keys, Searching on two keys, example-user-variables, Examples +@subsection Using Foreign Keys -You get the fastest executable when you link with @code{-static}. +You don't need foreign keys to join 2 tables. -On Linux, you will get the fastest code when compiling with @code{pgcc} -and @code{-O3}. To compile @file{sql_yacc.cc} with these options, you -need about 200M memory because @code{gcc/pgcc} needs a lot of memory to -make all functions inline. You should also set @code{CXX=gcc} when -configuring @strong{MySQL} to avoid inclusion of the @code{libstdc++} -library (it is not needed). Note that with some versions of @code{pgcc}, -the resulting code will only run on true Pentium processors, even if you -use the compiler option that you want the resulting code to be working on -all x586 type processors (like AMD). +The only thing @strong{MySQL} doesn't do is @code{CHECK} to make sure that +the keys you use really exist in the table(s) you're referencing and it +doesn't automatically delete rows from table with a foreign key +definition. If you use your keys like normal, it'll work just fine: -By just using a better compiler and/or better compiler options you can -get a 10-30 % speed increase in your application. This is particularly -important if you compile the SQL server yourself! -We have tested both the Cygnus CodeFusion and Fujitsu compilers, but -when we tested them, neither was sufficiently bug free to allow -@strong{MySQL} to be compiled with optimizations on. +@example +CREATE TABLE persons ( + id SMALLINT UNSIGNED NOT NULL AUTO_INCREMENT, + name CHAR(60) NOT NULL, + PRIMARY KEY (id) +); -When you compile @strong{MySQL} you should only include support for the -character sets that you are going to use. (Option @code{--with-charset=xxx}). -The standard @strong{MySQL} binary distributions are compiled with support -for all character sets. +CREATE TABLE shirts ( + id SMALLINT UNSIGNED NOT NULL AUTO_INCREMENT, + style ENUM('t-shirt', 'polo', 'dress') NOT NULL, + color ENUM('red', 'blue', 'orange', 'white', 'black') NOT NULL, + owner SMALLINT UNSIGNED NOT NULL REFERENCES persons, + PRIMARY KEY (id) +); -Here is a list of some measurements that we have done: -@itemize @bullet -@item -If you use @code{pgcc} and compile everything with @code{-O6}, the -@code{mysqld} server is 1% faster than with @code{gcc} 2.95.2. -@item -If you link dynamically (without @code{-static}), the result is 13% -slower on Linux. Note that you still can use a dynamic linked -@strong{MySQL} library. It is only the server that is critical for -performance. +INSERT INTO persons VALUES (NULL, 'Antonio Paz'); -@item -If you strip your @code{mysqld} binary with @code{strip libexec/mysqld}, -the resulting binary can be up to 4 % faster. +INSERT INTO shirts VALUES +(NULL, 'polo', 'blue', LAST_INSERT_ID()), +(NULL, 'dress', 'white', LAST_INSERT_ID()), +(NULL, 't-shirt', 'blue', LAST_INSERT_ID()); -@item -If you connect using TCP/IP rather than Unix sockets, the result is 7.5% -slower on the same computer. (If you are connection to @code{localhost}, -@strong{MySQL} will, by default, use sockets). -@item -If you connect using TCP/IP from another computer over a 100M Ethernet, -things will be 8-11 % slower. +INSERT INTO persons VALUES (NULL, 'Lilliana Angelovska'); -@item -If you compile with @code{--with-debug=full}, then you will loose 20 % -for most queries, but some queries may take substantially longer (The -@strong{MySQL} benchmarks ran 35 % slower) -If you use @code{--with-debug}, then you will only loose 15 %. -By starting a @code{mysqld} version compiled with @code{--with-debug=full} -with @code{--skip-safemalloc} the end result should be close to when -configuring with @code{--with-debug}. +INSERT INTO shirts VALUES +(NULL, 'dress', 'orange', LAST_INSERT_ID()), +(NULL, 'polo', 'red', LAST_INSERT_ID()), +(NULL, 'dress', 'blue', LAST_INSERT_ID()), +(NULL, 't-shirt', 'white', LAST_INSERT_ID()); -@item -On a Sun SPARCstation 20, SunPro C++ 4.2 is 5 % faster than @code{gcc} 2.95.2. -@item -Compiling with @code{gcc} 2.95.2 for ultrasparc with the option -@code{-mcpu=v8 -Wa,-xarch=v8plusa} gives 4 % more performance. +SELECT * FROM persons; ++----+---------------------+ +| id | name | ++----+---------------------+ +| 1 | Antonio Paz | +| 2 | Lilliana Angelovska | ++----+---------------------+ -@item -On Solaris 2.5.1, MIT-pthreads is 8-12% slower than Solaris native -threads on a single processor. With more load/CPUs the difference should -get bigger. +SELECT * FROM shirts; ++----+---------+--------+-------+ +| id | style | color | owner | ++----+---------+--------+-------+ +| 1 | polo | blue | 1 | +| 2 | dress | white | 1 | +| 3 | t-shirt | blue | 1 | +| 4 | dress | orange | 2 | +| 5 | polo | red | 2 | +| 6 | dress | blue | 2 | +| 7 | t-shirt | white | 2 | ++----+---------+--------+-------+ -@item -Running with @code{--log-bin} makes @strong{[MySQL} 1 % slower. -@item -Compiling on Linux-x86 using gcc without frame pointers -@code{-fomit-frame-pointer} or @code{-fomit-frame-pointer -ffixed-ebp} -@code{mysqld} 1-4% faster. -@end itemize +SELECT s.* FROM persons p, shirts s + WHERE p.name LIKE 'Lilliana%' + AND s.owner = p.id + AND s.color <> 'white'; -The @strong{MySQL}-Linux distribution provided by @strong{MySQL AB} used -to be compiled with @code{pgcc}, but we had to go back to regular gcc -because of a bug in @code{pgcc} that would generate the code that does -not run on AMD. We will continue using gcc until that bug is resolved. -In the meantime, if you have a non-AMD machine, you can get a faster -binary by compiling with @code{pgcc}. The standard @strong{MySQL} -Linux binary is linked statically to get it faster and more portable. ++----+-------+--------+-------+ +| id | style | color | owner | ++----+-------+--------+-------+ +| 4 | dress | orange | 2 | +| 5 | polo | red | 2 | +| 6 | dress | blue | 2 | ++----+-------+--------+-------+ +@end example -@cindex disk issues -@cindex performance, disk issues -@node Disk issues, Symbolic links, Compile and link options, System -@subsection Disk Issues +@findex UNION +@cindex searching, two keys +@cindex keys, searching on two +@node Searching on two keys, Calculating days, example-Foreign keys, Examples +@subsection Searching on Two Keys -@itemize @bullet -@item -As mentioned before, disks seeks are a big performance bottleneck. This -problems gets more and more apparent when the data starts to grow so -large that effective caching becomes impossible. For large databases, -where you access data more or less randomly, you can be sure that you -will need at least one disk seek to read and a couple of disk seeks to -write things. To minimize this problem, use disks with low seek times. -@item -Increase the number of available disk spindles (and thereby reduce -the seek overhead) by either symlink files to different disks or striping -the disks. -@table @strong -@item Using symbolic links -This means that you symlink the index and/or data file(s) from the -normal data directory to another disk (that may also be striped). This -makes both the seek and read times better (if the disks are not used for -other things). @xref{Symbolic links}. -@cindex striping, defined -@item Striping -Striping means that you have many disks and put the first block on the -first disk, the second block on the second disk, and the Nth on the -(N mod number_of_disks) disk, and so on. This means if your normal data -size is less than the stripe size (or perfectly aligned) you will get -much better performance. Note that striping is very dependent on the OS -and stripe-size. So benchmark your application with different -stripe-sizes. @xref{Benchmarks}. +@strong{MySQL} doesn't yet optimize when you search on two different +keys combined with @code{OR} (Searching on one key with different @code{OR} +parts is optimized quite good): -Note that the speed difference for striping is @strong{very} dependent -on the parameters. Depending on how you set the striping parameters and -number of disks you may get a difference in orders of magnitude. Note that -you have to choose to optimize for random or sequential access. -@end table -@item -For reliability you may want to use RAID 0+1 (striping + mirroring), but -in this case you will need 2*N drives to hold N drives of data. This is -probably the best option if you have the money for it! You may, however, -also have to invest in some volume-management software to handle it -efficiently. -@item -A good option is to have semi-important data (that can be regenerated) -on RAID 0 disk while storing really important data (like host information -and logs) on a RAID 0+1 or RAID N disk. RAID N can be a problem if you -have many writes because of the time to update the parity bits. -@item -You may also set the parameters for the file system that the database -uses. One easy change is to mount the file system with the noatime -option. That makes it skip the updating of the last access time in the -inode and by this will avoid some disk seeks. -@item -On Linux, you can get much more performance (up to 100 % under load is -not uncommon) by using hdpram to configure your disk's interface! The -following should be quite good hdparm options for @strong{MySQL} (and -probably many other applications): @example -hdparm -m 16 -d 1 - +SELECT field1_index, field2_index FROM test_table WHERE field1_index = '1' +OR field2_index = '1' @end example -Note that the performance/reliability when using the above depends on -your hardware, so we strongly suggest that you test your system -thoroughly after using @code{hdparm}! Please consult the @code{hdparm} -man page for more information! If @code{hdparm} is not used wisely, -filesystem corruption may result. Backup everything before experimenting! -@item -On many operating systems you can mount the disks with the 'async' flag to set the file -system to be updated asynchronously. If your computer is reasonable stable, -this should give you more performance without sacrificing too much reliability. -(This flag is on by default on Linux.) -@item -If you don't need to know when a file was last accessed (which is not -really useful on a database server), you can mount your file systems -with the noatime flag. -@end itemize - -@cindex symbolic links -@cindex links, symbolic -@node Symbolic links, Server parameters, Disk issues, System -@subsection Using Symbolic Links +The reason is that we haven't yet had time to come up with an efficient +way to handle this in the general case. (The @code{AND} handling is, +in comparison, now completely general and works very well). -You can move tables and databases from the database directory to other -locations and replace them with symbolic links to the new locations. -You might want to do this, for example, to move a database to a file -system with more free space or increase the speed of your system by -spreading your tables to different disk. +For the moment you can solve this very efficiently by using a +@code{TEMPORARY} table. This type of optimization is also very good if +you are using very complicated queries where the SQL server does the +optimizations in the wrong order. -The recommended may to do this, is to just symlink databases to different -disk and only symlink tables as a last resort. -. +@example +CREATE TEMPORARY TABLE tmp +SELECT field1_index, field2_index FROM test_table WHERE field1_index = '1'; +INSERT INTO tmp +SELECT field1_index, field2_index FROM test_table WHERE field2_index = '1'; +SELECT * from tmp; +DROP TABLE tmp; +@end example -@cindex databases, symbolic links -@menu -* Symbolic links to database:: -* Symbolic links to tables:: -@end menu +The above way to solve this query is in effect an @code{UNION} of two queries. -@node Symbolic links to database, Symbolic links to tables, Symbolic links, Symbolic links -@subsubsection Using Symbolic Links for Databases +@cindex bit_functions, example +@findex BIT_OR +@findex BIT_COUNT +@findex << +@node Calculating days, , Searching on two keys, Examples +@subsection Calculating visits per day -The way to symlink a database is to first create a directory on some -disk where you have free space and then create a symlink to it from -the @strong{MySQL} database directory. +The following shows an idea of how you can use the bit group functions +to calculate the number of days per month a user has visited a web page. @example -shell> mkdir /dr1/databases/test -shell> ln -s /dr1/databases/test mysqld-datadir -@end example +CREATE TABLE t1 (year YEAR(4), month INT(2) UNSIGNED ZEROFILL, day INT(2) UNSIGNED ZEROFILL); +INSERT INTO t1 VALUES(2000,1,1),(2000,1,20),(2000,1,30),(2000,2,2),(2000,2,23),(2000,2,23); -@strong{MySQL} doesn't support that you link one directory to multiple -databases. Replacing a database directory with a symbolic link will -work fine as long as you don't make a symbolic link between databases. -Suppose you have a database @code{db1} under the @strong{MySQL} data -directory, and then make a symlink @code{db2} that points to @code{db1}: +SELECT year,month,BIT_COUNT(BIT_OR(1<<day)) AS days FROM t1 GROUP BY year,month; -@example -shell> cd /path/to/datadir -shell> ln -s db1 db2 +Which returns: + ++------+-------+------+ +| year | month | days | ++------+-------+------+ +| 2000 | 01 | 3 | +| 2000 | 02 | 2 | ++------+-------+------+ @end example -Now, for any table @code{tbl_a} in @code{db1}, there also appears to be -a table @code{tbl_a} in @code{db2}. If one thread updates @code{db1.tbl_a} -and another thread updates @code{db2.tbl_a}, there will be problems. +The above calculates how many different days was used for a given +year/month combination, with automatic removal of duplicate entries. -If you really need this, you must change the following code in -@file{mysys/mf_format.c}: +@cindex modes, batch +@cindex batch mode +@cindex running, batch mode +@cindex script files +@cindex files, script +@node Batch mode, Twin, Examples, Tutorial +@section Using @code{mysql} in Batch Mode + +In the previous sections, you used @code{mysql} interactively to enter +queries and view the results. You can also run @code{mysql} in batch +mode. To do this, put the commands you want to run in a file, then +tell @code{mysql} to read its input from the file: @example -if (flag & 32 || (!lstat(to,&stat_buff) && S_ISLNK(stat_buff.st_mode))) +shell> mysql < batch-file @end example -to +If you need to specify connection parameters on the command line, the +command might look like this: @example -if (1) +shell> mysql -h host -u user -p < batch-file +Enter password: ******** @end example -On Windows you can use internal symbolic links to directories by compiling -@strong{MySQL} with @code{-DUSE_SYMDIR}. This allows you to put different -databases on different disks. @xref{Windows symbolic links}. +When you use @code{mysql} this way, you are creating a script file, then +executing the script. -@cindex databases, symbolic links -@node Symbolic links to tables, , Symbolic links to database, Symbolic links -@subsubsection Using Symbolic Links for Tables +Why use a script? Here are a few reasons: -Before @strong{MySQL} 4.0 you should not symlink tables, if you are not -very carefully with them. The problem is that if you run @code{ALTER -TABLE}, @code{REPAIR TABLE} or @code{OPTIMIZE TABLE} on a symlinked -table, the symlinks will be removed and replaced by the original -files. This happens because the above command works by creating a -temporary file in the database directory and when the command is -complete, replace the original file with the temporary file. +@itemize @bullet +@item +If you run a query repeatedly (say, every day or every week), making it a +script allows you to avoid retyping it each time you execute it. -You should not symlink tables on system that doesn't have a fully -working @code{realpath()} call. (At least Linux and Solaris support -@code{realpath()}) +@item +You can generate new queries from existing ones that are similar by copying +and editing script files. -In @strong{MySQL} 4.0 symlinks is only fully supported for @code{MyISAM} -tables. For other table types you will probably get strange problems -when doing any of the above mentioned commands. +@item +Batch mode can also be useful while you're developing a query, particularly +for multiple-line commands or multiple-statement sequences of commands. If +you make a mistake, you don't have to retype everything. Just edit your +script to correct the error, then tell @code{mysql} to execute it again. -The handling of symbolic links in @strong{MySQL} 4.0 works the following -way (this is mostly relevant only for @code{MyISAM} tables). +@item +If you have a query that produces a lot of output, you can run the output +through a pager rather than watching it scroll off the top of your screen: + +@example +shell> mysql < batch-file | more +@end example -@itemize @bullet @item -In the data directory you will always have the table definition file -and the data/index files. +You can catch the output in a file for further processing: + +@example +shell> mysql < batch-file > mysql.out +@end example + @item -You can symlink the index file and the data file to different directories -independent of the other. +You can distribute your script to other people so they can run the commands, +too. + @item -The symlinking can be done from the operating system (if @code{mysqld} is -not running) or with the @code{INDEX/DATA DIRECTORY="path-to-dir"} command -in @code{CREATE TABLE}. @xref{CREATE TABLE}. -@item -@code{myisamchk} will not replace a symlink with the index/file but -work directly on the files the symlinks points to. Any temporary files -will be created in the same directory where the data/index file is. -@item -When you drop a table that is using symlinks, both the symlink and the -file the symlink points to is dropped. This is a good reason to why you -should NOT run @code{mysqld} as root and not allow persons to have write -access to the @strong{MySQL} database directories. -@item -If you rename a table with @code{ALTER TABLE RENAME} and you don't change -database, the symlink in the database directory will be renamed to the new -name and the data/index file will be renamed accordingly. -@item -If you use @code{ALTER TABLE RENAME} to move a table to another database, -then the table will be moved to the other database directory and the old -symlinks and the files they pointed to will be deleted. -@item -If you are not using symlinks you should use the @code{--skip-symlink} -option to @code{mysqld} to ensure that no one can drop or rename a file -outside of the @code{mysqld} data directory. -@end itemize - -Things that are not yet supported: - -@cindex TODO, symlinks -@itemize @bullet -@item -@code{ALTER TABLE} ignores all @code{INDEX/DATA DIRECTORY="path"} options. -@item -@code{CREATE TABLE} doesn't report if the table has symbolic links. -@item -@code{mysqldump} doesn't include the symbolic links information in the output. -@item -@code{BACKUP TABLE} and @code{RESTORE TABLE} doesn't use symbolic links. +Some situations do not allow for interactive use, for example, when you run +a query from a @code{cron} job. In this case, you must use batch mode. @end itemize -@cindex parameters, server -@cindex @code{mysqld} server, buffer sizes -@cindex buffer sizes, @code{mysqld} server -@cindex startup parameters -@node Server parameters, Table cache, Symbolic links, System -@subsection Tuning Server Parameters - -You can get the default buffer sizes used by the @code{mysqld} server -with this command: +The default output format is different (more concise) when you run +@code{mysql} in batch mode than when you use it interactively. For +example, the output of @code{SELECT DISTINCT species FROM pet} looks like +this when run interactively: @example -shell> mysqld --help ++---------+ +| species | ++---------+ +| bird | +| cat | +| dog | +| hamster | +| snake | ++---------+ @end example -@cindex @code{mysqld} options -@cindex variables, @code{mysqld} -This command produces a list of all @code{mysqld} options and configurable -variables. The output includes the default values and looks something -like this: +But like this when run in batch mode: @example -Possible variables for option --set-variable (-O) are: -back_log current value: 5 -bdb_cache_size current value: 1048540 -binlog_cache_size current_value: 32768 -connect_timeout current value: 5 -delayed_insert_timeout current value: 300 -delayed_insert_limit current value: 100 -delayed_queue_size current value: 1000 -flush_time current value: 0 -interactive_timeout current value: 28800 -join_buffer_size current value: 131072 -key_buffer_size current value: 1048540 -lower_case_table_names current value: 0 -long_query_time current value: 10 -max_allowed_packet current value: 1048576 -max_binlog_cache_size current_value: 4294967295 -max_connections current value: 100 -max_connect_errors current value: 10 -max_delayed_threads current value: 20 -max_heap_table_size current value: 16777216 -max_join_size current value: 4294967295 -max_sort_length current value: 1024 -max_tmp_tables current value: 32 -max_write_lock_count current value: 4294967295 -myisam_sort_buffer_size current value: 8388608 -net_buffer_length current value: 16384 -net_retry_count current value: 10 -net_read_timeout current value: 30 -net_write_timeout current value: 60 -query_buffer_size current value: 0 -record_buffer current value: 131072 -slow_launch_time current value: 2 -sort_buffer current value: 2097116 -table_cache current value: 64 -thread_concurrency current value: 10 -tmp_table_size current value: 1048576 -thread_stack current value: 131072 -wait_timeout current value: 28800 +species +bird +cat +dog +hamster +snake @end example -If there is a @code{mysqld} server currently running, you can see what -values it actually is using for the variables by executing this command: +If you want to get the interactive output format in batch mode, use +@code{mysql -t}. To echo to the output the commands that are executed, use +@code{mysql -vvv}. + +@cindex Twin Studies, queries +@cindex queries, Twin Studeis project +@node Twin, , Batch mode, Tutorial +@section Queries from Twin Project + +At Analytikerna and Lentus, we have been doing the systems and field work +for a big research project. This project is a collaboration between the +Institute of Environmental Medicine at Karolinska Institutet Stockholm +and the Section on Clinical Research in Aging and Psychology at the +University of Southern California. + +The project involves a screening part where all twins in Sweden older +than 65 years are interviewed by telephone. Twins who meet certain +criteria are passed on to the next stage. In this latter stage, twins who +want to participate are visited by a doctor/nurse team. Some of the +examinations include physical and neuropsychological examination, +laboratory testing, neuroimaging, psychological status assessment, and family +history collection. In addition, data are collected on medical and +environmental risk factors. + +More information about Twin studies can be found at: @example -shell> mysqladmin variables +@url{http://www.imm.ki.se/TWIN/TWINUKW.HTM} @end example -You can find a full description for all variables in the @code{SHOW VARIABLES} -section in this manual. @xref{SHOW VARIABLES}. +The latter part of the project is administered with a Web interface +written using Perl and @strong{MySQL}. -You can also see some statistics from a running server by issuing the command -@code{SHOW STATUS}. @xref{SHOW STATUS}. +Each night all data from the interviews are moved into a @strong{MySQL} +database. -@strong{MySQL} uses algorithms that are very scalable, so you can usually -run with very little memory. If you, however, give @strong{MySQL} more -memory, you will normally also get better performance. +@menu +* Twin pool:: Find all non-distributed twins +* Twin event:: Show a table on twin pair status +@end menu -When tuning a @strong{MySQL} server, the two most important variables to use -are @code{key_buffer_size} and @code{table_cache}. You should first feel -confident that you have these right before trying to change any of the -other variables. +@node Twin pool, Twin event, Twin, Twin +@subsection Find all Non-distributed Twins -If you have much memory (>=256M) and many tables and want maximum performance -with a moderate number of clients, you should use something like this: +The following query is used to determine who goes into the second part of the +project: @example -shell> safe_mysqld -O key_buffer=64M -O table_cache=256 \ - -O sort_buffer=4M -O record_buffer=1M & +select + concat(p1.id, p1.tvab) + 0 as tvid, + concat(p1.christian_name, " ", p1.surname) as Name, + p1.postal_code as Code, + p1.city as City, + pg.abrev as Area, + if(td.participation = "Aborted", "A", " ") as A, + p1.dead as dead1, + l.event as event1, + td.suspect as tsuspect1, + id.suspect as isuspect1, + td.severe as tsevere1, + id.severe as isevere1, + p2.dead as dead2, + l2.event as event2, + h2.nurse as nurse2, + h2.doctor as doctor2, + td2.suspect as tsuspect2, + id2.suspect as isuspect2, + td2.severe as tsevere2, + id2.severe as isevere2, + l.finish_date +from + twin_project as tp + /* For Twin 1 */ + left join twin_data as td on tp.id = td.id and tp.tvab = td.tvab + left join informant_data as id on tp.id = id.id and tp.tvab = id.tvab + left join harmony as h on tp.id = h.id and tp.tvab = h.tvab + left join lentus as l on tp.id = l.id and tp.tvab = l.tvab + /* For Twin 2 */ + left join twin_data as td2 on p2.id = td2.id and p2.tvab = td2.tvab + left join informant_data as id2 on p2.id = id2.id and p2.tvab = id2.tvab + left join harmony as h2 on p2.id = h2.id and p2.tvab = h2.tvab + left join lentus as l2 on p2.id = l2.id and p2.tvab = l2.tvab, + person_data as p1, + person_data as p2, + postal_groups as pg +where + /* p1 gets main twin and p2 gets his/her twin. */ + /* ptvab is a field inverted from tvab */ + p1.id = tp.id and p1.tvab = tp.tvab and + p2.id = p1.id and p2.ptvab = p1.tvab and + /* Just the sceening survey */ + tp.survey_no = 5 and + /* Skip if partner died before 65 but allow emigration (dead=9) */ + (p2.dead = 0 or p2.dead = 9 or + (p2.dead = 1 and + (p2.death_date = 0 or + (((to_days(p2.death_date) - to_days(p2.birthday)) / 365) + >= 65)))) + and + ( + /* Twin is suspect */ + (td.future_contact = 'Yes' and td.suspect = 2) or + /* Twin is suspect - Informant is Blessed */ + (td.future_contact = 'Yes' and td.suspect = 1 and id.suspect = 1) or + /* No twin - Informant is Blessed */ + (ISNULL(td.suspect) and id.suspect = 1 and id.future_contact = 'Yes') or + /* Twin broken off - Informant is Blessed */ + (td.participation = 'Aborted' + and id.suspect = 1 and id.future_contact = 'Yes') or + /* Twin broken off - No inform - Have partner */ + (td.participation = 'Aborted' and ISNULL(id.suspect) and p2.dead = 0)) + and + l.event = 'Finished' + /* Get at area code */ + and substring(p1.postal_code, 1, 2) = pg.code + /* Not already distributed */ + and (h.nurse is NULL or h.nurse=00 or h.doctor=00) + /* Has not refused or been aborted */ + and not (h.status = 'Refused' or h.status = 'Aborted' + or h.status = 'Died' or h.status = 'Other') +order by + tvid; @end example -If you have only 128M and only a few tables, but you still do a lot of -sorting, you can use something like: +Some explanations: +@table @asis +@item @code{concat(p1.id, p1.tvab) + 0 as tvid} +We want to sort on the concatenated @code{id} and @code{tvab} in +numerical order. Adding @code{0} to the result causes @strong{MySQL} to +treat the result as a number. +@item column @code{id} +This identifies a pair of twins. It is a key in all tables. +@item column @code{tvab} +This identifies a twin in a pair. It has a value of @code{1} or @code{2}. +@item column @code{ptvab} +This is an inverse of @code{tvab}. When @code{tvab} is @code{1} this is +@code{2}, and vice versa. It exists to save typing and to make it easier for +@strong{MySQL} to optimize the query. +@end table + +This query demonstrates, among other things, how to do lookups on a +table from the same table with a join (@code{p1} and @code{p2}). In the example, this +is used to check whether a twin's partner died before the age of 65. If so, +the row is not returned. + +All of the above exist in all tables with twin-related information. We +have a key on both @code{id,tvab} (all tables), and @code{id,ptvab} +(@code{person_data}) to make queries faster. + +On our production machine (A 200MHz UltraSPARC), this query returns +about 150-200 rows and takes less than one second. + +The current number of records in the tables used above: +@multitable @columnfractions .3 .5 +@item @strong{Table} @tab @strong{Rows} +@item @code{person_data} @tab 71074 +@item @code{lentus} @tab 5291 +@item @code{twin_project} @tab 5286 +@item @code{twin_data} @tab 2012 +@item @code{informant_data} @tab 663 +@item @code{harmony} @tab 381 +@item @code{postal_groups} @tab 100 +@end multitable + +@node Twin event, , Twin pool, Twin +@subsection Show a Table on Twin Pair Status + +Each interview ends with a status code called @code{event}. The query +shown below is used to display a table over all twin pairs combined by +event. This indicates in how many pairs both twins are finished, in how many +pairs one twin is finished and the other refused, and so on. @example -shell> safe_mysqld -O key_buffer=16M -O sort_buffer=1M +select + t1.event, + t2.event, + count(*) +from + lentus as t1, + lentus as t2, + twin_project as tp +where + /* We are looking at one pair at a time */ + t1.id = tp.id + and t1.tvab=tp.tvab + and t1.id = t2.id + /* Just the sceening survey */ + and tp.survey_no = 5 + /* This makes each pair only appear once */ + and t1.tvab='1' and t2.tvab='2' +group by + t1.event, t2.event; + @end example -If you have little memory and lots of connections, use something like this: +@cindex functions, server +@cindex server functions +@node Server, Replication, Tutorial, Top +@chapter MySQL Server Functions + +@menu +* Languages:: What languages are supported by @strong{MySQL}? +@end menu + +This chapter describes the languages @strong{MySQL} supports, how sorting +works in @strong{MySQL}, and how to add new character sets to @strong{MySQL}. +You will also find information about maximum table sizes in this chapter. + +@cindex error messages, languages +@cindex messages, languages +@cindex files, error messages +@cindex language support +@node Languages, , Server, Server +@section What Languages Are Supported by MySQL? + +@code{mysqld} can issue error messages in the following languages: +Czech, Danish, Dutch, English (the default), Estonian, French, German, Greek, +Hungarian, Italian, Japanese, Korean, Norwegian, Norwegian-ny, Polish, +Portuguese, Romanian, Russian, Slovak, Spanish, and Swedish. + +To start @code{mysqld} with a particular language, use either the +@code{--language=lang} or @code{-L lang} options. For example: @example -shell> safe_mysqld -O key_buffer=512k -O sort_buffer=100k \ - -O record_buffer=100k & +shell> mysqld --language=swedish @end example -or even: +or: @example -shell> safe_mysqld -O key_buffer=512k -O sort_buffer=16k \ - -O table_cache=32 -O record_buffer=8k -O net_buffer=1K & +shell> mysqld --language=/usr/local/share/swedish @end example -When you have installed @strong{MySQL}, the @file{support-files} directory will -contain some different @code{my.cnf} example files, @file{my-huge.cnf}, -@file{my-large.cnf}, @file{my-medium.cnf}, and @file{my-small.cnf}, you can -use as a base to optimize your system. - -If there are very many connections, ``swapping problems'' may occur unless -@code{mysqld} has been configured to use very little memory for each -connection. @code{mysqld} performs better if you have enough memory for all -connections, of course. +Note that all language names are specified in lowercase. -Note that if you change an option to @code{mysqld}, it remains in effect only -for that instance of the server. +The language files are located (by default) in +@file{@var{mysql_base_dir}/share/@var{LANGUAGE}/}. -To see the effects of a parameter change, do something like this: +To update the error message file, you should edit the @file{errmsg.txt} file +and execute the following command to generate the @file{errmsg.sys} file: @example -shell> mysqld -O key_buffer=32m --help +shell> comp_err errmsg.txt errmsg.sys @end example -Make sure that the @code{--help} option is last; otherwise, the effect of any -options listed after it on the command line will not be reflected in the -output. +If you upgrade to a newer version of @strong{MySQL}, remember to repeat +your changes with the new @file{errmsg.txt} file. -@cindex tables, opening -@cindex tables, closing -@cindex opening, tables -@cindex closing, tables -@cindex table cache -@findex table_cache -@node Table cache, Creating many tables, Server parameters, System -@subsection How MySQL Opens and Closes Tables +@menu +* Character sets:: The character set used for data and sorting +* Adding character set:: Adding a new character set +* Character arrays:: The character definition arrays +* String collating:: String collating support +* Multi-byte characters:: Multi-byte character support +@end menu -@code{table_cache}, @code{max_connections}, and @code{max_tmp_tables} -affect the maximum number of files the server keeps open. If you -increase one or both of these values, you may run up against a limit -imposed by your operating system on the per-process number of open file -descriptors. However, you can increase the limit on many systems. -Consult your OS documentation to find out how to do this, because the -method for changing the limit varies widely from system to system. +@cindex character sets +@cindex data, character sets +@cindex sorting, character sets +@node Character sets, Adding character set, Languages, Languages +@subsection The Character Set Used for Data and Sorting -@code{table_cache} is related to @code{max_connections}. For example, -for 200 concurrent running connections, you should have a table cache of -at least @code{200 * n}, where @code{n} is the maximum number of tables -in a join. You also need to reserve some extra file descriptors for -temporary tables and files. +By default, @strong{MySQL} uses the ISO-8859-1 (Latin1) character set +with sorting according to Swedish/Finnish. This is the character set suitable +in the USA and western Europe. -The cache of open tables can grow to a maximum of @code{table_cache} -(default 64; this can be changed with the @code{-O table_cache=#} -option to @code{mysqld}). A table is never closed, except when the -cache is full and another thread tries to open a table or if you use -@code{mysqladmin refresh} or @code{mysqladmin flush-tables}. +All standard @strong{MySQL} binaries are compiled with +@code{--with-extra-charsets=complex}. This will add code to all +standard programs to be able to handle @code{latin1} and all multi-byte +character sets within the binary. Other character sets will be +loaded from a character-set definition file when needed. -When the table cache fills up, the server uses the following procedure -to locate a cache entry to use: +The character set determines what characters are allowed in names and how +things are sorted by the @code{ORDER BY} and @code{GROUP BY} clauses of +the @code{SELECT} statement. -@itemize @bullet -@item -Tables that are not currently in use are released, in least-recently-used -order. +You can change the character set with the @code{--default-character-set} +option when you start the server. The character sets available depend +on the @code{--with-charset=charset} and @code{--with-extra-charset= +list-of-charset | complex | all} options to @code{configure}, and the +character set configuration files listed in +@file{SHAREDIR/charsets/Index}. @xref{configure options}. -@item -If the cache is full and no tables can be released, but a new table needs to -be opened, the cache is temporarily extended as necessary. +If you change the character set when running @strong{MySQL} (which may +also change the sort order), you must run myisamchk -r -q on all +tables. Otherwise your indexes may not be ordered correctly. -@item -If the cache is in a temporarily-extended state and a table goes from in-use -to not-in-use state, the table is closed and released from the cache. -@end itemize +When a client connects to a @strong{MySQL} server, the server sends the +default character set in use to the client. The client will switch to +use this character set for this connection. -A table is opened for each concurrent access. This means that -if you have two threads accessing the same table or access the table -twice in the same query (with @code{AS}) the table needs to be opened twice. -The first open of any table takes two file descriptors; each additional -use of the table takes only one file descriptor. The extra descriptor -for the first open is used for the index file; this descriptor is shared -among all threads. +One should use @code{mysql_real_escape_string()} when escaping strings +for a SQL query. @code{mysql_real_escape_string()} is identical to the +old @code{mysql_escape_string()} function, except that it takes the MYSQL +connection handle as the first parameter. -You can check if your table cache is too small by checking the mysqld -variable @code{opened_tables}. If this is quite big, even if you -haven't done a lot of @code{FLUSH TABLES}, you should increase your table -cache. @xref{SHOW STATUS}. +If the client is compiled with different paths than where the server is +installed and the user who configured @strong{MySQL} didn't included all +character sets in the @strong{MySQL} binary, one must specify for +the client where it can find the additional character sets it will need +if the server runs with a different character set than the client. -@cindex tables, too many -@node Creating many tables, Open tables, Table cache, System -@subsection Drawbacks to Creating Large Numbers of Tables in the Same Database +One can specify this by putting in a @strong{MySQL} option file: -If you have many files in a directory, open, close, and create operations will -be slow. If you execute @code{SELECT} statements on many different tables, -there will be a little overhead when the table cache is full, because for -every table that has to be opened, another must be closed. You can reduce -this overhead by making the table cache larger. +@example +[client] +character-sets-dir=/usr/local/mysql/share/mysql/charsets +@end example -@cindex tables, open -@cindex open tables -@node Open tables, Memory use, Creating many tables, System -@subsection Why So Many Open tables? +where the path points to where the dynamic @strong{MySQL} character sets +are stored. -When you run @code{mysqladmin status}, you'll see something like this: +One can force the client to use specific character set by specifying: @example -Uptime: 426 Running threads: 1 Questions: 11082 Reloads: 1 Open tables: 12 +[client] +default-character-set=character-set-name @end example -This can be somewhat perplexing if you only have 6 tables. +but normally this is never needed. -@strong{MySQL} is multithreaded, so it may have many queries on the same -table simultaneously. To minimize the problem with two threads having -different states on the same file, the table is opened independently by -each concurrent thread. This takes some memory and one extra file -descriptor for the data file. The index file descriptor is shared -between all threads. +@cindex character sets, adding +@cindex adding, character sets +@node Adding character set, Character arrays, Character sets, Languages +@subsection Adding a New Character Set -@cindex memory use -@node Memory use, Internal locking, Open tables, System -@subsection How MySQL Uses Memory +To add another character set to @strong{MySQL}, use the following procedure. -The list below indicates some of the ways that the @code{mysqld} server -uses memory. Where applicable, the name of the server variable relevant -to the memory use is given: +Decide if the set is simple or complex. If the character set +does not need to use special string collating routines for +sorting and does not need multi-byte character support, it is +simple. If it needs either of those features, it is complex. -@itemize @bullet -@item -The key buffer (variable @code{key_buffer_size}) is shared by all -threads; Other buffers used by the server are allocated as -needed. @xref{Server parameters}. +For example, @code{latin1} and @code{danish} are simple charactersets while +@code{big5} or @code{czech} are complex character sets. -@item -Each connection uses some thread-specific space: A stack (default 64K, -variable @code{thread_stack}), a connection buffer (variable -@code{net_buffer_length}), and a result buffer (variable -@code{net_buffer_length}). The connection buffer and result buffer are -dynamically enlarged up to @code{max_allowed_packet} when needed. When -a query is running, a copy of the current query string is also allocated. +In the following section, we have assumed that you name your character +set @code{MYSET}. -@item -All threads share the same base memory. +For a simple character set do the following: +@enumerate @item -Only the compressed ISAM / MyISAM tables are memory mapped. This is -because the 32-bit memory space of 4GB is not large enough for most -big tables. When systems with a 64-bit address space become more -common we may add general support for memory mapping. +Add MYSET to the end of the @file{sql/share/charsets/Index} file +Assign an unique number to it. @item -Each request doing a sequential scan over a table allocates a read buffer -(variable @code{record_buffer}). +Create the file @file{sql/share/charsets/MYSET.conf}. +(You can use @file{sql/share/charsets/latin1.conf} as a base for this). +The syntax for the file very simple: + +@itemize @bullet @item -All joins are done in one pass, and most joins can be done without even -using a temporary table. Most temporary tables are memory-based (HEAP) -tables. Temporary tables with a big record length (calculated as the -sum of all column lengths) or that contain @code{BLOB} columns are -stored on disk. +Comments start with a '#' character and proceed to the end of the line. +@item +Words are separated by arbitrary amounts of whitespace. +@item +When defining the character set, every word must be a number in hexadecimal +format +@item +The @code{ctype} array takes up the first 257 words. The +@code{to_lower}, @code{to_upper} and @code{sort_order} arrays take up +256 words each after that. +@end itemize -One problem in @strong{MySQL} versions before Version 3.23.2 is that if a HEAP table -exceeds the size of @code{tmp_table_size}, you get the error @code{The -table tbl_name is full}. In newer versions this is handled by -automatically changing the in-memory (HEAP) table to a disk-based -(MyISAM) table as necessary. To work around this problem, you can -increase the temporary table size by setting the @code{tmp_table_size} -option to @code{mysqld}, or by setting the SQL option -@code{SQL_BIG_TABLES} in the client program. @xref{SET OPTION, , -@code{SET OPTION}}. In @strong{MySQL} Version 3.20, the maximum size of the -temporary table was @code{record_buffer*16}, so if you are using this -version, you have to increase the value of @code{record_buffer}. You can -also start @code{mysqld} with the @code{--big-tables} option to always -store temporary tables on disk. However, this will affect the speed of -many complicated queries. +@xref{Character arrays}. @item -Most requests doing a sort allocates a sort buffer and 0-2 temporary -files depending on the result set size. @xref{Temporary files}. +Add the character set name to the @code{CHARSETS_AVAILABLE} and +@code{COMPILED_CHARSETS} lists in @code{configure.in}. @item -Almost all parsing and calculating is done in a local memory store. No -memory overhead is needed for small items and the normal slow memory -allocation and freeing is avoided. Memory is allocated only for -unexpectedly large strings (this is done with @code{malloc()} and -@code{free()}). +Reconfigure, recompile, and test. -@item -Each index file is opened once and the data file is opened once for each -concurrently running thread. For each concurrent thread, a table structure, -column structures for each column, and a buffer of size @code{3 * n} is -allocated (where @code{n} is the maximum row length, not counting @code{BLOB} -columns). A @code{BLOB} uses 5 to 8 bytes plus the length of the @code{BLOB} -data. The @code{ISAM}/@code{MyISAM} table handlers will use one extra row -buffer for internal usage. +@end enumerate + +For a complex character set do the following: +@enumerate @item -For each table having @code{BLOB} columns, a buffer is enlarged dynamically -to read in larger @code{BLOB} values. If you scan a table, a buffer as large -as the largest @code{BLOB} value is allocated. +Create the file @file{strings/ctype-MYSET.c} in the @strong{MySQL} source +distribution. @item -Table handlers for all in-use tables are saved in a cache and managed as a -FIFO. Normally the cache has 64 entries. If a table has been used by two -running threads at the same time, the cache contains two entries for the -table. @xref{Table cache}. +Add MYSET to the end of the @file{sql/share/charsets/Index} file. +Assign an unique number to it. @item -A @code{mysqladmin flush-tables} command closes all tables that are not in -use and marks all in-use tables to be closed when the currently executing -thread finishes. This will effectively free most in-use memory. -@end itemize +Look at one of the existing @file{ctype-*.c} files to see what needs to +be defined, for example @file{strings/ctype-big5.c}. Note that the +arrays in your file must have names like @code{ctype_MYSET}, +@code{to_lower_MYSET}, and so on. This corresponds to the arrays +in the simple character set. @xref{Character arrays}. For a complex +character set -@code{ps} and other system status programs may report that @code{mysqld} -uses a lot of memory. This may be caused by thread-stacks on different -memory addresses. For example, the Solaris version of @code{ps} counts -the unused memory between stacks as used memory. You can verify this by -checking available swap with @code{swap -s}. We have tested -@code{mysqld} with commercial memory-leakage detectors, so there should -be no memory leaks. +@item +Near the top of the file, place a special comment like this: -@cindex internal locking -@cindex locking, tables -@cindex tables, locking -@node Internal locking, Table locking, Memory use, System -@subsection How MySQL Locks Tables +@example +/* + * This comment is parsed by configure to create ctype.c, + * so don't change it unless you know what you are doing. + * + * .configure. number_MYSET=MYNUMBER + * .configure. strxfrm_multiply_MYSET=N + * .configure. mbmaxlen_MYSET=N + */ +@end example -You can find a discussion about different locking methods in the appendix. -@xref{Locking methods}. +The @code{configure} program uses this comment to include +the character set into the @strong{MySQL} library automatically. -All locking in @strong{MySQL} is deadlock-free. This is managed by always -requesting all needed locks at once at the beginning of a query and always -locking the tables in the same order. +The strxfrm_multiply and mbmaxlen lines will be explained in +the following sections. Only include them if you the string +collating functions or the multi-byte character set functions, +respectively. -The locking method @strong{MySQL} uses for @code{WRITE} locks works as follows: +@item +You should then create some of the following functions: @itemize @bullet -@item -If there are no locks on the table, put a write lock on it. -@item -Otherwise, put the lock request in the write lock queue. +@item @code{my_strncoll_MYSET()} +@item @code{my_strcoll_MYSET()} +@item @code{my_strxfrm_MYSET()} +@item @code{my_like_range_MYSET()} @end itemize -The locking method @strong{MySQL} uses for @code{READ} locks works as follows: +@xref{String collating}. -@itemize @bullet @item -If there are no write locks on the table, put a read lock on it. +Add the character set name to the @code{CHARSETS_AVAILABLE} and +@code{COMPILED_CHARSETS} lists in @code{configure.in}. + @item -Otherwise, put the lock request in the read lock queue. -@end itemize +Reconfigure, recompile, and test. +@end enumerate -When a lock is released, the lock is made available to the threads -in the write lock queue, then to the threads in the read lock queue. +The file @file{sql/share/charsets/README} includes some more instructions. -This means that if you have many updates on a table, @code{SELECT} -statements will wait until there are no more updates. +If you want to have the character set included in the @strong{MySQL} +distribution, mail a patch to @email{internals@@lists.mysql.com}. -To work around this for the case where you want to do many @code{INSERT} and -@code{SELECT} operations on a table, you can insert rows in a temporary -table and update the real table with the records from the temporary table -once in a while. +@node Character arrays, String collating, Adding character set, Languages +@subsection The character definition arrays + +@code{to_lower[]} and @code{to_upper[]} are simple arrays that hold the +lowercase and uppercase characters corresponding to each member of the +character set. For example: -This can be done with the following code: @example -mysql> LOCK TABLES real_table WRITE, insert_table WRITE; -mysql> insert into real_table select * from insert_table; -mysql> TRUNCATE TABLE insert_table; -mysql> UNLOCK TABLES; +to_lower['A'] should contain 'a' +to_upper['a'] should contain 'A' @end example -You can use the @code{LOW_PRIORITY} options with @code{INSERT}, -@code{UPDATE} or @code{DELETE} or @code{HIGH_PRIORITY} with -@code{SELECT} if you want to prioritize retrieval in some specific -cases. You can also start @code{mysqld} with @code{--low-priority-updates} -to get the same behaveour. +@code{sort_order[]} is a map indicating how characters should be ordered for +comparison and sorting purposes. For many character sets, this is the same as +@code{to_upper[]} (which means sorting will be case insensitive). +@strong{MySQL} will sort characters based on the value of +@code{sort_order[character]}. For more complicated sorting rules, see +the discussion of string collating below. @xref{String collating}. -Using @code{SQL_BUFFER_RESULT} can also help making table locks shorter. -@xref{SELECT}. +@code{ctype[]} is an array of bit values, with one element for one character. +(Note that @code{to_lower[]}, @code{to_upper[]}, and @code{sort_order[]} +are indexed by character value, but @code{ctype[]} is indexed by character +value + 1. This is an old legacy to be able to handle EOF.) -You could also change the locking code in @file{mysys/thr_lock.c} to use a -single queue. In this case, write locks and read locks would have the same -priority, which might help some applications. +You can find the following bitmask definitions in @file{m_ctype.h}: -@cindex problems, table locking -@node Table locking, DNS, Internal locking, System -@subsection Table Locking Issues +@example +#define _U 01 /* Uppercase */ +#define _L 02 /* Lowercase */ +#define _N 04 /* Numeral (digit) */ +#define _S 010 /* Spacing character */ +#define _P 020 /* Punctuation */ +#define _C 040 /* Control character */ +#define _B 0100 /* Blank */ +#define _X 0200 /* heXadecimal digit */ +@end example -The table locking code in @strong{MySQL} is deadlock free. +The @code{ctype[]} entry for each character should be the union of the +applicable bitmask values that describe the character. For example, +@code{'A'} is an uppercase character (@code{_U}) as well as a +hexadecimal digit (@code{_X}), so @code{ctype['A'+1]} should contain the +value: -@strong{MySQL} uses table locking (instead of row locking or column -locking) on all table types, except @code{BDB} tables, to achieve a very -high lock speed. For large tables, table locking is MUCH better than -row locking for most applications, but there are, of course, some -pitfalls. +@example +_U + _X = 01 + 0200 = 0201 +@end example -For @code{BDB} and @code{InnoDB} tables, @strong{MySQL} only uses table -locking if you explicitely lock the table with @code{LOCK TABLES} or -execute a command that will modify every row in the table, like -@code{ALTER TABLE}. For these table types we recommend you to not use -@code{LOCK TABLES} at all. +@cindex collating, strings +@cindex string collating +@node String collating, Multi-byte characters, Character arrays, Languages +@subsection String Collating Support -In @strong{MySQL} Version 3.23.7 and above, you can insert rows into -@code{MyISAM} tables at the same time other threads are reading from the -table. Note that currently this only works if there are no holes after -deleted rows in the table at the time the insert is made. When all holes -has been filled with new data, concurrent inserts will automatically be -enabled again. +If the sorting rules for your language are too complex to be handled +with the simple @code{sort_order[]} table, you need to use the string +collating functions. -Table locking enables many threads to read from a table at the same -time, but if a thread wants to write to a table, it must first get -exclusive access. During the update, all other threads that want to -access this particular table will wait until the update is ready. +Right now the best documentation on this is the character sets that are +already implemented. Look at the big5, czech, gbk, sjis, and tis160 +character sets for examples. -As updates on tables normally are considered to be more important than -@code{SELECT}, all statements that update a table have higher priority -than statements that retrieve information from a table. This should -ensure that updates are not 'starved' because one issues a lot of heavy -queries against a specific table. (You can change this by using -LOW_PRIORITY with the statement that does the update or -@code{HIGH_PRIORITY} with the @code{SELECT} statement.) +You must specify the @code{strxfrm_multiply_MYSET=N} value in the +special comment at the top of the file. @code{N} should be set to +the maximum ratio the strings may grow during @code{my_strxfrm_MYSET} (it +must be a positive integer). -Starting from @strong{MySQL} Version 3.23.7 one can use the -@code{max_write_lock_count} variable to force @strong{MySQL} to -temporary give all @code{SELECT} statements, that wait for a table, a -higher priority after a specific number of inserts on a table. +@cindex characters, multi-byte +@cindex multi-byte characters +@node Multi-byte characters, , String collating, Languages +@subsection Multi-byte Character Support -Table locking is, however, not very good under the following senario: +If your want to add support for a new character set that includes +multi-byte characters, you need to use the multi-byte character +functions. -@itemize @bullet -@item -A client issues a @code{SELECT} that takes a long time to run. -@item -Another client then issues an @code{UPDATE} on a used table. This client -will wait until the @code{SELECT} is finished. -@item -Another client issues another @code{SELECT} statement on the same table. As -@code{UPDATE} has higher priority than @code{SELECT}, this @code{SELECT} -will wait for the @code{UPDATE} to finish. It will also wait for the first -@code{SELECT} to finish! -@item -A thread is waiting for something like @code{full disk}, in which case all -threads that wants to access the problem table will also be put in a waiting -state until more disk space is made available. -@end itemize +Right now the best documentation on this is the character sets that are +already implemented. Look at the euc_kr, gb2312, gbk, sjis and ujis +character sets for examples. These are implemented in the +@code{ctype-'charset'.c} files in the @file{strings} directory. -Some possible solutions to this problem are: +You must specify the @code{mbmaxlen_MYSET=N} value in the special +comment at the top of the source file. @code{N} should be set to the +size in bytes of the largest character in the set. -@itemize @bullet -@item -Try to get the @code{SELECT} statements to run faster. You may have to create -some summary tables to do this. -@item -Start @code{mysqld} with @code{--low-priority-updates}. This will give -all statements that update (modify) a table lower priority than a @code{SELECT} -statement. In this case the last @code{SELECT} statement in the previous -scenario would execute before the @code{INSERT} statement. +@cindex replication +@cindex increasing, speed +@cindex speed, increasing +@cindex databases, replicating +@node Replication, Fulltext Search, Server, Top +@chapter Replication in MySQL -@item -You can give a specific @code{INSERT}, @code{UPDATE}, or @code{DELETE} -statement lower priority with the @code{LOW_PRIORITY} attribute. +@menu +* Replication Intro:: Introduction +* Replication Implementation:: Replication Implementation Overview +* Replication HOWTO:: HOWTO +* Replication Features:: Replication Features +* Replication Options:: Replication Options in my.cnf +* Replication SQL:: SQL Commands related to replication +* Replication FAQ:: Frequently Asked Questions about replication +* Replication Problems:: Troubleshooting Replication. +@end menu -@item -Start @code{mysqld} with a low value for @strong{max_write_lock_count} to give -@code{READ} locks after a certain number of @code{WRITE} locks. +This chapter describes the various replication features in @strong{MySQL}. +It serves as a reference to the options available with replication. +You will be introduced to replication and learn how to implement it. +Towards the end, there are some frequently asked questions and descriptions +of problems and how to solve them. -@item -You can specify that all updates from a specific thread should be done with -low priority by using the SQL command: @code{SET SQL_LOW_PRIORITY_UPDATES=1}. -@xref{SET OPTION, , @code{SET OPTION}}. +@node Replication Intro, Replication Implementation, Replication, Replication +@section Introduction -@item -You can specify that a specific @code{SELECT} is very important with the -@code{HIGH_PRIORITY} attribute. @xref{SELECT, , @code{SELECT}}. +One way replication can be used is to increase both robustness and +speed. For robustness you can have two systems and can switch to the backup if +you have problems with the master. The extra speed is achieved by +sending a part of the non-updating queries to the replica server. Of +course this only works if non-updating queries dominate, but that is the +normal case. -@item -If you have problems with @code{INSERT} combined with @code{SELECT}, -switch to use the new @code{MyISAM} tables as these support concurrent -@code{SELECT}s and @code{INSERT}s. +Starting in Version 3.23.15, @strong{MySQL} supports one-way replication +internally. One server acts as the master, while the other acts as the +slave. Note that one server could play the roles of master in one pair +and slave in the other. The master server keeps a binary log of updates +(@xref{Binary log}.) and an index file to binary logs to keep track of +log rotation. The slave, upon connecting, informs the master where it +left off since the last successfully propagated update, catches up on +the updates, and then blocks and waits for the master to notify it of +the new updates. -@item -If you mainly mix @code{INSERT} and @code{SELECT} statements, the -@code{DELAYED} attribute to @code{INSERT} will probably solve your problems. -@xref{INSERT, , @code{INSERT}}. +Note that if you are replicating a database, all updates to this +database should be done through the master! -@item -If you have problems with @code{SELECT} and @code{DELETE}, the @code{LIMIT} -option to @code{DELETE} may help. @xref{DELETE, , @code{DELETE}}. -@end itemize +On older servers one can use the update log to do simple replication. +@xref{Log Replication}. -@cindex DNS -@cindex hostname caching -@node DNS, , Table locking, System -@subsection How MySQL uses DNS +Another benefit of using replication is that one can get live backups of +the system by doing a backup on a slave instead of doing it on the +master. @xref{Backup}. -When a new thread connects to @code{mysqld}, @code{mysqld} will span a -new thread to handle the request. This thread will first check if the -hostname is in the hostname cache. If not the thread will call -@code{gethostbyaddr_r()} and @code{gethostbyname_r()} to resolve the -hostname. +@cindex master-slave setup +@node Replication Implementation, Replication HOWTO, Replication Intro, Replication +@section Replication Implementation Overview -If the operating system doesn't support the above thread-safe calls, the -thread will lock a mutex and call @code{gethostbyaddr()} and -@code{gethostbyname()} instead. Note that in this case no other thread -can resolve other hostnames that is not in the hostname cache until the -first thread is ready. +@strong{MySQL} replication is based on the server keeping track of all +changes to your database (updates, deletes, etc) in the binary +log. (@xref{Binary log}.) and the slave server(s) reading the saved +queries from the master server's binary log so that the slave can +execute the same queries on its copy of the data. -You can disable DNS host lookup by starting @code{mysqld} with -@code{--skip-name-resolve}. In this case you can however only use IP -names in the @strong{MySQL} privilege tables. +It is @strong{very important} to realize that the binary log is simply a +record starting from a fixed point in time (the moment you enable binary +logging). Any slaves which you set up will need copies of all the data +from your master as it existed the moment that you enabled binary +logging on the master. If you start your slaves with data that doesn't +agree with what was on the master @strong{when the binary log was +started}, your slaves may fail. -If you have a very slow DNS and many hosts, you can get more performance by -either disabling DNS lookop with @code{--skip-name-resolve} or by -increasing the @code{HOST_CACHE_SIZE} define (default: 128) and recompile -@code{mysqld}. +A future version (4.0) of @strong{MySQL} will remove the need to keep a +(possibly large) snapshot of data for new slaves that you might wish to +set up through the live backup functionality with no locking required. +However, at this time, it is necessary to block all writes either with a +global read lock or by shutting down the master while taking a snapshot. -You can disable the hostname cache with @code{--skip-host-cache}. You -can clear the hostname cache with @code{FLUSH HOSTS} or @code{mysqladmin -flush-hosts}. +Once a slave is properly configured and running, it will simply connect +to the master and wait for updates to process. If the master goes away +or the slave loses connectivity with your master, it will keep trying to +connect every @code{master-connect-retry} seconds until it is able to +reconnect and resume listening for updates. -If you don't want to allow connections over @code{TCP/IP}, you can do this -by starting @code{mysqld} with @code{--skip-networking}. +Each slave keeps track of where it left off. The master server has no +knowledge of how many slaves there are or which ones are up-to-date at +any given time. -@cindex data, size -@cindex reducing, data size -@cindex storage space, minimizing -@cindex tables, improving performance -@cindex performance, improving -@node Data size, MySQL indexes, System, Performance -@section Get Your Data as Small as Possible +The next section explains the master/slave setup process in more detail. -One of the most basic optimization is to get your data (and indexes) to -take as little space on the disk (and in memory) as possible. This can -give huge improvements because disk reads are faster and normally less -main memory will be used. Indexing also takes less resources if -done on smaller columns. +@node Replication HOWTO, Replication Features, Replication Implementation, Replication +@section HOWTO -@strong{MySQL} supports a lot of different table types and row formats. -Choosing the right table format may give you a big performance gain. -@xref{Table types}. +Below is a quick description of how to set up complete replication on +your current @strong{MySQL} server. It assumes you want to replicate all +your databases and have not configured replication before. You will need +to shutdown your master server briefly to complete the steps outlined +below. -You can get better performance on a table and minimize storage space -using the techniques listed below: +@enumerate +@item +Make sure you have a recent version of @strong{MySQL} installed on the master +and slave(s). -@itemize @bullet +Use Version 3.23.29 or higher. Previous releases used a different binary +log format and had bugs which have been fixed in newer releases. Please, +do not report bugs until you have verified that the problem is present +in the latest release. @item -Use the most efficient (smallest) types possible. @strong{MySQL} has -many specialized types that save disk space and memory. +Set up special a replication user on the master with the @code{FILE} +privilege and permission to connect from all the slaves. If the user is +only doing replication (which is recommended), you don't need to grant any +additional privileges. + +For example, to create a user named @code{repl} which can access your +master from any host, you might use this command: + +@example +GRANT FILE ON *.* TO repl@@"%" IDENTIFIED BY '<password>'; +@end example + @item -Use the smaller integer types if possible to get smaller tables. For -example, @code{MEDIUMINT} is often better than @code{INT}. +Shut down @strong{MySQL} on the master. + +@example +mysqladmin -u root -p<password> shutdown +@end example + @item -Declare columns to be @code{NOT NULL} if possible. It makes everything -faster and you save one bit per column. Note that if you really need -@code{NULL} in your application you should definitely use it. Just avoid -having it on all columns by default. +Snapshot all the data on your master server. + +The easiest way to do this (on Unix) is to simply use @strong{tar} to +produce an archive of your entire data directory. The exact data +directory location depends on your installation. + +@example +tar -cvf /tmp/mysql-snapshot.tar /path/to/data-dir +@end example + +Windows users can use WinZip or similar software to create an archive of +the data directory. + @item -If you don't have any variable-length columns (@code{VARCHAR}, -@code{TEXT}, or @code{BLOB} columns), a fixed-size record format is -used. This is faster but unfortunately may waste some space. -@xref{MyISAM table formats}. -@item -The primary index of a table should be as short as possible. This makes -identification of one row easy and efficient. -@item -For each table, you have to decide which storage/index method to -use. @xref{Table types}. -@item -Only create the indexes that you really need. Indexes are good for -retrieval but bad when you need to store things fast. If you mostly -access a table by searching on a combination of columns, make an index -on them. The first index part should be the most used column. If you are -ALWAYS using many columns, you should use the column with more duplicates -first to get better compression of the index. -@item -If it's very likely that a column has a unique prefix on the first number -of characters, it's better to only index this prefix. @strong{MySQL} -supports an index on a part of a character column. Shorter indexes are -faster not only because they take less disk space but also because they -will give you more hits in the index cache and thus fewer disk -seeks. @xref{Server parameters}. -@item -In some circumstances it can be beneficial to split into two a table that is -scanned very often. This is especially true if it is a dynamic -format table and it is possible to use a smaller static format table that -can be used to find the relevant rows when scanning the table. -@end itemize - -@cindex indexes, uses for -@node MySQL indexes, Query Speed, Data size, Performance -@section How MySQL Uses Indexes - -Indexes are used to find rows with a specific value of one column -fast. Without an index @strong{MySQL} has to start with the first record -and then read through the whole table until it finds the relevant -rows. The bigger the table, the more this costs. If the table has an index -for the columns in question, @strong{MySQL} can quickly get a position to -seek to in the middle of the data file without having to look at all the -data. If a table has 1000 rows, this is at least 100 times faster than -reading sequentially. Note that if you need to access almost all 1000 -rows it is faster to read sequentially because we then avoid disk seeks. - -All @strong{MySQL} indexes (@code{PRIMARY}, @code{UNIQUE}, and -@code{INDEX}) are stored in B-trees. Strings are automatically prefix- -and end-space compressed. @xref{CREATE INDEX, , @code{CREATE INDEX}}. - -Indexes are used to: -@itemize @bullet -@item -Quickly find the rows that match a @code{WHERE} clause. - -@item -Retrieve rows from other tables when performing joins. - -@item -Find the @code{MAX()} or @code{MIN()} value for a specific indexed -column. This is optimized by a preprocessor that checks if you are -using @code{WHERE} key_part_# = constant on all key parts < N. In this case -@strong{MySQL} will do a single key lookup and replace the @code{MIN()} -expression with a constant. If all expressions are replaced with -constants, the query will return at once: +In @code{my.cnf} on the master add @code{log-bin} and +@code{server-id=unique number} to the @code{[mysqld]} section and +restart it. It is very important that the id of the slave is different from +the id of the master. Think of @code{server-id} as something similar +to the IP address - it uniquely identifies the server instance in the +community of replication partners. @example -SELECT MIN(key_part2),MAX(key_part2) FROM table_name where key_part1=10 +[mysqld] +log-bin +server-id=1 @end example @item -Sort or group a table if the sorting or grouping is done on a leftmost -prefix of a usable key (for example, @code{ORDER BY key_part_1,key_part_2 }). The -key is read in reverse order if all key parts are followed by @code{DESC}. - -The index can also be used even if the @code{ORDER BY} doesn't match the index -exactly, as long as all the unused index parts and all the extra -are @code{ORDER BY} columns are constants in the @code{WHERE} clause. The -following queries will use the index to resolve the @code{ORDER BY} part: - -@example -SELECT * FROM foo ORDER BY key_part1,key_part2,key_part3; -SELECT * FROM foo WHERE column=constant ORDER BY column, key_part1; -SELECT * FROM foo WHERE key_part1=const GROUP BY key_part2; -@end example +Restart @strong{MySQL} on the master. @item -In some cases a query can be optimized to retrieve values without -consulting the data file. If all used columns for some table are numeric -and form a leftmost prefix for some key, the values may be retrieved -from the index tree for greater speed: +Add the following to @code{my.cnf} on the slave(s): @example -SELECT key_part3 FROM table_name WHERE key_part1=1 +master-host=<hostname of the master> +master-user=<replication user name> +master-password=<replication user password> +master-port=<TCP/IP port for master> +server-id=<some unique number between 2 and 2^32-1> @end example -@end itemize - -Suppose you issue the following @code{SELECT} statement: +replacing the values in <> with what is relevant to your system. -@example -mysql> SELECT * FROM tbl_name WHERE col1=val1 AND col2=val2; -@end example +@code{server-id} must be different for each server participating in +replication. If you don't specify a server-id, it will be set to 1 if +you have not defined @code{master-host}, else it will be set to 2. Note +that in the case of @code{server-id} omission the master will refuse +connections from all slaves, and the slave will refuse to connect to a +master. Thus, omitting @code{server-id} is only good for backup with a +binary log. -If a multiple-column index exists on @code{col1} and @code{col2}, the -appropriate rows can be fetched directly. If separate single-column -indexes exist on @code{col1} and @code{col2}, the optimizer tries to -find the most restrictive index by deciding which index will find fewer -rows and using that index to fetch the rows. -@cindex indexes, leftmost prefix of -@cindex leftmost prefix of indexes -If the table has a multiple-column index, any leftmost prefix of the -index can be used by the optimizer to find rows. For example, if you -have a three-column index on @code{(col1,col2,col3)}, you have indexed -search capabilities on @code{(col1)}, @code{(col1,col2)}, and -@code{(col1,col2,col3)}. +@item +Copy the snapshot data into your data directory on your slave(s). Make +sure that the privileges on the files and directories are correct. The +user which @strong{MySQL} runs as needs to be able to read and write to +them, just as on the master. -@strong{MySQL} can't use a partial index if the columns don't form a -leftmost prefix of the index. Suppose you have the @code{SELECT} -statements shown below: +@item Restart the slave(s). -@example -mysql> SELECT * FROM tbl_name WHERE col1=val1; -mysql> SELECT * FROM tbl_name WHERE col2=val2; -mysql> SELECT * FROM tbl_name WHERE col2=val2 AND col3=val3; -@end example +@end enumerate -If an index exists on @code{(col1,col2,col3)}, only the first query -shown above uses the index. The second and third queries do involve -indexed columns, but @code{(col2)} and @code{(col2,col3)} are not -leftmost prefixes of @code{(col1,col2,col3)}. +After you have done the above, the slave(s) should connect to the master +and catch up on any updates which happened since the snapshot was taken. -@findex LIKE, and indexes -@findex LIKE, and wildcards -@cindex indexes, and @code{LIKE} -@cindex wildcards, and @code{LIKE} -@strong{MySQL} also uses indexes for @code{LIKE} comparisons if the argument -to @code{LIKE} is a constant string that doesn't start with a wild-card -character. For example, the following @code{SELECT} statements use indexes: +If you have forgotten to set @code{server-id} for the slave you will get +the following error in the error log file: @example -mysql> select * from tbl_name where key_col LIKE "Patrick%"; -mysql> select * from tbl_name where key_col LIKE "Pat%_ck%"; +Warning: one should set server_id to a non-0 value if master_host is set. +The server will not act as a slave. @end example -In the first statement, only rows with @code{"Patrick" <= key_col < -"Patricl"} are considered. In the second statement, only rows with -@code{"Pat" <= key_col < "Pau"} are considered. +If you have forgot to do this for the master, the slaves will not be +able to connect to the master. -The following @code{SELECT} statements will not use indexes: -@example -mysql> select * from tbl_name where key_col LIKE "%Patrick%"; -mysql> select * from tbl_name where key_col LIKE other_col; -@end example +If a slave is not able to replicate for any reason, you will find error +messages in the error log on the slave. -In the first statement, the @code{LIKE} value begins with a wild-card -character. In the second statement, the @code{LIKE} value is not a -constant. +Once a slave is replicating, you will find a file called +@code{master.info} in the same directory as your error log. The +@code{master.info} file is used by the slave to keep track of how much +of the master's binary log is has processed. @strong{Do not} remove or +edit the file, unless you really know what you are doing. Even in that case, +it is preferred that you use @code{CHANGE MASTER TO} command. -@findex IS NULL, and indexes -@cindex indexes, and @code{IS NULL} -Searching using @code{column_name IS NULL} will use indexes if column_name -is an index. +@cindex options, replication +@cindex @code{my.cnf} file +@cindex files,@code{my.cnf} +@node Replication Features, Replication Options, Replication HOWTO, Replication +@section Replication Features and known problems -@strong{MySQL} normally uses the index that finds the least number of rows. An -index is used for columns that you compare with the following operators: -@code{=}, @code{>}, @code{>=}, @code{<}, @code{<=}, @code{BETWEEN}, and a -@code{LIKE} with a non-wild-card prefix like @code{'something%'}. +Below is an explanation of what is supported and what is not: -Any index that doesn't span all @code{AND} levels in the @code{WHERE} clause -is not used to optimize the query. In other words: To be able to use an -index, a prefix of the index must be used in every @code{AND} group. +@itemize @bullet +@item +Replication will be done correctly with @code{AUTO_INCREMENT}, +@code{LAST_INSERT_ID}, and @code{TIMESTAMP} values. +@item +@code{RAND()} in updates does not replicate properly. Use +@code{RAND(some_non_rand_expr)} if you are replicating updates with +@code{RAND()}. You can, for example, use @code{UNIX_TIMESTAMP()} for the +argument to @code{RAND()}. +@item +You have to use the same character set (@code{--default-character-set}) +on the master and the slave. If not, you may get duplicate key errors on +the slave, because a key that is regarded as unique on the master may +not be that in the other character set. +@item +@code{LOAD DATA INFILE} will be handled properly as long as the file +still resides on the master server at the time of update +propagation. @code{LOAD LOCAL DATA INFILE} will be skipped. +@item +Update queries that use user variables are not replication-safe (yet). +@item +@code{FLUSH} commands are not stored in the binary log and are because +of this not replicated to the slaves. This is not normally a problem as +@code{FLUSH} doesn't change anything. This does however mean that if you +update the @code{MySQL} privilege tables directly without using +@code{GRANT} statement and you replicate the @code{MySQL} privilege +database, you must do a @code{FLUSH PRIVILEGES} on your slaves to put +the new privileges into effect. +@item +Temporary tables starting in 3.23.29 are replicated properly with the +exception of the case when you shut down slave server ( not just slave thread), +you have some temporary tables open, and the are used in subsequent updates. +To deal with this problem, to shut down the slave, do @code{SLAVE STOP}, then +check @code{Slave_open_temp_tables} variable to see if it is 0, then issue +@code{mysqladmin shutdown}. If the number is not 0, restart the slave thread +with @code{SLAVE START} and see +if you have better luck next time. There will be a cleaner solution, but it +has to wait until version 4.0. +In earlier versions temporary tables are not being replicated properly - we +recommend that you either upgrade, or execute @code{SET SQL_LOG_BIN=0} on +your clients before all queries with temp tables. +@item +@strong{MySQL} only supports one master and many slaves. We will in 4.x +add a voting algorithm to automatically change master if something goes +wrong with the current master. We will also introduce 'agent' processes +to help doing load balancing by sending select queries to different +slaves. +@item +Starting in Version 3.23.26, it is safe to connect servers in a circular +master-slave relationship with @code{log-slave-updates} enabled. +Note, however, that many queries will not work right in this kind of +setup unless your client code is written to take care of the potential +problems that can happen from updates that occur in different sequence +on different servers. -The following @code{WHERE} clauses use indexes: -@example -... WHERE index_part1=1 AND index_part2=2 AND other_column=3 -... WHERE index=1 OR A=10 AND index=2 /* index = 1 OR index = 2 */ -... WHERE index_part1='hello' AND index_part_3=5 - /* optimized like "index_part1='hello'" */ -... WHERE index1=1 and index2=2 or index1=3 and index3=3; - /* Can use index on index1 but not on index2 or index 3 */ -@end example +This means that you can do a setup like the following: -These @code{WHERE} clauses do @strong{NOT} use indexes: @example -... WHERE index_part2=1 AND index_part3=2 /* index_part_1 is not used */ -... WHERE index=1 OR A=10 /* Index is not used in both AND parts */ -... WHERE index_part1=1 OR index_part2=10 /* No index spans all rows */ +A -> B -> C -> A @end example -Note that in some cases @strong{MySQL} will not use an index, even if one -would be available. Some of the cases where this happens are: +This setup will only works if you only do non conflicting updates +between the tables. In other words, if you insert data in A and C, you +should never insert a row in A that may have a conflicting key with a +row insert in C. You should also not update the sam rows on two servers +if the order in which the updates are applied matters. -@itemize @bullet +Note that the log format has changed in Version 3.23.26 so that +pre-3.23.26 slaves will not be able to read it. @item -If the use of the index would require @strong{MySQL} to access more -than 30 % of the rows in the table. (In this case a table scan is -probably much faster, as this will require us to do much fewer seeks). -Note that if such a query uses @code{LIMIT} to only retrieve -part of the rows, @strong{MySQL} will use an index anyway, as it can -much more quickly find the few rows to return in the result. +If the query on the slave gets an error, the slave thread will +terminate, and a message will appear in the @code{.err} file. You should +then connect to the slave manually, fix the cause of the error (for +example, non-existent table), and then run @code{SLAVE START} sql +command (available starting in Version 3.23.16). In Version 3.23.15, you +will have to restart the server. +@item +If connection to the master is lost, the slave will retry immediately, +and then in case of failure every @code{master-connect-retry} (default +60) seconds. Because of this, it is safe to shut down the master, and +then restart it after a while. The slave will also be able to deal with +network connectivity outages. +@item +Shutting down the slave (cleanly) is also safe, as it keeps track of +where it left off. Unclean shutdowns might produce problems, especially +if disk cache was not synced before the system died. Your system fault +tolerance will be greatly increased if you have a good UPS. +@item +If the master is listening on a non-standard port, you will also need to +specify this with @code{master-port} parameter in @code{my.cnf} . +@item +In Version 3.23.15, all of the tables and databases will be +replicated. Starting in Version 3.23.16, you can restrict replication to +a set of databases with @code{replicate-do-db} directives in +@code{my.cnf} or just exclude a set of databases with +@code{replicate-ignore-db}. Note that up until Version 3.23.23, there was a bug +that did not properly deal with @code{LOAD DATA INFILE} if you did it in +a database that was excluded from replication. +@item +Starting in Version 3.23.16, @code{SET SQL_LOG_BIN = 0} will turn off +replication (binary) logging on the master, and @code{SET SQL_LOG_BIN = +1} will turn in back on - you must have the process privilege to do +this. +@item +Starting in Version 3.23.19, you can clean up stale replication leftovers when +something goes wrong and you want a clean start with @code{FLUSH MASTER} +and @code{FLUSH SLAVE} commands. In Version 3.23.26 we have renamed them to +@code{RESET MASTER} and @code{RESET SLAVE} respectively to clarify +what they do. The old @code{FLUSH} variants still work, though, for +compatibility. + +@item +Starting in Version 3.23.21, you can use @code{LOAD TABLE FROM MASTER} for +network backup and to set up replication initially. We have recently +received a number of bug reports concerning it that we are investigating, so +we recommend that you use it only in testing until we make it more stable. +@item +Starting in Version 3.23.23, you can change masters and adjust log position +with @code{CHANGE MASTER TO}. +@item +Starting in Version 3.23.23, you tell the master that updates in certain +databases should not be logged to the binary log with @code{binlog-ignore-db}. +@item +Starting in Version 3.23.26, you can use @code{replicate-rewrite-db} to tell +the slave to apply updates from one database on the master to the one +with a different name on the slave. +@item +Starting in Version 3.23.28, you can use @code{PURGE MASTER LOGS TO 'log-name'} +to get rid of old logs while the slave is running. @end itemize -@cindex queries, speed of -@cindex permission checks, effect on speed -@cindex speed, of queries -@node Query Speed, Tips, MySQL indexes, Performance -@section Speed of Queries that Access or Update Data +@node Replication Options, Replication SQL, Replication Features, Replication +@section Replication Options in my.cnf -First, one thing that affects all queries: The more complex permission -system setup you have, the more overhead you get. +If you are using replication, we recommend you to use @strong{MySQL} Version +3.23.30 or later. Older versions work, but they do have some bugs and are +missing some features. -If you do not have any @code{GRANT} statements done, @strong{MySQL} will -optimize the permission checking somewhat. So if you have a very high -volume it may be worth the time to avoid grants. Otherwise more -permission check results in a larger overhead. +On both master and slave you need to use the @code{server-id} option. +This sets an unique replication id. You should pick a unique value in the +range between 1 to 2^32-1 for each master and slave. +Example: @code{server-id=3} -If your problem is with some explicit @strong{MySQL} function, you can -always time this in the @strong{MySQL} client: +The following table has the options you can use for the @strong{MASTER}: -@example -mysql> select benchmark(1000000,1+1); -+------------------------+ -| benchmark(1000000,1+1) | -+------------------------+ -| 0 | -+------------------------+ -1 row in set (0.32 sec) -@end example +@multitable @columnfractions .3 .7 -The above shows that @strong{MySQL} can execute 1,000,000 @code{+} -expressions in 0.32 seconds on a @code{PentiumII 400MHz}. +@item @strong{Option} @tab @strong{Description} +@item @code{log-bin=filename} @tab +Write to a binary update log to the specified location. Note that if you +give it a parameter with an extension (for example, +@code{log-bin=/mysql/logs/replication.log} ) versions up to 3.23.24 will +not work right during replication if you do @code{FLUSH LOGS} . The +problem is fixed in Version 3.23.25. If you are using this kind of log +name, @code{FLUSH LOGS} will be ignored on binlog. To clear the log, run +@code{FLUSH MASTER}, and do not forget to run @code{FLUSH SLAVE} on all +slaves. In Version 3.23.26 and in later versions you should use +@code{RESET MASTER} and @code{RESET SLAVE} -All @strong{MySQL} functions should be very optimized, but there may be -some exceptions, and the @code{benchmark(loop_count,expression)} is a -great tool to find out if this is a problem with your query. +@item @code{log-bin-index=filename} @tab +Because the user could issue the @code{FLUSH LOGS} command, we need to +know which log is currently active and which ones have been rotated out +and in what sequence. This information is stored in the binary log index file. +The default is `hostname`.index. You can use this option if you want to +be a rebel. (Example: @code{log-bin-index=db.index}) -@menu -* Estimating performance:: Estimating query performance -* SELECT speed:: Speed of @code{SELECT} queries -* Where optimizations:: How MySQL optimizes @code{WHERE} clauses -* DISTINCT optimization:: How MySQL Optimizes @code{DISTINCT} -* LEFT JOIN optimization:: How MySQL optimizes @code{LEFT JOIN} -* LIMIT optimization:: How MySQL optimizes @code{LIMIT} -* Insert speed:: Speed of @code{INSERT} queries -* Update speed:: Speed of @code{UPDATE} queries -* Delete speed:: Speed of @code{DELETE} queries -@end menu +@item @code{sql-bin-update-same} @tab +If set, setting @code{SQL_LOG_BIN} to a value will automatically set +@code{SQL_LOG_UPDATE} to the same value and vice versa. -@cindex estimating, query performance -@cindex queries, estimating performance -@cindex performance, estimating -@node Estimating performance, SELECT speed, Query Speed, Query Speed -@subsection Estimating Query Performance +@item @code{binlog-do-db=database_name} @tab +Tells the master it should log updates for the specified database, and +exclude all others not explicitly mentioned. +(Example: @code{binlog-do-db=some_database}) -In most cases you can estimate the performance by counting disk seeks. -For small tables, you can usually find the row in 1 disk seek (as the -index is probably cached). For bigger tables, you can estimate that -(using B++ tree indexes) you will need: @code{log(row_count) / -log(index_block_length / 3 * 2 / (index_length + data_pointer_length)) + -1} seeks to find a row. +@item @code{binlog-ignore-db=database_name} @tab +Tells the master that updates to the given database should not be logged +to the binary log (Example: @code{binlog-ignore-db=some_database}) +@end multitable -In @strong{MySQL} an index block is usually 1024 bytes and the data -pointer is usually 4 bytes. A 500,000 row table with an -index length of 3 (medium integer) gives you: -@code{log(500,000)/log(1024/3*2/(3+4)) + 1} = 4 seeks. +The following table has the options you can use for the @strong{SLAVE}: -As the above index would require about 500,000 * 7 * 3/2 = 5.2M, -(assuming that the index buffers are filled to 2/3, which is typical) -you will probably have much of the index in memory and you will probably -only need 1-2 calls to read data from the OS to find the row. +@multitable @columnfractions .3 .7 -For writes, however, you will need 4 seek requests (as above) to find -where to place the new index and normally 2 seeks to update the index -and write the row. +@item @strong{Option} @tab @strong{Description} +@item @code{master-host=host} @tab +Master hostname or IP address for replication. If not set, the slave +thread will not be started. +(Example: @code{master-host=db-master.mycompany.com}) -Note that the above doesn't mean that your application will slowly -degenerate by N log N! As long as everything is cached by the OS or SQL -server things will only go marginally slower while the table gets -bigger. After the data gets too big to be cached, things will start to -go much slower until your applications is only bound by disk-seeks -(which increase by N log N). To avoid this, increase the index cache as -the data grows. @xref{Server parameters}. +@item @code{master-user=username} @tab +The user the slave thread will us for authentication when connecting to +the master. The user must have @code{FILE} privilege. If the master user +is not set, user @code{test} is assumed. (Example: +@code{master-user=scott}) -@cindex speed, of queries -@findex SELECT speed +@item @code{master-password=password} @tab +The password the slave thread will authenticate with when connecting to +the master. If not set, an empty password is assumed. (Example: +@code{master-password=tiger}) -@node SELECT speed, Where optimizations, Estimating performance, Query Speed -@subsection Speed of @code{SELECT} Queries +@item @code{master-port=portnumber} @tab +The port the master is listening on. If not set, the compiled setting of +@code{MYSQL_PORT} is assumed. If you have not tinkered with +@code{configure} options, this should be 3306. (Example: +@code{master-port=3306}) -In general, when you want to make a slow @code{SELECT ... WHERE} faster, the -first thing to check is whether or not you can add an index. @xref{MySQL -indexes, , @strong{MySQL} indexes}. All references between different tables -should usually be done with indexes. You can use the @code{EXPLAIN} command -to determine which indexes are used for a @code{SELECT}. -@xref{EXPLAIN, , @code{EXPLAIN}}. +@item @code{master-connect-retry=seconds} @tab +The number of seconds the slave thread will sleep before retrying to +connect to the master in case the master goes down or the connection is +lost. Default is 60. (Example: @code{master-connect-retry=60}) -Some general tips: +@item @code{master-info-file=filename} @tab +The location of the file that remembers where we left off on the master +during the replication process. The default is master.info in the data +directory. Sasha: The only reason I see for ever changing the default +is the desire to be rebelious. (Example: +@code{master-info-file=master.info}) -@itemize @bullet -@item -To help @strong{MySQL} optimize queries better, run @code{myisamchk ---analyze} on a table after it has been loaded with relevant data. This -updates a value for each index part that indicates the average number of -rows that have the same value. (For unique indexes, this is always 1, -of course.). @strong{MySQL} will use this to decide which index to -choose when you connect two tables with 'a non-constant expression'. -You can check the result from the @code{analyze} run by doing @code{SHOW -INDEX FROM table_name} and examining the @code{Cardinality} column. +@item @code{replicate-do-table=db_name.table_name} @tab +Tells the slave thread to restrict replication to the specified database. +To specify more than one table, use the directive multiple times, +once for each table. . +(Example: @code{replicate-do-table=some_db.some_table}) -@item -To sort an index and data according to an index, use @code{myisamchk ---sort-index --sort-records=1} (if you want to sort on index 1). If you -have a unique index from which you want to read all records in order -according to that index, this is a good way to make that faster. Note, -however, that this sorting isn't written optimally and will take a long -time for a large table! -@end itemize +@item @code{replicate-ignore-table=db_name.table_name} @tab +Tells the slave thread to not replicate to the specified table. To +specify more than one table to ignore, use the directive multiple +times, once for each table.(Example: +@code{replicate-ignore-table=db_name.some_table}) -@cindex optimizations -@findex WHERE -@node Where optimizations, DISTINCT optimization, SELECT speed, Query Speed -@subsection How MySQL Optimizes @code{WHERE} Clauses +@item @code{replicate-wild-do-table=db_name.table_name} @tab +Tells the slave thread to restrict replication to the tables that match the +specified wildcard pattern. . +To specify more than one table, use the directive multiple times, +once for each table. . +(Example: @code{replicate-do-table=foo%.bar%} will replicate only updates +to tables in all databases that start with foo and whose table names +start with bar) -The @code{WHERE} optimizations are put in the @code{SELECT} part here because -they are mostly used with @code{SELECT}, but the same optimizations apply for -@code{WHERE} in @code{DELETE} and @code{UPDATE} statements. +@item @code{replicate-wild-ignore-table=db_name.table_name} @tab +Tells the slave thread to not replicate to the tables that match the given +wild card pattern. To +specify more than one table to ignore, use the directive multiple +times, once for each table.(Example: +@code{replicate-ignore-table=foo%.bar%} - will not upates to tables in all databases that start with foo and whose table names +start with bar) -Also note that this section is incomplete. @strong{MySQL} does many -optimizations, and we have not had time to document them all. +@item @code{replicate-ignore-db=database_name} @tab +Tells the slave thread to not replicate to the specified database. To +specify more than one database to ignore, use the directive multiple +times, once for each database. This option will not work if you use cross +database updates. If you need cross database updates to work, make sure +you have 3.23.28 or later, and use +@code{replicate-wild-ignore-table=db_name.%}(Example: +@code{replicate-ignore-db=some_db}) -Some of the optimizations performed by @strong{MySQL} are listed below: +@item @code{replicate-do-db=database_name} @tab +Tells the slave thread to restrict replication to the specified database. +To specify more than one database, use the directive multiple times, +once for each database. Note that this will only work if you do not use +cross-database queries such as @code{UPDATE some_db.some_table SET +foo='bar'} while having selected a different or no database. If you need +cross database updates to work, make sure +you have 3.23.28 or later, and use +@code{replicate-wild-do-table=db_name.%} +(Example: @code{replicate-do-db=some_db}) -@itemize @bullet -@item -Removal of unnecessary parentheses: -@example - ((a AND b) AND c OR (((a AND b) AND (c AND d)))) --> (a AND b AND c) OR (a AND b AND c AND d) -@end example -@item -Constant folding: -@example - (a<b AND b=c) AND a=5 --> b>5 AND b=c AND a=5 -@end example -@item -Constant condition removal (needed because of constant folding): -@example - (B>=5 AND B=5) OR (B=6 AND 5=5) OR (B=7 AND 5=6) --> B=5 OR B=6 -@end example -@item -Constant expressions used by indexes are evaluated only once. -@item -@code{COUNT(*)} on a single table without a @code{WHERE} is retrieved -directly from the table information. This is also done for any @code{NOT NULL} -expression when used with only one table. -@item -Early detection of invalid constant expressions. @strong{MySQL} quickly -detects that some @code{SELECT} statements are impossible and returns no rows. -@item -@code{HAVING} is merged with @code{WHERE} if you don't use @code{GROUP BY} -or group functions (@code{COUNT()}, @code{MIN()}...). -@item -For each sub-join, a simpler @code{WHERE} is constructed to get a fast -@code{WHERE} evaluation for each sub-join and also to skip records as -soon as possible. -@cindex constant table -@cindex tables, constant -@item -All constant tables are read first, before any other tables in the query. -A constant table is: -@itemize @minus -@item -An empty table or a table with 1 row. -@item -A table that is used with a @code{WHERE} clause on a @code{UNIQUE} -index, or a @code{PRIMARY KEY}, where all index parts are used with constant -expressions and the index parts are defined as @code{NOT NULL}. -@end itemize -All the following tables are used as constant tables: -@example -mysql> SELECT * FROM t WHERE primary_key=1; -mysql> SELECT * FROM t1,t2 - WHERE t1.primary_key=1 AND t2.primary_key=t1.id; -@end example +@item @code{log-slave-updates} @tab +Tells the slave to log the updates from the slave thread to the binary +log. Off by default. You will need to turn it on if you plan to +daisy-chain the slaves. -@item -The best join combination to join the tables is found by trying all -possibilities. If all columns in @code{ORDER BY} and in @code{GROUP -BY} come from the same table, then this table is preferred first when -joining. -@item -If there is an @code{ORDER BY} clause and a different @code{GROUP BY} -clause, or if the @code{ORDER BY} or @code{GROUP BY} contains columns -from tables other than the first table in the join queue, a temporary -table is created. -@item -If you use @code{SQL_SMALL_RESULT}, @strong{MySQL} will use an in-memory -temporary table. -@item -Each table index is queried, and the best index that spans fewer than 30% of -the rows is used. If no such index can be found, a quick table scan is used. -@item -In some cases, @strong{MySQL} can read rows from the index without even -consulting the data file. If all columns used from the index are numeric, -then only the index tree is used to resolve the query. -@item -Before each record is output, those that do not match the @code{HAVING} clause -are skipped. -@end itemize +@item @code{replicate-rewrite-db=from_name->to_name} @tab +Updates to a database with a different name than the original (Example: +@code{replicate-rewrite-db=master_db_name->slave_db_name} -Some examples of queries that are very fast: +@item @code{skip-slave-start} @tab +Tells the slave server not to start the slave on the startup. The user +can start it later with @code{SLAVE START}. -@example -mysql> SELECT COUNT(*) FROM tbl_name; -mysql> SELECT MIN(key_part1),MAX(key_part1) FROM tbl_name; -mysql> SELECT MAX(key_part2) FROM tbl_name - WHERE key_part_1=constant; -mysql> SELECT ... FROM tbl_name - ORDER BY key_part1,key_part2,... LIMIT 10; -mysql> SELECT ... FROM tbl_name - ORDER BY key_part1 DESC,key_part2 DESC,... LIMIT 10; -@end example +@item @code{slave_read_timeout=#} +Number of seconds to wait for more data from the master before aborting +the read. +@end multitable -The following queries are resolved using only the index tree (assuming -the indexed columns are numeric): +@cindex SQL commands, replication +@cindex commands, replication +@cindex replication, commands +@node Replication SQL, Replication FAQ, Replication Options, Replication +@section SQL Commands Related to Replication -@example -mysql> SELECT key_part1,key_part2 FROM tbl_name WHERE key_part1=val; -mysql> SELECT COUNT(*) FROM tbl_name - WHERE key_part1=val1 AND key_part2=val2; -mysql> SELECT key_part2 FROM tbl_name GROUP BY key_part1; -@end example +Replication can be controlled through the SQL interface. Below is the +summary of commands: -The following queries use indexing to retrieve the rows in sorted -order without a separate sorting pass: +@multitable @columnfractions .30 .70 +@item @strong{Command} @tab @strong{Description} -@example -mysql> SELECT ... FROM tbl_name ORDER BY key_part1,key_part2,... ; -mysql> SELECT ... FROM tbl_name ORDER BY key_part1 DESC,key_part2 DESC,... ; -@end example +@item @code{SLAVE START} + @tab Starts the slave thread. (Slave) -@findex DISTINCT -@cindex optimizing, DISTINCT -@node DISTINCT optimization, LEFT JOIN optimization, Where optimizations, Query Speed -@subsection How MySQL Optimizes @code{DISTINCT} +@item @code{SLAVE STOP} + @tab Stops the slave thread. (Slave) -@code{DISTINCT} is converted to a @code{GROUP BY} on all columns, -@code{DISTINCT} combined with @code{ORDER BY} will in many cases also -need a temporary table. +@item @code{SET SQL_LOG_BIN=0} + @tab Disables update logging if the user has process privilege. + Ignored otherwise. (Master) -When combining @code{LIMIT #} with @code{DISTINCT}, @strong{MySQL} will stop -as soon as it finds @code{#} unique rows. +@item @code{SET SQL_LOG_BIN=1} + @tab Re-enables update logging if the user has process privilege. + Ignored otherwise. (Master) -If you don't use columns from all used tables, @strong{MySQL} will stop -the scanning of the not used tables as soon as it has found the first match. +@item @code{SET SQL_SLAVE_SKIP_COUNTER=n} + @tab Skip the next @code{n} events from the master. Only valid when +the slave thread is not running, otherwise, gives an error. Useful for +recovering from replication glitches. -@example -SELECT DISTINCT t1.a FROM t1,t2 where t1.a=t2.a; -@end example +@item @code{RESET MASTER} + @tab Deletes all binary logs listed in the index file, resetting the binlog +index file to be empty. In pre-3.23.26 versions, @code{FLUSH MASTER} (Master) -In the case, assuming t1 is used before t2 (check with @code{EXPLAIN}), then -@strong{MySQL} will stop reading from t2 (for that particular row in t1) -when the first row in t2 is found. +@item @code{RESET SLAVE} + @tab Makes the slave forget its replication position in the master +logs. In pre 3.23.26 versions the command was called +@code{FLUSH SLAVE}(Slave) -@findex LEFT JOIN -@cindex optimizing, LEFT JOIN -@node LEFT JOIN optimization, LIMIT optimization, DISTINCT optimization, Query Speed -@subsection How MySQL Optimizes @code{LEFT JOIN} and @code{RIGHT JOIN} +@item @code{LOAD TABLE tblname FROM MASTER} + @tab Downloads a copy of the table from master to the slave. (Slave) -@code{A LEFT JOIN B} in @strong{MySQL} is implemented as follows: +@item @code{CHANGE MASTER TO master_def_list} + @tab Changes the master parameters to the values specified in +@code{master_def_list} and restarts the slave thread. @code{master_def_list} +is a comma-separated list of @code{master_def} where @code{master_def} is +one of the following: @code{MASTER_HOST}, @code{MASTER_USER}, +@code{MASTER_PASSWORD}, @code{MASTER_PORT}, @code{MASTER_CONNECT_RETRY}, +@code{MASTER_LOG_FILE}, @code{MASTER_LOG_POS}. Example: -@itemize @bullet -@item -The table @code{B} is set to be dependent on table @code{A} and all tables -that @code{A} is dependent on. -@item -The table @code{A} is set to be dependent on all tables (except @code{B}) -that are used in the @code{LEFT JOIN} condition. -@item -All @code{LEFT JOIN} conditions are moved to the @code{WHERE} clause. -@item -All standard join optimizations are done, with the exception that a table is -always read after all tables it is dependent on. If there is a circular -dependence then @strong{MySQL} will issue an error. -@item -All standard @code{WHERE} optimizations are done. -@item -If there is a row in @code{A} that matches the @code{WHERE} clause, but there -wasn't any row in @code{B} that matched the @code{LEFT JOIN} condition, -then an extra @code{B} row is generated with all columns set to @code{NULL}. -@item -If you use @code{LEFT JOIN} to find rows that don't exist in some -table and you have the following test: @code{column_name IS NULL} in the -@code{WHERE} part, where column_name is a column that is declared as -@code{NOT NULL}, then @strong{MySQL} will stop searching after more rows -(for a particular key combination) after it has found one row that -matches the @code{LEFT JOIN} condition. -@end itemize +@example -@code{RIGHT JOIN} is implemented analogously as @code{LEFT JOIN}. +CHANGE MASTER TO + MASTER_HOST='master2.mycompany.com', + MASTER_USER='replication', + MASTER_PASSWORD='bigs3cret', + MASTER_PORT=3306, + MASTER_LOG_FILE='master2-bin.001', + MASTER_LOG_POS=4; -The table read order forced by @code{LEFT JOIN} and @code{STRAIGHT JOIN} -will help the join optimizer (which calculates in which order tables -should be joined) to do its work much more quickly, as there are fewer -table permutations to check. +@end example -Note that the above means that if you do a query of type: +You only need to specify the values that need to be changed. The values that +you omit will stay the same with the exception of when you change the host or +the port. In that case, the slave will assume that since you are connecting to +a different host or a different port, the master is different. Therefore, the +old values of log and position are not applicable anymore, and will +automatically be reset to an empty string and 0, respectively (the start +values). Note that if you restart the slave, it will remember its last master. +If this is not desirable, you should delete the @file{master.info} file before +restarting, and the slave will read its master from @code{my.cnf} or the +command line. (Slave) -@example -SELECT * FROM a,b LEFT JOIN c ON (c.key=a.key) LEFT JOIN d (d.key=a.key) WHERE b.key=d.key -@end example +@item @code{SHOW MASTER STATUS} @tab Provides status information on the binlog of the master. (Master) -@strong{MySQL} will do a full scan on @code{b} as the @code{LEFT -JOIN} will force it to be read before @code{d}. +@item @code{SHOW SLAVE STATUS} @tab Provides status information on essential parameters of the slave thread. (Slave) +@item @code{SHOW MASTER LOGS} @tab Only available starting in Version 3.23.28. Lists the binary logs on the master. You should use this command prior to @code{PURGE MASTER LOGS TO} to find out how far you should go. -The fix in this case is to change the query to: +@item @code{PURGE MASTER LOGS TO 'logname'} + @tab Available starting in Version 3.23.28. Deletes all the +replication logs that are listed in the log +index as being prior to the specified log, and removed them from the +log index, so that the given log now becomes first. Example: @example -SELECT * FROM b,a LEFT JOIN c ON (c.key=a.key) LEFT JOIN d (d.key=a.key) WHERE b.key=d.key +PURGE MASTER LOGS TO 'mysql-bin.010' @end example -@cindex optimizing, LIMIT -@findex LIMIT -@node LIMIT optimization, Insert speed, LEFT JOIN optimization, Query Speed -@subsection How MySQL Optimizes @code{LIMIT} +This command will do nothing and fail with an error if you have an +active slave that is currently reading one of the logs you are trying to +delete. However, if you have a dormant slave, and happen to purge one of +the logs it wants to read, the slave will be unable to replicate once it +comes up. The command is safe to run while slaves are replicating - you +do not need to stop them. -In some cases @strong{MySQL} will handle the query differently when you are -using @code{LIMIT #} and not using @code{HAVING}: +You must first check all the slaves with @code{SHOW SLAVE STATUS} to +see which log they are on, then do a listing of the logs on the +master with @code{SHOW MASTER LOGS}, find the earliest log among all +the slaves (if all the slaves are up to date, this will be the +last log on the list), backup all the logs you are about to delete +(optional) and purge up to the target log. -@itemize @bullet -@item -If you are selecting only a few rows with @code{LIMIT}, @strong{MySQL} -will use indexes in some cases when it normally would prefer to do a -full table scan. -@item -If you use @code{LIMIT #} with @code{ORDER BY}, @strong{MySQL} will end the -sorting as soon as it has found the first @code{#} lines instead of sorting -the whole table. -@item -When combining @code{LIMIT #} with @code{DISTINCT}, @strong{MySQL} will stop -as soon as it finds @code{#} unique rows. -@item -In some cases a @code{GROUP BY} can be resolved by reading the key in order -(or do a sort on the key) and then calculate summaries until the -key value changes. In this case @code{LIMIT #} will not calculate any -unnecessary @code{GROUP BY}'s. -@item -As soon as @strong{MySQL} has sent the first @code{#} rows to the client, it -will abort the query. -@item -@code{LIMIT 0} will always quickly return an empty set. This is useful -to check the query and to get the column types of the result columns. -@item -The size of temporary tables uses the @code{LIMIT #} to calculate how much -space is needed to resolve the query. -@end itemize +@end multitable -@cindex speed, inserting -@cindex inserting, speed of -@node Insert speed, Update speed, LIMIT optimization, Query Speed -@subsection Speed of @code{INSERT} Queries +@node Replication FAQ, Replication Problems, Replication SQL, Replication +@section Replication FAQ -The time to insert a record consists approximately of: +@cindex @code{Binlog_Dump} +@strong{Q}: Why do I sometimes see more than one @code{Binlog_Dump} thread on +the master after I have restarted the slave? + +@strong{A}: @code{Binlog_Dump} is a continuous process that is handled by the +server in the following way: @itemize @bullet @item -Connect: (3) -@item -Sending query to server: (2) -@item -Parsing query: (2) +Catch up on the updates. @item -Inserting record: (1 x size of record) +Once there are no more updates left, go into @code{pthread_cond_wait()}, +from which we can be awakened either by an update or a kill. @item -Inserting indexes: (1 x number of indexes) +On wake up, check the reason. If we are not supposed to die, continue +the @code{Binlog_dump} loop. @item -Close: (1) +If there is some fatal error, such as detecting a dead client, +terminate the loop. @end itemize -where the numbers are somewhat proportional to the overall time. This -does not take into consideration the initial overhead to open tables -(which is done once for each concurrently running query). - -The size of the table slows down the insertion of indexes by N log N -(B-trees). +So if the slave thread stops on the slave, the corresponding +@code{Binlog_Dump} thread on the master will not notice it until after +at least one update to the master (or a kill), which is needed to wake +it up from @code{pthread_cond_wait()}. In the meantime, the slave +could have opened another connection, which resulted in another +@code{Binlog_Dump} thread. -Some ways to speed up inserts: +The above problem should not be present in Version 3.23.26 and later +versions. In Version 3.23.26 we added @code{server-id} to each +replication server, and now all the old zombie threads are killed on the +master when a new replication thread connects from the same slave -@itemize @bullet -@item -If you are inserting many rows from the same client at the same time, use -multiple value lists @code{INSERT} statements. This is much faster (many -times in some cases) than using separate @code{INSERT} statements. -@item -If you are inserting a lot of rows from different clients, you can get -higher speed by using the @code{INSERT DELAYED} statement. @xref{INSERT, -, @code{INSERT}}. -@item -Note that with @code{MyISAM} you can insert rows at the same time -@code{SELECT}s are running if there are no deleted rows in the tables. -@item -When loading a table from a text file, use @code{LOAD DATA INFILE}. This -is usually 20 times faster than using a lot of @code{INSERT} statements. -@xref{LOAD DATA, , @code{LOAD DATA}}. -@item -It is possible with some extra work to make @code{LOAD DATA INFILE} run even -faster when the table has many indexes. Use the following procedure: +@strong{Q}: How do I rotate replication logs? -@enumerate -@item -Optionally create the table with @code{CREATE TABLE}. For example, using -@code{mysql} or Perl-DBI. +@strong{A}: In Version 3.23.28 you should use @code{PURGE MASTER LOGS +TO} command after determining which logs can be deleted, and optionally +backing them up first. In earlier versions the process is much more +painful, and cannot be safely done without stopping all the slaves in +the case that you plan to re-use log names. You will need to stop the +slave threads, edit the binary log index file, delete all the old logs, +restart the master, start slave threads, and then remove the old log files. -@item -Execute a @code{FLUSH TABLES} statement or the shell command @code{mysqladmin -flush-tables}. -@item -Use @code{myisamchk --keys-used=0 -rq /path/to/db/tbl_name}. This will -remove all usage of all indexes from the table. +@strong{Q}: How do I upgrade on a hot replication setup? -@item -Insert data into the table with @code{LOAD DATA INFILE}. This will not -update any indexes and will therefore be very fast. +@strong{A}: If you are upgrading pre-3.23.26 versions, you should just +lock the master tables, let the slave catch up, then run @code{FLUSH +MASTER} on the master, and @code{FLUSH SLAVE} on the slave to reset the +logs, then restart new versions of the master and the slave. Note that +the slave can stay down for some time - since the master is logging +all the updates, the slave will be able to catch up once it is up and +can connect. -@item -If you are going to only read the table in the future, run @code{myisampack} -on it to make it smaller. @xref{Compressed format}. +After 3.23.26, we have locked the replication protocol for modifications, so +you can upgrade masters and slave on the fly to a newer 3.23 version and you +can have different versions of @strong{MySQL} running on the slave and the +master, as long as they are both newer than 3.23.26. -@item -Re-create the indexes with @code{myisamchk -r -q -/path/to/db/tbl_name}. This will create the index tree in memory before -writing it to disk, which is much faster because it avoids lots of disk -seeks. The resulting index tree is also perfectly balanced. +@cindex replication, two-way +@strong{Q}: What issues should I be aware of when setting up two-way +replication? -@item -Execute a @code{FLUSH TABLES} statement or the shell command @code{mysqladmin -flush-tables}. -@end enumerate +@strong{A}: @strong{MySQL} replication currently does not support any +locking protocol between master and slave to guarantee the atomicity of +a distributed (cross-server) update. In in other words, it is possible +for client A to make an update to co-master 1, and in the meantime, +before it propagates to co-master 2, client B could make an update to +co-master 2 that will make the update of client A work differently than +it did on co-master 1. Thus when the update of client A will make it +to co-master 2, it will produce tables that will be different than +what you have on co-master 1, even after all the updates from co-master +2 have also propagated. So you should not co-chain two servers in a +two-way replication relationship, unless you are sure that you updates +can safely happen in any order, or unless you take care of mis-ordered +updates somehow in the client code. -This procedure will be built into @code{LOAD DATA INFILE} in some future -version of @strong{MySQL}. -@item -You can speed up insertions by locking your tables: -@example -mysql> LOCK TABLES a WRITE; -mysql> INSERT INTO a VALUES (1,23),(2,34),(4,33); -mysql> INSERT INTO a VALUES (8,26),(6,29); -mysql> UNLOCK TABLES; -@end example +You must also realize that two-way replication actually does not improve +performance very much, if at all, as far as updates are concerned. Both +servers need to do the same amount of updates each, as you would have +one server do. The only difference is that there will be a little less +lock contention, because the updates originating on another server will +be serialized in one slave thread. This benefit, though, might be +offset by network delays. -The main speed difference is that the index buffer is flushed to disk only -once, after all @code{INSERT} statements have completed. Normally there would -be as many index buffer flushes as there are different @code{INSERT} -statements. Locking is not needed if you can insert all rows with a single -statement. +@cindex performance, improving +@cindex increasing, performance +@strong{Q}: How can I use replication to improve performance of my system? -Locking will also lower the total time of multi-connection tests, but the -maximum wait time for some threads will go up (because they wait for -locks). For example: +@strong{A}: You should set up one server as the master, and direct all +writes to it, and configure as many slaves as you have the money and +rackspace for, distributing the reads among the master and the slaves. +You can also start the slaves with @code{--skip-bdb}, +@code{--low-priority-updates} and @code{--delay-key-write-for-all-tables} +to get speed improvements for the slave. In this case the slave will +use non-transactional @code{MyISAM} tables instead of @code{BDB} tables +to get more speed. -@example -thread 1 does 1000 inserts -thread 2, 3, and 4 does 1 insert -thread 5 does 1000 inserts -@end example +@strong{Q}: What should I do to prepare my client code to use +performance-enhancing replication? -If you don't use locking, 2, 3, and 4 will finish before 1 and 5. If you -use locking, 2, 3, and 4 probably will not finish before 1 or 5, but the -total time should be about 40% faster. - -As @code{INSERT}, @code{UPDATE}, and @code{DELETE} operations are very -fast in @strong{MySQL}, you will obtain better overall performance by -adding locks around everything that does more than about 5 inserts or -updates in a row. If you do very many inserts in a row, you could do a -@code{LOCK TABLES} followed by an @code{UNLOCK TABLES} once in a while -(about each 1000 rows) to allow other threads access to the table. This -would still result in a nice performance gain. +@strong{A}: +If the part of your code that is responsible for database access has +been properly abstracted/modularized, converting it to run with the +replicated setup should be very smooth and easy - just change the +implementation of your database access to read from some slave or the +master, and to always write to the master. If your code does not have +this level of abstraction, +setting up a replicated system will give you an opportunity/motivation +to it clean up. + You should start by creating a wrapper library +/module with the following functions: -Of course, @code{LOAD DATA INFILE} is much faster for loading data. +@itemize +@item +@code{safe_writer_connect()} +@item +@code{safe_reader_connect()} +@item +@code{safe_reader_query()} +@item +@code{safe_writer_query()} @end itemize -To get some more speed for both @code{LOAD DATA INFILE} and -@code{INSERT}, enlarge the key buffer. @xref{Server parameters}. +@code{safe_} means that the function will take care of handling all +the error conditions. -@node Update speed, Delete speed, Insert speed, Query Speed -@subsection Speed of @code{UPDATE} Queries +You should then convert your client code to use the wrapper library. +It may be a painful and scary process at first, but it will pay off in +the long run. All applications that follow the above pattern will be +able to take advantage of one-master/many slaves solution. The +code will be a lot easier to maintain, and adding troubleshooting +options will be trivial. You will just need to modify one or two +functions, for example, to log how long each query took, or which +query, among your many thousands, gave you an error. If you have written a lot of code already, +you may want to automate the conversion task by using Monty's +@code{replace} utility, which comes with the standard distribution of +@strong{MySQL}, or just write your own Perl script. Hopefully, your +code follows some recognizable pattern. If not, then you are probably +better off re-writing it anyway, or at least going through and manually +beating it into a pattern. -Update queries are optimized as a @code{SELECT} query with the additional -overhead of a write. The speed of the write is dependent on the size of -the data that is being updated and the number of indexes that are -updated. Indexes that are not changed will not be updated. +Note that, of course, you can use different names for the +functions. What is important is having unified interface for connecting +for reads, connecting for writes, doing a read, and doing a write. -Also, another way to get fast updates is to delay updates and then do -many updates in a row later. Doing many updates in a row is much quicker -than doing one at a time if you lock the table. -Note that, with dynamic record format, updating a record to -a longer total length may split the record. So if you do this often, -it is very important to @code{OPTIMIZE TABLE} sometimes. -@xref{OPTIMIZE TABLE, , @code{OPTIMIZE TABLE}}. +@strong{Q}: When and how much can @strong{MySQL} replication improve the performance +of my system? -@node Delete speed, , Update speed, Query Speed -@subsection Speed of @code{DELETE} Queries +@strong{A}: @strong{MySQL} replication is most beneficial for a system +with frequent reads and not so frequent writes. In theory, by using a +one master/many slaves setup you can scale by adding more slaves until +you either run out of network bandwidth, or your update +load grows to the point +that the master cannot handle it. -If you want to delete all rows in the table, you should use -@code{TRUNCATE TABLE table_name}. @xref{TRUNCATE}. +In order to determine how many slaves you can get before the added +benefits begin to level out, and how much you can improve performance +of your site, you need to know your query patterns, and empirically + (by benchmarking) determine the relationship between the throughput +on reads (reads per second, or @code{max_reads}) and on writes +@code{max_writes}) on a typical master and a typical slave. The +example below will show you a rather simplified calculation of what you +can get with replication for our imagined system. -The time to delete a record is exactly proportional to the number of -indexes. To delete records more quickly, you can increase the size of -the index cache. @xref{Server parameters}. +Let's say our system load consists of 10% writes and 90% reads, and we +have determined that @code{max_reads} = 1200 - 2 * @code{max_writes}, +or in other words, our system can do 1200 reads per second with no +writes, our average write is twice as slow as average read, +and the relationship is +linear. Let us suppose that our master and slave are of the same +capacity, and we have N slaves and 1 master. Then we have for each +server (master or slave): -@cindex optimization, tips -@cindex tips, optimization -@node Tips, Benchmarks, Query Speed, Performance -@section Other Optimization Tips +@code{reads = 1200 - 2 * writes} (from bencmarks) -Unsorted tips for faster systems: +@code{reads = 9* writes / (N + 1) } (reads split, but writes go +to all servers) + +@code{9*writes/(N+1) + 2 * writes = 1200} + +@code{writes = 1200/(2 + 9/(N+1)} + +So if N = 0, which means we have no replication, our system can handle +1200/11, about 109 writes per second (which means we will have 9 times +as many reads due to the nature of our application). + +If N = 1, we can get up to 184 writes per second. + +If N = 8, we get up to 400. + +If N = 17, 480 writes. + +Eventually as N approaches infinity (and our budget negative infinity), +we can get very close to 600 writes per second, increasing system +throughput about 5.5 times. However, with only 8 servers, we increased +it almost 4 times already. + +Note that our computations assumed infinite network bandwidth, and +neglected several other factors that could turn out to be significant on +your system. In many cases, you may not be able to make a computation +similar to the one above that will accurately predict what will happen +on your system if you add N replication slaves. However, answering the +following questions should help you decided whether and how much, if at +all, the replication will improve the performance of your system: @itemize @bullet @item -Use persistent connections to the database to avoid the connection -overhead. If you can't use persistent connections and you are doing a -lot of new connections to the database, you may want to change the value -of the @code{thread_cache_size} variable. @xref{Server parameters}. -@item -Always check that all your queries really use the indexes you have created -in the tables. In @strong{MySQL} you can do this with the @code{EXPLAIN} -command. @xref{EXPLAIN, Explain, Explain, manual}. -@item -Try to avoid complex @code{SELECT} queries on tables that are updated a -lot. This is to avoid problems with table locking. -@item -The new @code{MyISAM} tables can insert rows in a table without deleted -rows at the same time another table is reading from it. If this is important -for you, you should consider methods where you don't have to delete rows -or run @code{OPTIMIZE TABLE} after you have deleted a lot of rows. -@item -Use @code{ALTER TABLE ... ORDER BY expr1,expr2...} if you mostly -retrieve rows in expr1,expr2.. order. By using this option after big -changes to the table, you may be able to get higher performance. -@item -In some cases it may make sense to introduce a column that is 'hashed' -based on information from other columns. If this column is short and -reasonably unique it may be much faster than a big index on many -columns. In @strong{MySQL} it's very easy to use this extra column: -@code{SELECT * FROM table_name WHERE hash=MD5(concat(col1,col2)) -AND col_1='constant' AND col_2='constant'} -@item -For tables that change a lot you should try to avoid all @code{VARCHAR} -or @code{BLOB} columns. You will get dynamic row length as soon as you -are using a single @code{VARCHAR} or @code{BLOB} column. @xref{Table -types}. +What is the read/write ratio on your system? @item -It's not normally useful to split a table into different tables just -because the rows gets 'big'. To access a row, the biggest performance -hit is the disk seek to find the first byte of the row. After finding -the data most new disks can read the whole row fast enough for most -applications. The only cases where it really matters to split up a table is if -it's a dynamic row size table (see above) that you can change to a fixed -row size, or if you very often need to scan the table and don't need -most of the columns. @xref{Table types}. +How much more write load can one server handle if you reduce the reads? @item -If you very often need to calculate things based on information from a -lot of rows (like counts of things), it's probably much better to -introduce a new table and update the counter in real time. An update of -type @code{UPDATE table set count=count+1 where index_column=constant} -is very fast! +How many slaves do you have bandwidth for on your network? +@end itemize -This is really important when you use databases like @strong{MySQL} that -only have table locking (multiple readers / single writers). This will -also give better performance with most databases, as the row locking -manager in this case will have less to do. +@strong{Q}: How can I use replication to provide redundancy/high +availability? + +@strong{A}: With the currently available features, you would have to +set up a master and a slave (or several slaves), and write a script +that will monitor the +master to see if it is up, and instruct your applications and +the slaves of the master change in case of failure. Some suggestions: + +@itemize @bullet @item -If you need to collect statistics from big log tables, use summary tables -instead of scanning the whole table. Maintaining the summaries should be -much faster than trying to do statistics 'live'. It's much faster to -regenerate new summary tables from the logs when things change -(depending on business decisions) than to have to change the running -application! +To tell a slave to change the master use the @code{CHANGE MASTER TO} command. @item -If possible, one should classify reports as 'live' or 'statistical', -where data needed for statistical reports are only generated based on -summary tables that are generated from the actual data. +A good way to keep your applications informed where the master is by +having a dynamic DNS entry for the master. With @strong{bind} you can +use @code{nsupdate} to dynamically update your DNS. @item -Take advantage of the fact that columns have default values. Insert -values explicitly only when the value to be inserted differs from the -default. This reduces the parsing that @strong{MySQL} need to do and -improves the insert speed. +You should run your slaves with the @code{log-bin} option and without +@code{log-slave-updates}. This way the slave will be ready to become a +master as soon as you issue @code{STOP SLAVE}; @code{RESET MASTER}, and +@code{CHANGE MASTER TO} on the other slaves. It will also help you catch +spurious updates that may happen because of misconfiguration of the +slave (ideally, you want to configure access rights so that no client +can update the slave, except for the slave thread) combined with the +bugs in your client programs (they should never update the slave +directly). + +@end itemize + +We are currently working on integrating an automatic master election +system into @strong{MySQL}, but until it is ready, you will have to +create your own monitoring tools. + +@node Replication Problems, , Replication FAQ, Replication +@section Troubleshooting Replication + +If you have followed the instructions, and your replication setup is not +working, first eliminate the user error factor by checking the following: + +@itemize @bullet @item -In some cases it's convenient to pack and store data into a blob. In this -case you have to add some extra code in your application to pack/unpack -things in the blob, but this may save a lot of accesses at some stage. -This is practical when you have data that doesn't conform to a static -table structure. +Is the master logging to the binary log? Check with @code{SHOW MASTER STATUS}. +If it is, @code{Position} will be non-zero. If not, verify that you have +given the master @code{log-bin} option and have set @code{server-id}. @item -Normally you should try to keep all data non-redundant (what -is called 3rd normal form in database theory), but you should not be -afraid of duplicating things or creating summary tables if you need these -to gain more speed. +Is the slave running? Check with @code{SHOW SLAVE STATUS}. The answer is found +in @code{Slave_running} column. If not, verify slave options and check the +error log for messages. @item -Stored procedures or UDF (user-defined functions) may be a good way to -get more performance. In this case you should, however, always have a way -to do this some other (slower) way if you use some database that doesn't -support this. +If the slave is running, did it establish connection with the master? Do +@code{SHOW PROCESSLIST}, find the thread with @code{system user} value in +@code{User} column and @code{none} in the @code{Host} column, and check the +@code{State} column. If it says @code{connecting to master}, verify the +privileges for the replication user on the master, master host name, your +DNS setup, whether the master is actually running, whether it is reachable +from the slave, and if all that seems ok, read the error logs. @item -You can always gain something by caching queries/answers in your -application and trying to do many inserts/updates at the same time. If -your database supports lock tables (like @strong{MySQL} and Oracle), -this should help to ensure that the index cache is only flushed once -after all updates. +If the slave was running, but then stopped, look at SHOW SLAVE STATUS +output and check the error logs. It usually +happens when some query that succeeded on the master fails on the slave. This +should never happen if you have taken a proper snapshot of the master, and +never modify the data on the slave outside of the slave thread. If it does, +it is a bug, read below on how to report it. @item -Use @code{INSERT /*! DELAYED */} when you do not need to know when your -data is written. This speeds things up because many records can be written -with a single disk write. +If a query on that succeeded on the master refuses to run on the slave, and +a full database resync ( the proper thing to do ) does not seem feasible, +try the following: +@itemize bullet @item -Use @code{INSERT /*! LOW_PRIORITY */} when you want your selects to be -more important. +First see if there is some stray record in the way. Understand how it got +there, then delete it and run @code{SLAVE START} @item -Use @code{SELECT /*! HIGH_PRIORITY */} to get selects that jump the -queue. That is, the select is done even if there is somebody waiting to -do a write. +If the above does not work or does not apply, try to understand if it would +be safe to make the update manually ( if needed) and then ignore the next +query from the master. @item -Use the multi-line @code{INSERT} statement to store many rows with one -SQL command (many SQL servers supports this). +If you have decided you can skip the next query, do +@code{SET SQL_SLAVE_SKIP_COUNTER=1; SLAVE START;} to skip a query that +does not use auto_increment, or last_insert_id or +@code{SET SQL_SLAVE_SKIP_COUNTER=2; SLAVE START;} otherwise. The reason +auto_increment/last_insert_id queries are different is that they take +two events in the binary log of the master. + @item -Use @code{LOAD DATA INFILE} to load bigger amounts of data. This is -faster than normal inserts and will be even faster when @code{myisamchk} -is integrated in @code{mysqld}. +If you are sure the slave started out perfectly in sync with the master, +and no one has updated the tables involved outside of slave thread, +report the bug, so +you will not have to do the above tricks again. +@end itemize @item -Use @code{AUTO_INCREMENT} columns to make unique values. +Make sure you are not running into an old bug by upgrading to the most recent +version. @item -Use @code{OPTIMIZE TABLE} once in a while to avoid fragmentation when -using dynamic table format. @xref{OPTIMIZE TABLE, , @code{OPTIMIZE TABLE}}. +If all else fails, read the error logs. If they are big, +@code{grep -i slave /path/to/your-log.err} on the slave. There is no +generic pattern to search for on the master, as the only errors it logs +are general system errors - if it can, it will send the error to the slave +when things go wrong. +@end itemize + +When you have determined that there is no user error involved, and replication +still either does not work at all or is unstable, it is time to start working +on a bug report. We need to get as much info as possible from you to be able +to track down the bug. Please do spend some time and effort preparing a good +bug report. Ideally, we would like to have a test case in the format found in +@code{mysql-test/t/rpl*} directory of the source tree. If you submit a test +case like that, you can expect a patch within a day or two in most cases, +although, of course, you mileage may vary depending on a number of factors. + +Second best option is a just program with easily configurable connection +arguments for the master and the slave that will demonstrate the problem on our +systems. You can write one in Perl or in C, depending on which language you +know better. +If you have one of the above ways to demonstrate the bug, use +@code{mysqlbug} to prepare a bug report and send it to +@email{bugs@@lists.mysql.com}. If you have a phantom - a problem that +does occur but you cannot duplicate "at will": + +@itemize @bullet @item -Use @code{HEAP} tables to get more speed when possible. @xref{Table -types}. +Verify that there is no user error involved. For example, if you update the +slave outside of the slave thread, the data will be out of sync, and you can +have unique key violations on updates, in which case the slave thread will +stop and wait for you to clean up the tables manually to bring them in sync. @item -When using a normal Web server setup, images should be stored as -files. That is, store only a file reference in the database. The main -reason for this is that a normal Web server is much better at caching -files than database contents. So it it's much easier to get a fast -system if you are using files. +Run slave with @code{log-slave-updates} and @code{log-bin} - this will keep +a log of all updates on the slave. @item -Use in memory tables for non-critical data that are accessed often (like -information about the last shown banner for users that don't have -cookies). +Save all evidence before resetting the replication. If we have no or only +sketchy information, it would take us a while to track down the problem. The +evidence you should collect is: +@itemize @bullet @item -Columns with identical information in different tables should be -declared identical and have identical names. Before Version 3.23 you -got slow joins otherwise. - -Try to keep the names simple (use @code{name} instead of -@code{customer_name} in the customer table). To make your names portable -to other SQL servers you should keep them shorter than 18 characters. +All binary logs on the master @item -If you need REALLY high speed, you should take a look at the low-level -interfaces for data storage that the different SQL servers support! For -example, by accessing the @strong{MySQL} @code{MyISAM} directly, you could -get a speed increase of 2-5 times compared to using the SQL interface. -To be able to do this the data must be on the same server as -the application, and usually it should only be accessed by one process -(because external file locking is really slow). One could eliminate the -above problems by introducing low-level @code{MyISAM} commands in the -@strong{MySQL} server (this could be one easy way to get more -performance if needed). By carefully designing the database interface, -it should be quite easy to support this types of optimization. +All binary log on the slave @item -In many cases it's faster to access data from a database (using a live -connection) than accessing a text file, just because the database is -likely to be more compact than the text file (if you are using numerical -data), and this will involve fewer disk accesses. You will also save -code because you don't have to parse your text files to find line and -column boundaries. +The output of @code{SHOW MASTER STATUS} on the master at the time +you have discovered the problem @item -You can also use replication to speed things up. @xref{Replication}. +The output of @code{SHOW SLAVE STATUS} on the master at the time +you have discovered the problem @item -Declaring a table with @code{DELAY_KEY_WRITE=1} will make the updating of -indexes faster, as these are not logged to disk until the file is closed. -The downside is that you should run @code{myisamchk} on these tables before -you start @code{mysqld} to ensure that they are okay if something killed -@code{mysqld} in the middle. As the key information can always be generated -from the data, you should not lose anything by using @code{DELAY_KEY_WRITE}. +Error logs on the master and on the slave +@end itemize +@item +Use @code{mysqlbinlog} to examine the binary logs. The following should +be helpful +to find the trouble query, for example: +@example +mysqlbinlog -j pos_from_slave_status /path/to/log_from_slave_status | head +@end example @end itemize -@cindex benchmarks -@cindex performance, benchmarks -@node Benchmarks, Design, Tips, Performance -@section Using Your Own Benchmarks +Once you have collected the evidence on the phantom problem, try hard to +isolate it into a separate test case first. Then report the problem to +@email{bugs@@lists.mysql.com} with as much info as possible. -You should definitely benchmark your application and database to find -out where the bottlenecks are. By fixing it (or by replacing the -bottleneck with a 'dummy module') you can then easily identify the next -bottleneck (and so on). Even if the overall performance for your -application is sufficient, you should at least make a plan for each -bottleneck, and decide how to solve it if someday you really need the -extra performance. +@cindex searching, full-text +@cindex full-text search +@cindex FULLTEXT +@node Fulltext Search, Performance, Replication, Top +@chapter MySQL Full-text Search -For an example of portable benchmark programs, look at the @strong{MySQL} -benchmark suite. @xref{MySQL Benchmarks, , @strong{MySQL} Benchmarks}. You -can take any program from this suite and modify it for your needs. By doing this, -you can try different solutions to your problem and test which is really the -fastest solution for you. +Since Version 3.23.23, @strong{MySQL} has support for full-text indexing +and searching. Full-text indexes in @strong{MySQL} are an index of type +@code{FULLTEXT}. @code{FULLTEXT} indexes can be created from @code{VARCHAR} +and @code{TEXT} columns at @code{CREATE TABLE} time or added later with +@code{ALTER TABLE} or @code{CREATE INDEX}. For large datasets, adding +@code{FULLTEXT} index with @code{ALTER TABLE} (or @code{CREATE INDEX}) would +be much faster than inserting rows into the empty table with a @code{FULLTEXT} +index. -It is very common that some problems only occur when the system is very -heavily loaded. We have had many customers who contact us when they -have a (tested) system in production and have encountered load problems. In -every one of these cases so far, it has been problems with basic design -(table scans are NOT good at high load) or OS/Library issues. Most of -this would be a @strong{LOT} easier to fix if the systems were not -already in production. +Full-text search is performed with the @code{MATCH} function. -To avoid problems like this, you should put some effort into benchmarking -your whole application under the worst possible load! You can use Sasha's -recent hack for this - -@uref{http://www.mysql.com/Downloads/super-smack/super-smack-1.0.tar.gz, -super-smack}. -As the name suggests, it can bring your system down to its knees if you ask it, -so make sure to use it only on your development systems. +@example +mysql> CREATE TABLE articles ( + -> id INT UNSIGNED AUTO_INCREMENT NOT NULL PRIMARY KEY, + -> title VARCHAR(200), + -> body TEXT, + -> FULLTEXT (title,body) + -> ); +Query OK, 0 rows affected (0.00 sec) -@cindex design, choices -@cindex database design -@cindex storage of data -@node Design, Design Limitations, Benchmarks, Performance -@section Design Choices +mysql> INSERT INTO articles VALUES + -> (0,'MySQL Tutorial', 'DBMS stands for DataBase Management ...'), + -> (0,'How To Use MySQL Efficiently', 'After you went through a ...'), + -> (0,'Optimizing MySQL','In this tutorial we will show how to ...'), + -> (0,'1001 MySQL Trick','1. Never run mysqld as root. 2. Normalize ...'), + -> (0,'MySQL vs. YourSQL', 'In the following database comparison we ...'), + -> (0,'MySQL Security', 'When configured properly, MySQL could be ...'); +Query OK, 5 rows affected (0.00 sec) +Records: 5 Duplicates: 0 Warnings: 0 -@strong{MySQL} keeps row data and index data in separate files. Many (almost -all) other databases mix row and index data in the same file. We believe that -the @strong{MySQL} choice is better for a very wide range of modern systems. +mysql> SELECT * FROM articles WHERE MATCH (title,body) AGAINST ('database'); ++----+-------------------+---------------------------------------------+ +| id | title | body | ++----+-------------------+---------------------------------------------+ +| 5 | MySQL vs. YourSQL | In the following database comparison we ... | +| 1 | MySQL Tutorial | DBMS stands for DataBase Management ... | ++----+-------------------+---------------------------------------------+ +2 rows in set (0.00 sec) +@end example -Another way to store the row data is to keep the information for each -column in a separate area (examples are SDBM and Focus). This will cause a -performance hit for every query that accesses more than one column. Because -this degenerates so quickly when more than one column is accessed, -we believe that this model is not good for general purpose databases. +The function @code{MATCH} matches a natural language query @code{AGAINST} +a text collection (which is simply the set of columns covered by a +@code{FULLTEXT} index). For every row in a table it returns relevance - +a similarity measure between the text in that row (in the columns that are +part of the collection) and the query. When it is used in a @code{WHERE} +clause (see example above) the rows returned are automatically sorted with +relevance decreasing. Relevance is a non-negative floating-point number. +Zero relevance means no similarity. Relevance is computed based on the +number of words in the row, the number of unique words in that row, the +total number of words in the collection, and the number of documents (rows) +that contain a particular word. -The more common case is that the index and data are stored together -(like in Oracle/Sybase et al). In this case you will find the row -information at the leaf page of the index. The good thing with this -layout is that it, in many cases, depending on how well the index is -cached, saves a disk read. The bad things with this layout are: +The above is a basic example of using @code{MATCH} function. Rows are +returned with relevance decreasing. -@itemize @bullet -@item -Table scanning is much slower because you have to read through the indexes -to get at the data. -@item -You can't use only the index table to retrieve data for a query. -@item -You lose a lot of space, as you must duplicate indexes from the nodes -(as you can't store the row in the nodes). -@item -Deletes will degenerate the table over time (as indexes in nodes are -usually not updated on delete). -@item -It's harder to cache ONLY the index data. -@end itemize - -@cindex design, limitations -@node Design Limitations, Portability, Design, Performance -@section MySQL Design Limitations/Tradeoffs - -Because @strong{MySQL} uses extremely fast table locking (multiple readers / -single writers) the biggest remaining problem is a mix of a steady stream of -inserts and slow selects on the same table. +@example +mysql> SELECT id,MATCH (title,body) AGAINST ('Tutorial') FROM articles; ++----+-----------------------------------------+ +| id | MATCH (title,body) AGAINST ('Tutorial') | ++----+-----------------------------------------+ +| 1 | 0.64840710366884 | +| 2 | 0 | +| 3 | 0.66266459031789 | +| 4 | 0 | +| 5 | 0 | +| 6 | 0 | ++----+-----------------------------------------+ +5 rows in set (0.00 sec) +@end example -We believe that for a huge number of systems the extremely fast -performance in other cases make this choice a win. This case is usually -also possible to solve by having multiple copies of the table, but it -takes more effort and hardware. +This example shows how to retrieve the relevances. As neither @code{WHERE} +nor @code{ORDER BY} clauses are present, returned rows are not ordered. -We are also working on some extensions to solve this problem for some -common application niches. +@example +mysql> SELECT id, body, MATCH (title,body) AGAINST ( + -> 'Security implications of running MySQL as root') AS score + -> FROM articles WHERE MATCH (title,body) AGAINST + -> ('Security implications of running MySQL as root'); ++----+-----------------------------------------------+-----------------+ +| id | body | score | ++----+-----------------------------------------------+-----------------+ +| 4 | 1. Never run mysqld as root. 2. Normalize ... | 1.5055546709332 | +| 6 | When configured properly, MySQL could be ... | 1.31140957288 | ++----+-----------------------------------------------+-----------------+ +2 rows in set (0.00 sec) +@end example -@cindex portability -@cindex crash-me program -@cindex programs, crash-me -@node Portability, Internal use, Design Limitations, Performance -@section Portability +This is more complex example - the query returns the relevance and still +sorts the rows with relevance decreasing. To achieve it one should specify +@code{MATCH} twice. Note, that this will cause no additional overhead, as +@strong{MySQL} optimizer will notice that these two @code{MATCH} calls are +identical and will call full-text search code only once. -Because all SQL servers implement different parts of SQL, it takes work to -write portable SQL applications. For very simple selects/inserts it is -very easy, but the more you need the harder it gets. If you want an -application that is fast with many databases it becomes even harder! +@strong{MySQL} uses a very simple parser to split text into words. A +``word'' is any sequence of letters, numbers, @samp{'}, and @samp{_}. Any +``word'' that is present in the stopword list or just too short (3 +characters or less) is ignored. -To make a complex application portable you need to choose a number of -SQL servers that it should work with. +Every correct word in the collection and in the query is weighted, +according to its significance in the query or collection. This way, a +word that is present in many documents will have lower weight (and may +even have a zero weight), because it has lower semantic value in this +particular collection. Otherwise, if the word is rare, it will receive a +higher weight. The weights of the words are then combined to compute the +relevance of the row. -You can use the @strong{MySQL} crash-me program/web-page -@uref{http://www.mysql.com/information/crash-me.php} to find functions, -types, and limits you can use with a selection of database -servers. Crash-me now tests far from everything possible, but it -is still comprehensive with about 450 things tested. +Such a technique works best with large collections (in fact, it was +carefully tuned this way). For very small tables, word distribution +does not reflect adequately their semantical value, and this model +may sometimes produce bizarre results. -For example, you shouldn't have column names longer than 18 characters -if you want to be able to use Informix or DB2. +@example +mysql> SELECT * FROM articles WHERE MATCH (title,body) AGAINST ('MySQL'); +Empty set (0.00 sec) +@end example -Both the @strong{MySQL} benchmarks and crash-me programs are very -database-independent. By taking a look at how we have handled this, you -can get a feeling for what you have to do to write your application -database-independent. The benchmarks themselves can be found in the -@file{sql-bench} directory in the @strong{MySQL} source -distribution. They are written in Perl with DBI database interface -(which solves the access part of the problem). +Search for the word @code{MySQL} produces no results in the above example. +Word @code{MySQL} is present in more than half of rows, and as such, is +effectively treated as a stopword (that is, with semantical value zero). +It is, really, the desired behavior - a natural language query should not +return every second row in 1GB table. -See @uref{http://www.mysql.com/information/benchmarks.html} for the results -from this benchmark. +A word that matches half of rows in a table is less likely to locate relevant +documents. In fact, it will most likely find plenty of irrelevant documents. +We all know this happens far too often when we are trying to find something on +the Internet with a search engine. It is with this reasoning that such rows +have been assigned a low semantical value in @strong{this particular dataset}. -As you can see in these results, all databases have some weak points. That -is, they have different design compromises that lead to different -behavior. +@menu +* Fulltext Fine-tuning:: +* Fulltext Features to Appear in MySQL 4.0:: +* Fulltext TODO:: +@end menu -If you strive for database independence, you need to get a good feeling -for each SQL server's bottlenecks. @strong{MySQL} is VERY fast in -retrieving and updating things, but will have a problem in mixing slow -readers/writers on the same table. Oracle, on the other hand, has a big -problem when you try to access rows that you have recently updated -(until they are flushed to disk). Transaction databases in general are -not very good at generating summary tables from log tables, as in this -case row locking is almost useless. +@node Fulltext Fine-tuning, Fulltext Features to Appear in MySQL 4.0, Fulltext Search, Fulltext Search +@section Fine-tuning MySQL Full-text Search -To get your application @emph{really} database-independent, you need to define -an easy extendable interface through which you manipulate your data. As -C++ is available on most systems, it makes sense to use a C++ classes -interface to the databases. +Unfortunately, full-text search has no user-tunable parameters yet, +although adding some is very high on the TODO. However, if you have a +@strong{MySQL} source distribution (@xref{Installing source}.), you can +somewhat alter the full-text search behavior. -If you use some specific feature for some database (like the -@code{REPLACE} command in @strong{MySQL}), you should code a method for -the other SQL servers to implement the same feature (but slower). With -@strong{MySQL} you can use the @code{/*! */} syntax to add -@strong{MySQL}-specific keywords to a query. The code inside -@code{/**/} will be treated as a comment (ignored) by most other SQL -servers. +Note that full-text search was carefully tuned for the best searching +effectiveness. Modifying the default behavior will, in most cases, +only make the search results worse. Do not alter the @strong{MySQL} sources +unless you know what you are doing! -If REAL high performance is more important than exactness, as in some -Web applications, a possibility is to create an application layer that -caches all results to give you even higher performance. By letting -old results 'expire' after a while, you can keep the cache reasonably -fresh. This is quite nice in case of extremely high load, in which case -you can dynamically increase the cache and set the expire timeout higher -until things get back to normal. +@itemize -In this case the table creation information should contain information -of the initial size of the cache and how often the table should normally -be refreshed. +@item +Minimal length of word to be indexed is defined in +@code{myisam/ftdefs.h} file by the line +@example +#define MIN_WORD_LEN 4 +@end example +Change it to the value you prefer, recompile @strong{MySQL}, and rebuild +your @code{FULLTEXT} indexes. -@cindex uses, of MySQL -@cindex customers, of MySQL -@node Internal use, , Portability, Performance -@section What Have We Used MySQL For? +@item +The stopword list is defined in @code{myisam/ft_static.c} +Modify it to your taste, recompile @strong{MySQL} and rebuild +your @code{FULLTEXT} indexes. -During @strong{MySQL} initial development, the features of @strong{MySQL} were made to fit -our largest customer. They handle data warehousing for a couple of the -biggest retailers in Sweden. +@item +The 50% threshold is caused by the particular weighting scheme chosen. To +disable it, change the following line in @code{myisam/ftdefs.h}: +@example +#define GWS_IN_USE GWS_PROB +@end example +to +@example +#define GWS_IN_USE GWS_FREQ +@end example +and recompile @strong{MySQL}. +There is no need to rebuild the indexes in this case. -From all stores, we get weekly summaries of all bonus card transactions, -and we are expected to provide useful information for the store owners -to help them find how their advertisement campaigns are affecting their -customers. +@end itemize -The data is quite huge (about 7 million summary transactions per month), -and we have data for 4-10 years that we need to present to the users. -We got weekly requests from the customers that they want to get -'instant' access to new reports from this data. +@node Fulltext Features to Appear in MySQL 4.0, Fulltext TODO, Fulltext Fine-tuning, Fulltext Search +@section New Features of Full-text Search to Appear in MySQL 4.0 -We solved this by storing all information per month in compressed -'transaction' tables. We have a set of simple macros (script) that -generates summary tables grouped by different criteria (product group, -customer id, store ...) from the transaction tables. The reports are -Web pages that are dynamically generated by a small Perl script that -parses a Web page, executes the SQL statements in it, and inserts the -results. We would have used PHP or mod_perl instead but they were -not available at that time. +This section includes a list of the fulltext features that are already +implemented in the 4.0 tree. It explains +@strong{More functions for full-text search} entry of @ref{TODO MySQL 4.0}. -For graphical data we wrote a simple tool in @code{C} that can produce -GIFs based on the result of a SQL query (with some processing of the -result). This is also dynamically executed from the Perl script that -parses the @code{HTML} files. +@itemize @bullet +@item @code{REPAIR TABLE} with @code{FULLTEXT} indexes, +@code{ALTER TABLE} with @code{FULLTEXT} indexes, and +@code{OPTIMIZE TABLE} with @code{FULLTEXT} indexes are now +up to 100 times faster. -In most cases a new report can simply be done by copying an existing -script and modifying the SQL query in it. In some cases, we will need to -add more fields to an existing summary table or generate a new one, but -this is also quite simple, as we keep all transactions tables on disk. -(Currently we have at least 50G of transactions tables and 200G of other -customer data.) +@item @code{MATCH ... AGAINST} now supports the following +@strong{boolean operators}: -We also let our customers access the summary tables directly with ODBC -so that the advanced users can themselves experiment with the data. +@itemize @bullet +@item @code{+}word means the that word @strong{must} be present in every +row returned. +@item @code{-}word means the that word @strong{must not} be present in every +row returned. +@item @code{<} and @code{>} can be used to decrease and increase word +weight in the query. +@item @code{~} can be used to assign a @strong{negative} weight to a noise +word. +@item @code{*} is a truncation operator. +@end itemize -We haven't had any problems handling this with quite modest Sun Ultra -SPARCstation (2x200 Mhz). We recently upgraded one of our servers to a 2 -CPU 400 Mhz UltraSPARC, and we are now planning to start handling -transactions on the product level, which would mean a ten-fold increase -of data. We think we can keep up with this by just adding more disk to -our systems. +Boolean search utilizes a more simplistic way of calculating the relevance, +that does not have a 50% threshold. -We are also experimenting with Intel-Linux to be able to get more CPU -power cheaper. Now that we have the binary portable database format (new -in Version 3.23), we will start to use this for some parts of the application. +@item Searches are now up to 2 times faster due to optimized search algorithm. -Our initial feelings are that Linux will perform much better on -low-to-medium load and Solaris will perform better when you start to get a -high load because of extreme disk IO, but we don't yet have anything -conclusive about this. After some discussion with a Linux Kernel -developer, this might be a side effect of Linux giving so much resources -to the batch job that the interactive performance gets very low. This -makes the machine feel very slow and unresponsive while big batches are -going. Hopefully this will be better handled in future Linux Kernels. +@item Utility program @code{ft_dump} added for low-level @code{FULLTEXT} +index operations (querying/dumping/statistics). -@cindex benchmark suite -@cindex crash-me program -@node MySQL Benchmarks, Tools, Performance, Top -@chapter The MySQL Benchmark Suite +@end itemize -This should contain a technical description of the @strong{MySQL} -benchmark suite (and @code{crash-me}), but that description is not -written yet. Currently, you can get a good idea of the benchmark by -looking at the code and results in the @file{sql-bench} directory in any -@strong{MySQL} source distributions. +@node Fulltext TODO, , Fulltext Features to Appear in MySQL 4.0, Fulltext Search +@section Full-text Search TODO -This benchmark suite is meant to be a benchmark that will tell any user -what things a given SQL implementation performs well or poorly at. +@itemize @bullet +@item Make all operations with @code{FULLTEXT} index @strong{faster}. +@item Support for braces @code{()} in boolean full-text search. +@item Phrase search, proximity operators +@item Boolean search can work without @code{FULLTEXT} index +(yes, @strong{very} slow). +@item Support for "always-index words". They could be any strings +the user wants to treat as words, examples are "C++", "AS/400", "TCP/IP", etc. +@item Support for full-text search in @code{MERGE} tables. +@item Support for multi-byte charsets. +@item Make stopword list to depend of the language of the data. +@item Stemming (dependent of the language of the data, of course). +@item Generic user-supplyable UDF (?) preparser. +@item Make the model more flexible (by adding some adjustable +parameters to @code{FULLTEXT} in @code{CREATE/ALTER TABLE}). +@end itemize -Note that this benchmark is single threaded, so it measures the minimum -time for the operations. We plan to in the future add a lot of -multi-threaded tests to the benchmark suite. +@cindex performance, maximizing +@cindex optimization +@node Performance, MySQL Benchmarks, Fulltext Search, Top +@chapter Getting Maximum Performance from MySQL -For example, (run on the same NT 4.0 machine): +Optimization is a complicated task because it ultimately requires +understanding of the whole system. While it may be possible to do some +local optimizations with small knowledge of your system/application, the +more optimal you want your system to become the more you will have to +know about it. -@multitable @columnfractions .6 .2 .2 -@strong{Reading 2000000 rows by index} @tab @strong{Seconds} @tab @strong{Seconds} -@item mysql @tab 367 @tab 249 -@item mysql_odbc @tab 464 -@item db2_odbc @tab 1206 -@item informix_odbc @tab 121126 -@item ms-sql_odbc @tab 1634 -@item oracle_odbc @tab 20800 -@item solid_odbc @tab 877 -@item sybase_odbc @tab 17614 -@end multitable +So this chapter will try to explain and give some examples of different +ways to optimize @strong{MySQL}. But remember that there are always some +(increasingly harder) additional ways to make the system even faster. -@multitable @columnfractions .6 .2 .2 -@strong{Inserting (350768) rows} @tab @strong{Seconds} @tab @strong{Seconds} -@item mysql @tab 381 @tab 206 -@item mysql_odbc @tab 619 -@item db2_odbc @tab 3460 -@item informix_odbc @tab 2692 -@item ms-sql_odbc @tab 4012 -@item oracle_odbc @tab 11291 -@item solid_odbc @tab 1801 -@item sybase_odbc @tab 4802 -@end multitable +@menu +* Optimize Basics:: Optimization overview +* System:: System/Compile time and startup parameter tuning +* Data size:: Get your data as small as possible +* MySQL indexes:: How @strong{MySQL} uses indexes +* Query Speed:: Speed of queries that access or update data +* Tips:: Other optimization tips +* Benchmarks:: Using your own benchmarks +* Design:: Design choices +* Design Limitations:: MySQL design limitations/tradeoffs +* Portability:: Portability +* Internal use:: What have we used MySQL for? +@end menu -In the above test @strong{MySQL} was run with a 8M index cache. +@node Optimize Basics, System, Performance, Performance +@section Optimization Overview -We have gather some more benchmark results at -@uref{http://www.mysql.com/information/benchmarks.html}. +The most important part for getting a system fast is of course the basic +design. You also need to know what kinds of things your system will be +doing, and what your bottlenecks are. -Note that Oracle is not included because they asked to be removed. All -Oracle benchmarks have to be passed by Oracle! We believe that makes -Oracle benchmarks @strong{VERY} biased because the above benchmarks are -supposed to show what a standard installation can do for a single -client. +The most common bottlenecks are: +@itemize @bullet +@item Disk seeks. +It takes time for the disk to find a piece of data. With modern disks in +1999, the mean time for this is usually lower than 10ms, so we can in +theory do about 1000 seeks a second. This time improves slowly with new +disks and is very hard to optimize for a single table. The way to +optimize this is to spread the data on more than one disk. +@item Disk reading/writing. +When the disk is at the correct position we need to read the data. With +modern disks in 1999, one disk delivers something like 10-20Mb/s. This +is easier to optimize than seeks because you can read in parallel from +multiple disks. +@item CPU cycles. +When we have the data in main memory (or if it already were +there) we need to process it to get to our result. Having small +tables compared to the memory is the most common limiting +factor. But then, with small tables speed is usually not the problem. +@item Memory bandwidth. +When the CPU needs more data than can fit in the CPU cache the main +memory bandwidth becomes a bottleneck. This is an uncommon bottleneck +for most systems, but one should be aware of it. +@end itemize -To run the benchmark suite, you have to download a @strong{MySQL} source -distribution, install the perl DBI driver, the perl DBD driver for the -database you want to test and then do: +@cindex compiling, optimizing +@cindex system optimization +@cindex startup parameters, tuning +@node System, Data size, Optimize Basics, Performance +@section System/Compile Time and Startup Parameter Tuning -@example -cd sql-bench -perl run-all-tests --server=# -@end example +We start with the system level things since some of these decisions have +to be made very early. In other cases a fast look at this part may +suffice because it not that important for the big gains. However, it is always +nice to have a feeling about how much one could gain by changing things +at this level. -where # is one of supported servers. You can get a list of all options -and supported servers by doing @code{run-all-tests --help}. +The default OS to use is really important! To get the most use of +multiple CPU machines one should use Solaris (because the threads works +really nice) or Linux (because the 2.2 kernel has really good SMP +support). Also on 32-bit machines Linux has a 2G file size limit by +default. Hopefully this will be fixed soon when new filesystems are +released (XFS/Reiserfs). If you have a desperate need for files bigger +than 2G on Linux-intel 32 bit, you should get the LFS patch for the ext2 +file system. -@cindex crash-me -@code{crash-me} tries to determine what features a database supports and -what its capabilities and limitations are by actually running -queries. For example, it determines: +Because we have not run @strong{MySQL} in production on that many platforms, we +advice you to test your intended platform before choosing it, if possible. +@cindex locking +Other tips: @itemize @bullet @item -What column types are supported -@item -How many indexes are supported -@item -What functions are supported -@item -How big a query can be +If you have enough RAM, you could remove all swap devices. Some +operating systems will use a swap device in some contexts even if you +have free memory. @item -How big a @code{VARCHAR} column can be -@end itemize +Use the @code{--skip-locking} @strong{MySQL} option to avoid external +locking. Note that this will not impact @strong{MySQL}'s functionality as +long as you only run one server. Just remember to take down the server (or +lock relevant parts) before you run @code{myisamchk}. On some system +this switch is mandatory because the external locking does not work in any +case. -We can find the result from crash-me on a lot of different databases at -@uref{http://www.mysql.com/information/crash-me.php}. +The @code{--skip-locking} option is on by default when compiling with +MIT-pthreads, because @code{flock()} isn't fully supported by +MIT-pthreads on all platforms. It's also on default for Linux +as Linux file locking are not yet safe. -@cindex utilities -@node Tools, Maintenance, MySQL Benchmarks, Top -@chapter MySQL Utilites +The only case when you can't use @code{--skip-locking} is if you run +multiple @strong{MySQL} @emph{servers} (not clients) on the same data, +or run @code{myisamchk} on the table without first flushing and locking +the @code{mysqld} server tables first. + +You can still use @code{LOCK TABLES}/@code{UNLOCK TABLES} even if you +are using @code{--skip-locking} +@end itemize @menu -* Programs:: What do the executables do? -* mysqld-max:: mysqld-max, An extended mysqld server -* safe_mysqld:: safe_mysqld, the wrapper around mysqld -* mysqld_multi:: Program for managing multiple @strong{MySQL} servers -* mysql:: The command line tool -* mysqladmin:: Administering a @strong{MySQL} server -* mysqldump:: Dumping the structure and data from @strong{MySQL} databases and tables -* mysqlhotcopy:: Copying @strong{MySQL} Databases and Tables -* mysqlimport:: Importing data from text files -* perror:: Displaying error messages -* mysqlshow:: Showing databases, tables and columns -* myisampack:: The @strong{MySQL} compressed read-only table generator +* Compile and link options:: How compiling and linking affects the speed of MySQL +* Disk issues:: Disk issues +* Symbolic links:: Using Symbolic Links +* Server parameters:: Tuning server parameters +* Table cache:: How MySQL opens and closes tables +* Creating many tables:: Drawbacks of creating large numbers of tables in the same database +* Open tables:: Why so many open tables? +* Memory use:: How MySQL uses memory +* Internal locking:: How MySQL locks tables +* Table locking:: Table locking issues +* DNS:: @end menu -In this chapter you will learn about the @strong{MySQL} Utilities that -come in a given distribution. You will learn what each of them does, how -to use it, and what you should use it for. - -@cindex environment variables -@cindex programs, list of -@node Programs, mysqld-max, Tools, Tools -@section Overview of the Different MySQL Programs +@node Compile and link options, Disk issues, System, System +@subsection How Compiling and Linking Affects the Speed of MySQL -All @strong{MySQL} clients that communicate with the server using the -@code{mysqlclient} library use the following environment variables: +Most of the following tests are done on Linux with the +@strong{MySQL} benchmarks, but they should give some indication for +other operating systems and workloads. -@tindex MYSQL_UNIX_PORT environment variable -@tindex Environment variable, MYSQL_UNIX_PORT -@tindex MYSQL_TCP_PORT environment variable -@tindex Environment variable, MYSQL_TCP_PORT -@tindex MYSQL_PWD environment variable -@tindex Environment variable, MYSQL_PWD -@tindex MYSQL_DEBUG environment variable -@tindex Environment variable, MYSQL_DEBUG -@multitable @columnfractions .25 .75 -@item @strong{Name} @tab @strong{Description} -@item @code{MYSQL_UNIX_PORT} @tab The default socket; used for connections to @code{localhost} -@item @code{MYSQL_TCP_PORT} @tab The default TCP/IP port -@item @code{MYSQL_PWD} @tab The default password -@item @code{MYSQL_DEBUG} @tab Debug-trace options when debugging -@item @code{TMPDIR} @tab The directory where temporary tables/files are created -@end multitable +@cindex linking, speed +@cindex compiling, speed +@cindex speed, compiling +@cindex speed, linking -Use of @code{MYSQL_PWD} is insecure. -@xref{Connecting}. +You get the fastest executable when you link with @code{-static}. -@tindex MYSQL_HISTFILE environment variable -@tindex Environment variable, MYSQL_HISTFILE -@tindex HOME environment variable -@tindex Environment variable, HOME -@cindex history file -@cindex command line history -@tindex .mysql_history file -The @file{mysql} client uses the file named in the @code{MYSQL_HISTFILE} -environment variable to save the command-line history. The default value for -the history file is @file{$HOME/.mysql_history}, where @code{$HOME} is the -value of the @code{HOME} environment variable. @xref{Environment variables}. +On Linux, you will get the fastest code when compiling with @code{pgcc} +and @code{-O3}. To compile @file{sql_yacc.cc} with these options, you +need about 200M memory because @code{gcc/pgcc} needs a lot of memory to +make all functions inline. You should also set @code{CXX=gcc} when +configuring @strong{MySQL} to avoid inclusion of the @code{libstdc++} +library (it is not needed). Note that with some versions of @code{pgcc}, +the resulting code will only run on true Pentium processors, even if you +use the compiler option that you want the resulting code to be working on +all x586 type processors (like AMD). -All @strong{MySQL} programs take many different options. However, every -@strong{MySQL} program provides a @code{--help} option that you can use -to get a full description of the program's different options. For example, try -@code{mysql --help}. +By just using a better compiler and/or better compiler options you can +get a 10-30 % speed increase in your application. This is particularly +important if you compile the SQL server yourself! -You can override default options for all standard client programs with an -option file. @ref{Option files}. +We have tested both the Cygnus CodeFusion and Fujitsu compilers, but +when we tested them, neither was sufficiently bug free to allow +@strong{MySQL} to be compiled with optimizations on. -The list below briefly describes the @strong{MySQL} programs: +When you compile @strong{MySQL} you should only include support for the +character sets that you are going to use. (Option @code{--with-charset=xxx}). +The standard @strong{MySQL} binary distributions are compiled with support +for all character sets. -@table @code +Here is a list of some measurements that we have done: +@itemize @bullet +@item +If you use @code{pgcc} and compile everything with @code{-O6}, the +@code{mysqld} server is 1% faster than with @code{gcc} 2.95.2. -@cindex @code{myisamchk} -@item myisamchk -Utility to describe, check, optimize, and repair @strong{MySQL} tables. -Because @code{myisamchk} has many functions, it is described in its own -chapter. @xref{Maintenance}. +@item +If you link dynamically (without @code{-static}), the result is 13% +slower on Linux. Note that you still can use a dynamic linked +@strong{MySQL} library. It is only the server that is critical for +performance. -@cindex @code{make_binary_distribution} -@item make_binary_distribution -Makes a binary release of a compiled @strong{MySQL}. This could be sent -by FTP to @file{/pub/mysql/Incoming} on @code{support.mysql.com} for the -convenience of other @strong{MySQL} users. +@item +If you strip your @code{mysqld} binary with @code{strip libexec/mysqld}, +the resulting binary can be up to 4 % faster. -@cindex @code{msql2mysql} -@item msql2mysql -A shell script that converts @code{mSQL} programs to @strong{MySQL}. It doesn't -handle all cases, but it gives a good start when converting. +@item +If you connect using TCP/IP rather than Unix sockets, the result is 7.5% +slower on the same computer. (If you are connection to @code{localhost}, +@strong{MySQL} will, by default, use sockets). -@cindex @code{mysqlaccess} -@item mysqlaccess -A script that checks the access privileges for a host, user, and database -combination. +@item +If you connect using TCP/IP from another computer over a 100M Ethernet, +things will be 8-11 % slower. -@cindex @code{mysqladmin} -@item mysqladmin -Utility for performing administrative operations, such as creating or -dropping databases, reloading the grant tables, flushing tables to disk, and -reopening log files. @code{mysqladmin} can also be used to retrieve version, -process, and status information from the server. -@xref{mysqladmin, , @code{mysqladmin}}. +@item +If you compile with @code{--with-debug=full}, then you will loose 20 % +for most queries, but some queries may take substantially longer (The +@strong{MySQL} benchmarks ran 35 % slower) +If you use @code{--with-debug}, then you will only loose 15 %. +By starting a @code{mysqld} version compiled with @code{--with-debug=full} +with @code{--skip-safemalloc} the end result should be close to when +configuring with @code{--with-debug}. -@cindex @code{mysqlbug} -@item mysqlbug -The @strong{MySQL} bug report script. This script should always be used when -filing a bug report to the @strong{MySQL} list. +@item +On a Sun SPARCstation 20, SunPro C++ 4.2 is 5 % faster than @code{gcc} 2.95.2. -@cindex @code{mysqld} -@item mysqld -The SQL daemon. This should always be running. +@item +Compiling with @code{gcc} 2.95.2 for ultrasparc with the option +@code{-mcpu=v8 -Wa,-xarch=v8plusa} gives 4 % more performance. -@cindex @code{mysqldump} -@item mysqldump -Dumps a @strong{MySQL} database into a file as SQL statements or -as tab-separated text files. Enhanced freeware originally by Igor Romanenko. -@xref{mysqldump, , @code{mysqldump}}. +@item +On Solaris 2.5.1, MIT-pthreads is 8-12% slower than Solaris native +threads on a single processor. With more load/CPUs the difference should +get bigger. -@cindex @code{mysqlimport} -@item mysqlimport -Imports text files into their respective tables using @code{LOAD DATA -INFILE}. @xref{mysqlimport, , @code{mysqlimport}}. +@item +Running with @code{--log-bin} makes @strong{[MySQL} 1 % slower. -@cindex @code{mysqlshow} -@item mysqlshow -Displays information about databases, tables, columns, and indexes. +@item +Compiling on Linux-x86 using gcc without frame pointers +@code{-fomit-frame-pointer} or @code{-fomit-frame-pointer -ffixed-ebp} +@code{mysqld} 1-4% faster. +@end itemize -@cindex @code{mysql_install_db} -@item mysql_install_db -Creates the @strong{MySQL} grant tables with default privileges. This is -usually executed only once, when first installing @strong{MySQL} -on a system. +The @strong{MySQL}-Linux distribution provided by @strong{MySQL AB} used +to be compiled with @code{pgcc}, but we had to go back to regular gcc +because of a bug in @code{pgcc} that would generate the code that does +not run on AMD. We will continue using gcc until that bug is resolved. +In the meantime, if you have a non-AMD machine, you can get a faster +binary by compiling with @code{pgcc}. The standard @strong{MySQL} +Linux binary is linked statically to get it faster and more portable. -@cindex @code{replace} -@item replace -A utility program that is used by @code{msql2mysql}, but that has more -general applicability as well. @code{replace} changes strings in place in -files or on the standard input. Uses a finite state machine to match longer -strings first. Can be used to swap strings. For example, this command -swaps @code{a} and @code{b} in the given files: +@cindex disk issues +@cindex performance, disk issues +@node Disk issues, Symbolic links, Compile and link options, System +@subsection Disk Issues + +@itemize @bullet +@item +As mentioned before, disks seeks are a big performance bottleneck. This +problems gets more and more apparent when the data starts to grow so +large that effective caching becomes impossible. For large databases, +where you access data more or less randomly, you can be sure that you +will need at least one disk seek to read and a couple of disk seeks to +write things. To minimize this problem, use disks with low seek times. +@item +Increase the number of available disk spindles (and thereby reduce +the seek overhead) by either symlink files to different disks or striping +the disks. +@table @strong +@item Using symbolic links +This means that you symlink the index and/or data file(s) from the +normal data directory to another disk (that may also be striped). This +makes both the seek and read times better (if the disks are not used for +other things). @xref{Symbolic links}. +@cindex striping, defined +@item Striping +Striping means that you have many disks and put the first block on the +first disk, the second block on the second disk, and the Nth on the +(N mod number_of_disks) disk, and so on. This means if your normal data +size is less than the stripe size (or perfectly aligned) you will get +much better performance. Note that striping is very dependent on the OS +and stripe-size. So benchmark your application with different +stripe-sizes. @xref{Benchmarks}. +Note that the speed difference for striping is @strong{very} dependent +on the parameters. Depending on how you set the striping parameters and +number of disks you may get a difference in orders of magnitude. Note that +you have to choose to optimize for random or sequential access. +@end table +@item +For reliability you may want to use RAID 0+1 (striping + mirroring), but +in this case you will need 2*N drives to hold N drives of data. This is +probably the best option if you have the money for it! You may, however, +also have to invest in some volume-management software to handle it +efficiently. +@item +A good option is to have semi-important data (that can be regenerated) +on RAID 0 disk while storing really important data (like host information +and logs) on a RAID 0+1 or RAID N disk. RAID N can be a problem if you +have many writes because of the time to update the parity bits. +@item +You may also set the parameters for the file system that the database +uses. One easy change is to mount the file system with the noatime +option. That makes it skip the updating of the last access time in the +inode and by this will avoid some disk seeks. +@item +On Linux, you can get much more performance (up to 100 % under load is +not uncommon) by using hdpram to configure your disk's interface! The +following should be quite good hdparm options for @strong{MySQL} (and +probably many other applications): @example -shell> replace a b b a -- file1 file2 ... +hdparm -m 16 -d 1 + @end example -@end table -@cindex @code{mysqld-max} -@node mysqld-max, safe_mysqld, Programs, Tools -@section mysqld-max, An extended mysqld server +Note that the performance/reliability when using the above depends on +your hardware, so we strongly suggest that you test your system +thoroughly after using @code{hdparm}! Please consult the @code{hdparm} +man page for more information! If @code{hdparm} is not used wisely, +filesystem corruption may result. Backup everything before experimenting! +@item +On many operating systems you can mount the disks with the 'async' flag to set the file +system to be updated asynchronously. If your computer is reasonable stable, +this should give you more performance without sacrificing too much reliability. +(This flag is on by default on Linux.) +@item +If you don't need to know when a file was last accessed (which is not +really useful on a database server), you can mount your file systems +with the noatime flag. +@end itemize -@code{mysqld-max} is the MySQL server (@code{mysqld}) configured with -the following configure options: +@cindex symbolic links +@cindex links, symbolic +@node Symbolic links, Server parameters, Disk issues, System +@subsection Using Symbolic Links -@multitable @columnfractions .3 .7 -@item @strong{Option} @tab @strong{Comment} -@item --with-server-suffix=-max @tab Add a suffix to the @code{mysqld} version string. -@item --with-bdb @tab Support for Berkeley DB (BDB) tables -@item --with-innodb @tab Support for InnoDB tables. -@item CFLAGS=-DUSE_SYMDIR @tab Symbolic links support for Windows. -@end multitable +You can move tables and databases from the database directory to other +locations and replace them with symbolic links to the new locations. +You might want to do this, for example, to move a database to a file +system with more free space or increase the speed of your system by +spreading your tables to different disk. -You can find the @strong{MySQL}-max binaries at -@uref{http://www.mysql.com/downloads/mysql-max-3.23.html}. +The recommended may to do this, is to just symlink databases to different +disk and only symlink tables as a last resort. +. -The Windows @strong{MySQL} 3.23 binary distribution includes both the -standard @strong{mysqld.exe} binary and the @code{mysqld-max.exe} binary. -@uref{http://www.mysql.com/downloads/mysql-3.23.html}. -@xref{Windows installation}. +@cindex databases, symbolic links +@menu +* Symbolic links to database:: +* Symbolic links to tables:: +@end menu -Note that as Berkeley DB and InnoDB are not available for all platforms, -some of the @code{Max} binaries may not have support for both of these. -You can check which table types are supported by doing the following -query: +@node Symbolic links to database, Symbolic links to tables, Symbolic links, Symbolic links +@subsubsection Using Symbolic Links for Databases + +The way to symlink a database is to first create a directory on some +disk where you have free space and then create a symlink to it from +the @strong{MySQL} database directory. @example -mysql> show variables like "have_%"; -+---------------+-------+ -| Variable_name | Value | -+---------------+-------+ -| have_bdb | YES | -| have_innodb | NO | -| have_isam | YES | -| have_raid | YES | -| have_ssl | NO | -+---------------+-------+ +shell> mkdir /dr1/databases/test +shell> ln -s /dr1/databases/test mysqld-datadir @end example -The meaning of the values are: - -@multitable @columnfractions .3 .7 -@item @strong{Value} @tab @strong{Meaning}. -@item YES @tab The option is activated and usable. -@item NO @tab @strong{MySQL} is not compiled with support for this option. -@item DISABLED @tab The xxxx option is disabled because one started @code{mysqld} with @code{--skip-xxxx} or because one didn't start @code{mysqld} with all needed options to enable the option. In this case the @code{hostname.err} file should contain a reason for why the option is disabled. -@end multitable - -@strong{NOTE}: To be able to create InnoDB tables you @strong{MUST} edit -your startup options to include at least the @code{innodb_data_file_path} -option. @xref{InnoDB start}. +@strong{MySQL} doesn't support that you link one directory to multiple +databases. Replacing a database directory with a symbolic link will +work fine as long as you don't make a symbolic link between databases. +Suppose you have a database @code{db1} under the @strong{MySQL} data +directory, and then make a symlink @code{db2} that points to @code{db1}: -To get better performance for BDB tables, you should add some configuration -options for these too. @xref{BDB start}. +@example +shell> cd /path/to/datadir +shell> ln -s db1 db2 +@end example -@code{safe_mysqld} will automatically try to start any @code{mysqld} binary -with the @code{-max} prefix. This makes it very easy to test out a -another @code{mysqld} binary in an existing installation. Just -run @code{configure} with the options you want and then install the -new @code{mysqld} binary as @code{mysqld-max} in the same directory -where your old @code{mysqld} binary is. @xref{safe_mysqld, , @code{safe_mysqld}}. +Now, for any table @code{tbl_a} in @code{db1}, there also appears to be +a table @code{tbl_a} in @code{db2}. If one thread updates @code{db1.tbl_a} +and another thread updates @code{db2.tbl_a}, there will be problems. -The @code{mysqld-max} RPM uses the above mentioned @code{safe_mysqld} -feature. It just installs the @code{mysqld-max} executable and -@code{safe_mysqld} will automatically use this executable when -@code{safe_mysqld} is restarted. +If you really need this, you must change the following code in +@file{mysys/mf_format.c}: -The following table shows which table types our standard @strong{MySQL-Max} -binaries includes: +@example +if (flag & 32 || (!lstat(to,&stat_buff) && S_ISLNK(stat_buff.st_mode))) +@end example -@multitable @columnfractions .4 .3 .3 -@item @strong{System} @tab @strong{BDB} @tab @strong{InnoDB} -@item AIX 4.3 @tab N @tab Y -@item HP-UX 11.0 @tab N @tab Y -@item Linux-Alpha @tab N @tab Y -@item Linux-Intel @tab Y @tab Y -@item Linux-Ia64 @tab N @tab Y -@item Solaris-intel @tab N @tab Y -@item Solaris-sparc @tab Y @tab Y -@item SCO OSR5 @tab Y @tab Y -@item UnixWare @tab Y @tab Y -@item Windows/NT @tab Y @tab Y -@end multitable +to -@cindex tools, safe_mysqld -@cindex scripts -@cindex @code{safe_mysqld} -@node safe_mysqld, mysqld_multi, mysqld-max, Tools -@section safe_mysqld, the wrapper around mysqld +@example +if (1) +@end example -@code{safe_mysqld} is the recommended way to start a @code{mysqld} -daemon on Unix. @code{safe_mysqld} adds some safety features such as -restarting the server when an error occurs and logging run-time -information to a log file. +On Windows you can use internal symbolic links to directories by compiling +@strong{MySQL} with @code{-DUSE_SYMDIR}. This allows you to put different +databases on different disks. @xref{Windows symbolic links}. -If you don't use @code{--mysqld=#} or @code{--mysqld-version=#} -@code{safe_mysqld} will use an executable named @code{mysqld-max} if it -exists. If not, @code{safe_mysqld} will start @code{mysqld}. -This makes it very easy to test to use @code{mysqld-max} instead of -@code{mysqld}; Just copy @code{mysqld-max} to where you have -@code{mysqld} and it will be used. +@cindex databases, symbolic links +@node Symbolic links to tables, , Symbolic links to database, Symbolic links +@subsubsection Using Symbolic Links for Tables -Normally one should never edit the @code{safe_mysqld} script, but -instead put the options to @code{safe_mysqld} in the -@code{[safe_mysqld]} section in the @code{my.cnf} -file. @code{safe_mysqld} will read all options from the @code{[mysqld]}, -@code{[server]} and @code{[safe_mysqld]} sections from the option files. -@xref{Option files}. +Before @strong{MySQL} 4.0 you should not symlink tables, if you are not +very carefully with them. The problem is that if you run @code{ALTER +TABLE}, @code{REPAIR TABLE} or @code{OPTIMIZE TABLE} on a symlinked +table, the symlinks will be removed and replaced by the original +files. This happens because the above command works by creating a +temporary file in the database directory and when the command is +complete, replace the original file with the temporary file. -Note that all options on the command line to @code{safe_mysqld} are passed -to @code{mysqld}. If you wants to use any options in @code{safe_mysqld} that -@code{mysqld} doesn't support, you must specify these in the option file. +You should not symlink tables on system that doesn't have a fully +working @code{realpath()} call. (At least Linux and Solaris support +@code{realpath()}) -Most of the options to @code{safe_mysqld} are the same as the options to -@code{mysqld}. @xref{Command-line options}. +In @strong{MySQL} 4.0 symlinks is only fully supported for @code{MyISAM} +tables. For other table types you will probably get strange problems +when doing any of the above mentioned commands. -@code{safe_mysqld} supports the following options: +The handling of symbolic links in @strong{MySQL} 4.0 works the following +way (this is mostly relevant only for @code{MyISAM} tables). -@table @code -@item --basedir=path -@item --core-file-size=# -Size of the core file @code{mysqld} should be able to create. Passed to @code{ulimit -c}. -@item --datadir=path -@item --defaults-extra-file=path -@item --defaults-file=path -@item --err-log=path -@item --ledir=path -Path to @code{mysqld} -@item --log=path -@item --mysqld=mysqld-version -Name of the @code{mysqld} version in the @code{ledir} directory you want to start. -@item --mysqld-version=version -Similar to @code{--mysqld=} but here you only give the suffix for @code{mysqld}. -For example if you use @code{--mysqld-version=max}, @code{safe_mysqld} will -start the @code{ledir/mysqld-max} version. If the argument to -@code{--mysqld-version} is empty, @code{ledir/mysqld} will be used. -@item --no-defaults -@item --open-files-limit=# -Number of files @code{mysqld} should be able to open. Passed to @code{ulimit -n}. Note that you need to start @code{safe_mysqld} as root for this to work properly! -@item --pid-file=path -@item --port=# -@item --socket=path -@item --timezone=# -Set the timezone (the @code{TZ}) variable to the value of this parameter. -@item --user=# -@end table +@itemize @bullet +@item +In the data directory you will always have the table definition file +and the data/index files. +@item +You can symlink the index file and the data file to different directories +independent of the other. +@item +The symlinking can be done from the operating system (if @code{mysqld} is +not running) or with the @code{INDEX/DATA DIRECTORY="path-to-dir"} command +in @code{CREATE TABLE}. @xref{CREATE TABLE}. +@item +@code{myisamchk} will not replace a symlink with the index/file but +work directly on the files the symlinks points to. Any temporary files +will be created in the same directory where the data/index file is. +@item +When you drop a table that is using symlinks, both the symlink and the +file the symlink points to is dropped. This is a good reason to why you +should NOT run @code{mysqld} as root and not allow persons to have write +access to the @strong{MySQL} database directories. +@item +If you rename a table with @code{ALTER TABLE RENAME} and you don't change +database, the symlink in the database directory will be renamed to the new +name and the data/index file will be renamed accordingly. +@item +If you use @code{ALTER TABLE RENAME} to move a table to another database, +then the table will be moved to the other database directory and the old +symlinks and the files they pointed to will be deleted. +@item +If you are not using symlinks you should use the @code{--skip-symlink} +option to @code{mysqld} to ensure that no one can drop or rename a file +outside of the @code{mysqld} data directory. +@end itemize -The @code{safe_mysqld} script is written so that it normally is able to start -a server that was installed from either a source or a binary version of -@strong{MySQL}, even if these install the server in slightly different -locations. @code{safe_mysqld} expects one of these conditions to be true: +Things that are not yet supported: +@cindex TODO, symlinks @itemize @bullet @item -The server and databases can be found relative to the directory from which -@code{safe_mysqld} is invoked. @code{safe_mysqld} looks under its working -directory for @file{bin} and @file{data} directories (for binary -distributions) or for @file{libexec} and @file{var} directories (for source -distributions). This condition should be met if you execute -@code{safe_mysqld} from your @strong{MySQL} installation directory (for -example, @file{/usr/local/mysql} for a binary distribution). - +@code{ALTER TABLE} ignores all @code{INDEX/DATA DIRECTORY="path"} options. @item -If the server and databases cannot be found relative to the working directory, -@code{safe_mysqld} attempts to locate them by absolute pathnames. Typical -locations are @file{/usr/local/libexec} and @file{/usr/local/var}. -The actual locations are determined when the distribution was built from which -@code{safe_mysqld} comes. They should be correct if -@strong{MySQL} was installed in a standard location. +@code{CREATE TABLE} doesn't report if the table has symbolic links. +@item +@code{mysqldump} doesn't include the symbolic links information in the output. +@item +@code{BACKUP TABLE} and @code{RESTORE TABLE} doesn't use symbolic links. @end itemize -Because @code{safe_mysqld} will try to find the server and databases relative -to its own working directory, you can install a binary distribution of -@strong{MySQL} anywhere, as long as you start @code{safe_mysqld} from the -@strong{MySQL} installation directory: +@cindex parameters, server +@cindex @code{mysqld} server, buffer sizes +@cindex buffer sizes, @code{mysqld} server +@cindex startup parameters +@node Server parameters, Table cache, Symbolic links, System +@subsection Tuning Server Parameters + +You can get the default buffer sizes used by the @code{mysqld} server +with this command: @example -shell> cd mysql_installation_directory -shell> bin/safe_mysqld & +shell> mysqld --help @end example -If @code{safe_mysqld} fails, even when invoked from the @strong{MySQL} -installation directory, you can modify it to use the path to @code{mysqld} -and the pathname options that are correct for your system. Note that if you -upgrade @strong{MySQL} in the future, your modified version of -@code{safe_mysqld} will be overwritten, so you should make a copy of your -edited version that you can reinstall. - -@cindex tools, mysqld_multi -@cindex scripts -@cindex multi mysqld -@cindex @code{mysqld_multi} -@node mysqld_multi, mysql, safe_mysqld, Tools -@section mysqld_multi, program for managing multiple @strong{MySQL} servers - -@code{mysqld_multi} is meant for managing several @code{mysqld} -processes running in different UNIX sockets and TCP/IP ports. - -The program will search for group(s) named [mysqld#] from my.cnf (or the -given --config-file=...), where # can be any positive number starting -from 1. These groups should be the same as the usual @code{[mysqld]} -group (e.g. options to mysqld, see @strong{MySQL} manual for detailed -information about this group), but with those port, socket etc. options -that are wanted for each separate @code{mysqld} processes. The number in -the group name has another function; it can be used for starting, -stopping, or reporting some specific @code{mysqld} servers with this -program. See the usage and options below for more information. +@cindex @code{mysqld} options +@cindex variables, @code{mysqld} +This command produces a list of all @code{mysqld} options and configurable +variables. The output includes the default values and looks something +like this: @example -Usage: mysqld_multi [OPTIONS] @{start|stop|report@} [GNR,GNR,GNR...] -or mysqld_multi [OPTIONS] @{start|stop|report@} [GNR-GNR,GNR,GNR-GNR,...] +Possible variables for option --set-variable (-O) are: +back_log current value: 5 +bdb_cache_size current value: 1048540 +binlog_cache_size current_value: 32768 +connect_timeout current value: 5 +delayed_insert_timeout current value: 300 +delayed_insert_limit current value: 100 +delayed_queue_size current value: 1000 +flush_time current value: 0 +interactive_timeout current value: 28800 +join_buffer_size current value: 131072 +key_buffer_size current value: 1048540 +lower_case_table_names current value: 0 +long_query_time current value: 10 +max_allowed_packet current value: 1048576 +max_binlog_cache_size current_value: 4294967295 +max_connections current value: 100 +max_connect_errors current value: 10 +max_delayed_threads current value: 20 +max_heap_table_size current value: 16777216 +max_join_size current value: 4294967295 +max_sort_length current value: 1024 +max_tmp_tables current value: 32 +max_write_lock_count current value: 4294967295 +myisam_sort_buffer_size current value: 8388608 +net_buffer_length current value: 16384 +net_retry_count current value: 10 +net_read_timeout current value: 30 +net_write_timeout current value: 60 +query_buffer_size current value: 0 +record_buffer current value: 131072 +slow_launch_time current value: 2 +sort_buffer current value: 2097116 +table_cache current value: 64 +thread_concurrency current value: 10 +tmp_table_size current value: 1048576 +thread_stack current value: 131072 +wait_timeout current value: 28800 @end example -The GNR above means the group number. You can start, stop or report -any GNR, or several of them at the same time. (See --example) The GNRs -list can be comma separated, or a dash combined, of which the latter -means that all the GNRs between GNR1-GNR2 will be affected. Without -GNR argument all the found groups will be either started, stopped, or -reported. Note that you must not have any white spaces in the GNR -list. Anything after a white space is ignored. - -@code{mysqld_multi} supports the following options: - -@table @code -@cindex config-file option -@item --config-file=... -Alternative config file. NOTE: This will not affect this program's own -options (group @code{[mysqld_multi]}), but only groups -[mysqld#]. Without this option everything will be searched from the -ordinary my.cnf file. -@cindex example option -@item --example -Give an example of a config file. -@cindex help option -@item --help -Print this help and exit. -@cindex log option -@item --log=... -Log file. Full path to and the name for the log file. NOTE: If the file -exists, everything will be appended. -@cindex mysqladmin option -@item --mysqladmin=... -@code{mysqladmin} binary to be used for a server shutdown. -@cindex mysqld option -@item --mysqld=... -@code{mysqld} binary to be used. Note that you can give -@code{safe_mysqld} to this option also. The options are passed to -@code{mysqld}. Just make sure you have @code{mysqld} in your environment -variable @code{PATH} or fix @code{safe_mysqld}. -@cindex no-log option -@item --no-log -Print to stdout instead of the log file. By default the log file is -turned on. -@cindex password option -@item --password=... -Password for user for @code{mysqladmin}. -@cindex tcp-ip option -@item --tcp-ip -Connect to the @strong{MySQL} server(s) via the TCP/IP port instead of -the UNIX socket. This affects stopping and reporting. If a socket file -is missing, the server may still be running, but can be accessed only -via the TCP/IP port. By default connecting is done via the UNIX socket. -@cindex user option -@item --user=... -@strong{MySQL} user for @code{mysqladmin}. -@cindex version option -@item --version -Print the version number and exit. -@end table - -Some notes about @code{mysqld_multi}: +If there is a @code{mysqld} server currently running, you can see what +values it actually is using for the variables by executing this command: -@itemize @bullet -@item -Make sure that the @strong{MySQL} user, who is stopping the -@code{mysqld} services (e.g using the @code{mysqladmin}) have the same -password and username for all the data directories accessed (to the -'mysql' database) And make sure that the user has the 'Shutdown_priv' -privilege! If you have many data- directories and many different 'mysql' -databases with different passwords for the @strong{MySQL} 'root' user, -you may want to create a common 'multi_admin' user for each using the -same password (see below). Example how to do it: -@example -shell> mysql -u root -S /tmp/mysql.sock -proot_password -e -"GRANT SHUTDOWN ON *.* TO multi_admin@@localhost IDENTIFIED BY 'multipass'" -@xref{Privileges}. -@end example -You will have to do the above for each @code{mysqld} running in each -data directory, that you have (just change the socket, -S=...) -@item -@code{pid-file} is very important, if you are using @code{safe_mysqld} -to start @code{mysqld} (e.g. --mysqld=safe_mysqld) Every @code{mysqld} -should have its own @code{pid-file}. The advantage using -@code{safe_mysqld} instead of @code{mysqld} directly here is, that -@code{safe_mysqld} 'guards' every @code{mysqld} process and will restart -it, if a @code{mysqld} process fails due to signal kill -9, or -similar. (Like segmentation fault, which @strong{MySQL} should never do, -of course ;) Please note that @code{safe_mysqld} script may require that -you start it from a certain place. This means that you may have to CD to -a certain directory, before you start the @code{mysqld_multi}. If -you have problems starting, please see the @code{safe_mysqld} -script. Check especially the lines: @example --------------------------------------------------------------------------- -MY_PWD=`pwd` Check if we are starting this relative (for the binary -release) if test -d /data/mysql -a -f ./share/mysql/english/errmsg.sys --a -x ./bin/mysqld --------------------------------------------------------------------------- -@xref{safe_mysqld, , @code{safe_mysqld}}. +shell> mysqladmin variables @end example -The above test should be successful, or you may encounter problems. -@item -Beware of the dangers starting multiple @code{mysqlds} in the same data -directory. Use separate data directories, unless you @strong{KNOW} what -you are doing! -@item -The socket file and the TCP/IP port must be different for every @code{mysqld}. -@item -The first and fifth @code{mysqld} group were intentionally left out from -the example. You may have 'gaps' in the config file. This gives you -more flexibility. The order in which the @code{mysqlds} are started or -stopped depends on the order in which they appear in the config file. -@item -When you want to refer to a certain group using GNR with this program, -just use the number in the end of the group name ( [mysqld# <== ). -@item -You may want to use option '--user' for @code{mysqld}, but in order to -do this you need to be root when you start the @code{mysqld_multi} -script. Having the option in the config file doesn't matter; you will -just get a warning, if you are not the superuser and the @code{mysqlds} -are started under @strong{YOUR} UNIX account. @strong{IMPORTANT}: Make -sure that the @code{pid-file} and the data directory are -read+write(+execute for the latter one) accessible for @strong{THAT} -UNIX user, who the specific @code{mysqld} process is started -as. @strong{DON'T} use the UNIX root account for this, unless you -@strong{KNOW} what you are doing! -@item -@strong{MOST IMPORTANT}: Make sure that you understand the meanings of -the options that are passed to the @code{mysqlds} and why @strong{WOULD -YOU WANT} to have separate @code{mysqld} processes. Starting multiple -@code{mysqlds} in one data directory @strong{WILL NOT} give you extra -performance in a threaded system! -@end itemize - -@xref{Multiple servers}. - -This is an example of the config file on behalf of @code{mysqld_multi}. -@example -# This file should probably be in your home dir (~/.my.cnf) or /etc/my.cnf -# Version 2.1 by Jani Tolonen +You can find a full description for all variables in the @code{SHOW VARIABLES} +section in this manual. @xref{SHOW VARIABLES}. -[mysqld_multi] -mysqld = /usr/local/bin/safe_mysqld -mysqladmin = /usr/local/bin/mysqladmin -user = multi_admin -password = multipass +You can also see some statistics from a running server by issuing the command +@code{SHOW STATUS}. @xref{SHOW STATUS}. -[mysqld2] -socket = /tmp/mysql.sock2 -port = 3307 -pid-file = /usr/local/mysql/var2/hostname.pid2 -datadir = /usr/local/mysql/var2 -language = /usr/local/share/mysql/english -user = john +@strong{MySQL} uses algorithms that are very scalable, so you can usually +run with very little memory. If you, however, give @strong{MySQL} more +memory, you will normally also get better performance. -[mysqld3] -socket = /tmp/mysql.sock3 -port = 3308 -pid-file = /usr/local/mysql/var3/hostname.pid3 -datadir = /usr/local/mysql/var3 -language = /usr/local/share/mysql/swedish -user = monty +When tuning a @strong{MySQL} server, the two most important variables to use +are @code{key_buffer_size} and @code{table_cache}. You should first feel +confident that you have these right before trying to change any of the +other variables. -[mysqld4] -socket = /tmp/mysql.sock4 -port = 3309 -pid-file = /usr/local/mysql/var4/hostname.pid4 -datadir = /usr/local/mysql/var4 -language = /usr/local/share/mysql/estonia -user = tonu +If you have much memory (>=256M) and many tables and want maximum performance +with a moderate number of clients, you should use something like this: -[mysqld6] -socket = /tmp/mysql.sock6 -port = 3311 -pid-file = /usr/local/mysql/var6/hostname.pid6 -datadir = /usr/local/mysql/var6 -language = /usr/local/share/mysql/japanese -user = jani +@example +shell> safe_mysqld -O key_buffer=64M -O table_cache=256 \ + -O sort_buffer=4M -O record_buffer=1M & @end example -@xref{Option files}. +If you have only 128M and only a few tables, but you still do a lot of +sorting, you can use something like: -@cindex command line tool -@cindex tools, command line -@cindex scripts -@cindex @code{mysql} -@node mysql, mysqladmin, mysqld_multi, Tools -@section The Command-line Tool +@example +shell> safe_mysqld -O key_buffer=16M -O sort_buffer=1M +@end example -@code{mysql} is a simple SQL shell (with GNU @code{readline} capabilities). -It supports interactive and non-interactive use. When used interactively, -query results are presented in an ASCII-table format. When used -non-interactively (for example, as a filter), the result is presented in -tab-separated format. (The output format can be changed using command-line -options.) You can run scripts simply like this: +If you have little memory and lots of connections, use something like this: @example -shell> mysql database < script.sql > output.tab +shell> safe_mysqld -O key_buffer=512k -O sort_buffer=100k \ + -O record_buffer=100k & @end example -If you have problems due to insufficient memory in the client, use the -@code{--quick} option! This forces @code{mysql} to use -@code{mysql_use_result()} rather than @code{mysql_store_result()} to -retrieve the result set. - -Using @code{mysql} is very easy. Just start it as follows: -@code{mysql database} or @code{mysql --user=user_name --password=your_password database}. Type a SQL statement, end it with @samp{;}, @samp{\g}, or @samp{\G} -and press RETURN/ENTER. +or even: -@cindex command line options -@cindex options, command line -@cindex startup parameters -@code{mysql} supports the following options: +@example +shell> safe_mysqld -O key_buffer=512k -O sort_buffer=16k \ + -O table_cache=32 -O record_buffer=8k -O net_buffer=1K & +@end example -@table @code -@cindex help option -@item -?, --help -Display this help and exit. -@cindex automatic rehash option -@item -A, --no-auto-rehash -No automatic rehashing. One has to use 'rehash' to get table and field -completion. This gives a quicker start of mysql. -@cindex batch option -@item -B, --batch -Print results with a tab as separator, each row on a new line. Doesn't use -history file. -@cindex character sets option -@item --character-sets-dir=... -Directory where character sets are located. -@cindex compress option. -@item -C, --compress -Use compression in server/client protocol. -@cindex debug option -@item -#, --debug[=...] -Debug log. Default is 'd:t:o,/tmp/mysql.trace'. -@cindex database option -@item -D, --database=... -Database to use. This is mainly useful in the @code{my.cnf} file. -@cindex default character set option -@item --default-character-set=... -Set the default character set. -@cindex execute option -@item -e, --execute=... -Execute command and quit. (Output like with --batch) -@cindex vertical option -@item -E, --vertical -Print the output of a query (rows) vertically. Without this option you -can also force this output by ending your statements with @code{\G}. -@cindex force option -@item -f, --force -Continue even if we get a SQL error. -@cindex no-named-commands option -@item -g, --no-named-commands -Named commands are disabled. Use \* form only, or use named commands -only in the beginning of a line ending with a semicolon (;). Since -Version 10.9, the client now starts with this option ENABLED by default! -With the -g option, long format commands will still work from the first -line, however. -@cindex enable-named-commands option -@item -G, --enable-named-commands -Named commands are @strong{enabled}. Long format commands are allowed as -well as shortened \* commands. -@cindex ignore space option. -@item -i, --ignore-space -Ignore space after function names. -@cindex host option -@item -h, --host=... -Connect to the given host. -@cindex html option -@item -H, --html -Produce HTML output. -@cindex skip line numbers option -@item -L, --skip-line-numbers -Don't write line number for errors. Useful when one wants to compare result -files that includes error messages -@cindex no pager option -@item --no-pager -Disable pager and print to stdout. See interactive help (\h) also. -@cindex no tee option -@item --no-tee -Disable outfile. See interactive help (\h) also. -@cindex unbuffered option. -@item -n, --unbuffered -Flush buffer after each query. -@cindex skip column names option -@item -N, --skip-column-names -Don't write column names in results. -@cindex set variable option -@item -O, --set-variable var=option -Give a variable a value. @code{--help} lists variables. -@cindex one database option -@item -o, --one-database -Only update the default database. This is useful for skipping updates to -other database in the update log. -@cindex pager option -@item @code{--pager[=...]} -Output type. Default is your @code{ENV} variable @code{PAGER}. Valid -pagers are less, more, cat [> filename], etc. See interactive help (\h) -also. This option does not work in batch mode. Pager works only in UNIX. -@cindex password option -@item -p[password], --password[=...] -Password to use when connecting to server. If a password is not given on -the command line, you will be prompted for it. Note that if you use the -short form @code{-p} you can't have a space between the option and the -password. -@item -P --port=... -TCP/IP port number to use for connection. -@cindex quick option -@item -q, --quick -Don't cache result, print it row-by-row. This may slow down the server -if the output is suspended. Doesn't use history file. -@cindex raw option -@item -r, --raw -Write column values without escape conversion. Used with @code{--batch} -@cindex silent option -@item -s, --silent -Be more silent. -@item -S --socket=... -Socket file to use for connection. -@cindex table option -@item -t --table -Output in table format. This is default in non-batch mode. -@item -T, --debug-info -Print some debug information at exit. -@cindex tee option -@item --tee=... -Append everything into outfile. See interactive help (\h) also. Does not -work in batch mode. -@cindex user option -@item -u, --user=# -User for login if not current user. -@cindex safe updates option -@item -U, --safe-updates[=#], --i-am-a-dummy[=#] -Only allow @code{UPDATE} and @code{DELETE} that uses keys. See below for -more information about this option. You can reset this option if you have -it in your @code{my.cnf} file by using @code{--safe-updates=0}. -@cindex verbose option -@item -v, --verbose -More verbose output (-v -v -v gives the table output format). -@cindex version option -@item -V, --version -Output version information and exit. -@cindex wait option -@item -w, --wait -Wait and retry if connection is down instead of aborting. -@end table +When you have installed @strong{MySQL}, the @file{support-files} directory will +contain some different @code{my.cnf} example files, @file{my-huge.cnf}, +@file{my-large.cnf}, @file{my-medium.cnf}, and @file{my-small.cnf}, you can +use as a base to optimize your system. -You can also set the following variables with @code{-O} or -@code{--set-variable}: +If there are very many connections, ``swapping problems'' may occur unless +@code{mysqld} has been configured to use very little memory for each +connection. @code{mysqld} performs better if you have enough memory for all +connections, of course. -@cindex timeout -@multitable @columnfractions .3 .2 .5 -@item Variablename @tab Default @tab Description -@item connect_timeout @tab 0 @tab Number of seconds before timeout connection. -@item max_allowed_packet @tab 16777216 @tab Max packetlength to send/receive from to server -@item net_buffer_length @tab 16384 @tab Buffer for TCP/IP and socket communication -@item select_limit @tab 1000 @tab Automatic limit for SELECT when using --i-am-a-dummy -@item max_join_size @tab 1000000 @tab Automatic limit for rows in a join when using --i-am-a-dummy. -@end multitable +Note that if you change an option to @code{mysqld}, it remains in effect only +for that instance of the server. -If you type 'help' on the command line, @code{mysql} will print out the -commands that it supports: +To see the effects of a parameter change, do something like this: -@cindex commands, list of @example -mysql> help - -MySQL commands: -help (\h) Display this text. -? (\h) Synonym for `help'. -clear (\c) Clear command. -connect (\r) Reconnect to the server. Optional arguments are db and host. -edit (\e) Edit command with $EDITOR. -ego (\G) Send command to mysql server, display result vertically. -exit (\q) Exit mysql. Same as quit. -go (\g) Send command to mysql server. -nopager (\n) Disable pager, print to stdout. -notee (\t) Don't write into outfile. -pager (\P) Set PAGER [to_pager]. Print the query results via PAGER. -print (\p) Print current command. -quit (\q) Quit mysql. -rehash (\#) Rebuild completion hash. -source (\.) Execute a SQL script file. Takes a file name as an argument. -status (\s) Get status information from the server. -tee (\T) Set outfile [to_outfile]. Append everything into given outfile. -use (\u) Use another database. Takes database name as argument. +shell> mysqld -O key_buffer=32m --help @end example -From the above, pager only works in UNIX. +Make sure that the @code{--help} option is last; otherwise, the effect of any +options listed after it on the command line will not be reflected in the +output. -@cindex status command -The @code{status} command gives you some information about the -connection and the server you are using. If you are running in the -@code{--safe-updates} mode, @code{status} will also print the values for -the @code{mysql} variables that affect your queries. +@cindex tables, opening +@cindex tables, closing +@cindex opening, tables +@cindex closing, tables +@cindex table cache +@findex table_cache +@node Table cache, Creating many tables, Server parameters, System +@subsection How MySQL Opens and Closes Tables -@cindex @code{safe-mode} command -A useful startup option for beginners (introduced in @strong{MySQL} -Version 3.23.11) is @code{--safe-updates} (or @code{--i-am-a-dummy} for -users that has at some time done a @code{DELETE FROM table_name} but -forgot the @code{WHERE} clause). When using this option, @code{mysql} -sends the following command to the @strong{MySQL} server when opening -the connection: +@code{table_cache}, @code{max_connections}, and @code{max_tmp_tables} +affect the maximum number of files the server keeps open. If you +increase one or both of these values, you may run up against a limit +imposed by your operating system on the per-process number of open file +descriptors. However, you can increase the limit on many systems. +Consult your OS documentation to find out how to do this, because the +method for changing the limit varies widely from system to system. -@example -SET SQL_SAFE_UPDATES=1,SQL_SELECT_LIMIT=#select_limit#, - SQL_MAX_JOIN_SIZE=#max_join_size#" -@end example +@code{table_cache} is related to @code{max_connections}. For example, +for 200 concurrent running connections, you should have a table cache of +at least @code{200 * n}, where @code{n} is the maximum number of tables +in a join. You also need to reserve some extra file descriptors for +temporary tables and files. -where @code{#select_limit#} and @code{#max_join_size#} are variables that -can be set from the @code{mysql} command line. @xref{SET OPTION, @code{SET}}. +The cache of open tables can grow to a maximum of @code{table_cache} +(default 64; this can be changed with the @code{-O table_cache=#} +option to @code{mysqld}). A table is never closed, except when the +cache is full and another thread tries to open a table or if you use +@code{mysqladmin refresh} or @code{mysqladmin flush-tables}. -The effect of the above is: +When the table cache fills up, the server uses the following procedure +to locate a cache entry to use: @itemize @bullet @item -You are not allowed to do an @code{UPDATE} or @code{DELETE} statement -if you don't have a key constraint in the @code{WHERE} part. One can, -however, force an @code{UPDATE/DELETE} by using @code{LIMIT}: -@example -UPDATE table_name SET not_key_column=# WHERE not_key_column=# LIMIT 1; -@end example +Tables that are not currently in use are released, in least-recently-used +order. + @item -All big results are automatically limited to @code{#select_limit#} rows. +If the cache is full and no tables can be released, but a new table needs to +be opened, the cache is temporarily extended as necessary. + @item -@code{SELECT}'s that will probably need to examine more than -@code{#max_join_size} row combinations will be aborted. +If the cache is in a temporarily-extended state and a table goes from in-use +to not-in-use state, the table is closed and released from the cache. @end itemize -Some useful hints about the @code{mysql} client: +A table is opened for each concurrent access. This means that +if you have two threads accessing the same table or access the table +twice in the same query (with @code{AS}) the table needs to be opened twice. +The first open of any table takes two file descriptors; each additional +use of the table takes only one file descriptor. The extra descriptor +for the first open is used for the index file; this descriptor is shared +among all threads. -Some data is much more readable when displayed vertically, instead of -the usual horizontal box type output. For example longer text, which -includes new lines, is often much easier to be read with vertical -output. +You can check if your table cache is too small by checking the mysqld +variable @code{opened_tables}. If this is quite big, even if you +haven't done a lot of @code{FLUSH TABLES}, you should increase your table +cache. @xref{SHOW STATUS}. -@example -mysql> select * from mails where length(txt) < 300 limit 300,1\G -*************************** 1. row *************************** - msg_nro: 3068 - date: 2000-03-01 23:29:50 -time_zone: +0200 -mail_from: Monty - reply: monty@@no.spam.com - mail_to: "Thimble Smith" <tim@@no.spam.com> - sbj: UTF-8 - txt: >>>>> "Thimble" == Thimble Smith writes: +@cindex tables, too many +@node Creating many tables, Open tables, Table cache, System +@subsection Drawbacks to Creating Large Numbers of Tables in the Same Database -Thimble> Hi. I think this is a good idea. Is anyone familiar with UTF-8 -Thimble> or Unicode? Otherwise I'll put this on my TODO list and see what -Thimble> happens. +If you have many files in a directory, open, close, and create operations will +be slow. If you execute @code{SELECT} statements on many different tables, +there will be a little overhead when the table cache is full, because for +every table that has to be opened, another must be closed. You can reduce +this overhead by making the table cache larger. -Yes, please do that. +@cindex tables, open +@cindex open tables +@node Open tables, Memory use, Creating many tables, System +@subsection Why So Many Open tables? -Regards, -Monty - file: inbox-jani-1 - hash: 190402944 -1 row in set (0.09 sec) -@end example +When you run @code{mysqladmin status}, you'll see something like this: -@itemize @bullet -@item -For logging, you can use the @code{tee} option. The @code{tee} can be -started with option @code{--tee=...}, or from the command line -interactively with command @code{tee}. All the data displayed on the -screen will also be appended into a given file. This can be very useful -for debugging purposes also. The @code{tee} can be disabled from the -command line with command @code{notee}. Executing @code{tee} again -starts logging again. Without a parameter the previous file will be -used. Note that @code{tee} will flush the results into the file after -each command, just before the command line appears again waiting for the -next command. -@item -Browsing, or searching the results in the interactive mode in UNIX less, -more, or any other similar program, is now possible with option -@code{--pager[=...]}. Without argument, @code{mysql} client will look -for environment variable PAGER and set @code{pager} to that. -@code{pager} can be started from the interactive command line with -command @code{pager} and disabled with command @code{nopager}. The -command takes an argument optionally and the @code{pager} will be set to -that. Command @code{pager} can be called without an argument, but this -requires that the option @code{--pager} was used, or the @code{pager} -will default to stdout. @code{pager} works only in UNIX, since it uses -the popen() function, which doesn't exist in Windows. In Windows, the -@code{tee} option can be used instead, although it may not be as handy -as @code{pager} can be in some situations. -@item -A few tips about @code{pager}: You can use it to write to a file: -@example -mysql> pager cat > /tmp/log.txt -@end example -and the results will only go to a file. You can also pass any options -for the programs that you want to use with the @code{pager}: -@example -mysql> pager less -n -i -S -@end example -From the above do note the option '-S'. You may find it very useful when -browsing the results; try the option with horizontal output (end -commands with '\g', or ';') and with vertical output (end commands with -'\G'). Sometimes a very wide result set is hard to be read from the screen, -with option -S to less you can browse the results within the interactive -less from left to right, preventing lines longer than your screen from -being continued to the next line. This can make the result set much more -readable. You can swith the mode between on and off within the interactive -less with '-S'. See the 'h' for more help about less. -@item -Last (unless you already understood this from the above examples ;) you -can combine very complex ways to handle the results, for example the -following would send the results to two files in two different -directories, on two different hard-disks mounted on /dr1 and /dr2, yet -let the results still be seen on the screen via less: @example -mysql> pager cat | tee /dr1/tmp/res.txt | tee /dr2/tmp/res2.txt | less -n -i -S +Uptime: 426 Running threads: 1 Questions: 11082 Reloads: 1 Open tables: 12 @end example -@item -You can also combine the two functions above; have the @code{tee} -enabled, @code{pager} set to 'less' and you will be able to browse the -results in unix 'less' and still have everything appended into a file -the same time. The difference between @code{UNIX tee} used with the -@code{pager} and the @code{mysql} client in-built @code{tee}, is that -the in-built @code{tee} works even if you don't have the @code{UNIX tee} -available. The in-built @code{tee} also logs everything that is printed -on the screen, where the @code{UNIX tee} used with @code{pager} doesn't -log quite that much. Last, but not least, the interactive @code{tee} is -more handy to switch on and off, when you want to log something into a -file, but want to be able to turn the feature off sometimes. -@end itemize +This can be somewhat perplexing if you only have 6 tables. -@cindex administration, server -@cindex server administration -@cindex @code{mysladmn} -@node mysqladmin, mysqldump, mysql, Tools -@section Administering a MySQL Server +@strong{MySQL} is multithreaded, so it may have many queries on the same +table simultaneously. To minimize the problem with two threads having +different states on the same file, the table is opened independently by +each concurrent thread. This takes some memory and one extra file +descriptor for the data file. The index file descriptor is shared +between all threads. -A utility for performing administrative operations. The syntax is: +@cindex memory use +@node Memory use, Internal locking, Open tables, System +@subsection How MySQL Uses Memory -@example -shell> mysqladmin [OPTIONS] command [command-option] command ... -@end example +The list below indicates some of the ways that the @code{mysqld} server +uses memory. Where applicable, the name of the server variable relevant +to the memory use is given: -You can get a list of the options your version of @code{mysqladmin} supports -by executing @code{mysqladmin --help}. +@itemize @bullet +@item +The key buffer (variable @code{key_buffer_size}) is shared by all +threads; Other buffers used by the server are allocated as +needed. @xref{Server parameters}. -The current @code{mysqladmin} supports the following commands: +@item +Each connection uses some thread-specific space: A stack (default 64K, +variable @code{thread_stack}), a connection buffer (variable +@code{net_buffer_length}), and a result buffer (variable +@code{net_buffer_length}). The connection buffer and result buffer are +dynamically enlarged up to @code{max_allowed_packet} when needed. When +a query is running, a copy of the current query string is also allocated. -@multitable @columnfractions .3 .7 -@item create databasename @tab Create a new database. -@item drop databasename @tab Delete a database and all its tables. -@item extended-status @tab Gives an extended status message from the server. -@item flush-hosts @tab Flush all cached hosts. -@item flush-logs @tab Flush all logs. -@item flush-tables @tab Flush all tables. -@item flush-privileges @tab Reload grant tables (same as reload). -@item kill id,id,... @tab Kill mysql threads. -@item password @tab New-password. Change old password to new-password. -@item ping @tab Check if @code{mysqld} is alive. -@item processlist @tab Show list of active threads in server. -@item reload @tab Reload grant tables. -@item refresh @tab Flush all tables and close and open logfiles. -@item shutdown @tab Take server down. -@item slave-start @tab Start slave replication thread. -@item slave-stop @tab Stop slave replication thread. -@item status @tab Gives a short status message from the server. -@item variables @tab Prints variables available. -@item version @tab Get version info from server. -@end multitable +@item +All threads share the same base memory. -All commands can be shortened to their unique prefix. For example: +@item +Only the compressed ISAM / MyISAM tables are memory mapped. This is +because the 32-bit memory space of 4GB is not large enough for most +big tables. When systems with a 64-bit address space become more +common we may add general support for memory mapping. -@example -shell> mysqladmin proc stat -+----+-------+-----------+----+-------------+------+-------+------+ -| Id | User | Host | db | Command | Time | State | Info | -+----+-------+-----------+----+-------------+------+-------+------+ -| 6 | monty | localhost | | Processlist | 0 | | | -+----+-------+-----------+----+-------------+------+-------+------+ -Uptime: 10077 Threads: 1 Questions: 9 Slow queries: 0 Opens: 6 Flush tables: 1 Open tables: 2 Memory in use: 1092K Max memory used: 1116K -@end example +@item +Each request doing a sequential scan over a table allocates a read buffer +(variable @code{record_buffer}). -@cindex status command, results -The @code{mysqladmin status} command result has the following columns: +@item +All joins are done in one pass, and most joins can be done without even +using a temporary table. Most temporary tables are memory-based (HEAP) +tables. Temporary tables with a big record length (calculated as the +sum of all column lengths) or that contain @code{BLOB} columns are +stored on disk. -@cindex uptime -@multitable @columnfractions .3 .7 -@item Uptime @tab Number of seconds the @strong{MySQL} server has been up. -@cindex threads -@item Threads @tab Number of active threads (clients). -@cindex questions -@item Questions @tab Number of questions from clients since @code{mysqld} was started. -@cindex slow queries -@item Slow queries @tab Queries that have taken more than @code{long_query_time} seconds. @xref{Slow query log}. -@cindex opens -@item Opens @tab How many tables @code{mysqld} has opened. -@cindex flush tables -@cindex tables, flush -@item Flush tables @tab Number of @code{flush ...}, @code{refresh}, and @code{reload} commands. -@cindex open tables -@item Open tables @tab Number of tables that are open now. -@cindex memory use -@item Memory in use @tab Memory allocated directly by the @code{mysqld} code (only available when @strong{MySQL} is compiled with --with-debug=full). -@cindex max memory used -@item Max memory used @tab Maximum memory allocated directly by the @code{mysqld} code (only available when @strong{MySQL} is compiled with --with-debug=full). -@end multitable +One problem in @strong{MySQL} versions before Version 3.23.2 is that if a HEAP table +exceeds the size of @code{tmp_table_size}, you get the error @code{The +table tbl_name is full}. In newer versions this is handled by +automatically changing the in-memory (HEAP) table to a disk-based +(MyISAM) table as necessary. To work around this problem, you can +increase the temporary table size by setting the @code{tmp_table_size} +option to @code{mysqld}, or by setting the SQL option +@code{SQL_BIG_TABLES} in the client program. @xref{SET OPTION, , +@code{SET OPTION}}. In @strong{MySQL} Version 3.20, the maximum size of the +temporary table was @code{record_buffer*16}, so if you are using this +version, you have to increase the value of @code{record_buffer}. You can +also start @code{mysqld} with the @code{--big-tables} option to always +store temporary tables on disk. However, this will affect the speed of +many complicated queries. -If you do @code{myslqadmin shutdown} on a socket (in other words, on a -the computer where @code{mysqld} is running), @code{mysqladmin} will -wait until the @strong{MySQL} @code{pid-file} is removed to ensure that -the @code{mysqld} server has stopped properly. +@item +Most requests doing a sort allocates a sort buffer and 0-2 temporary +files depending on the result set size. @xref{Temporary files}. -@cindex dumping, databases -@cindex databases, dumping -@cindex tables, dumping -@cindex backing up, databases -@node mysqldump, mysqlhotcopy, mysqladmin, Tools -@section Dumping the Structure and Data from MySQL Databases and Tables +@item +Almost all parsing and calculating is done in a local memory store. No +memory overhead is needed for small items and the normal slow memory +allocation and freeing is avoided. Memory is allocated only for +unexpectedly large strings (this is done with @code{malloc()} and +@code{free()}). -@cindex @code{mysqldump} -Utility to dump a database or a collection of database for backup or for -transferring the data to another SQL server (not necessarily a @strong{MySQL} -server). The dump will contain SQL statements to create the table -and/or populate the table. +@item +Each index file is opened once and the data file is opened once for each +concurrently running thread. For each concurrent thread, a table structure, +column structures for each column, and a buffer of size @code{3 * n} is +allocated (where @code{n} is the maximum row length, not counting @code{BLOB} +columns). A @code{BLOB} uses 5 to 8 bytes plus the length of the @code{BLOB} +data. The @code{ISAM}/@code{MyISAM} table handlers will use one extra row +buffer for internal usage. -If you are doing a backup on the server, you should consider using -the @code{mysqlhotcopy} instead. @xref{mysqlhotcopy, , @code{mysqlhotcopy}}. +@item +For each table having @code{BLOB} columns, a buffer is enlarged dynamically +to read in larger @code{BLOB} values. If you scan a table, a buffer as large +as the largest @code{BLOB} value is allocated. -@example -shell> mysqldump [OPTIONS] database [tables] -OR mysqldump [OPTIONS] --databases [OPTIONS] DB1 [DB2 DB3...] -OR mysqldump [OPTIONS] --all-databases [OPTIONS] -@end example +@item +Table handlers for all in-use tables are saved in a cache and managed as a +FIFO. Normally the cache has 64 entries. If a table has been used by two +running threads at the same time, the cache contains two entries for the +table. @xref{Table cache}. -If you don't give any tables or use the @code{--databases} or -@code{--all-databases}, the whole database(s) will be dumped. +@item +A @code{mysqladmin flush-tables} command closes all tables that are not in +use and marks all in-use tables to be closed when the currently executing +thread finishes. This will effectively free most in-use memory. +@end itemize -You can get a list of the options your version of @code{mysqldump} supports -by executing @code{mysqldump --help}. +@code{ps} and other system status programs may report that @code{mysqld} +uses a lot of memory. This may be caused by thread-stacks on different +memory addresses. For example, the Solaris version of @code{ps} counts +the unused memory between stacks as used memory. You can verify this by +checking available swap with @code{swap -s}. We have tested +@code{mysqld} with commercial memory-leakage detectors, so there should +be no memory leaks. -Note that if you run @code{mysqldump} without @code{--quick} or -@code{--opt}, @code{mysqldump} will load the whole result set into -memory before dumping the result. This will probably be a problem if -you are dumping a big database. +@cindex internal locking +@cindex locking, tables +@cindex tables, locking +@node Internal locking, Table locking, Memory use, System +@subsection How MySQL Locks Tables -Note that if you are using a new copy of the @code{mysqldump} program -and you are going to do a dump that will be read into a very old @strong{MySQL} -server, you should not use the @code{--opt} or @code{-e} options. +You can find a discussion about different locking methods in the appendix. +@xref{Locking methods}. -@code{mysqldump} supports the following options: +All locking in @strong{MySQL} is deadlock-free. This is managed by always +requesting all needed locks at once at the beginning of a query and always +locking the tables in the same order. -@table @code -@item --add-locks -Add @code{LOCK TABLES} before and @code{UNLOCK TABLE} after each table dump. -(To get faster inserts into @strong{MySQL}.) -@item --add-drop-table -Add a @code{drop table} before each create statement. -@item -A, --all-databases -Dump all the databases. This will be same as @code{--databases} with all -databases selected. -@item -a, --all -Include all @strong{MySQL}-specific create options. -@item --allow-keywords -Allow creation of column names that are keywords. This works by -prefixing each column name with the table name. -@item -c, --complete-insert -Use complete insert statements (with column names). -@item -C, --compress -Compress all information between the client and the server if both support -compression. -@item -B, --databases -To dump several databases. Note the difference in usage. In this case -no tables are given. All name arguments are regarded as database names. -@code{USE db_name;} will be included in the output before each new database. -@item --delayed -Insert rows with the @code{INSERT DELAYED} command. -@item -e, --extended-insert -Use the new multiline @code{INSERT} syntax. (Gives more compact and -faster inserts statements.) -@item -#, --debug[=option_string] -Trace usage of the program (for debugging). -@item --help -Display a help message and exit. -@item --fields-terminated-by=... -@itemx --fields-enclosed-by=... -@itemx --fields-optionally-enclosed-by=... -@itemx --fields-escaped-by=... -@itemx --lines-terminated-by=... -These options are used with the @code{-T} option and have the same -meaning as the corresponding clauses for @code{LOAD DATA INFILE}. -@xref{LOAD DATA, , @code{LOAD DATA}}. -@item -F, --flush-logs -Flush log file in the @strong{MySQL} server before starting the dump. -@item -f, --force, -Continue even if we get a SQL error during a table dump. -@item -h, --host=.. -Dump data from the @strong{MySQL} server on the named host. The default host -is @code{localhost}. -@item -l, --lock-tables. -Lock all tables before starting the dump. The tables are locked with -@code{READ LOCAL} to allow concurrent inserts in the case of @code{MyISAM} -tables. -@item -n, --no-create-db -'CREATE DATABASE /*!32312 IF NOT EXISTS*/ db_name;' will not be put in the -output. The above line will be added otherwise, if --databases or ---all-databases option was given. -@item -t, --no-create-info -Don't write table creation information (The @code{CREATE TABLE} statement.) -@item -d, --no-data -Don't write any row information for the table. This is very useful if you -just want to get a dump of the structure for a table! -@item --opt -Same as @code{--quick --add-drop-table --add-locks --extended-insert ---lock-tables}. Should give you the fastest possible dump for reading -into a @strong{MySQL} server. -@item -pyour_pass, --password[=your_pass] -The password to use when connecting to the server. If you specify -no @samp{=your_pass} part, -@code{mysqldump} you will be prompted for a password. -@item -P port_num, --port=port_num -The TCP/IP port number to use for connecting to a host. (This is used for -connections to hosts other than @code{localhost}, for which Unix sockets are -used.) -@item -q, --quick -Don't buffer query, dump directly to stdout. Uses @code{mysql_use_result()} -to do this. -@item -r, --result-file=... -Direct output to a given file. This option should be used in MSDOS, -because it prevents new line '\n' from being converted to '\n\r' (new -line + carriage return). -@item -S /path/to/socket, --socket=/path/to/socket -The socket file to use when connecting to @code{localhost} (which is the -default host). -@item --tables -Overrides option --databases (-B). -@item -T, --tab=path-to-some-directory -Creates a @code{table_name.sql} file, that contains the SQL CREATE commands, -and a @code{table_name.txt} file, that contains the data, for each give table. -@strong{NOTE}: This only works if @code{mysqldump} is run on the same -machine as the @code{mysqld} daemon. The format of the @code{.txt} file -is made according to the @code{--fields-xxx} and @code{--lines--xxx} options. -@item -u user_name, --user=user_name -The @strong{MySQL} user name to use when connecting to the server. The -default value is your Unix login name. -@item -O var=option, --set-variable var=option -Set the value of a variable. The possible variables are listed below. -@item -v, --verbose -Verbose mode. Print out more information on what the program does. -@item -V, --version -Print version information and exit. -@item -w, --where='where-condition' -Dump only selected records. Note that QUOTES are mandatory: +The locking method @strong{MySQL} uses for @code{WRITE} locks works as follows: -@example -"--where=user='jimf'" "-wuserid>1" "-wuserid<1" -@end example -@item -O net_buffer_length=#, where # < 16M -When creating multi-row-insert statements (as with option -@code{--extended-insert} or @code{--opt}), @code{mysqldump} will create -rows up to @code{net_buffer_length} length. If you increase this -variable, you should also ensure that the @code{max_allowed_packet} -variable in the @strong{MySQL} server is bigger than the -@code{net_buffer_length}. -@end table +@itemize @bullet +@item +If there are no locks on the table, put a write lock on it. +@item +Otherwise, put the lock request in the write lock queue. +@end itemize -The most normal use of @code{mysqldump} is probably for making a backup of -whole databases. @xref{Backup}. +The locking method @strong{MySQL} uses for @code{READ} locks works as follows: -@example -mysqldump --opt database > backup-file.sql -@end example +@itemize @bullet +@item +If there are no write locks on the table, put a read lock on it. +@item +Otherwise, put the lock request in the read lock queue. +@end itemize -You can read this back into @strong{MySQL} with: +When a lock is released, the lock is made available to the threads +in the write lock queue, then to the threads in the read lock queue. -@example -mysql database < backup-file.sql -@end example +This means that if you have many updates on a table, @code{SELECT} +statements will wait until there are no more updates. -or +To work around this for the case where you want to do many @code{INSERT} and +@code{SELECT} operations on a table, you can insert rows in a temporary +table and update the real table with the records from the temporary table +once in a while. +This can be done with the following code: @example -mysql -e "source /patch-to-backup/backup-file.sql" database +mysql> LOCK TABLES real_table WRITE, insert_table WRITE; +mysql> insert into real_table select * from insert_table; +mysql> TRUNCATE TABLE insert_table; +mysql> UNLOCK TABLES; @end example -However, it's also very useful to populate another @strong{MySQL} server with -information from a database: +You can use the @code{LOW_PRIORITY} options with @code{INSERT}, +@code{UPDATE} or @code{DELETE} or @code{HIGH_PRIORITY} with +@code{SELECT} if you want to prioritize retrieval in some specific +cases. You can also start @code{mysqld} with @code{--low-priority-updates} +to get the same behaveour. -@example -mysqldump --opt database | mysql ---host=remote-host -C database -@end example +Using @code{SQL_BUFFER_RESULT} can also help making table locks shorter. +@xref{SELECT}. -It is possible to dump several databases with one command: +You could also change the locking code in @file{mysys/thr_lock.c} to use a +single queue. In this case, write locks and read locks would have the same +priority, which might help some applications. -@example -mysqldump --databases database1 [database2 database3...] > my_databases.sql -@end example +@cindex problems, table locking +@node Table locking, DNS, Internal locking, System +@subsection Table Locking Issues -If all the databases are wanted, one can use: +The table locking code in @strong{MySQL} is deadlock free. -@example -mysqldump --all-databases > all_databases.sql -@end example +@strong{MySQL} uses table locking (instead of row locking or column +locking) on all table types, except @code{BDB} tables, to achieve a very +high lock speed. For large tables, table locking is MUCH better than +row locking for most applications, but there are, of course, some +pitfalls. -@cindex dumping, databases -@cindex databases, dumping -@cindex tables, dumping -@cindex backing up, databases -@node mysqlhotcopy, mysqlimport, mysqldump, Tools -@section Copying MySQL Databases and Tables - -@code{mysqlhotcopy} is a perl script that uses @code{LOCK TABLES}, -@code{FLUSH TABLES} and @code{cp} or @code{scp} to quickly make a backup -of a database. It's the fastest way to make a backup of the database, -of single tables but it can only be run on the same machine where the -database directories are. +For @code{BDB} and @code{InnoDB} tables, @strong{MySQL} only uses table +locking if you explicitely lock the table with @code{LOCK TABLES} or +execute a command that will modify every row in the table, like +@code{ALTER TABLE}. For these table types we recommend you to not use +@code{LOCK TABLES} at all. -@example -mysqlhotcopy db_name [/path/to/new_directory] +In @strong{MySQL} Version 3.23.7 and above, you can insert rows into +@code{MyISAM} tables at the same time other threads are reading from the +table. Note that currently this only works if there are no holes after +deleted rows in the table at the time the insert is made. When all holes +has been filled with new data, concurrent inserts will automatically be +enabled again. -mysqlhotcopy db_name_1 ... db_name_n /path/to/new_directory +Table locking enables many threads to read from a table at the same +time, but if a thread wants to write to a table, it must first get +exclusive access. During the update, all other threads that want to +access this particular table will wait until the update is ready. -mysqlhotcopy db_name./regex/ -@end example +As updates on tables normally are considered to be more important than +@code{SELECT}, all statements that update a table have higher priority +than statements that retrieve information from a table. This should +ensure that updates are not 'starved' because one issues a lot of heavy +queries against a specific table. (You can change this by using +LOW_PRIORITY with the statement that does the update or +@code{HIGH_PRIORITY} with the @code{SELECT} statement.) -@code{mysqlhotcopy} supports the following options: +Starting from @strong{MySQL} Version 3.23.7 one can use the +@code{max_write_lock_count} variable to force @strong{MySQL} to +temporary give all @code{SELECT} statements, that wait for a table, a +higher priority after a specific number of inserts on a table. -@table @code -@item -?, --help -Display a help screen and exit -@item -u, --user=# -User for database login -@item -p, --password=# -Password to use when connecting to server -@item -P, --port=# -Port to use when connecting to local server -@item -S, --socket=# -Socket to use when connecting to local server -@item --allowold -Don't abort if target already exists (rename it _old) -@item --keepold -Don't delete previous (now renamed) target when done -@item --noindices -Don't include full index files in copy to make the backup smaller and faster -The indexes can later be reconstructed with @code{myisamchk -rq.}. -@item --method=# -Method for copy (@code{cp} or @code{scp}). -@item -q, --quiet -Be silent except for errors -@item --debug -Enable debug -@item -n, --dryrun -Report actions without doing them -@item --regexp=# -Copy all databases with names matching regexp -@item --suffix=# -Suffix for names of copied databases -@item --checkpoint=# -Insert checkpoint entry into specified db.table -@item --flushlog -Flush logs once all tables are locked. -@item --tmpdir=# -Temporary directory (instead of /tmp). -@end table +Table locking is, however, not very good under the following senario: -You can use @code{perldoc mysqlhotcopy} to get a more complete -documentation for @code{mysqlhotcopy}. +@itemize @bullet +@item +A client issues a @code{SELECT} that takes a long time to run. +@item +Another client then issues an @code{UPDATE} on a used table. This client +will wait until the @code{SELECT} is finished. +@item +Another client issues another @code{SELECT} statement on the same table. As +@code{UPDATE} has higher priority than @code{SELECT}, this @code{SELECT} +will wait for the @code{UPDATE} to finish. It will also wait for the first +@code{SELECT} to finish! +@item +A thread is waiting for something like @code{full disk}, in which case all +threads that wants to access the problem table will also be put in a waiting +state until more disk space is made available. +@end itemize -@code{mysqlhotcopy} reads the groups @code{[client]} and @code{[mysqlhotcopy]} -from the option files. +Some possible solutions to this problem are: -To be able to execute @code{mysqlhotcopy} you need write access to the -backup directory, @code{SELECT} privilege to the tables you are about to -copy and the @strong{MySQL} @code{Reload} privilege (to be able to -execute @code{FLUSH TABLES}). +@itemize @bullet +@item +Try to get the @code{SELECT} statements to run faster. You may have to create +some summary tables to do this. -@cindex importing, data -@cindex data, importing -@cindex files, text -@cindex text files, importing -@cindex @code{mysqlimport} -@node mysqlimport, perror, mysqlhotcopy, Tools -@section Importing Data from Text Files +@item +Start @code{mysqld} with @code{--low-priority-updates}. This will give +all statements that update (modify) a table lower priority than a @code{SELECT} +statement. In this case the last @code{SELECT} statement in the previous +scenario would execute before the @code{INSERT} statement. -@code{mysqlimport} provides a command-line interface to the @code{LOAD DATA -INFILE} SQL statement. Most options to @code{mysqlimport} correspond -directly to the same options to @code{LOAD DATA INFILE}. -@xref{LOAD DATA, , @code{LOAD DATA}}. +@item +You can give a specific @code{INSERT}, @code{UPDATE}, or @code{DELETE} +statement lower priority with the @code{LOW_PRIORITY} attribute. -@code{mysqlimport} is invoked like this: +@item +Start @code{mysqld} with a low value for @strong{max_write_lock_count} to give +@code{READ} locks after a certain number of @code{WRITE} locks. -@example -shell> mysqlimport [options] database textfile1 [textfile2....] -@end example +@item +You can specify that all updates from a specific thread should be done with +low priority by using the SQL command: @code{SET SQL_LOW_PRIORITY_UPDATES=1}. +@xref{SET OPTION, , @code{SET OPTION}}. -For each text file named on the command line, -@code{mysqlimport} strips any extension from the filename and uses the result -to determine which table to import the file's contents into. For example, -files named @file{patient.txt}, @file{patient.text}, and @file{patient} would -all be imported into a table named @code{patient}. +@item +You can specify that a specific @code{SELECT} is very important with the +@code{HIGH_PRIORITY} attribute. @xref{SELECT, , @code{SELECT}}. -@code{mysqlimport} supports the following options: +@item +If you have problems with @code{INSERT} combined with @code{SELECT}, +switch to use the new @code{MyISAM} tables as these support concurrent +@code{SELECT}s and @code{INSERT}s. -@table @code -@item -c, --columns=... -This option takes a comma-separated list of field names as an argument. -The field list is used to create a proper @code{LOAD DATA INFILE} command, -which is then passed to @strong{MySQL}. @xref{LOAD DATA, , @code{LOAD DATA}}. +@item +If you mainly mix @code{INSERT} and @code{SELECT} statements, the +@code{DELAYED} attribute to @code{INSERT} will probably solve your problems. +@xref{INSERT, , @code{INSERT}}. -@item -C, --compress -Compress all information between the client and the server if both support -compression. +@item +If you have problems with @code{SELECT} and @code{DELETE}, the @code{LIMIT} +option to @code{DELETE} may help. @xref{DELETE, , @code{DELETE}}. +@end itemize -@item -#, --debug[=option_string] -Trace usage of the program (for debugging). +@cindex DNS +@cindex hostname caching +@node DNS, , Table locking, System +@subsection How MySQL uses DNS -@item -d, --delete -Empty the table before importing the text file. +When a new thread connects to @code{mysqld}, @code{mysqld} will span a +new thread to handle the request. This thread will first check if the +hostname is in the hostname cache. If not the thread will call +@code{gethostbyaddr_r()} and @code{gethostbyname_r()} to resolve the +hostname. -@item --fields-terminated-by=... -@itemx --fields-enclosed-by=... -@itemx --fields-optionally-enclosed-by=... -@itemx --fields-escaped-by=... -@itemx --lines-terminated-by=... -These options have the same meaning as the corresponding clauses for -@code{LOAD DATA INFILE}. @xref{LOAD DATA, , @code{LOAD DATA}}. +If the operating system doesn't support the above thread-safe calls, the +thread will lock a mutex and call @code{gethostbyaddr()} and +@code{gethostbyname()} instead. Note that in this case no other thread +can resolve other hostnames that is not in the hostname cache until the +first thread is ready. -@item -f, --force -Ignore errors. For example, if a table for a text file doesn't exist, -continue processing any remaining files. Without @code{--force}, -@code{mysqlimport} exits if a table doesn't exist. +You can disable DNS host lookup by starting @code{mysqld} with +@code{--skip-name-resolve}. In this case you can however only use IP +names in the @strong{MySQL} privilege tables. -@item --help -Display a help message and exit. +If you have a very slow DNS and many hosts, you can get more performance by +either disabling DNS lookop with @code{--skip-name-resolve} or by +increasing the @code{HOST_CACHE_SIZE} define (default: 128) and recompile +@code{mysqld}. -@item -h host_name, --host=host_name -Import data to the @strong{MySQL} server on the named host. The default host -is @code{localhost}. +You can disable the hostname cache with @code{--skip-host-cache}. You +can clear the hostname cache with @code{FLUSH HOSTS} or @code{mysqladmin +flush-hosts}. -@item -i, --ignore -See the description for the @code{--replace} option. +If you don't want to allow connections over @code{TCP/IP}, you can do this +by starting @code{mysqld} with @code{--skip-networking}. -@item -l, --lock-tables -Lock @strong{ALL} tables for writing before processing any text files. This -ensures that all tables are synchronized on the server. +@cindex data, size +@cindex reducing, data size +@cindex storage space, minimizing +@cindex tables, improving performance +@cindex performance, improving +@node Data size, MySQL indexes, System, Performance +@section Get Your Data as Small as Possible -@item -L, --local -Read input files from the client. By default, text files are assumed to be on -the server if you connect to @code{localhost} (which is the default host). +One of the most basic optimization is to get your data (and indexes) to +take as little space on the disk (and in memory) as possible. This can +give huge improvements because disk reads are faster and normally less +main memory will be used. Indexing also takes less resources if +done on smaller columns. -@item -pyour_pass, --password[=your_pass] -The password to use when connecting to the server. If you specify -no @samp{=your_pass} part, -@code{mysqlimport} you will be prompted for a password. +@strong{MySQL} supports a lot of different table types and row formats. +Choosing the right table format may give you a big performance gain. +@xref{Table types}. -@item -P port_num, --port=port_num -The TCP/IP port number to use for connecting to a host. (This is used for -connections to hosts other than @code{localhost}, for which Unix sockets are -used.) +You can get better performance on a table and minimize storage space +using the techniques listed below: -@item -r, --replace -The @code{--replace} and @code{--ignore} options control handling of input -records that duplicate existing records on unique key values. If you specify -@code{--replace}, new rows replace existing rows that have the same unique key -value. If you specify @code{--ignore}, input rows that duplicate an existing -row on a unique key value are skipped. If you don't specify either option, an -error occurs when a duplicate key value is found, and the rest of the text -file is ignored. +@itemize @bullet +@item +Use the most efficient (smallest) types possible. @strong{MySQL} has +many specialized types that save disk space and memory. +@item +Use the smaller integer types if possible to get smaller tables. For +example, @code{MEDIUMINT} is often better than @code{INT}. +@item +Declare columns to be @code{NOT NULL} if possible. It makes everything +faster and you save one bit per column. Note that if you really need +@code{NULL} in your application you should definitely use it. Just avoid +having it on all columns by default. +@item +If you don't have any variable-length columns (@code{VARCHAR}, +@code{TEXT}, or @code{BLOB} columns), a fixed-size record format is +used. This is faster but unfortunately may waste some space. +@xref{MyISAM table formats}. +@item +The primary index of a table should be as short as possible. This makes +identification of one row easy and efficient. +@item +For each table, you have to decide which storage/index method to +use. @xref{Table types}. +@item +Only create the indexes that you really need. Indexes are good for +retrieval but bad when you need to store things fast. If you mostly +access a table by searching on a combination of columns, make an index +on them. The first index part should be the most used column. If you are +ALWAYS using many columns, you should use the column with more duplicates +first to get better compression of the index. +@item +If it's very likely that a column has a unique prefix on the first number +of characters, it's better to only index this prefix. @strong{MySQL} +supports an index on a part of a character column. Shorter indexes are +faster not only because they take less disk space but also because they +will give you more hits in the index cache and thus fewer disk +seeks. @xref{Server parameters}. +@item +In some circumstances it can be beneficial to split into two a table that is +scanned very often. This is especially true if it is a dynamic +format table and it is possible to use a smaller static format table that +can be used to find the relevant rows when scanning the table. +@end itemize -@item -s, --silent -Silent mode. Write output only when errors occur. +@cindex indexes, uses for +@node MySQL indexes, Query Speed, Data size, Performance +@section How MySQL Uses Indexes -@item -S /path/to/socket, --socket=/path/to/socket -The socket file to use when connecting to @code{localhost} (which is the -default host). +Indexes are used to find rows with a specific value of one column +fast. Without an index @strong{MySQL} has to start with the first record +and then read through the whole table until it finds the relevant +rows. The bigger the table, the more this costs. If the table has an index +for the columns in question, @strong{MySQL} can quickly get a position to +seek to in the middle of the data file without having to look at all the +data. If a table has 1000 rows, this is at least 100 times faster than +reading sequentially. Note that if you need to access almost all 1000 +rows it is faster to read sequentially because we then avoid disk seeks. -@item -u user_name, --user=user_name -The @strong{MySQL} user name to use when connecting to the server. The -default value is your Unix login name. +All @strong{MySQL} indexes (@code{PRIMARY}, @code{UNIQUE}, and +@code{INDEX}) are stored in B-trees. Strings are automatically prefix- +and end-space compressed. @xref{CREATE INDEX, , @code{CREATE INDEX}}. -@item -v, --verbose -Verbose mode. Print out more information what the program does. +Indexes are used to: +@itemize @bullet +@item +Quickly find the rows that match a @code{WHERE} clause. -@item -V, --version -Print version information and exit. -@end table +@item +Retrieve rows from other tables when performing joins. -Here is a sample run using @code{mysqlimport}: +@item +Find the @code{MAX()} or @code{MIN()} value for a specific indexed +column. This is optimized by a preprocessor that checks if you are +using @code{WHERE} key_part_# = constant on all key parts < N. In this case +@strong{MySQL} will do a single key lookup and replace the @code{MIN()} +expression with a constant. If all expressions are replaced with +constants, the query will return at once: @example -$ mysql --version -mysql Ver 9.33 Distrib 3.22.25, for pc-linux-gnu (i686) -$ uname -a -Linux xxx.com 2.2.5-15 #1 Mon Apr 19 22:21:09 EDT 1999 i586 unknown -$ mysql -e 'CREATE TABLE imptest(id INT, n VARCHAR(30))' test -$ ed -a -100 Max Sydow -101 Count Dracula -. -w imptest.txt -32 -q -$ od -c imptest.txt -0000000 1 0 0 \t M a x S y d o w \n 1 0 -0000020 1 \t C o u n t D r a c u l a \n -0000040 -$ mysqlimport --local test imptest.txt -test.imptest: Records: 2 Deleted: 0 Skipped: 0 Warnings: 0 -$ mysql -e 'SELECT * FROM imptest' test -+------+---------------+ -| id | n | -+------+---------------+ -| 100 | Max Sydow | -| 101 | Count Dracula | -+------+---------------+ +SELECT MIN(key_part2),MAX(key_part2) FROM table_name where key_part1=10 @end example -@cindex error messages, displaying -@cindex perror -@node perror, mysqlshow, mysqlimport, Tools -@section Converting an error code to the corresponding error message +@item +Sort or group a table if the sorting or grouping is done on a leftmost +prefix of a usable key (for example, @code{ORDER BY key_part_1,key_part_2 }). The +key is read in reverse order if all key parts are followed by @code{DESC}. -@code{perror} can be used to print error message(s). @code{perror} can -be invoked like this: +The index can also be used even if the @code{ORDER BY} doesn't match the index +exactly, as long as all the unused index parts and all the extra +are @code{ORDER BY} columns are constants in the @code{WHERE} clause. The +following queries will use the index to resolve the @code{ORDER BY} part: @example -shell> perror [OPTIONS] [ERRORCODE [ERRORCODE...]] - -For example: - -shell> perror 64 79 -Error code 64: Machine is not on the network -Error code 79: Can not access a needed shared library +SELECT * FROM foo ORDER BY key_part1,key_part2,key_part3; +SELECT * FROM foo WHERE column=constant ORDER BY column, key_part1; +SELECT * FROM foo WHERE key_part1=const GROUP BY key_part2; @end example -@code{perror} can be used to display a description for a system error -code, or an MyISAM/ISAM table handler error code. The error messages -are mostly system dependent. - -@cindex databases, displaying -@cindex displaying, database information -@cindex tables, displaying -@cindex columns, displaying -@cindex showing, database information -@node mysqlshow, myisampack, perror, Tools -@section Showing Databases, Tables, and Columns - -@code{mysqlshow} can be used to quickly look at which databases exist, -their tables, and the table's columns. - -With the @code{mysql} program you can get the same information with the -@code{SHOW} commands. @xref{SHOW}. - -@code{mysqlshow} is invoked like this: +@item +In some cases a query can be optimized to retrieve values without +consulting the data file. If all used columns for some table are numeric +and form a leftmost prefix for some key, the values may be retrieved +from the index tree for greater speed: @example -shell> mysqlshow [OPTIONS] [database [table [column]]] +SELECT key_part3 FROM table_name WHERE key_part1=1 @end example -@itemize @bullet -@item -If no database is given, all matching databases are shown. -@item -If no table is given, all matching tables in the database are shown. -@item -If no column is given, all matching columns and column types in the table -are shown. @end itemize -Note that in newer @strong{MySQL} versions, you only see those -database/tables/columns for which you have some privileges. - -If the last argument contains a shell or SQL wild-card (@code{*}, @code{?}, -@code{%} or @code{_}) then only what's matched by the wild card is shown. -This may cause some confusion when you try to display the columns for a -table with a @code{_} as in this case @code{mysqlshow} only shows you -the table names that match the pattern. This is easily fixed by -adding an extra @code{%} last on the command line (as a separate -argument). +Suppose you issue the following @code{SELECT} statement: -@cindex compressed tables -@cindex tables, compressed -@cindex MyISAM, compressed tables -@cindex @code{myisampack} -@cindex @code{pack_isam} -@node myisampack, , mysqlshow, Tools -@section The MySQL Compressed Read-only Table Generator +@example +mysql> SELECT * FROM tbl_name WHERE col1=val1 AND col2=val2; +@end example -@code{myisampack} is used to compress MyISAM tables, and @code{pack_isam} -is used to compress ISAM tables. Because ISAM tables are deprecated, we -will only discuss @code{myisampack} here, but everything said about -@code{myisampack} should also be true for @code{pack_isam}. +If a multiple-column index exists on @code{col1} and @code{col2}, the +appropriate rows can be fetched directly. If separate single-column +indexes exist on @code{col1} and @code{col2}, the optimizer tries to +find the most restrictive index by deciding which index will find fewer +rows and using that index to fetch the rows. -@code{myisampack} works by compressing each column in the table separately. -The information needed to decompress columns is read into memory when the -table is opened. This results in much better performance when accessing -individual records, because you only have to uncompress exactly one record, not -a much larger disk block as when using Stacker on MS-DOS. -Usually, @code{myisampack} packs the data file 40%-70%. +@cindex indexes, leftmost prefix of +@cindex leftmost prefix of indexes +If the table has a multiple-column index, any leftmost prefix of the +index can be used by the optimizer to find rows. For example, if you +have a three-column index on @code{(col1,col2,col3)}, you have indexed +search capabilities on @code{(col1)}, @code{(col1,col2)}, and +@code{(col1,col2,col3)}. -@strong{MySQL} uses memory mapping (@code{mmap()}) on compressed tables and -falls back to normal read/write file usage if @code{mmap()} doesn't work. +@strong{MySQL} can't use a partial index if the columns don't form a +leftmost prefix of the index. Suppose you have the @code{SELECT} +statements shown below: -There are currently two limitations with @code{myisampack}: -@itemize @bullet -@item -After packing, the table is read-only. -@item -@code{myisampack} can also pack @code{BLOB} or @code{TEXT} columns. The -older @code{pack_isam} could not do this. -@end itemize +@example +mysql> SELECT * FROM tbl_name WHERE col1=val1; +mysql> SELECT * FROM tbl_name WHERE col2=val2; +mysql> SELECT * FROM tbl_name WHERE col2=val2 AND col3=val3; +@end example -Fixing these limitations is on our TODO list but with low priority. +If an index exists on @code{(col1,col2,col3)}, only the first query +shown above uses the index. The second and third queries do involve +indexed columns, but @code{(col2)} and @code{(col2,col3)} are not +leftmost prefixes of @code{(col1,col2,col3)}. -@code{myisampack} is invoked like this: +@findex LIKE, and indexes +@findex LIKE, and wildcards +@cindex indexes, and @code{LIKE} +@cindex wildcards, and @code{LIKE} +@strong{MySQL} also uses indexes for @code{LIKE} comparisons if the argument +to @code{LIKE} is a constant string that doesn't start with a wild-card +character. For example, the following @code{SELECT} statements use indexes: @example -shell> myisampack [options] filename ... +mysql> select * from tbl_name where key_col LIKE "Patrick%"; +mysql> select * from tbl_name where key_col LIKE "Pat%_ck%"; @end example -Each filename should be the name of an index (@file{.MYI}) file. If you -are not in the database directory, you should specify the pathname to the -file. It is permissible to omit the @file{.MYI} extension. - -@code{myisampack} supports the following options: +In the first statement, only rows with @code{"Patrick" <= key_col < +"Patricl"} are considered. In the second statement, only rows with +@code{"Pat" <= key_col < "Pau"} are considered. -@table @code -@item -b, --backup -Make a backup of the table as @code{tbl_name.OLD}. - -@item -#, --debug=debug_options -Output debug log. The @code{debug_options} string often is -@code{'d:t:o,filename'}. +The following @code{SELECT} statements will not use indexes: +@example +mysql> select * from tbl_name where key_col LIKE "%Patrick%"; +mysql> select * from tbl_name where key_col LIKE other_col; +@end example -@item -f, --force -Force packing of the table even if it becomes bigger or if the temporary file -exists. @code{myisampack} creates a temporary file named @file{tbl_name.TMD} -while it compresses the table. If you kill @code{myisampack}, the @file{.TMD} -file may not be deleted. Normally, @code{myisampack} exits with an error if -it finds that @file{tbl_name.TMD} exists. With @code{--force}, -@code{myisampack} packs the table anyway. +In the first statement, the @code{LIKE} value begins with a wild-card +character. In the second statement, the @code{LIKE} value is not a +constant. -@item -?, --help -Display a help message and exit. +@findex IS NULL, and indexes +@cindex indexes, and @code{IS NULL} +Searching using @code{column_name IS NULL} will use indexes if column_name +is an index. -@item -j big_tbl_name, --join=big_tbl_name -Join all tables named on the command line into a single table -@code{big_tbl_name}. All tables that are to be combined -MUST be identical (same column names and types, same indexes, etc.). +@strong{MySQL} normally uses the index that finds the least number of rows. An +index is used for columns that you compare with the following operators: +@code{=}, @code{>}, @code{>=}, @code{<}, @code{<=}, @code{BETWEEN}, and a +@code{LIKE} with a non-wild-card prefix like @code{'something%'}. -@item -p #, --packlength=# -Specify the record length storage size, in bytes. The value should be 1, 2, -or 3. (@code{myisampack} stores all rows with length pointers of 1, 2, or 3 -bytes. In most normal cases, @code{myisampack} can determine the right length -value before it begins packing the file, but it may notice during the packing -process that it could have used a shorter length. In this case, -@code{myisampack} will print a note that the next time you pack the same file, -you could use a shorter record length.) +Any index that doesn't span all @code{AND} levels in the @code{WHERE} clause +is not used to optimize the query. In other words: To be able to use an +index, a prefix of the index must be used in every @code{AND} group. -@item -s, --silent -Silent mode. Write output only when errors occur. +The following @code{WHERE} clauses use indexes: +@example +... WHERE index_part1=1 AND index_part2=2 AND other_column=3 +... WHERE index=1 OR A=10 AND index=2 /* index = 1 OR index = 2 */ +... WHERE index_part1='hello' AND index_part_3=5 + /* optimized like "index_part1='hello'" */ +... WHERE index1=1 and index2=2 or index1=3 and index3=3; + /* Can use index on index1 but not on index2 or index 3 */ +@end example -@item -t, --test -Don't actually pack table, just test packing it. +These @code{WHERE} clauses do @strong{NOT} use indexes: +@example +... WHERE index_part2=1 AND index_part3=2 /* index_part_1 is not used */ +... WHERE index=1 OR A=10 /* Index is not used in both AND parts */ +... WHERE index_part1=1 OR index_part2=10 /* No index spans all rows */ +@end example -@item -T dir_name, --tmp_dir=dir_name -Use the named directory as the location in which to write the temporary table. +Note that in some cases @strong{MySQL} will not use an index, even if one +would be available. Some of the cases where this happens are: -@item -v, --verbose -Verbose mode. Write information about progress and packing result. +@itemize @bullet +@item +If the use of the index would require @strong{MySQL} to access more +than 30 % of the rows in the table. (In this case a table scan is +probably much faster, as this will require us to do much fewer seeks). +Note that if such a query uses @code{LIMIT} to only retrieve +part of the rows, @strong{MySQL} will use an index anyway, as it can +much more quickly find the few rows to return in the result. +@end itemize -@item -V, --version -Display version information and exit. +@cindex queries, speed of +@cindex permission checks, effect on speed +@cindex speed, of queries +@node Query Speed, Tips, MySQL indexes, Performance +@section Speed of Queries that Access or Update Data -@item -w, --wait +First, one thing that affects all queries: The more complex permission +system setup you have, the more overhead you get. -Wait and retry if table is in use. If the @code{mysqld} server was -invoked with the @code{--skip-locking} option, it is not a good idea to -invoke @code{myisampack} if the table might be updated during the -packing process. -@end table +If you do not have any @code{GRANT} statements done, @strong{MySQL} will +optimize the permission checking somewhat. So if you have a very high +volume it may be worth the time to avoid grants. Otherwise more +permission check results in a larger overhead. -@cindex examples, compressed tables -The sequence of commands shown below illustrates a typical table compression -session: +If your problem is with some explicit @strong{MySQL} function, you can +always time this in the @strong{MySQL} client: @example -shell> ls -l station.* --rw-rw-r-- 1 monty my 994128 Apr 17 19:00 station.MYD --rw-rw-r-- 1 monty my 53248 Apr 17 19:00 station.MYI --rw-rw-r-- 1 monty my 5767 Apr 17 19:00 station.frm - -shell> myisamchk -dvv station +mysql> select benchmark(1000000,1+1); ++------------------------+ +| benchmark(1000000,1+1) | ++------------------------+ +| 0 | ++------------------------+ +1 row in set (0.32 sec) +@end example -MyISAM file: station -Isam-version: 2 -Creation time: 1996-03-13 10:08:58 -Recover time: 1997-02-02 3:06:43 -Data records: 1192 Deleted blocks: 0 -Datafile: Parts: 1192 Deleted data: 0 -Datafile pointer (bytes): 2 Keyfile pointer (bytes): 2 -Max datafile length: 54657023 Max keyfile length: 33554431 -Recordlength: 834 -Record format: Fixed length +The above shows that @strong{MySQL} can execute 1,000,000 @code{+} +expressions in 0.32 seconds on a @code{PentiumII 400MHz}. -table description: -Key Start Len Index Type Root Blocksize Rec/key -1 2 4 unique unsigned long 1024 1024 1 -2 32 30 multip. text 10240 1024 1 +All @strong{MySQL} functions should be very optimized, but there may be +some exceptions, and the @code{benchmark(loop_count,expression)} is a +great tool to find out if this is a problem with your query. -Field Start Length Type -1 1 1 -2 2 4 -3 6 4 -4 10 1 -5 11 20 -6 31 1 -7 32 30 -8 62 35 -9 97 35 -10 132 35 -11 167 4 -12 171 16 -13 187 35 -14 222 4 -15 226 16 -16 242 20 -17 262 20 -18 282 20 -19 302 30 -20 332 4 -21 336 4 -22 340 1 -23 341 8 -24 349 8 -25 357 8 -26 365 2 -27 367 2 -28 369 4 -29 373 4 -30 377 1 -31 378 2 -32 380 8 -33 388 4 -34 392 4 -35 396 4 -36 400 4 -37 404 1 -38 405 4 -39 409 4 -40 413 4 -41 417 4 -42 421 4 -43 425 4 -44 429 20 -45 449 30 -46 479 1 -47 480 1 -48 481 79 -49 560 79 -50 639 79 -51 718 79 -52 797 8 -53 805 1 -54 806 1 -55 807 20 -56 827 4 -57 831 4 +@menu +* Estimating performance:: Estimating query performance +* SELECT speed:: Speed of @code{SELECT} queries +* Where optimizations:: How MySQL optimizes @code{WHERE} clauses +* DISTINCT optimization:: How MySQL Optimizes @code{DISTINCT} +* LEFT JOIN optimization:: How MySQL optimizes @code{LEFT JOIN} +* LIMIT optimization:: How MySQL optimizes @code{LIMIT} +* Insert speed:: Speed of @code{INSERT} queries +* Update speed:: Speed of @code{UPDATE} queries +* Delete speed:: Speed of @code{DELETE} queries +@end menu -shell> myisampack station.MYI -Compressing station.MYI: (1192 records) -- Calculating statistics +@cindex estimating, query performance +@cindex queries, estimating performance +@cindex performance, estimating +@node Estimating performance, SELECT speed, Query Speed, Query Speed +@subsection Estimating Query Performance -normal: 20 empty-space: 16 empty-zero: 12 empty-fill: 11 -pre-space: 0 end-space: 12 table-lookups: 5 zero: 7 -Original trees: 57 After join: 17 -- Compressing file -87.14% +In most cases you can estimate the performance by counting disk seeks. +For small tables, you can usually find the row in 1 disk seek (as the +index is probably cached). For bigger tables, you can estimate that +(using B++ tree indexes) you will need: @code{log(row_count) / +log(index_block_length / 3 * 2 / (index_length + data_pointer_length)) + +1} seeks to find a row. -shell> ls -l station.* --rw-rw-r-- 1 monty my 127874 Apr 17 19:00 station.MYD --rw-rw-r-- 1 monty my 55296 Apr 17 19:04 station.MYI --rw-rw-r-- 1 monty my 5767 Apr 17 19:00 station.frm +In @strong{MySQL} an index block is usually 1024 bytes and the data +pointer is usually 4 bytes. A 500,000 row table with an +index length of 3 (medium integer) gives you: +@code{log(500,000)/log(1024/3*2/(3+4)) + 1} = 4 seeks. -shell> myisamchk -dvv station +As the above index would require about 500,000 * 7 * 3/2 = 5.2M, +(assuming that the index buffers are filled to 2/3, which is typical) +you will probably have much of the index in memory and you will probably +only need 1-2 calls to read data from the OS to find the row. -MyISAM file: station -Isam-version: 2 -Creation time: 1996-03-13 10:08:58 -Recover time: 1997-04-17 19:04:26 -Data records: 1192 Deleted blocks: 0 -Datafile: Parts: 1192 Deleted data: 0 -Datafilepointer (bytes): 3 Keyfile pointer (bytes): 1 -Max datafile length: 16777215 Max keyfile length: 131071 -Recordlength: 834 -Record format: Compressed +For writes, however, you will need 4 seek requests (as above) to find +where to place the new index and normally 2 seeks to update the index +and write the row. -table description: -Key Start Len Index Type Root Blocksize Rec/key -1 2 4 unique unsigned long 10240 1024 1 -2 32 30 multip. text 54272 1024 1 +Note that the above doesn't mean that your application will slowly +degenerate by N log N! As long as everything is cached by the OS or SQL +server things will only go marginally slower while the table gets +bigger. After the data gets too big to be cached, things will start to +go much slower until your applications is only bound by disk-seeks +(which increase by N log N). To avoid this, increase the index cache as +the data grows. @xref{Server parameters}. -Field Start Length Type Huff tree Bits -1 1 1 constant 1 0 -2 2 4 zerofill(1) 2 9 -3 6 4 no zeros, zerofill(1) 2 9 -4 10 1 3 9 -5 11 20 table-lookup 4 0 -6 31 1 3 9 -7 32 30 no endspace, not_always 5 9 -8 62 35 no endspace, not_always, no empty 6 9 -9 97 35 no empty 7 9 -10 132 35 no endspace, not_always, no empty 6 9 -11 167 4 zerofill(1) 2 9 -12 171 16 no endspace, not_always, no empty 5 9 -13 187 35 no endspace, not_always, no empty 6 9 -14 222 4 zerofill(1) 2 9 -15 226 16 no endspace, not_always, no empty 5 9 -16 242 20 no endspace, not_always 8 9 -17 262 20 no endspace, no empty 8 9 -18 282 20 no endspace, no empty 5 9 -19 302 30 no endspace, no empty 6 9 -20 332 4 always zero 2 9 -21 336 4 always zero 2 9 -22 340 1 3 9 -23 341 8 table-lookup 9 0 -24 349 8 table-lookup 10 0 -25 357 8 always zero 2 9 -26 365 2 2 9 -27 367 2 no zeros, zerofill(1) 2 9 -28 369 4 no zeros, zerofill(1) 2 9 -29 373 4 table-lookup 11 0 -30 377 1 3 9 -31 378 2 no zeros, zerofill(1) 2 9 -32 380 8 no zeros 2 9 -33 388 4 always zero 2 9 -34 392 4 table-lookup 12 0 -35 396 4 no zeros, zerofill(1) 13 9 -36 400 4 no zeros, zerofill(1) 2 9 -37 404 1 2 9 -38 405 4 no zeros 2 9 -39 409 4 always zero 2 9 -40 413 4 no zeros 2 9 -41 417 4 always zero 2 9 -42 421 4 no zeros 2 9 -43 425 4 always zero 2 9 -44 429 20 no empty 3 9 -45 449 30 no empty 3 9 -46 479 1 14 4 -47 480 1 14 4 -48 481 79 no endspace, no empty 15 9 -49 560 79 no empty 2 9 -50 639 79 no empty 2 9 -51 718 79 no endspace 16 9 -52 797 8 no empty 2 9 -53 805 1 17 1 -54 806 1 3 9 -55 807 20 no empty 3 9 -56 827 4 no zeros, zerofill(2) 2 9 -57 831 4 no zeros, zerofill(1) 2 9 -@end example +@cindex speed, of queries +@findex SELECT speed -The information printed by @code{myisampack} is described below: +@node SELECT speed, Where optimizations, Estimating performance, Query Speed +@subsection Speed of @code{SELECT} Queries -@table @code -@item normal -The number of columns for which no extra packing is used. +In general, when you want to make a slow @code{SELECT ... WHERE} faster, the +first thing to check is whether or not you can add an index. @xref{MySQL +indexes, , @strong{MySQL} indexes}. All references between different tables +should usually be done with indexes. You can use the @code{EXPLAIN} command +to determine which indexes are used for a @code{SELECT}. +@xref{EXPLAIN, , @code{EXPLAIN}}. -@item empty-space -The number of columns containing -values that are only spaces; these will occupy 1 bit. +Some general tips: -@item empty-zero -The number of columns containing -values that are only binary 0's; these will occupy 1 bit. +@itemize @bullet +@item +To help @strong{MySQL} optimize queries better, run @code{myisamchk +--analyze} on a table after it has been loaded with relevant data. This +updates a value for each index part that indicates the average number of +rows that have the same value. (For unique indexes, this is always 1, +of course.). @strong{MySQL} will use this to decide which index to +choose when you connect two tables with 'a non-constant expression'. +You can check the result from the @code{analyze} run by doing @code{SHOW +INDEX FROM table_name} and examining the @code{Cardinality} column. -@item empty-fill -The number of integer columns that don't occupy the full byte range of their -type; these are changed to a smaller type (for example, an @code{INTEGER} -column may be changed to @code{MEDIUMINT}). +@item +To sort an index and data according to an index, use @code{myisamchk +--sort-index --sort-records=1} (if you want to sort on index 1). If you +have a unique index from which you want to read all records in order +according to that index, this is a good way to make that faster. Note, +however, that this sorting isn't written optimally and will take a long +time for a large table! +@end itemize -@item pre-space -The number of decimal columns that are stored with leading spaces. In this -case, each value will contain a count for the number of leading spaces. +@cindex optimizations +@findex WHERE +@node Where optimizations, DISTINCT optimization, SELECT speed, Query Speed +@subsection How MySQL Optimizes @code{WHERE} Clauses -@item end-space -The number of columns that have a lot of trailing spaces. In this case, each -value will contain a count for the number of trailing spaces. +The @code{WHERE} optimizations are put in the @code{SELECT} part here because +they are mostly used with @code{SELECT}, but the same optimizations apply for +@code{WHERE} in @code{DELETE} and @code{UPDATE} statements. -@item table-lookup -The column had only a small number of different values, which were -converted to an @code{ENUM} before Huffman compression. +Also note that this section is incomplete. @strong{MySQL} does many +optimizations, and we have not had time to document them all. -@item zero -The number of columns for which all values are zero. +Some of the optimizations performed by @strong{MySQL} are listed below: -@item Original trees -The initial number of Huffman trees. +@itemize @bullet +@item +Removal of unnecessary parentheses: +@example + ((a AND b) AND c OR (((a AND b) AND (c AND d)))) +-> (a AND b AND c) OR (a AND b AND c AND d) +@end example +@item +Constant folding: +@example + (a<b AND b=c) AND a=5 +-> b>5 AND b=c AND a=5 +@end example +@item +Constant condition removal (needed because of constant folding): +@example + (B>=5 AND B=5) OR (B=6 AND 5=5) OR (B=7 AND 5=6) +-> B=5 OR B=6 +@end example +@item +Constant expressions used by indexes are evaluated only once. +@item +@code{COUNT(*)} on a single table without a @code{WHERE} is retrieved +directly from the table information. This is also done for any @code{NOT NULL} +expression when used with only one table. +@item +Early detection of invalid constant expressions. @strong{MySQL} quickly +detects that some @code{SELECT} statements are impossible and returns no rows. +@item +@code{HAVING} is merged with @code{WHERE} if you don't use @code{GROUP BY} +or group functions (@code{COUNT()}, @code{MIN()}...). +@item +For each sub-join, a simpler @code{WHERE} is constructed to get a fast +@code{WHERE} evaluation for each sub-join and also to skip records as +soon as possible. +@cindex constant table +@cindex tables, constant +@item +All constant tables are read first, before any other tables in the query. +A constant table is: +@itemize @minus +@item +An empty table or a table with 1 row. +@item +A table that is used with a @code{WHERE} clause on a @code{UNIQUE} +index, or a @code{PRIMARY KEY}, where all index parts are used with constant +expressions and the index parts are defined as @code{NOT NULL}. +@end itemize +All the following tables are used as constant tables: +@example +mysql> SELECT * FROM t WHERE primary_key=1; +mysql> SELECT * FROM t1,t2 + WHERE t1.primary_key=1 AND t2.primary_key=t1.id; +@end example -@item After join -The number of distinct Huffman trees left after joining -trees to save some header space. -@end table +@item +The best join combination to join the tables is found by trying all +possibilities. If all columns in @code{ORDER BY} and in @code{GROUP +BY} come from the same table, then this table is preferred first when +joining. +@item +If there is an @code{ORDER BY} clause and a different @code{GROUP BY} +clause, or if the @code{ORDER BY} or @code{GROUP BY} contains columns +from tables other than the first table in the join queue, a temporary +table is created. +@item +If you use @code{SQL_SMALL_RESULT}, @strong{MySQL} will use an in-memory +temporary table. +@item +Each table index is queried, and the best index that spans fewer than 30% of +the rows is used. If no such index can be found, a quick table scan is used. +@item +In some cases, @strong{MySQL} can read rows from the index without even +consulting the data file. If all columns used from the index are numeric, +then only the index tree is used to resolve the query. +@item +Before each record is output, those that do not match the @code{HAVING} clause +are skipped. +@end itemize -After a table has been compressed, @code{myisamchk -dvv} prints additional -information about each field: +Some examples of queries that are very fast: -@table @code -@item Type -The field type may contain the following descriptors: +@example +mysql> SELECT COUNT(*) FROM tbl_name; +mysql> SELECT MIN(key_part1),MAX(key_part1) FROM tbl_name; +mysql> SELECT MAX(key_part2) FROM tbl_name + WHERE key_part_1=constant; +mysql> SELECT ... FROM tbl_name + ORDER BY key_part1,key_part2,... LIMIT 10; +mysql> SELECT ... FROM tbl_name + ORDER BY key_part1 DESC,key_part2 DESC,... LIMIT 10; +@end example -@table @code -@item constant -All rows have the same value. +The following queries are resolved using only the index tree (assuming +the indexed columns are numeric): -@item no endspace -Don't store endspace. +@example +mysql> SELECT key_part1,key_part2 FROM tbl_name WHERE key_part1=val; +mysql> SELECT COUNT(*) FROM tbl_name + WHERE key_part1=val1 AND key_part2=val2; +mysql> SELECT key_part2 FROM tbl_name GROUP BY key_part1; +@end example -@item no endspace, not_always -Don't store endspace and don't do end space compression for all values. +The following queries use indexing to retrieve the rows in sorted +order without a separate sorting pass: -@item no endspace, no empty -Don't store endspace. Don't store empty values. +@example +mysql> SELECT ... FROM tbl_name ORDER BY key_part1,key_part2,... ; +mysql> SELECT ... FROM tbl_name ORDER BY key_part1 DESC,key_part2 DESC,... ; +@end example -@item table-lookup -The column was converted to an @code{ENUM}. +@findex DISTINCT +@cindex optimizing, DISTINCT +@node DISTINCT optimization, LEFT JOIN optimization, Where optimizations, Query Speed +@subsection How MySQL Optimizes @code{DISTINCT} -@item zerofill(n) -The most significant @code{n} bytes in the value are always 0 and are not -stored. +@code{DISTINCT} is converted to a @code{GROUP BY} on all columns, +@code{DISTINCT} combined with @code{ORDER BY} will in many cases also +need a temporary table. -@item no zeros -Don't store zeros. +When combining @code{LIMIT #} with @code{DISTINCT}, @strong{MySQL} will stop +as soon as it finds @code{#} unique rows. -@item always zero -0 values are stored in 1 bit. -@end table +If you don't use columns from all used tables, @strong{MySQL} will stop +the scanning of the not used tables as soon as it has found the first match. -@item Huff tree -The Huffman tree associated with the field. +@example +SELECT DISTINCT t1.a FROM t1,t2 where t1.a=t2.a; +@end example -@item Bits -The number of bits used in the Huffman tree. -@end table +In the case, assuming t1 is used before t2 (check with @code{EXPLAIN}), then +@strong{MySQL} will stop reading from t2 (for that particular row in t1) +when the first row in t2 is found. -After you have run @code{pack_isam}/@code{myisampack} you must run -@code{isamchk}/@code{myisamchk} to re-create the index. At this time you -can also sort the index blocks and create statistics needed for -the @strong{MySQL} optimizer to work more efficiently: +@findex LEFT JOIN +@cindex optimizing, LEFT JOIN +@node LEFT JOIN optimization, LIMIT optimization, DISTINCT optimization, Query Speed +@subsection How MySQL Optimizes @code{LEFT JOIN} and @code{RIGHT JOIN} -@example -myisamchk -rq --analyze --sort-index table_name.MYI -isamchk -rq --analyze --sort-index table_name.ISM -@end example +@code{A LEFT JOIN B} in @strong{MySQL} is implemented as follows: -After you have installed the packed table into the @strong{MySQL} database -directory you should do @code{mysqladmin flush-tables} to force @code{mysqld} -to start using the new table. +@itemize @bullet +@item +The table @code{B} is set to be dependent on table @code{A} and all tables +that @code{A} is dependent on. +@item +The table @code{A} is set to be dependent on all tables (except @code{B}) +that are used in the @code{LEFT JOIN} condition. +@item +All @code{LEFT JOIN} conditions are moved to the @code{WHERE} clause. +@item +All standard join optimizations are done, with the exception that a table is +always read after all tables it is dependent on. If there is a circular +dependence then @strong{MySQL} will issue an error. +@item +All standard @code{WHERE} optimizations are done. +@item +If there is a row in @code{A} that matches the @code{WHERE} clause, but there +wasn't any row in @code{B} that matched the @code{LEFT JOIN} condition, +then an extra @code{B} row is generated with all columns set to @code{NULL}. +@item +If you use @code{LEFT JOIN} to find rows that don't exist in some +table and you have the following test: @code{column_name IS NULL} in the +@code{WHERE} part, where column_name is a column that is declared as +@code{NOT NULL}, then @strong{MySQL} will stop searching after more rows +(for a particular key combination) after it has found one row that +matches the @code{LEFT JOIN} condition. +@end itemize -If you want to unpack a packed table, you can do this with the -@code{--unpack} option to @code{isamchk} or @code{myisamchk}. +@code{RIGHT JOIN} is implemented analogously as @code{LEFT JOIN}. +The table read order forced by @code{LEFT JOIN} and @code{STRAIGHT JOIN} +will help the join optimizer (which calculates in which order tables +should be joined) to do its work much more quickly, as there are fewer +table permutations to check. -@cindex installation maintenance -@cindex maintaining, tables -@cindex tables, maintaining -@cindex databases, maintaining -@cindex @code{myisamchk} -@cindex @code{mysqlcheck} -@cindex crash, recovery -@cindex recovery, from crash -@node Maintenance, Adding functions, Tools, Top -@chapter Maintaining a MySQL Installation +Note that the above means that if you do a query of type: -@menu -* Table maintenance:: Table maintenance and crash recovery -* Using mysqlcheck:: Using mysqlcheck for maintenance and recovery -* Maintenance regimen:: Setting up a table maintenance regimen -* Table-info:: Getting information about a table -* Crash recovery:: Using @code{myisamchk} for crash recovery -* Log file maintenance:: Log file maintenance -@end menu +@example +SELECT * FROM a,b LEFT JOIN c ON (c.key=a.key) LEFT JOIN d (d.key=a.key) WHERE b.key=d.key +@end example -This chapter covers what you should know about maintaining a @strong{MySQL} -distribution. You will learn how to care for your tables on a regular -basis, and what to do when disaster strikes. +@strong{MySQL} will do a full scan on @code{b} as the @code{LEFT +JOIN} will force it to be read before @code{d}. -@node Table maintenance, Using mysqlcheck, Maintenance, Maintenance -@section Using @code{myisamchk} for Table Maintenance and Crash Recovery +The fix in this case is to change the query to: -Starting with @strong{MySQL} Version 3.23.13, you can check MyISAM -tables with the @code{CHECK TABLE} command. @xref{CHECK TABLE}. You can -repair tables with the @code{REPAIR TABLE} command. @xref{REPAIR TABLE}. +@example +SELECT * FROM b,a LEFT JOIN c ON (c.key=a.key) LEFT JOIN d (d.key=a.key) WHERE b.key=d.key +@end example -To check/repair MyISAM tables (@code{.MYI} and @code{.MYD}) you should -use the @code{myisamchk} utility. To check/repair ISAM tables -(@code{.ISM} and @code{.ISD}) you should use the @code{isamchk} -utility. @xref{Table types}. +@cindex optimizing, LIMIT +@findex LIMIT +@node LIMIT optimization, Insert speed, LEFT JOIN optimization, Query Speed +@subsection How MySQL Optimizes @code{LIMIT} -In the following text we will talk about @code{myisamchk}, but everything -also applies to the old @code{isamchk}. +In some cases @strong{MySQL} will handle the query differently when you are +using @code{LIMIT #} and not using @code{HAVING}: -You can use the @code{myisamchk} utility to get information about your -database tables, check and repair them, or optimize them. The following -sections describe how to invoke @code{myisamchk} (including a -description of its options), how to set up a table maintenance schedule, -and how to use @code{myisamchk} to perform its various functions. +@itemize @bullet +@item +If you are selecting only a few rows with @code{LIMIT}, @strong{MySQL} +will use indexes in some cases when it normally would prefer to do a +full table scan. +@item +If you use @code{LIMIT #} with @code{ORDER BY}, @strong{MySQL} will end the +sorting as soon as it has found the first @code{#} lines instead of sorting +the whole table. +@item +When combining @code{LIMIT #} with @code{DISTINCT}, @strong{MySQL} will stop +as soon as it finds @code{#} unique rows. +@item +In some cases a @code{GROUP BY} can be resolved by reading the key in order +(or do a sort on the key) and then calculate summaries until the +key value changes. In this case @code{LIMIT #} will not calculate any +unnecessary @code{GROUP BY}'s. +@item +As soon as @strong{MySQL} has sent the first @code{#} rows to the client, it +will abort the query. +@item +@code{LIMIT 0} will always quickly return an empty set. This is useful +to check the query and to get the column types of the result columns. +@item +The size of temporary tables uses the @code{LIMIT #} to calculate how much +space is needed to resolve the query. +@end itemize -You can, in most cases, also use the command @code{OPTIMIZE TABLES} to -optimize and repair tables, but this is not as fast or reliable (in case -of real fatal errors) as @code{myisamchk}. On the other hand, -@code{OPTIMIZE TABLE} is easier to use and you don't have to worry about -flushing tables. -@xref{OPTIMIZE TABLE, , @code{OPTIMIZE TABLE}}. +@cindex speed, inserting +@cindex inserting, speed of +@node Insert speed, Update speed, LIMIT optimization, Query Speed +@subsection Speed of @code{INSERT} Queries -Even that the repair in @code{myisamchk} is quite secure, it's always a -good idea to make a backup BEFORE doing a repair (or anything that could -make a lot of changes to a table) +The time to insert a record consists approximately of: -@menu -* myisamchk syntax:: @code{myisamchk} invocation syntax -* myisamchk memory:: @code{myisamchk} memory usage -@end menu +@itemize @bullet +@item +Connect: (3) +@item +Sending query to server: (2) +@item +Parsing query: (2) +@item +Inserting record: (1 x size of record) +@item +Inserting indexes: (1 x number of indexes) +@item +Close: (1) +@end itemize -@node myisamchk syntax, myisamchk memory, Table maintenance, Table maintenance -@subsection @code{myisamchk} Invocation Syntax +where the numbers are somewhat proportional to the overall time. This +does not take into consideration the initial overhead to open tables +(which is done once for each concurrently running query). -@code{myisamchk} is invoked like this: +The size of the table slows down the insertion of indexes by N log N +(B-trees). -@example -shell> myisamchk [options] tbl_name -@end example +Some ways to speed up inserts: -The @code{options} specify what you want @code{myisamchk} to do. They are -described below. (You can also get a list of options by invoking -@code{myisamchk --help}.) With no options, @code{myisamchk} simply checks your -table. To get more information or to tell @code{myisamchk} to take corrective -action, specify options as described below and in the following sections. +@itemize @bullet +@item +If you are inserting many rows from the same client at the same time, use +multiple value lists @code{INSERT} statements. This is much faster (many +times in some cases) than using separate @code{INSERT} statements. +@item +If you are inserting a lot of rows from different clients, you can get +higher speed by using the @code{INSERT DELAYED} statement. @xref{INSERT, +, @code{INSERT}}. +@item +Note that with @code{MyISAM} you can insert rows at the same time +@code{SELECT}s are running if there are no deleted rows in the tables. +@item +When loading a table from a text file, use @code{LOAD DATA INFILE}. This +is usually 20 times faster than using a lot of @code{INSERT} statements. +@xref{LOAD DATA, , @code{LOAD DATA}}. +@item +It is possible with some extra work to make @code{LOAD DATA INFILE} run even +faster when the table has many indexes. Use the following procedure: -@code{tbl_name} is the database table you want to check/repair. If you run -@code{myisamchk} somewhere other than in the database directory, you must -specify the path to the file, because @code{myisamchk} has no idea where your -database is located. Actually, @code{myisamchk} doesn't care whether or not -the files you are working on are located in a database directory; you can -copy the files that correspond to a database table into another location and -perform recovery operations on them there. +@enumerate +@item +Optionally create the table with @code{CREATE TABLE}. For example, using +@code{mysql} or Perl-DBI. -You can name several tables on the @code{myisamchk} command line if you -wish. You can also specify a name as an index file -name (with the @file{.MYI} suffix), which allows you to specify all -tables in a directory by using the pattern @file{*.MYI}. -For example, if you are in a database directory, you can check all the -tables in the directory like this: +@item +Execute a @code{FLUSH TABLES} statement or the shell command @code{mysqladmin +flush-tables}. -@example -shell> myisamchk *.MYI -@end example +@item +Use @code{myisamchk --keys-used=0 -rq /path/to/db/tbl_name}. This will +remove all usage of all indexes from the table. -If you are not in the database directory, you can check all the tables there -by specifying the path to the directory: +@item +Insert data into the table with @code{LOAD DATA INFILE}. This will not +update any indexes and will therefore be very fast. -@example -shell> myisamchk /path/to/database_dir/*.MYI -@end example +@item +If you are going to only read the table in the future, run @code{myisampack} +on it to make it smaller. @xref{Compressed format}. -You can even check all tables in all databases by specifying a wild card -with the path to the @strong{MySQL} data directory: +@item +Re-create the indexes with @code{myisamchk -r -q +/path/to/db/tbl_name}. This will create the index tree in memory before +writing it to disk, which is much faster because it avoids lots of disk +seeks. The resulting index tree is also perfectly balanced. -@example -shell> myisamchk /path/to/datadir/*/*.MYI -@end example +@item +Execute a @code{FLUSH TABLES} statement or the shell command @code{mysqladmin +flush-tables}. +@end enumerate -The recommended way to quickly check all tables is: +This procedure will be built into @code{LOAD DATA INFILE} in some future +version of @strong{MySQL}. +@item +You can speed up insertions by locking your tables: @example -myisamchk --silent --fast /path/to/datadir/*/*.MYI -isamchk --silent /path/to/datadir/*/*.ISM +mysql> LOCK TABLES a WRITE; +mysql> INSERT INTO a VALUES (1,23),(2,34),(4,33); +mysql> INSERT INTO a VALUES (8,26),(6,29); +mysql> UNLOCK TABLES; @end example -If you want to check all tables and repair all tables that are corrupted, -you can use the following line: +The main speed difference is that the index buffer is flushed to disk only +once, after all @code{INSERT} statements have completed. Normally there would +be as many index buffer flushes as there are different @code{INSERT} +statements. Locking is not needed if you can insert all rows with a single +statement. + +Locking will also lower the total time of multi-connection tests, but the +maximum wait time for some threads will go up (because they wait for +locks). For example: @example -myisamchk --silent --force --fast --update-state -O key_buffer=64M -O sort_buffer=64M -O read_buffer=1M -O write_buffer=1M /path/to/datadir/*/*.MYI -isamchk --silent --force -O key_buffer=64M -O sort_buffer=64M -O read_buffer=1M -O write_buffer=1M /path/to/datadir/*/*.ISM +thread 1 does 1000 inserts +thread 2, 3, and 4 does 1 insert +thread 5 does 1000 inserts @end example -The above assumes that you have more than 64 M free. +If you don't use locking, 2, 3, and 4 will finish before 1 and 5. If you +use locking, 2, 3, and 4 probably will not finish before 1 or 5, but the +total time should be about 40% faster. -Note that if you get an error like: +As @code{INSERT}, @code{UPDATE}, and @code{DELETE} operations are very +fast in @strong{MySQL}, you will obtain better overall performance by +adding locks around everything that does more than about 5 inserts or +updates in a row. If you do very many inserts in a row, you could do a +@code{LOCK TABLES} followed by an @code{UNLOCK TABLES} once in a while +(about each 1000 rows) to allow other threads access to the table. This +would still result in a nice performance gain. -@example -myisamchk: warning: 1 clients is using or hasn't closed the table properly -@end example +Of course, @code{LOAD DATA INFILE} is much faster for loading data. +@end itemize -This means that you are trying to check a table that has been updated by -the another program (like the @code{mysqld} server) that hasn't yet closed -the file or that has died without closing the file properly. +To get some more speed for both @code{LOAD DATA INFILE} and +@code{INSERT}, enlarge the key buffer. @xref{Server parameters}. -If you @code{mysqld} is running, you must force a sync/close of all -tables with @code{FLUSH TABLES} and ensure that no one is using the -tables while you are running @code{myisamchk}. In @strong{MySQL} Version 3.23 -the easiest way to avoid this problem is to use @code{CHECK TABLE} -instead of @code{myisamchk} to check tables. +@node Update speed, Delete speed, Insert speed, Query Speed +@subsection Speed of @code{UPDATE} Queries -@menu -* myisamchk general options:: -* myisamchk check options:: -* myisamchk repair options:: -* myisamchk other options:: -@end menu +Update queries are optimized as a @code{SELECT} query with the additional +overhead of a write. The speed of the write is dependent on the size of +the data that is being updated and the number of indexes that are +updated. Indexes that are not changed will not be updated. -@cindex options, @code{myisamchk} -@cindex @code{myisamchk}, options -@node myisamchk general options, myisamchk check options, myisamchk syntax, myisamchk syntax -@subsubsection General Options for @code{myisamchk} +Also, another way to get fast updates is to delay updates and then do +many updates in a row later. Doing many updates in a row is much quicker +than doing one at a time if you lock the table. -@code{myisamchk} supports the following options. +Note that, with dynamic record format, updating a record to +a longer total length may split the record. So if you do this often, +it is very important to @code{OPTIMIZE TABLE} sometimes. +@xref{OPTIMIZE TABLE, , @code{OPTIMIZE TABLE}}. -@table @code -@item -# or --debug=debug_options -Output debug log. The @code{debug_options} string often is -@code{'d:t:o,filename'}. -@item -? or --help -Display a help message and exit. -@item -O var=option, --set-variable var=option -Set the value of a variable. The possible variables and their default values -for myisamchk can be examined with @code{myisamchk --help}: -@multitable @columnfractions .3 .7 -@item key_buffer_size @tab 523264 -@item read_buffer_size @tab 262136 -@item write_buffer_size @tab 262136 -@item sort_buffer_size @tab 2097144 -@item sort_key_blocks @tab 16 -@item decode_bits @tab 9 -@end multitable +@node Delete speed, , Update speed, Query Speed +@subsection Speed of @code{DELETE} Queries -@code{sort_buffer_size} is used when the keys are repaired by sorting -keys, which is the normal case when you use @code{--recover}. +If you want to delete all rows in the table, you should use +@code{TRUNCATE TABLE table_name}. @xref{TRUNCATE}. -@code{key_buffer_size} is used when you are checking the table with -@code{--extended-check} or when the keys are repaired by inserting key -row by row in to the table (like when doing normal inserts). Repairing -through the key buffer is used in the following cases: +The time to delete a record is exactly proportional to the number of +indexes. To delete records more quickly, you can increase the size of +the index cache. @xref{Server parameters}. + +@cindex optimization, tips +@cindex tips, optimization +@node Tips, Benchmarks, Query Speed, Performance +@section Other Optimization Tips + +Unsorted tips for faster systems: @itemize @bullet @item -If you use @code{--safe-recover}. +Use persistent connections to the database to avoid the connection +overhead. If you can't use persistent connections and you are doing a +lot of new connections to the database, you may want to change the value +of the @code{thread_cache_size} variable. @xref{Server parameters}. @item -If you are using a @code{FULLTEXT} index. +Always check that all your queries really use the indexes you have created +in the tables. In @strong{MySQL} you can do this with the @code{EXPLAIN} +command. @xref{EXPLAIN, Explain, Explain, manual}. @item -If the temporary files needed to sort the keys would be more than twice -as big as when creating the key file directly. This is often the case -when you have big @code{CHAR}, @code{VARCHAR} or @code{TEXT} keys as the -sort needs to store the whole keys during sorting. If you have lots -of temporary space and you can force @code{myisamchk} to repair by sorting -you can use the @code{--sort-recover} option. -@end itemize +Try to avoid complex @code{SELECT} queries on tables that are updated a +lot. This is to avoid problems with table locking. +@item +The new @code{MyISAM} tables can insert rows in a table without deleted +rows at the same time another table is reading from it. If this is important +for you, you should consider methods where you don't have to delete rows +or run @code{OPTIMIZE TABLE} after you have deleted a lot of rows. +@item +Use @code{ALTER TABLE ... ORDER BY expr1,expr2...} if you mostly +retrieve rows in expr1,expr2.. order. By using this option after big +changes to the table, you may be able to get higher performance. +@item +In some cases it may make sense to introduce a column that is 'hashed' +based on information from other columns. If this column is short and +reasonably unique it may be much faster than a big index on many +columns. In @strong{MySQL} it's very easy to use this extra column: +@code{SELECT * FROM table_name WHERE hash=MD5(concat(col1,col2)) +AND col_1='constant' AND col_2='constant'} +@item +For tables that change a lot you should try to avoid all @code{VARCHAR} +or @code{BLOB} columns. You will get dynamic row length as soon as you +are using a single @code{VARCHAR} or @code{BLOB} column. @xref{Table +types}. +@item +It's not normally useful to split a table into different tables just +because the rows gets 'big'. To access a row, the biggest performance +hit is the disk seek to find the first byte of the row. After finding +the data most new disks can read the whole row fast enough for most +applications. The only cases where it really matters to split up a table is if +it's a dynamic row size table (see above) that you can change to a fixed +row size, or if you very often need to scan the table and don't need +most of the columns. @xref{Table types}. +@item +If you very often need to calculate things based on information from a +lot of rows (like counts of things), it's probably much better to +introduce a new table and update the counter in real time. An update of +type @code{UPDATE table set count=count+1 where index_column=constant} +is very fast! -Reparing through the key buffer takes much less disk space than using -sorting, but is also much slower. +This is really important when you use databases like @strong{MySQL} that +only have table locking (multiple readers / single writers). This will +also give better performance with most databases, as the row locking +manager in this case will have less to do. +@item +If you need to collect statistics from big log tables, use summary tables +instead of scanning the whole table. Maintaining the summaries should be +much faster than trying to do statistics 'live'. It's much faster to +regenerate new summary tables from the logs when things change +(depending on business decisions) than to have to change the running +application! +@item +If possible, one should classify reports as 'live' or 'statistical', +where data needed for statistical reports are only generated based on +summary tables that are generated from the actual data. +@item +Take advantage of the fact that columns have default values. Insert +values explicitly only when the value to be inserted differs from the +default. This reduces the parsing that @strong{MySQL} need to do and +improves the insert speed. +@item +In some cases it's convenient to pack and store data into a blob. In this +case you have to add some extra code in your application to pack/unpack +things in the blob, but this may save a lot of accesses at some stage. +This is practical when you have data that doesn't conform to a static +table structure. +@item +Normally you should try to keep all data non-redundant (what +is called 3rd normal form in database theory), but you should not be +afraid of duplicating things or creating summary tables if you need these +to gain more speed. +@item +Stored procedures or UDF (user-defined functions) may be a good way to +get more performance. In this case you should, however, always have a way +to do this some other (slower) way if you use some database that doesn't +support this. +@item +You can always gain something by caching queries/answers in your +application and trying to do many inserts/updates at the same time. If +your database supports lock tables (like @strong{MySQL} and Oracle), +this should help to ensure that the index cache is only flushed once +after all updates. +@item +Use @code{INSERT /*! DELAYED */} when you do not need to know when your +data is written. This speeds things up because many records can be written +with a single disk write. +@item +Use @code{INSERT /*! LOW_PRIORITY */} when you want your selects to be +more important. +@item +Use @code{SELECT /*! HIGH_PRIORITY */} to get selects that jump the +queue. That is, the select is done even if there is somebody waiting to +do a write. +@item +Use the multi-line @code{INSERT} statement to store many rows with one +SQL command (many SQL servers supports this). +@item +Use @code{LOAD DATA INFILE} to load bigger amounts of data. This is +faster than normal inserts and will be even faster when @code{myisamchk} +is integrated in @code{mysqld}. +@item +Use @code{AUTO_INCREMENT} columns to make unique values. +@item +Use @code{OPTIMIZE TABLE} once in a while to avoid fragmentation when +using dynamic table format. @xref{OPTIMIZE TABLE, , @code{OPTIMIZE TABLE}}. -If you want a faster repair, set the above variables to about 1/4 of your -available memory. You can set both variables to big values, as only one -of the above buffers will be used at a time. - -@item -s or --silent -Silent mode. Write output only when errors occur. You can use @code{-s} -twice (@code{-ss}) to make @code{myisamchk} very silent. -@item -v or --verbose -Verbose mode. Print more information. This can be used with @code{-d} and -@code{-e}. Use @code{-v} multiple times (@code{-vv}, @code{-vvv}) for more -verbosity! -@item -V or --version -Print the @code{myisamchk} version and exit. -@item -w or, --wait -Instead of giving an error if the table is locked, wait until the table -is unlocked before continuing. Note that if you are running @code{mysqld} -on the table with @code{--skip-locking}, the table can only be locked -by another @code{myisamchk} command. -@end table - -@cindex check options, myisamchk -@cindex tables, checking -@node myisamchk check options, myisamchk repair options, myisamchk general options, myisamchk syntax -@subsubsection Check Options for @code{myisamchk} - -@table @code -@item -c or --check -Check table for errors. This is the default operation if you are not -giving @code{myisamchk} any options that override this. - -@item -e or --extend-check -Check the table VERY thoroughly (which is quite slow if you have many -indexes). This option should only be used in extreme cases. Normally, -@code{myisamchk} or @code{myisamchk --medium-check} should, in most -cases, be able to find out if there are any errors in the table. - -If you are using @code{--extended-check} and have much memory, you should -increase the value of @code{key_buffer_size} a lot! - -@item -F or --fast -Check only tables that haven't been closed properly. -@item -C or --check-only-changed -Check only tables that have changed since the last check. -@item -f or --force -Restart @code{myisamchk} with @code{-r} (repair) on the table, if -@code{myisamchk} finds any errors in the table. -@item -i or --information -Print informational statistics about the table that is checked. -@item -m or --medium-check -Faster than extended-check, but only finds 99.99% of all errors. -Should, however, be good enough for most cases. -@item -U or --update-state -Store in the @file{.MYI} file when the table was checked and if the table crashed. This should be used to get full benefit of the -@code{--check-only-changed} option, but you shouldn't use this -option if the @code{mysqld} server is using the table and you are -running @code{mysqld} with @code{--skip-locking}. -@item -T or --read-only -Don't mark table as checked. This is useful if you use @code{myisamchk} -to check a table that is in use by some other application that doesn't -use locking (like @code{mysqld --skip-locking}). -@end table - -@cindex repair options, myisamchk -@cindex files, repairing -@node myisamchk repair options, myisamchk other options, myisamchk check options, myisamchk syntax -@subsubsection Repair Options for myisamchk - -The following options are used if you start @code{myisamchk} with -@code{-r} or @code{-o}: - -@table @code -@item -D # or --data-file-length=# -Max length of data file (when re-creating data file when it's 'full'). -@item -e or --extend-check -Try to recover every possible row from the data file. -Normally this will also find a lot of garbage rows. Don't use this option -if you are not totally desperate. -@item -f or --force -Overwrite old temporary files (@code{table_name.TMD}) instead of aborting. -@item -k # or keys-used=# -If you are using ISAM, tells the ISAM table handler to update only the -first @code{#} indexes. If you are using @code{MyISAM}, tells which keys -to use, where each binary bit stands for one key (first key is bit 0). -This can be used to get faster inserts! Deactivated indexes can be -reactivated by using @code{myisamchk -r}. keys. -@item -l or --no-symlinks -Do not follow symbolic links. Normally @code{myisamchk} repairs the -table a symlink points at. This option doesn't exist in MySQL 4.0, -as MySQL 4.0 will not remove symlinks during repair. -@item -r or --recover -Can fix almost anything except unique keys that aren't unique -(which is an extremely unlikely error with ISAM/MyISAM tables). -If you want to recover a table, this is the option to try first. Only if -myisamchk reports that the table can't be recovered by @code{-r}, you -should then try @code{-o}. (Note that in the unlikely case that @code{-r} -fails, the data file is still intact.) -If you have lots of memory, you should increase the size of -@code{sort_buffer_size}! -@item -o or --safe-recover -Uses an old recovery method (reads through all rows in order and updates -all index trees based on the found rows); this is a magnitude slower -than @code{-r}, but can handle a couple of very unlikely cases that -@code{-r} cannot handle. This recovery method also uses much less disk -space than @code{-r}. Normally one should always first repair with -@code{-r}, and only if this fails use @code{-o}. - -If you have lots of memory, you should increase the size of -@code{key_buffer_size}! -@item -n or --sort-recover -Force @code{myisamchk} to use sorting to resolve the keys even if the -temporary files should be very big. This will not have any effect if you have -fulltext keys in the table. +@item +Use @code{HEAP} tables to get more speed when possible. @xref{Table +types}. +@item +When using a normal Web server setup, images should be stored as +files. That is, store only a file reference in the database. The main +reason for this is that a normal Web server is much better at caching +files than database contents. So it it's much easier to get a fast +system if you are using files. +@item +Use in memory tables for non-critical data that are accessed often (like +information about the last shown banner for users that don't have +cookies). +@item +Columns with identical information in different tables should be +declared identical and have identical names. Before Version 3.23 you +got slow joins otherwise. -@item --character-sets-dir=... -Directory where character sets are stored. -@item --set-character-set=name -Change the character set used by the index -@item .t or --tmpdir=path -Path for storing temporary files. If this is not set, @code{myisamchk} will -use the environment variable @code{TMPDIR} for this. -@item -q or --quick -Faster repair by not modifying the data file. One can give a second -@code{-q} to force @code{myisamchk} to modify the original datafile in case -of duplicate keys -@item -u or --unpack -Unpack file packed with myisampack. -@end table +Try to keep the names simple (use @code{name} instead of +@code{customer_name} in the customer table). To make your names portable +to other SQL servers you should keep them shorter than 18 characters. +@item +If you need REALLY high speed, you should take a look at the low-level +interfaces for data storage that the different SQL servers support! For +example, by accessing the @strong{MySQL} @code{MyISAM} directly, you could +get a speed increase of 2-5 times compared to using the SQL interface. +To be able to do this the data must be on the same server as +the application, and usually it should only be accessed by one process +(because external file locking is really slow). One could eliminate the +above problems by introducing low-level @code{MyISAM} commands in the +@strong{MySQL} server (this could be one easy way to get more +performance if needed). By carefully designing the database interface, +it should be quite easy to support this types of optimization. +@item +In many cases it's faster to access data from a database (using a live +connection) than accessing a text file, just because the database is +likely to be more compact than the text file (if you are using numerical +data), and this will involve fewer disk accesses. You will also save +code because you don't have to parse your text files to find line and +column boundaries. +@item +You can also use replication to speed things up. @xref{Replication}. +@item +Declaring a table with @code{DELAY_KEY_WRITE=1} will make the updating of +indexes faster, as these are not logged to disk until the file is closed. +The downside is that you should run @code{myisamchk} on these tables before +you start @code{mysqld} to ensure that they are okay if something killed +@code{mysqld} in the middle. As the key information can always be generated +from the data, you should not lose anything by using @code{DELAY_KEY_WRITE}. +@end itemize -@node myisamchk other options, , myisamchk repair options, myisamchk syntax -@subsubsection Other Options for @code{myisamchk} +@cindex benchmarks +@cindex performance, benchmarks +@node Benchmarks, Design, Tips, Performance +@section Using Your Own Benchmarks -Other actions that @code{myisamchk} can do, besides repair and check tables: +You should definitely benchmark your application and database to find +out where the bottlenecks are. By fixing it (or by replacing the +bottleneck with a 'dummy module') you can then easily identify the next +bottleneck (and so on). Even if the overall performance for your +application is sufficient, you should at least make a plan for each +bottleneck, and decide how to solve it if someday you really need the +extra performance. -@table @code -@item -a or --analyze -Analyze the distribution of keys. This improves join performance by -enabling the join optimizer to better choose in which order it should -join the tables and which keys it should use: -@code{myisamchk --describe --verbose table_name'} or using @code{SHOW KEYS} in -@strong{MySQL}. -@item -d or --description -Prints some information about table. -@item -A or --set-auto-increment[=value] -Force auto_increment to start at this or higher value. If no value is -given, then sets the next auto_increment value to the highest used value -for the auto key + 1. -@item -S or --sort-index -Sort the index tree blocks in high-low order. -This will optimize seeks and will make table scanning by key faster. -@item -R or --sort-records=# -Sorts records according to an index. This makes your data much more localized -and may speed up ranged @code{SELECT} and @code{ORDER BY} operations on -this index. (It may be VERY slow to do a sort the first time!) -To find out a table's index numbers, use @code{SHOW INDEX}, which shows a -table's indexes in the same order that @code{myisamchk} sees them. Indexes are -numbered beginning with 1. -@end table +For an example of portable benchmark programs, look at the @strong{MySQL} +benchmark suite. @xref{MySQL Benchmarks, , @strong{MySQL} Benchmarks}. You +can take any program from this suite and modify it for your needs. By doing this, +you can try different solutions to your problem and test which is really the +fastest solution for you. -@cindex memory usage, myisamchk -@node myisamchk memory, , myisamchk syntax, Table maintenance -@subsection @code{myisamchk} Memory Usage +It is very common that some problems only occur when the system is very +heavily loaded. We have had many customers who contact us when they +have a (tested) system in production and have encountered load problems. In +every one of these cases so far, it has been problems with basic design +(table scans are NOT good at high load) or OS/Library issues. Most of +this would be a @strong{LOT} easier to fix if the systems were not +already in production. -Memory allocation is important when you run @code{myisamchk}. -@code{myisamchk} uses no more memory than you specify with the @code{-O} -options. If you are going to use @code{myisamchk} on very large files, -you should first decide how much memory you want it to use. The default -is to use only about 3M to fix things. By using larger values, you can -get @code{myisamchk} to operate faster. For example, if you have more -than 32M RAM, you could use options such as these (in addition to any -other options you might specify): +To avoid problems like this, you should put some effort into benchmarking +your whole application under the worst possible load! You can use Sasha's +recent hack for this - +@uref{http://www.mysql.com/Downloads/super-smack/super-smack-1.0.tar.gz, +super-smack}. +As the name suggests, it can bring your system down to its knees if you ask it, +so make sure to use it only on your development systems. -@example -shell> myisamchk -O sort=16M -O key=16M -O read=1M -O write=1M ... -@end example +@cindex design, choices +@cindex database design +@cindex storage of data +@node Design, Design Limitations, Benchmarks, Performance +@section Design Choices -Using @code{-O sort=16M} should probably be enough for most cases. +@strong{MySQL} keeps row data and index data in separate files. Many (almost +all) other databases mix row and index data in the same file. We believe that +the @strong{MySQL} choice is better for a very wide range of modern systems. -Be aware that @code{myisamchk} uses temporary files in @code{TMPDIR}. If -@code{TMPDIR} points to a memory file system, you may easily get out of -memory errors. If this happens, set @code{TMPDIR} to point at some directory -with more space and restart @code{myisamchk}. +Another way to store the row data is to keep the information for each +column in a separate area (examples are SDBM and Focus). This will cause a +performance hit for every query that accesses more than one column. Because +this degenerates so quickly when more than one column is accessed, +we believe that this model is not good for general purpose databases. -When repairing, @code{myisamchk} will also need a lot of disk space: +The more common case is that the index and data are stored together +(like in Oracle/Sybase et al). In this case you will find the row +information at the leaf page of the index. The good thing with this +layout is that it, in many cases, depending on how well the index is +cached, saves a disk read. The bad things with this layout are: @itemize @bullet @item -Double the size of the record file (the original one and a copy). This -space is not needed if one does a repair with @code{--quick}, as in this -case only the index file will be re-created. This space is needed on the -same disk as the original record file! +Table scanning is much slower because you have to read through the indexes +to get at the data. @item -Space for the new index file that replaces the old one. The old -index file is truncated at start, so one usually ignore this space. -This space is needed on the same disk as the original index file! +You can't use only the index table to retrieve data for a query. @item -When using @code{--recover} or @code{--sort-recover} -(but not when using @code{--safe-recover}, you will need space for a -sort buffer for: -@code{(largest_key + row_pointer_length)*number_of_rows * 2}. -You can check the length of the keys and the row_pointer_length with -@code{myisamchk -dv table}. -This space is allocated on the temporary disk (specified by @code{TMPDIR} or -@code{--tmpdir=#}). +You lose a lot of space, as you must duplicate indexes from the nodes +(as you can't store the row in the nodes). +@item +Deletes will degenerate the table over time (as indexes in nodes are +usually not updated on delete). +@item +It's harder to cache ONLY the index data. @end itemize -If you have a problem with disk space during repair, you can try to use -@code{--safe-recover} instead of @code{--recover}. +@cindex design, limitations +@node Design Limitations, Portability, Design, Performance +@section MySQL Design Limitations/Tradeoffs -@node Using mysqlcheck, Maintenance regimen, Table maintenance, Maintenance -@section Using @code{mysqlcheck} for Table Maintenance and Crash Recovery +Because @strong{MySQL} uses extremely fast table locking (multiple readers / +single writers) the biggest remaining problem is a mix of a steady stream of +inserts and slow selects on the same table. -Since @strong{MySQL} version 3.23.38 you will be able to use a new -checking and repairing tool for @code{MyISAM} tables. The difference to -@code{myisamchk} is that @code{mysqlcheck} should be used when the -@code{mysqld} server is running, where as @code{myisamchk} should be used -when it is not. The benefit is that you no longer have to take the -server down for checking or repairing your tables. +We believe that for a huge number of systems the extremely fast +performance in other cases make this choice a win. This case is usually +also possible to solve by having multiple copies of the table, but it +takes more effort and hardware. -@code{mysqlcheck} uses @strong{MySQL} server commands @code{CHECK}, -@code{REPAIR}, @code{ANALYZE} and @code{OPTIMIZE} in a convenient way -for the user. +We are also working on some extensions to solve this problem for some +common application niches. -There are three alternative ways to invoke @code{mysqlcheck}: +@cindex portability +@cindex crash-me program +@cindex programs, crash-me +@node Portability, Internal use, Design Limitations, Performance +@section Portability -@example -shell> mysqlcheck [OPTIONS] database [tables] -shell> mysqlcheck [OPTIONS] --databases DB1 [DB2 DB3...] -shell> mysqlcheck [OPTIONS] --all-databases -@end example +Because all SQL servers implement different parts of SQL, it takes work to +write portable SQL applications. For very simple selects/inserts it is +very easy, but the more you need the harder it gets. If you want an +application that is fast with many databases it becomes even harder! -So it can be used in a similar way as @code{mysqldump} when it -comes to what databases and tables you want to choose. +To make a complex application portable you need to choose a number of +SQL servers that it should work with. -@code{mysqlcheck} does have a special feature compared to the other -clients; the default behavior, checking tables (-c), can be changed by -renaming the binary. So if you want to have a tool that repairs tables -by default, you should just copy @code{mysqlcheck} to your harddrive -with a new name, @code{mysqlrepair}, or alternatively make a symbolic -link to @code{mysqlrepair} and name the symbolic link as -@code{mysqlrepair}. If you invoke @code{mysqlrepair} now, it will repair -tables by default. +You can use the @strong{MySQL} crash-me program/web-page +@uref{http://www.mysql.com/information/crash-me.php} to find functions, +types, and limits you can use with a selection of database +servers. Crash-me now tests far from everything possible, but it +is still comprehensive with about 450 things tested. -The names that you can use to change @code{mysqlcheck} default behavior -are here: +For example, you shouldn't have column names longer than 18 characters +if you want to be able to use Informix or DB2. -@example -mysqlrepair: The default option will be -r -mysqlanalyze: The default option will be -a -mysqloptimize: The default option will be -o -@end example +Both the @strong{MySQL} benchmarks and crash-me programs are very +database-independent. By taking a look at how we have handled this, you +can get a feeling for what you have to do to write your application +database-independent. The benchmarks themselves can be found in the +@file{sql-bench} directory in the @strong{MySQL} source +distribution. They are written in Perl with DBI database interface +(which solves the access part of the problem). -The options available for @code{mysqlcheck} are listed here, please -check what your version supports with @code{mysqlcheck --help}. +See @uref{http://www.mysql.com/information/benchmarks.html} for the results +from this benchmark. -@table @code -@item -A, --all-databases -Check all the databases. This will be same as --databases with all -databases selected -@item -1, --all-in-1 -Instead of making one query for each table, execute all queries in 1 -query separately for each database. Table names will be in a comma -separated list. -@item -a, --analyze -Analyze given tables. -@item --auto-repair -If a checked table is corrupted, automatically fix it. Repairing will be -done after all tables have been checked, if corrupted ones were found. -@item -#, --debug=... -Output debug log. Often this is 'd:t:o,filename' -@item --character-sets-dir=... -Directory where character sets are -@item -c, --check -Check table for errors -@item -C, --check-only-changed -Check only tables that have changed since last check or haven't been -closed properly. -@item --compress -Use compression in server/client protocol. -@item -?, --help -Display this help message and exit. -@item -B, --databases -To check several databases. Note the difference in usage; In this case -no tables are given. All name arguments are regarded as database names. -@item --default-character-set=... -Set the default character set -@item -F, --fast -Check only tables that hasn't been closed properly -@item -f, --force -Continue even if we get an sql-error. -@item -e, --extended -If you are using this option with CHECK TABLE, it will ensure that the -table is 100 percent consistent, but will take a long time. +As you can see in these results, all databases have some weak points. That +is, they have different design compromises that lead to different +behavior. -If you are using this option with REPAIR TABLE, it will run an extended -repair on the table, which may not only take a long time to execute, but -may produce a lot of garbage rows also! -@item -h, --host=... -Connect to host. -@item -m, --medium-check -Faster than extended-check, but only finds 99.99 percent of all -errors. Should be good enough for most cases. -@item -o, --optimize -Optimize table -@item -p, --password[=...] -Password to use when connecting to server. If password is not given -it's solicited on the tty. -@item -P, --port=... -Port number to use for connection. -@item -q, --quick -If you are using this option with CHECK TABLE, it prevents the check -from scanning the rows to check for wrong links. This is the fastest -check. +If you strive for database independence, you need to get a good feeling +for each SQL server's bottlenecks. @strong{MySQL} is VERY fast in +retrieving and updating things, but will have a problem in mixing slow +readers/writers on the same table. Oracle, on the other hand, has a big +problem when you try to access rows that you have recently updated +(until they are flushed to disk). Transaction databases in general are +not very good at generating summary tables from log tables, as in this +case row locking is almost useless. -If you are using this option with REPAIR TABLE, it will try to repair -only the index tree. This is the fastest repair method for a table. -@item -r, --repair -Can fix almost anything except unique keys that aren't unique. -@item -s, --silent -Print only error messages. -@item -S, --socket=... -Socket file to use for connection. -@item --tables -Overrides option --databases (-B). -@item -u, --user=# -User for login if not current user. -@item -v, --verbose -Print info about the various stages. -@item -V, --version -Output version information and exit. -@end table +To get your application @emph{really} database-independent, you need to define +an easy extendable interface through which you manipulate your data. As +C++ is available on most systems, it makes sense to use a C++ classes +interface to the databases. -@cindex maintaining, tables -@cindex tables, maintenance regimen -@node Maintenance regimen, Table-info, Using mysqlcheck, Maintenance -@section Setting Up a Table Maintenance Regimen +If you use some specific feature for some database (like the +@code{REPLACE} command in @strong{MySQL}), you should code a method for +the other SQL servers to implement the same feature (but slower). With +@strong{MySQL} you can use the @code{/*! */} syntax to add +@strong{MySQL}-specific keywords to a query. The code inside +@code{/**/} will be treated as a comment (ignored) by most other SQL +servers. -Starting with @strong{MySQL} Version 3.23.13, you can check MyISAM -tables with the @code{CHECK TABLE} command. @xref{CHECK TABLE}. You can -repair tables with the @code{REPAIR TABLE} command. @xref{REPAIR TABLE}. +If REAL high performance is more important than exactness, as in some +Web applications, a possibility is to create an application layer that +caches all results to give you even higher performance. By letting +old results 'expire' after a while, you can keep the cache reasonably +fresh. This is quite nice in case of extremely high load, in which case +you can dynamically increase the cache and set the expire timeout higher +until things get back to normal. -It is a good idea to perform table checks on a regular basis rather than -waiting for problems to occur. For maintenance purposes, you can use -@code{myisamchk -s} to check tables. The @code{-s} option (short for -@code{--silent}) causes @code{myisamchk} to run in silent mode, printing -messages only when errors occur. +In this case the table creation information should contain information +of the initial size of the cache and how often the table should normally +be refreshed. -@tindex .pid (process ID) file -It's also a good idea to check tables when the server starts up. -For example, whenever the machine has done a reboot in the middle of an -update, you usually need to check all the tables that could have been -affected. (This is an ``expected crashed table''.) You could add a test to -@code{safe_mysqld} that runs @code{myisamchk} to check all tables that have -been modified during the last 24 hours if there is an old @file{.pid} -(process ID) file left after a reboot. (The @file{.pid} file is created by -@code{mysqld} when it starts up and removed when it terminates normally. The -presence of a @file{.pid} file at system startup time indicates that -@code{mysqld} terminated abnormally.) +@cindex uses, of MySQL +@cindex customers, of MySQL +@node Internal use, , Portability, Performance +@section What Have We Used MySQL For? -An even better test would be to check any table whose last-modified time -is more recent than that of the @file{.pid} file. +During @strong{MySQL} initial development, the features of @strong{MySQL} were made to fit +our largest customer. They handle data warehousing for a couple of the +biggest retailers in Sweden. -You should also check your tables regularly during normal system -operation. At @strong{MySQL AB}, we run a @code{cron} job to check all -our important tables once a week, using a line like this in a @file{crontab} -file: - -@example -35 0 * * 0 /path/to/myisamchk --fast --silent /path/to/datadir/*/*.MYI -@end example +From all stores, we get weekly summaries of all bonus card transactions, +and we are expected to provide useful information for the store owners +to help them find how their advertisement campaigns are affecting their +customers. -This prints out information about crashed tables so we can examine and repair -them when needed. +The data is quite huge (about 7 million summary transactions per month), +and we have data for 4-10 years that we need to present to the users. +We got weekly requests from the customers that they want to get +'instant' access to new reports from this data. -As we haven't had any unexpectedly crashed tables (tables that become -corrupted for reasons other than hardware trouble) -for a couple of years now (this is really true), once a week is -more than enough for us. +We solved this by storing all information per month in compressed +'transaction' tables. We have a set of simple macros (script) that +generates summary tables grouped by different criteria (product group, +customer id, store ...) from the transaction tables. The reports are +Web pages that are dynamically generated by a small Perl script that +parses a Web page, executes the SQL statements in it, and inserts the +results. We would have used PHP or mod_perl instead but they were +not available at that time. -We recommend that to start with, you execute @code{myisamchk -s} each -night on all tables that have been updated during the last 24 hours, -until you come to trust @strong{MySQL} as much as we do. +For graphical data we wrote a simple tool in @code{C} that can produce +GIFs based on the result of a SQL query (with some processing of the +result). This is also dynamically executed from the Perl script that +parses the @code{HTML} files. -@cindex tables, defragment -Normally you don't need to maintain @strong{MySQL} tables that much. If -you are changing tables with dynamic size rows (tables with @code{VARCHAR}, -@code{BLOB} or @code{TEXT} columns) or have tables with many deleted rows -you may want to from time to time (once a month?) defragment/reclaim space -from the tables. +In most cases a new report can simply be done by copying an existing +script and modifying the SQL query in it. In some cases, we will need to +add more fields to an existing summary table or generate a new one, but +this is also quite simple, as we keep all transactions tables on disk. +(Currently we have at least 50G of transactions tables and 200G of other +customer data.) -You can do this by using @code{OPTIMIZE TABLE} on the tables in question or -if you can take the @code{mysqld} server down for a while do: +We also let our customers access the summary tables directly with ODBC +so that the advanced users can themselves experiment with the data. -@example -isamchk -r --silent --sort-index -O sort_buffer_size=16M */*.ISM -myisamchk -r --silent --sort-index -O sort_buffer_size=16M */*.MYI -@end example +We haven't had any problems handling this with quite modest Sun Ultra +SPARCstation (2x200 Mhz). We recently upgraded one of our servers to a 2 +CPU 400 Mhz UltraSPARC, and we are now planning to start handling +transactions on the product level, which would mean a ten-fold increase +of data. We think we can keep up with this by just adding more disk to +our systems. -@cindex tables, information -@node Table-info, Crash recovery, Maintenance regimen, Maintenance -@section Getting Information About a Table +We are also experimenting with Intel-Linux to be able to get more CPU +power cheaper. Now that we have the binary portable database format (new +in Version 3.23), we will start to use this for some parts of the application. -To get a description of a table or statistics about it, use the commands shown -below. We explain some of the information in more detail later: +Our initial feelings are that Linux will perform much better on +low-to-medium load and Solaris will perform better when you start to get a +high load because of extreme disk IO, but we don't yet have anything +conclusive about this. After some discussion with a Linux Kernel +developer, this might be a side effect of Linux giving so much resources +to the batch job that the interactive performance gets very low. This +makes the machine feel very slow and unresponsive while big batches are +going. Hopefully this will be better handled in future Linux Kernels. -@table @code -@item myisamchk -d tbl_name -Runs @code{myisamchk} in ``describe mode'' to produce a description of -your table. If you start the @strong{MySQL} server using the -@code{--skip-locking} option, @code{myisamchk} may report an error for a -table that is updated while it runs. However, because @code{myisamchk} -doesn't change the table in describe mode, there isn't any risk of -destroying data. +@cindex benchmark suite +@cindex crash-me program +@node MySQL Benchmarks, Tools, Performance, Top +@chapter The MySQL Benchmark Suite -@item myisamchk -d -v tbl_name -To produce more information about what @code{myisamchk} is doing, add @code{-v} -to tell it to run in verbose mode. +This should contain a technical description of the @strong{MySQL} +benchmark suite (and @code{crash-me}), but that description is not +written yet. Currently, you can get a good idea of the benchmark by +looking at the code and results in the @file{sql-bench} directory in any +@strong{MySQL} source distributions. -@item myisamchk -eis tbl_name -Shows only the most important information from a table. It is slow because it -must read the whole table. +This benchmark suite is meant to be a benchmark that will tell any user +what things a given SQL implementation performs well or poorly at. -@item myisamchk -eiv tbl_name -This is like @code{-eis}, but tells you what is being done. -@end table +Note that this benchmark is single threaded, so it measures the minimum +time for the operations. We plan to in the future add a lot of +multi-threaded tests to the benchmark suite. -@cindex examples, @code{myisamchk} output -@cindex @code{myisamchk}, example output -Example of @code{myisamchk -d} output: -@example -MyISAM file: company.MYI -Record format: Fixed length -Data records: 1403698 Deleted blocks: 0 -Recordlength: 226 +For example, (run on the same NT 4.0 machine): -table description: -Key Start Len Index Type -1 2 8 unique double -2 15 10 multip. text packed stripped -3 219 8 multip. double -4 63 10 multip. text packed stripped -5 167 2 multip. unsigned short -6 177 4 multip. unsigned long -7 155 4 multip. text -8 138 4 multip. unsigned long -9 177 4 multip. unsigned long - 193 1 text -@end example +@multitable @columnfractions .6 .2 .2 +@strong{Reading 2000000 rows by index} @tab @strong{Seconds} @tab @strong{Seconds} +@item mysql @tab 367 @tab 249 +@item mysql_odbc @tab 464 +@item db2_odbc @tab 1206 +@item informix_odbc @tab 121126 +@item ms-sql_odbc @tab 1634 +@item oracle_odbc @tab 20800 +@item solid_odbc @tab 877 +@item sybase_odbc @tab 17614 +@end multitable -Example of @code{myisamchk -d -v} output: -@example -MyISAM file: company -Record format: Fixed length -File-version: 1 -Creation time: 1999-10-30 12:12:51 -Recover time: 1999-10-31 19:13:01 -Status: checked -Data records: 1403698 Deleted blocks: 0 -Datafile parts: 1403698 Deleted data: 0 -Datafilepointer (bytes): 3 Keyfile pointer (bytes): 3 -Max datafile length: 3791650815 Max keyfile length: 4294967294 -Recordlength: 226 +@multitable @columnfractions .6 .2 .2 +@strong{Inserting (350768) rows} @tab @strong{Seconds} @tab @strong{Seconds} +@item mysql @tab 381 @tab 206 +@item mysql_odbc @tab 619 +@item db2_odbc @tab 3460 +@item informix_odbc @tab 2692 +@item ms-sql_odbc @tab 4012 +@item oracle_odbc @tab 11291 +@item solid_odbc @tab 1801 +@item sybase_odbc @tab 4802 +@end multitable -table description: -Key Start Len Index Type Rec/key Root Blocksize -1 2 8 unique double 1 15845376 1024 -2 15 10 multip. text packed stripped 2 25062400 1024 -3 219 8 multip. double 73 40907776 1024 -4 63 10 multip. text packed stripped 5 48097280 1024 -5 167 2 multip. unsigned short 4840 55200768 1024 -6 177 4 multip. unsigned long 1346 65145856 1024 -7 155 4 multip. text 4995 75090944 1024 -8 138 4 multip. unsigned long 87 85036032 1024 -9 177 4 multip. unsigned long 178 96481280 1024 - 193 1 text -@end example +In the above test @strong{MySQL} was run with a 8M index cache. -Example of @code{myisamchk -eis} output: -@example -Checking MyISAM file: company -Key: 1: Keyblocks used: 97% Packed: 0% Max levels: 4 -Key: 2: Keyblocks used: 98% Packed: 50% Max levels: 4 -Key: 3: Keyblocks used: 97% Packed: 0% Max levels: 4 -Key: 4: Keyblocks used: 99% Packed: 60% Max levels: 3 -Key: 5: Keyblocks used: 99% Packed: 0% Max levels: 3 -Key: 6: Keyblocks used: 99% Packed: 0% Max levels: 3 -Key: 7: Keyblocks used: 99% Packed: 0% Max levels: 3 -Key: 8: Keyblocks used: 99% Packed: 0% Max levels: 3 -Key: 9: Keyblocks used: 98% Packed: 0% Max levels: 4 -Total: Keyblocks used: 98% Packed: 17% +We have gather some more benchmark results at +@uref{http://www.mysql.com/information/benchmarks.html}. -Records: 1403698 M.recordlength: 226 Packed: 0% -Recordspace used: 100% Empty space: 0% Blocks/Record: 1.00 -Record blocks: 1403698 Delete blocks: 0 -Recorddata: 317235748 Deleted data: 0 -Lost space: 0 Linkdata: 0 +Note that Oracle is not included because they asked to be removed. All +Oracle benchmarks have to be passed by Oracle! We believe that makes +Oracle benchmarks @strong{VERY} biased because the above benchmarks are +supposed to show what a standard installation can do for a single +client. -User time 1626.51, System time 232.36 -Maximum resident set size 0, Integral resident set size 0 -Non physical pagefaults 0, Physical pagefaults 627, Swaps 0 -Blocks in 0 out 0, Messages in 0 out 0, Signals 0 -Voluntary context switches 639, Involuntary context switches 28966 -@end example +To run the benchmark suite, you have to download a @strong{MySQL} source +distribution, install the perl DBI driver, the perl DBD driver for the +database you want to test and then do: -Example of @code{myisamchk -eiv} output: @example -Checking MyISAM file: company -Data records: 1403698 Deleted blocks: 0 -- check file-size -- check delete-chain -block_size 1024: -index 1: -index 2: -index 3: -index 4: -index 5: -index 6: -index 7: -index 8: -index 9: -No recordlinks -- check index reference -- check data record references index: 1 -Key: 1: Keyblocks used: 97% Packed: 0% Max levels: 4 -- check data record references index: 2 -Key: 2: Keyblocks used: 98% Packed: 50% Max levels: 4 -- check data record references index: 3 -Key: 3: Keyblocks used: 97% Packed: 0% Max levels: 4 -- check data record references index: 4 -Key: 4: Keyblocks used: 99% Packed: 60% Max levels: 3 -- check data record references index: 5 -Key: 5: Keyblocks used: 99% Packed: 0% Max levels: 3 -- check data record references index: 6 -Key: 6: Keyblocks used: 99% Packed: 0% Max levels: 3 -- check data record references index: 7 -Key: 7: Keyblocks used: 99% Packed: 0% Max levels: 3 -- check data record references index: 8 -Key: 8: Keyblocks used: 99% Packed: 0% Max levels: 3 -- check data record references index: 9 -Key: 9: Keyblocks used: 98% Packed: 0% Max levels: 4 -Total: Keyblocks used: 9% Packed: 17% +cd sql-bench +perl run-all-tests --server=# +@end example -- check records and index references -[LOTS OF ROW NUMBERS DELETED] +where # is one of supported servers. You can get a list of all options +and supported servers by doing @code{run-all-tests --help}. -Records: 1403698 M.recordlength: 226 Packed: 0% -Recordspace used: 100% Empty space: 0% Blocks/Record: 1.00 -Record blocks: 1403698 Delete blocks: 0 -Recorddata: 317235748 Deleted data: 0 -Lost space: 0 Linkdata: 0 +@cindex crash-me +@code{crash-me} tries to determine what features a database supports and +what its capabilities and limitations are by actually running +queries. For example, it determines: -User time 1639.63, System time 251.61 -Maximum resident set size 0, Integral resident set size 0 -Non physical pagefaults 0, Physical pagefaults 10580, Swaps 0 -Blocks in 4 out 0, Messages in 0 out 0, Signals 0 -Voluntary context switches 10604, Involuntary context switches 122798 -@end example +@itemize @bullet +@item +What column types are supported +@item +How many indexes are supported +@item +What functions are supported +@item +How big a query can be +@item +How big a @code{VARCHAR} column can be +@end itemize -Here are the sizes of the data and index files for the table used in the -preceding examples: +We can find the result from crash-me on a lot of different databases at +@uref{http://www.mysql.com/information/crash-me.php}. -@example --rw-rw-r-- 1 monty tcx 317235748 Jan 12 17:30 company.MYD --rw-rw-r-- 1 davida tcx 96482304 Jan 12 18:35 company.MYM -@end example +@cindex utilities +@node Tools, Maintenance, MySQL Benchmarks, Top +@chapter MySQL Utilites -Explanations for the types of information @code{myisamchk} produces are -given below. The ``keyfile'' is the index file. ``Record'' and ``row'' -are synonymous: +@menu +* Programs:: What do the executables do? +* mysqld-max:: mysqld-max, An extended mysqld server +* safe_mysqld:: safe_mysqld, the wrapper around mysqld +* mysqld_multi:: Program for managing multiple @strong{MySQL} servers +* mysql:: The command line tool +* mysqladmin:: Administering a @strong{MySQL} server +* mysqldump:: Dumping the structure and data from @strong{MySQL} databases and tables +* mysqlhotcopy:: Copying @strong{MySQL} Databases and Tables +* mysqlimport:: Importing data from text files +* perror:: Displaying error messages +* mysqlshow:: Showing databases, tables and columns +* myisampack:: The @strong{MySQL} compressed read-only table generator +@end menu -@table @code -@item ISAM file -Name of the ISAM (index) file. +In this chapter you will learn about the @strong{MySQL} Utilities that +come in a given distribution. You will learn what each of them does, how +to use it, and what you should use it for. -@item Isam-version -Version of ISAM format. Currently always 2. +@cindex environment variables +@cindex programs, list of +@node Programs, mysqld-max, Tools, Tools +@section Overview of the Different MySQL Programs -@item Creation time -When the data file was created. +All @strong{MySQL} clients that communicate with the server using the +@code{mysqlclient} library use the following environment variables: -@item Recover time -When the index/data file was last reconstructed. +@tindex MYSQL_UNIX_PORT environment variable +@tindex Environment variable, MYSQL_UNIX_PORT +@tindex MYSQL_TCP_PORT environment variable +@tindex Environment variable, MYSQL_TCP_PORT +@tindex MYSQL_PWD environment variable +@tindex Environment variable, MYSQL_PWD +@tindex MYSQL_DEBUG environment variable +@tindex Environment variable, MYSQL_DEBUG +@multitable @columnfractions .25 .75 +@item @strong{Name} @tab @strong{Description} +@item @code{MYSQL_UNIX_PORT} @tab The default socket; used for connections to @code{localhost} +@item @code{MYSQL_TCP_PORT} @tab The default TCP/IP port +@item @code{MYSQL_PWD} @tab The default password +@item @code{MYSQL_DEBUG} @tab Debug-trace options when debugging +@item @code{TMPDIR} @tab The directory where temporary tables/files are created +@end multitable -@item Data records -How many records are in the table. +Use of @code{MYSQL_PWD} is insecure. +@xref{Connecting}. -@item Deleted blocks -How many deleted blocks still have reserved space. -You can optimize your table to minimize this space. -@xref{Optimization}. +@tindex MYSQL_HISTFILE environment variable +@tindex Environment variable, MYSQL_HISTFILE +@tindex HOME environment variable +@tindex Environment variable, HOME +@cindex history file +@cindex command line history +@tindex .mysql_history file +The @file{mysql} client uses the file named in the @code{MYSQL_HISTFILE} +environment variable to save the command-line history. The default value for +the history file is @file{$HOME/.mysql_history}, where @code{$HOME} is the +value of the @code{HOME} environment variable. @xref{Environment variables}. -@item Datafile: Parts -For dynamic record format, this indicates how many data blocks there are. For -an optimized table without fragmented records, this is the same as @code{Data -records}. +All @strong{MySQL} programs take many different options. However, every +@strong{MySQL} program provides a @code{--help} option that you can use +to get a full description of the program's different options. For example, try +@code{mysql --help}. -@item Deleted data -How many bytes of non-reclaimed deleted data there are. -You can optimize your table to minimize this space. -@xref{Optimization}. +You can override default options for all standard client programs with an +option file. @ref{Option files}. -@item Datafile pointer -The size of the data file pointer, in bytes. It is usually 2, 3, 4, or 5 -bytes. Most tables manage with 2 bytes, but this cannot be controlled -from @strong{MySQL} yet. For fixed tables, this is a record address. For -dynamic tables, this is a byte address. +The list below briefly describes the @strong{MySQL} programs: -@item Keyfile pointer -The size of the index file pointer, in bytes. It is usually 1, 2, or 3 -bytes. Most tables manage with 2 bytes, but this is calculated -automatically by @strong{MySQL}. It is always a block address. +@table @code -@item Max datafile length -How long the table's data file (@code{.MYD} file) can become, in bytes. +@cindex @code{myisamchk} +@item myisamchk +Utility to describe, check, optimize, and repair @strong{MySQL} tables. +Because @code{myisamchk} has many functions, it is described in its own +chapter. @xref{Maintenance}. -@item Max keyfile length -How long the table's key file (@code{.MYI} file) can become, in bytes. +@cindex @code{make_binary_distribution} +@item make_binary_distribution +Makes a binary release of a compiled @strong{MySQL}. This could be sent +by FTP to @file{/pub/mysql/Incoming} on @code{support.mysql.com} for the +convenience of other @strong{MySQL} users. -@item Recordlength -How much space each record takes, in bytes. +@cindex @code{msql2mysql} +@item msql2mysql +A shell script that converts @code{mSQL} programs to @strong{MySQL}. It doesn't +handle all cases, but it gives a good start when converting. -@item Record format -The format used to store table rows. -The examples shown above use @code{Fixed length}. -Other possible values are @code{Compressed} and @code{Packed}. +@cindex @code{mysqlaccess} +@item mysqlaccess +A script that checks the access privileges for a host, user, and database +combination. -@item table description -A list of all keys in the table. For each key, some low-level information -is presented: +@cindex @code{mysqladmin} +@item mysqladmin +Utility for performing administrative operations, such as creating or +dropping databases, reloading the grant tables, flushing tables to disk, and +reopening log files. @code{mysqladmin} can also be used to retrieve version, +process, and status information from the server. +@xref{mysqladmin, , @code{mysqladmin}}. -@table @code -@item Key -This key's number. +@cindex @code{mysqlbug} +@item mysqlbug +The @strong{MySQL} bug report script. This script should always be used when +filing a bug report to the @strong{MySQL} list. -@item Start -Where in the record this index part starts. +@cindex @code{mysqld} +@item mysqld +The SQL daemon. This should always be running. -@item Len -How long this index part is. For packed numbers, this should always be -the full length of the column. For strings, it may be shorter than the full -length of the indexed column, because you can index a prefix of a string -column. +@cindex @code{mysqldump} +@item mysqldump +Dumps a @strong{MySQL} database into a file as SQL statements or +as tab-separated text files. Enhanced freeware originally by Igor Romanenko. +@xref{mysqldump, , @code{mysqldump}}. -@item Index -@code{unique} or @code{multip.} (multiple). Indicates whether or not one value -can exist multiple times in this index. +@cindex @code{mysqlimport} +@item mysqlimport +Imports text files into their respective tables using @code{LOAD DATA +INFILE}. @xref{mysqlimport, , @code{mysqlimport}}. -@item Type -What data-type this index part has. This is an ISAM data-type -with the options @code{packed}, @code{stripped} or @code{empty}. +@cindex @code{mysqlshow} +@item mysqlshow +Displays information about databases, tables, columns, and indexes. -@item Root -Address of the root index block. +@cindex @code{mysql_install_db} +@item mysql_install_db +Creates the @strong{MySQL} grant tables with default privileges. This is +usually executed only once, when first installing @strong{MySQL} +on a system. -@item Blocksize -The size of each index block. By default this is 1024, but the value may be -changed at compile time. +@cindex @code{replace} +@item replace +A utility program that is used by @code{msql2mysql}, but that has more +general applicability as well. @code{replace} changes strings in place in +files or on the standard input. Uses a finite state machine to match longer +strings first. Can be used to swap strings. For example, this command +swaps @code{a} and @code{b} in the given files: -@item Rec/key -This is a statistical value used by the optimizer. It tells how many -records there are per value for this key. A unique key always has a -value of 1. This may be updated after a table is loaded (or greatly -changed) with @code{myisamchk -a}. If this is not updated at all, a default -value of 30 is given. +@example +shell> replace a b b a -- file1 file2 ... +@end example @end table -@item -In the first example above, the 9th key is a multi-part key with two parts. +@cindex @code{mysqld-max} +@node mysqld-max, safe_mysqld, Programs, Tools +@section mysqld-max, An extended mysqld server -@item Keyblocks used -What percentage of the keyblocks are used. Because the table used in the -examples had just been reorganized with @code{myisamchk}, the values are very -high (very near the theoretical maximum). +@code{mysqld-max} is the MySQL server (@code{mysqld}) configured with +the following configure options: -@item Packed -@strong{MySQL} tries to pack keys with a common suffix. This can only be used -for @code{CHAR}/@code{VARCHAR}/@code{DECIMAL} keys. For long strings like -names, this can significantly reduce the space used. In the third example -above, the 4th key is 10 characters long and a 60% reduction in space is -achieved. +@multitable @columnfractions .3 .7 +@item @strong{Option} @tab @strong{Comment} +@item --with-server-suffix=-max @tab Add a suffix to the @code{mysqld} version string. +@item --with-bdb @tab Support for Berkeley DB (BDB) tables +@item --with-innodb @tab Support for InnoDB tables. +@item CFLAGS=-DUSE_SYMDIR @tab Symbolic links support for Windows. +@end multitable -@item Max levels -How deep the B-tree for this key is. Large tables with long keys get high -values. - -@item Records -How many rows are in the table. +You can find the @strong{MySQL}-max binaries at +@uref{http://www.mysql.com/downloads/mysql-max-3.23.html}. -@item M.recordlength -The average record length. For tables with fixed-length records, this is the -exact record length. +The Windows @strong{MySQL} 3.23 binary distribution includes both the +standard @strong{mysqld.exe} binary and the @code{mysqld-max.exe} binary. +@uref{http://www.mysql.com/downloads/mysql-3.23.html}. +@xref{Windows installation}. -@item Packed -@strong{MySQL} strips spaces from the end of strings. The @code{Packed} -value indicates the percentage of savings achieved by doing this. +Note that as Berkeley DB and InnoDB are not available for all platforms, +some of the @code{Max} binaries may not have support for both of these. +You can check which table types are supported by doing the following +query: -@item Recordspace used -What percentage of the data file is used. +@example +mysql> show variables like "have_%"; ++---------------+-------+ +| Variable_name | Value | ++---------------+-------+ +| have_bdb | YES | +| have_innodb | NO | +| have_isam | YES | +| have_raid | YES | +| have_ssl | NO | ++---------------+-------+ +@end example -@item Empty space -What percentage of the data file is unused. +The meaning of the values are: -@item Blocks/Record -Average number of blocks per record (that is, how many links a fragmented -record is composed of). This is always 1 for fixed-format tables. This value -should stay as close to 1.0 as possible. If it gets too big, you can -reorganize the table with @code{myisamchk}. -@xref{Optimization}. +@multitable @columnfractions .3 .7 +@item @strong{Value} @tab @strong{Meaning}. +@item YES @tab The option is activated and usable. +@item NO @tab @strong{MySQL} is not compiled with support for this option. +@item DISABLED @tab The xxxx option is disabled because one started @code{mysqld} with @code{--skip-xxxx} or because one didn't start @code{mysqld} with all needed options to enable the option. In this case the @code{hostname.err} file should contain a reason for why the option is disabled. +@end multitable -@item Recordblocks -How many blocks (links) are used. For fixed format, this is the same as the number -of records. +@strong{NOTE}: To be able to create InnoDB tables you @strong{MUST} edit +your startup options to include at least the @code{innodb_data_file_path} +option. @xref{InnoDB start}. -@item Deleteblocks -How many blocks (links) are deleted. +To get better performance for BDB tables, you should add some configuration +options for these too. @xref{BDB start}. -@item Recorddata -How many bytes in the data file are used. +@code{safe_mysqld} will automatically try to start any @code{mysqld} binary +with the @code{-max} prefix. This makes it very easy to test out a +another @code{mysqld} binary in an existing installation. Just +run @code{configure} with the options you want and then install the +new @code{mysqld} binary as @code{mysqld-max} in the same directory +where your old @code{mysqld} binary is. @xref{safe_mysqld, , @code{safe_mysqld}}. -@item Deleted data -How many bytes in the data file are deleted (unused). +The @code{mysqld-max} RPM uses the above mentioned @code{safe_mysqld} +feature. It just installs the @code{mysqld-max} executable and +@code{safe_mysqld} will automatically use this executable when +@code{safe_mysqld} is restarted. -@item Lost space -If a record is updated to a shorter length, some space is lost. This is -the sum of all such losses, in bytes. +The following table shows which table types our standard @strong{MySQL-Max} +binaries includes: -@item Linkdata -When the dynamic table format is used, record fragments are linked with -pointers (4 to 7 bytes each). @code{Linkdata} is the sum of the amount of -storage used by all such pointers. -@end table +@multitable @columnfractions .4 .3 .3 +@item @strong{System} @tab @strong{BDB} @tab @strong{InnoDB} +@item AIX 4.3 @tab N @tab Y +@item HP-UX 11.0 @tab N @tab Y +@item Linux-Alpha @tab N @tab Y +@item Linux-Intel @tab Y @tab Y +@item Linux-Ia64 @tab N @tab Y +@item Solaris-intel @tab N @tab Y +@item Solaris-sparc @tab Y @tab Y +@item SCO OSR5 @tab Y @tab Y +@item UnixWare @tab Y @tab Y +@item Windows/NT @tab Y @tab Y +@end multitable -If a table has been compressed with @code{myisampack}, @code{myisamchk --d} prints additional information about each table column. See -@ref{myisampack, , @code{myisampack}}, for an example of this -information and a description of what it means. +@cindex tools, safe_mysqld +@cindex scripts +@cindex @code{safe_mysqld} +@node safe_mysqld, mysqld_multi, mysqld-max, Tools +@section safe_mysqld, the wrapper around mysqld -@cindex crash, recovery -@cindex recovery, from crash -@node Crash recovery, Log file maintenance, Table-info, Maintenance -@section Using @code{myisamchk} for Crash Recovery +@code{safe_mysqld} is the recommended way to start a @code{mysqld} +daemon on Unix. @code{safe_mysqld} adds some safety features such as +restarting the server when an error occurs and logging run-time +information to a log file. -If you run @code{mysqld} with @code{--skip-locking} (which is the default on -some systems, like Linux), you can't reliably use @code{myisamchk} to -check a table when @code{mysqld} is using the same table. If you -can be sure that no one is accessing the tables through @code{mysqld} -while you run @code{myisamchk}, you only have to do @code{mysqladmin -flush-tables} before you start checking the tables. If you can't -guarantee the above, then you must take down @code{mysqld} while you -check the tables. If you run @code{myisamchk} while @code{mysqld} is updating -the tables, you may get a warning that a table is corrupt even if it -isn't. +If you don't use @code{--mysqld=#} or @code{--mysqld-version=#} +@code{safe_mysqld} will use an executable named @code{mysqld-max} if it +exists. If not, @code{safe_mysqld} will start @code{mysqld}. +This makes it very easy to test to use @code{mysqld-max} instead of +@code{mysqld}; Just copy @code{mysqld-max} to where you have +@code{mysqld} and it will be used. -If you are not using @code{--skip-locking}, you can use @code{myisamchk} -to check tables at any time. While you do this, all clients that try -to update the table will wait until @code{myisamchk} is ready before -continuing. +Normally one should never edit the @code{safe_mysqld} script, but +instead put the options to @code{safe_mysqld} in the +@code{[safe_mysqld]} section in the @code{my.cnf} +file. @code{safe_mysqld} will read all options from the @code{[mysqld]}, +@code{[server]} and @code{[safe_mysqld]} sections from the option files. +@xref{Option files}. -If you use @code{myisamchk} to repair or optimize tables, you -@strong{MUST} always ensure that the @code{mysqld} server is not using -the table (this also applies if you are using @code{--skip-locking}). -If you don't take down @code{mysqld} you should at least do a -@code{mysqladmin flush-tables} before you run @code{myisamchk}. +Note that all options on the command line to @code{safe_mysqld} are passed +to @code{mysqld}. If you wants to use any options in @code{safe_mysqld} that +@code{mysqld} doesn't support, you must specify these in the option file. -This chapter describes how to check for and deal with data corruption -in @strong{MySQL} databases. If your tables get corrupted a lot you should -try to find the reason for this! @xref{Crashing}. +Most of the options to @code{safe_mysqld} are the same as the options to +@code{mysqld}. @xref{Command-line options}. -The @code{MyISAM} table section contains reason for why a table could be -corrupted. @xref{MyISAM table problems}. +@code{safe_mysqld} supports the following options: -When performing crash recovery, it is important to understand that each table -@code{tbl_name} in a database corresponds to three files in the database -directory: +@table @code +@item --basedir=path +@item --core-file-size=# +Size of the core file @code{mysqld} should be able to create. Passed to @code{ulimit -c}. +@item --datadir=path +@item --defaults-extra-file=path +@item --defaults-file=path +@item --err-log=path +@item --ledir=path +Path to @code{mysqld} +@item --log=path +@item --mysqld=mysqld-version +Name of the @code{mysqld} version in the @code{ledir} directory you want to start. +@item --mysqld-version=version +Similar to @code{--mysqld=} but here you only give the suffix for @code{mysqld}. +For example if you use @code{--mysqld-version=max}, @code{safe_mysqld} will +start the @code{ledir/mysqld-max} version. If the argument to +@code{--mysqld-version} is empty, @code{ledir/mysqld} will be used. +@item --no-defaults +@item --open-files-limit=# +Number of files @code{mysqld} should be able to open. Passed to @code{ulimit -n}. Note that you need to start @code{safe_mysqld} as root for this to work properly! +@item --pid-file=path +@item --port=# +@item --socket=path +@item --timezone=# +Set the timezone (the @code{TZ}) variable to the value of this parameter. +@item --user=# +@end table -@multitable @columnfractions .2 .8 -@item @strong{File} @tab @strong{Purpose} -@item @file{tbl_name.frm} @tab Table definition (form) file -@item @file{tbl_name.MYD} @tab Data file -@item @file{tbl_name.MYI} @tab Index file -@end multitable +The @code{safe_mysqld} script is written so that it normally is able to start +a server that was installed from either a source or a binary version of +@strong{MySQL}, even if these install the server in slightly different +locations. @code{safe_mysqld} expects one of these conditions to be true: -Each of these three file types is subject to corruption in various ways, but -problems occur most often in data files and index files. +@itemize @bullet +@item +The server and databases can be found relative to the directory from which +@code{safe_mysqld} is invoked. @code{safe_mysqld} looks under its working +directory for @file{bin} and @file{data} directories (for binary +distributions) or for @file{libexec} and @file{var} directories (for source +distributions). This condition should be met if you execute +@code{safe_mysqld} from your @strong{MySQL} installation directory (for +example, @file{/usr/local/mysql} for a binary distribution). -@code{myisamchk} works by creating a copy of the @file{.MYD} (data) file -row by row. It ends the repair stage by removing the old @file{.MYD} -file and renaming the new file to the original file name. If you use -@code{--quick}, @code{myisamchk} does not create a temporary @file{.MYD} -file, but instead assumes that the @file{.MYD} file is correct and only -generates a new index file without touching the @file{.MYD} file. This -is safe, because @code{myisamchk} automatically detects if the -@file{.MYD} file is corrupt and aborts the repair in this case. You can -also give two @code{--quick} options to @code{myisamchk}. In this case, -@code{myisamchk} does not abort on some errors (like duplicate key) but -instead tries to resolve them by modifying the @file{.MYD} -file. Normally the use of two @code{--quick} options is useful only if -you have too little free disk space to perform a normal repair. In this -case you should at least make a backup before running @code{myisamchk}. +@item +If the server and databases cannot be found relative to the working directory, +@code{safe_mysqld} attempts to locate them by absolute pathnames. Typical +locations are @file{/usr/local/libexec} and @file{/usr/local/var}. +The actual locations are determined when the distribution was built from which +@code{safe_mysqld} comes. They should be correct if +@strong{MySQL} was installed in a standard location. +@end itemize -@menu -* Check:: How to check tables for errors -* Repair:: How to repair tables -* Optimization:: Table optimization -@end menu +Because @code{safe_mysqld} will try to find the server and databases relative +to its own working directory, you can install a binary distribution of +@strong{MySQL} anywhere, as long as you start @code{safe_mysqld} from the +@strong{MySQL} installation directory: -@cindex checking, tables for errors -@cindex tables, error checking -@cindex errors, checking tables for -@node Check, Repair, Crash recovery, Crash recovery -@subsection How to Check Tables for Errors +@example +shell> cd mysql_installation_directory +shell> bin/safe_mysqld & +@end example -To check a MyISAM table, use the following commands: +If @code{safe_mysqld} fails, even when invoked from the @strong{MySQL} +installation directory, you can modify it to use the path to @code{mysqld} +and the pathname options that are correct for your system. Note that if you +upgrade @strong{MySQL} in the future, your modified version of +@code{safe_mysqld} will be overwritten, so you should make a copy of your +edited version that you can reinstall. -@table @code -@item myisamchk tbl_name -This finds 99.99% of all errors. What it can't find is corruption that -involves @strong{ONLY} the data file (which is very unusual). If you want -to check a table, you should normally run @code{myisamchk} without options or -with either the @code{-s} or @code{--silent} option. +@cindex tools, mysqld_multi +@cindex scripts +@cindex multi mysqld +@cindex @code{mysqld_multi} +@node mysqld_multi, mysql, safe_mysqld, Tools +@section mysqld_multi, program for managing multiple @strong{MySQL} servers -@item myisamchk -m tbl_name -This finds 99.999% of all errors. It checks first all index entries for errors and -then it reads through all rows. It calculates a checksum for all keys in -the rows and verifies that they checksum matches the checksum for the keys -in the index tree. +@code{mysqld_multi} is meant for managing several @code{mysqld} +processes running in different UNIX sockets and TCP/IP ports. -@item myisamchk -e tbl_name -This does a complete and thorough check of all data (@code{-e} means -``extended check''). It does a check-read of every key for each row to verify -that they indeed point to the correct row. This may take a LONG time on a -big table with many keys. @code{myisamchk} will normally stop after the first -error it finds. If you want to obtain more information, you can add the -@code{--verbose} (@code{-v}) option. This causes @code{myisamchk} to keep -going, up through a maximum of 20 errors. In normal usage, a simple -@code{myisamchk} (with no arguments other than the table name) is sufficient. +The program will search for group(s) named [mysqld#] from my.cnf (or the +given --config-file=...), where # can be any positive number starting +from 1. These groups should be the same as the usual @code{[mysqld]} +group (e.g. options to mysqld, see @strong{MySQL} manual for detailed +information about this group), but with those port, socket etc. options +that are wanted for each separate @code{mysqld} processes. The number in +the group name has another function; it can be used for starting, +stopping, or reporting some specific @code{mysqld} servers with this +program. See the usage and options below for more information. -@item myisamchk -e -i tbl_name -Like the previous command, but the @code{-i} option tells @code{myisamchk} to -print some informational statistics, too. -@end table +@example +Usage: mysqld_multi [OPTIONS] @{start|stop|report@} [GNR,GNR,GNR...] +or mysqld_multi [OPTIONS] @{start|stop|report@} [GNR-GNR,GNR,GNR-GNR,...] +@end example -@cindex tables, repairing -@cindex repairing, tables -@node Repair, Optimization, Check, Crash recovery -@subsection How to Repair Tables +The GNR above means the group number. You can start, stop or report +any GNR, or several of them at the same time. (See --example) The GNRs +list can be comma separated, or a dash combined, of which the latter +means that all the GNRs between GNR1-GNR2 will be affected. Without +GNR argument all the found groups will be either started, stopped, or +reported. Note that you must not have any white spaces in the GNR +list. Anything after a white space is ignored. -In the following section we only talk about using @code{myisamchk} on -@code{MyISAM} tables (extensions @code{.MYI} and @code{.MYD}). If you -are using @code{ISAM} tables (extensions @code{.ISM} and @code{.ISD}), -you should use @code{isamchk} instead. +@code{mysqld_multi} supports the following options: -Starting with @strong{MySQL} Version 3.23.14, you can repair MyISAM -tables with the @code{REPAIR TABLE} command. @xref{REPAIR TABLE}. +@table @code +@cindex config-file option +@item --config-file=... +Alternative config file. NOTE: This will not affect this program's own +options (group @code{[mysqld_multi]}), but only groups +[mysqld#]. Without this option everything will be searched from the +ordinary my.cnf file. +@cindex example option +@item --example +Give an example of a config file. +@cindex help option +@item --help +Print this help and exit. +@cindex log option +@item --log=... +Log file. Full path to and the name for the log file. NOTE: If the file +exists, everything will be appended. +@cindex mysqladmin option +@item --mysqladmin=... +@code{mysqladmin} binary to be used for a server shutdown. +@cindex mysqld option +@item --mysqld=... +@code{mysqld} binary to be used. Note that you can give +@code{safe_mysqld} to this option also. The options are passed to +@code{mysqld}. Just make sure you have @code{mysqld} in your environment +variable @code{PATH} or fix @code{safe_mysqld}. +@cindex no-log option +@item --no-log +Print to stdout instead of the log file. By default the log file is +turned on. +@cindex password option +@item --password=... +Password for user for @code{mysqladmin}. +@cindex tcp-ip option +@item --tcp-ip +Connect to the @strong{MySQL} server(s) via the TCP/IP port instead of +the UNIX socket. This affects stopping and reporting. If a socket file +is missing, the server may still be running, but can be accessed only +via the TCP/IP port. By default connecting is done via the UNIX socket. +@cindex user option +@item --user=... +@strong{MySQL} user for @code{mysqladmin}. +@cindex version option +@item --version +Print the version number and exit. +@end table -The symptoms of a corrupted table include queries that abort unexpectedly -and observable errors such as these: +Some notes about @code{mysqld_multi}: @itemize @bullet @item -@file{tbl_name.frm} is locked against change +Make sure that the @strong{MySQL} user, who is stopping the +@code{mysqld} services (e.g using the @code{mysqladmin}) have the same +password and username for all the data directories accessed (to the +'mysql' database) And make sure that the user has the 'Shutdown_priv' +privilege! If you have many data- directories and many different 'mysql' +databases with different passwords for the @strong{MySQL} 'root' user, +you may want to create a common 'multi_admin' user for each using the +same password (see below). Example how to do it: +@example +shell> mysql -u root -S /tmp/mysql.sock -proot_password -e +"GRANT SHUTDOWN ON *.* TO multi_admin@@localhost IDENTIFIED BY 'multipass'" +@xref{Privileges}. +@end example +You will have to do the above for each @code{mysqld} running in each +data directory, that you have (just change the socket, -S=...) @item -Can't find file @file{tbl_name.MYI} (Errcode: ###) +@code{pid-file} is very important, if you are using @code{safe_mysqld} +to start @code{mysqld} (e.g. --mysqld=safe_mysqld) Every @code{mysqld} +should have its own @code{pid-file}. The advantage using +@code{safe_mysqld} instead of @code{mysqld} directly here is, that +@code{safe_mysqld} 'guards' every @code{mysqld} process and will restart +it, if a @code{mysqld} process fails due to signal kill -9, or +similar. (Like segmentation fault, which @strong{MySQL} should never do, +of course ;) Please note that @code{safe_mysqld} script may require that +you start it from a certain place. This means that you may have to CD to +a certain directory, before you start the @code{mysqld_multi}. If +you have problems starting, please see the @code{safe_mysqld} +script. Check especially the lines: +@example +-------------------------------------------------------------------------- +MY_PWD=`pwd` Check if we are starting this relative (for the binary +release) if test -d /data/mysql -a -f ./share/mysql/english/errmsg.sys +-a -x ./bin/mysqld +-------------------------------------------------------------------------- +@xref{safe_mysqld, , @code{safe_mysqld}}. +@end example +The above test should be successful, or you may encounter problems. @item -Unexpected end of file +Beware of the dangers starting multiple @code{mysqlds} in the same data +directory. Use separate data directories, unless you @strong{KNOW} what +you are doing! @item -Record file is crashed +The socket file and the TCP/IP port must be different for every @code{mysqld}. @item -Got error ### from table handler - -To get more information about the error you can run @code{perror ###}. Here -is the most common errors that indicates a problem with the table: +The first and fifth @code{mysqld} group were intentionally left out from +the example. You may have 'gaps' in the config file. This gives you +more flexibility. The order in which the @code{mysqlds} are started or +stopped depends on the order in which they appear in the config file. +@item +When you want to refer to a certain group using GNR with this program, +just use the number in the end of the group name ( [mysqld# <== ). +@item +You may want to use option '--user' for @code{mysqld}, but in order to +do this you need to be root when you start the @code{mysqld_multi} +script. Having the option in the config file doesn't matter; you will +just get a warning, if you are not the superuser and the @code{mysqlds} +are started under @strong{YOUR} UNIX account. @strong{IMPORTANT}: Make +sure that the @code{pid-file} and the data directory are +read+write(+execute for the latter one) accessible for @strong{THAT} +UNIX user, who the specific @code{mysqld} process is started +as. @strong{DON'T} use the UNIX root account for this, unless you +@strong{KNOW} what you are doing! +@item +@strong{MOST IMPORTANT}: Make sure that you understand the meanings of +the options that are passed to the @code{mysqlds} and why @strong{WOULD +YOU WANT} to have separate @code{mysqld} processes. Starting multiple +@code{mysqlds} in one data directory @strong{WILL NOT} give you extra +performance in a threaded system! +@end itemize -@example -shell> perror 126 127 132 134 135 136 141 144 145 -126 = Index file is crashed / Wrong file format -127 = Record-file is crashed -132 = Old database file -134 = Record was already deleted (or record file crashed) -135 = No more room in record file -136 = No more room in index file -141 = Duplicate unique key or constraint on write or update -144 = Table is crashed and last repair failed -145 = Table was marked as crashed and should be repaired -@end example +@xref{Multiple servers}. -Note that error 135, no more room in record file, is not an error that -can be fixed by a simple repair. In this case you have to do: +This is an example of the config file on behalf of @code{mysqld_multi}. @example -ALTER TABLE table MAX_ROWS=xxx AVG_ROW_LENGTH=yyy; -@end example +# This file should probably be in your home dir (~/.my.cnf) or /etc/my.cnf +# Version 2.1 by Jani Tolonen -@end itemize +[mysqld_multi] +mysqld = /usr/local/bin/safe_mysqld +mysqladmin = /usr/local/bin/mysqladmin +user = multi_admin +password = multipass -In the other cases, you must repair your tables. @code{myisamchk} -can usually detect and fix most things that go wrong. +[mysqld2] +socket = /tmp/mysql.sock2 +port = 3307 +pid-file = /usr/local/mysql/var2/hostname.pid2 +datadir = /usr/local/mysql/var2 +language = /usr/local/share/mysql/english +user = john -The repair process involves up to four stages, described below. Before you -begin, you should @code{cd} to the database directory and check the -permissions of the table files. Make sure they are readable by the Unix user -that @code{mysqld} runs as (and to you, because you need to access the files -you are checking). If it turns out you need to modify files, they must also -be writable by you. +[mysqld3] +socket = /tmp/mysql.sock3 +port = 3308 +pid-file = /usr/local/mysql/var3/hostname.pid3 +datadir = /usr/local/mysql/var3 +language = /usr/local/share/mysql/swedish +user = monty -If you are using @strong{MySQL} Version 3.23.16 and above, you can (and -should) use the @code{CHECK} and @code{REPAIR} commands to check and repair -@code{MyISAM} tables. @xref{CHECK TABLE}. @xref{REPAIR TABLE}. +[mysqld4] +socket = /tmp/mysql.sock4 +port = 3309 +pid-file = /usr/local/mysql/var4/hostname.pid4 +datadir = /usr/local/mysql/var4 +language = /usr/local/share/mysql/estonia +user = tonu -The manual section about table maintenance includes the options to -@code{isamchk}/@code{myisamchk}. @xref{Table maintenance}. +[mysqld6] +socket = /tmp/mysql.sock6 +port = 3311 +pid-file = /usr/local/mysql/var6/hostname.pid6 +datadir = /usr/local/mysql/var6 +language = /usr/local/share/mysql/japanese +user = jani +@end example -The following section is for the cases where the above command fails or -if you want to use the extended features that @code{isamchk}/@code{myisamchk} provides. +@xref{Option files}. -If you are going to repair a table from the command line, you must first -take down the @code{mysqld} server. Note that when you do -@code{mysqladmin shutdown} on a remote server, the @code{mysqld} server -will still be alive for a while after @code{mysqladmin} returns, until -all queries are stopped and all keys have been flushed to disk. - -@noindent -@strong{Stage 1: Checking your tables} - -Run @code{myisamchk *.MYI} or @code{myisamchk -e *.MYI} if you have -more time. Use the @code{-s} (silent) option to suppress unnecessary -information. +@cindex command line tool +@cindex tools, command line +@cindex scripts +@cindex @code{mysql} +@node mysql, mysqladmin, mysqld_multi, Tools +@section The Command-line Tool -If the @code{mysqld} server is done you should use the --update option to tell -@code{myisamchk} to mark the table as 'checked'. +@code{mysql} is a simple SQL shell (with GNU @code{readline} capabilities). +It supports interactive and non-interactive use. When used interactively, +query results are presented in an ASCII-table format. When used +non-interactively (for example, as a filter), the result is presented in +tab-separated format. (The output format can be changed using command-line +options.) You can run scripts simply like this: -You have to repair only those tables for which @code{myisamchk} announces an -error. For such tables, proceed to Stage 2. +@example +shell> mysql database < script.sql > output.tab +@end example -If you get weird errors when checking (such as @code{out of -memory} errors), or if @code{myisamchk} crashes, go to Stage 3. +If you have problems due to insufficient memory in the client, use the +@code{--quick} option! This forces @code{mysql} to use +@code{mysql_use_result()} rather than @code{mysql_store_result()} to +retrieve the result set. -@noindent -@strong{Stage 2: Easy safe repair} +Using @code{mysql} is very easy. Just start it as follows: +@code{mysql database} or @code{mysql --user=user_name --password=your_password database}. Type a SQL statement, end it with @samp{;}, @samp{\g}, or @samp{\G} +and press RETURN/ENTER. -NOTE: If you want repairing to go much faster, you should add: @code{-O -sort_buffer=# -O key_buffer=#} (where # is about 1/4 of the available -memory) to all @code{isamchk/myisamchk} commands. +@cindex command line options +@cindex options, command line +@cindex startup parameters +@code{mysql} supports the following options: -First, try @code{myisamchk -r -q tbl_name} (@code{-r -q} means ``quick -recovery mode''). This will attempt to repair the index file without -touching the data file. If the data file contains everything that it -should and the delete links point at the correct locations within the -data file, this should work, and the table is fixed. Start repairing the -next table. Otherwise, use the following procedure: +@table @code +@cindex help option +@item -?, --help +Display this help and exit. +@cindex automatic rehash option +@item -A, --no-auto-rehash +No automatic rehashing. One has to use 'rehash' to get table and field +completion. This gives a quicker start of mysql. +@cindex batch option +@item -B, --batch +Print results with a tab as separator, each row on a new line. Doesn't use +history file. +@cindex character sets option +@item --character-sets-dir=... +Directory where character sets are located. +@cindex compress option. +@item -C, --compress +Use compression in server/client protocol. +@cindex debug option +@item -#, --debug[=...] +Debug log. Default is 'd:t:o,/tmp/mysql.trace'. +@cindex database option +@item -D, --database=... +Database to use. This is mainly useful in the @code{my.cnf} file. +@cindex default character set option +@item --default-character-set=... +Set the default character set. +@cindex execute option +@item -e, --execute=... +Execute command and quit. (Output like with --batch) +@cindex vertical option +@item -E, --vertical +Print the output of a query (rows) vertically. Without this option you +can also force this output by ending your statements with @code{\G}. +@cindex force option +@item -f, --force +Continue even if we get a SQL error. +@cindex no-named-commands option +@item -g, --no-named-commands +Named commands are disabled. Use \* form only, or use named commands +only in the beginning of a line ending with a semicolon (;). Since +Version 10.9, the client now starts with this option ENABLED by default! +With the -g option, long format commands will still work from the first +line, however. +@cindex enable-named-commands option +@item -G, --enable-named-commands +Named commands are @strong{enabled}. Long format commands are allowed as +well as shortened \* commands. +@cindex ignore space option. +@item -i, --ignore-space +Ignore space after function names. +@cindex host option +@item -h, --host=... +Connect to the given host. +@cindex html option +@item -H, --html +Produce HTML output. +@cindex skip line numbers option +@item -L, --skip-line-numbers +Don't write line number for errors. Useful when one wants to compare result +files that includes error messages +@cindex no pager option +@item --no-pager +Disable pager and print to stdout. See interactive help (\h) also. +@cindex no tee option +@item --no-tee +Disable outfile. See interactive help (\h) also. +@cindex unbuffered option. +@item -n, --unbuffered +Flush buffer after each query. +@cindex skip column names option +@item -N, --skip-column-names +Don't write column names in results. +@cindex set variable option +@item -O, --set-variable var=option +Give a variable a value. @code{--help} lists variables. +@cindex one database option +@item -o, --one-database +Only update the default database. This is useful for skipping updates to +other database in the update log. +@cindex pager option +@item @code{--pager[=...]} +Output type. Default is your @code{ENV} variable @code{PAGER}. Valid +pagers are less, more, cat [> filename], etc. See interactive help (\h) +also. This option does not work in batch mode. Pager works only in UNIX. +@cindex password option +@item -p[password], --password[=...] +Password to use when connecting to server. If a password is not given on +the command line, you will be prompted for it. Note that if you use the +short form @code{-p} you can't have a space between the option and the +password. +@item -P --port=... +TCP/IP port number to use for connection. +@cindex quick option +@item -q, --quick +Don't cache result, print it row-by-row. This may slow down the server +if the output is suspended. Doesn't use history file. +@cindex raw option +@item -r, --raw +Write column values without escape conversion. Used with @code{--batch} +@cindex silent option +@item -s, --silent +Be more silent. +@item -S --socket=... +Socket file to use for connection. +@cindex table option +@item -t --table +Output in table format. This is default in non-batch mode. +@item -T, --debug-info +Print some debug information at exit. +@cindex tee option +@item --tee=... +Append everything into outfile. See interactive help (\h) also. Does not +work in batch mode. +@cindex user option +@item -u, --user=# +User for login if not current user. +@cindex safe updates option +@item -U, --safe-updates[=#], --i-am-a-dummy[=#] +Only allow @code{UPDATE} and @code{DELETE} that uses keys. See below for +more information about this option. You can reset this option if you have +it in your @code{my.cnf} file by using @code{--safe-updates=0}. +@cindex verbose option +@item -v, --verbose +More verbose output (-v -v -v gives the table output format). +@cindex version option +@item -V, --version +Output version information and exit. +@cindex wait option +@item -w, --wait +Wait and retry if connection is down instead of aborting. +@end table -@enumerate -@item -Make a backup of the data file before continuing. +You can also set the following variables with @code{-O} or +@code{--set-variable}: -@item -Use @code{myisamchk -r tbl_name} (@code{-r} means ``recovery mode''). This will -remove incorrect records and deleted records from the data file and -reconstruct the index file. +@cindex timeout +@multitable @columnfractions .3 .2 .5 +@item Variablename @tab Default @tab Description +@item connect_timeout @tab 0 @tab Number of seconds before timeout connection. +@item max_allowed_packet @tab 16777216 @tab Max packetlength to send/receive from to server +@item net_buffer_length @tab 16384 @tab Buffer for TCP/IP and socket communication +@item select_limit @tab 1000 @tab Automatic limit for SELECT when using --i-am-a-dummy +@item max_join_size @tab 1000000 @tab Automatic limit for rows in a join when using --i-am-a-dummy. +@end multitable -@item -If the preceding step fails, use @code{myisamchk --safe-recover tbl_name}. -Safe recovery mode uses an old recovery method that handles a few cases that -regular recovery mode doesn't (but is slower). -@end enumerate +If you type 'help' on the command line, @code{mysql} will print out the +commands that it supports: -If you get weird errors when repairing (such as @code{out of -memory} errors), or if @code{myisamchk} crashes, go to Stage 3. +@cindex commands, list of +@example +mysql> help -@noindent -@strong{Stage 3: Difficult repair} +MySQL commands: +help (\h) Display this text. +? (\h) Synonym for `help'. +clear (\c) Clear command. +connect (\r) Reconnect to the server. Optional arguments are db and host. +edit (\e) Edit command with $EDITOR. +ego (\G) Send command to mysql server, display result vertically. +exit (\q) Exit mysql. Same as quit. +go (\g) Send command to mysql server. +nopager (\n) Disable pager, print to stdout. +notee (\t) Don't write into outfile. +pager (\P) Set PAGER [to_pager]. Print the query results via PAGER. +print (\p) Print current command. +quit (\q) Quit mysql. +rehash (\#) Rebuild completion hash. +source (\.) Execute a SQL script file. Takes a file name as an argument. +status (\s) Get status information from the server. +tee (\T) Set outfile [to_outfile]. Append everything into given outfile. +use (\u) Use another database. Takes database name as argument. +@end example -You should only reach this stage if the first 16K block in the index file is -destroyed or contains incorrect information, or if the index file is -missing. In this case, it's necessary to create a new index file. Do so as -follows: +From the above, pager only works in UNIX. -@enumerate -@item -Move the data file to some safe place. +@cindex status command +The @code{status} command gives you some information about the +connection and the server you are using. If you are running in the +@code{--safe-updates} mode, @code{status} will also print the values for +the @code{mysql} variables that affect your queries. -@item -Use the table description file to create new (empty) data and index files: +@cindex @code{safe-mode} command +A useful startup option for beginners (introduced in @strong{MySQL} +Version 3.23.11) is @code{--safe-updates} (or @code{--i-am-a-dummy} for +users that has at some time done a @code{DELETE FROM table_name} but +forgot the @code{WHERE} clause). When using this option, @code{mysql} +sends the following command to the @strong{MySQL} server when opening +the connection: @example -shell> mysql db_name -mysql> SET AUTOCOMMIT=1; -mysql> TRUNCATE TABLE table_name; -mysql> quit +SET SQL_SAFE_UPDATES=1,SQL_SELECT_LIMIT=#select_limit#, + SQL_MAX_JOIN_SIZE=#max_join_size#" @end example -If your SQL version doesn't have @code{TRUNCATE TABLE}, use @code{DELETE FROM -table_name} instead. - -@item -Copy the old data file back onto the newly created data file. -(Don't just move the old file back onto the new file; you want to retain -a copy in case something goes wrong.) -@end enumerate - -Go back to Stage 2. @code{myisamchk -r -q} should work now. (This shouldn't -be an endless loop.) - -@noindent -@strong{Stage 4: Very difficult repair} - -You should reach this stage only if the description file has also -crashed. That should never happen, because the description file isn't changed -after the table is created: +where @code{#select_limit#} and @code{#max_join_size#} are variables that +can be set from the @code{mysql} command line. @xref{SET OPTION, @code{SET}}. -@enumerate -@item -Restore the description file from a backup and go back to Stage 3. You can -also restore the index file and go back to Stage 2. In the latter case, you -should start with @code{myisamchk -r}. +The effect of the above is: +@itemize @bullet @item -If you don't have a backup but know exactly how the table was created, create -a copy of the table in another database. Remove the new data file, then move -the description and index files from the other database to your crashed -database. This gives you new description and index files, but leaves -the data file alone. Go back to Stage 2 and attempt to reconstruct -the index file. -@end enumerate - -@cindex tables, optimizing -@cindex optimizing, tables -@node Optimization, , Repair, Crash recovery -@subsection Table Optimization - -To coalesce fragmented records and eliminate wasted space resulting from -deleting or updating records, run @code{myisamchk} in recovery mode: - +You are not allowed to do an @code{UPDATE} or @code{DELETE} statement +if you don't have a key constraint in the @code{WHERE} part. One can, +however, force an @code{UPDATE/DELETE} by using @code{LIMIT}: @example -shell> myisamchk -r tbl_name +UPDATE table_name SET not_key_column=# WHERE not_key_column=# LIMIT 1; @end example +@item +All big results are automatically limited to @code{#select_limit#} rows. +@item +@code{SELECT}'s that will probably need to examine more than +@code{#max_join_size} row combinations will be aborted. +@end itemize -You can optimize a table in the same way using the SQL @code{OPTIMIZE TABLE} -statement. @code{OPTIMIZE TABLE} does a repair of the table, a key -analyzes and also sorts the index tree to give faster key lookups. -There is also no possibility of unwanted interaction between a utility -and the server, because the server does all the work when you use -@code{OPTIMIZE TABLE}. @xref{OPTIMIZE TABLE}. - -@code{myisamchk} also has a number of other options you can use to improve -the performance of a table: - -@table @code -@item -S, --sort-index -@item -R index_num, --sort-records=index_num -@item -a, --analyze -@end table - -For a full description of the option. @xref{myisamchk syntax}. - -@cindex files, log -@cindex maintaining, log files -@cindex log files, maintaining -@node Log file maintenance, , Crash recovery, Maintenance -@section Log file Maintenance - -@strong{MySQL} has a lot of log files which make it easy to see what is -going. @xref{Log files}. One must however from time to time clean up -after @code{MysQL} to ensure that the logs don't take up too much disk -space. +Some useful hints about the @code{mysql} client: -When using @strong{MySQL} with log files, you will, from time to time, -want to remove/backup old log files and tell @strong{MySQL} to start -logging on new files. @xref{Backup}. +Some data is much more readable when displayed vertically, instead of +the usual horizontal box type output. For example longer text, which +includes new lines, is often much easier to be read with vertical +output. -On a Linux (@code{Redhat}) installation, you can use the -@code{mysql-log-rotate} script for this. If you installed @strong{MySQL} -from an RPM distribution, the script should have been installed -automatically. Note that you should be careful with this if you are using -the log for replication! +@example +mysql> select * from mails where length(txt) < 300 limit 300,1\G +*************************** 1. row *************************** + msg_nro: 3068 + date: 2000-03-01 23:29:50 +time_zone: +0200 +mail_from: Monty + reply: monty@@no.spam.com + mail_to: "Thimble Smith" <tim@@no.spam.com> + sbj: UTF-8 + txt: >>>>> "Thimble" == Thimble Smith writes: -On other systems you must install a short script yourself that you -start from @code{cron} to handle log files. +Thimble> Hi. I think this is a good idea. Is anyone familiar with UTF-8 +Thimble> or Unicode? Otherwise I'll put this on my TODO list and see what +Thimble> happens. -You can force @strong{MySQL} to start using new log files by using -@code{mysqladmin flush-logs} or by using the SQL command @code{FLUSH LOGS}. -If you are using @strong{MySQL} Version 3.21 you must use @code{mysqladmin refresh}. +Yes, please do that. -The above command does the following: +Regards, +Monty + file: inbox-jani-1 + hash: 190402944 +1 row in set (0.09 sec) +@end example @itemize @bullet @item -If standard logging (@code{--log}) or slow query logging -(@code{--log-slow-queries}) is used, closes and reopens the log file. -(@file{mysql.log} and @file{`hostname`-slow.log} as default). +For logging, you can use the @code{tee} option. The @code{tee} can be +started with option @code{--tee=...}, or from the command line +interactively with command @code{tee}. All the data displayed on the +screen will also be appended into a given file. This can be very useful +for debugging purposes also. The @code{tee} can be disabled from the +command line with command @code{notee}. Executing @code{tee} again +starts logging again. Without a parameter the previous file will be +used. Note that @code{tee} will flush the results into the file after +each command, just before the command line appears again waiting for the +next command. @item -If update logging (@code{--log-update}) is used, closes the update log and -opens a new log file with a higher sequence number. +Browsing, or searching the results in the interactive mode in UNIX less, +more, or any other similar program, is now possible with option +@code{--pager[=...]}. Without argument, @code{mysql} client will look +for environment variable PAGER and set @code{pager} to that. +@code{pager} can be started from the interactive command line with +command @code{pager} and disabled with command @code{nopager}. The +command takes an argument optionally and the @code{pager} will be set to +that. Command @code{pager} can be called without an argument, but this +requires that the option @code{--pager} was used, or the @code{pager} +will default to stdout. @code{pager} works only in UNIX, since it uses +the popen() function, which doesn't exist in Windows. In Windows, the +@code{tee} option can be used instead, although it may not be as handy +as @code{pager} can be in some situations. +@item +A few tips about @code{pager}: You can use it to write to a file: +@example +mysql> pager cat > /tmp/log.txt +@end example +and the results will only go to a file. You can also pass any options +for the programs that you want to use with the @code{pager}: +@example +mysql> pager less -n -i -S +@end example +From the above do note the option '-S'. You may find it very useful when +browsing the results; try the option with horizontal output (end +commands with '\g', or ';') and with vertical output (end commands with +'\G'). Sometimes a very wide result set is hard to be read from the screen, +with option -S to less you can browse the results within the interactive +less from left to right, preventing lines longer than your screen from +being continued to the next line. This can make the result set much more +readable. You can swith the mode between on and off within the interactive +less with '-S'. See the 'h' for more help about less. +@item +Last (unless you already understood this from the above examples ;) you +can combine very complex ways to handle the results, for example the +following would send the results to two files in two different +directories, on two different hard-disks mounted on /dr1 and /dr2, yet +let the results still be seen on the screen via less: +@example +mysql> pager cat | tee /dr1/tmp/res.txt | tee /dr2/tmp/res2.txt | less -n -i -S +@end example +@item +You can also combine the two functions above; have the @code{tee} +enabled, @code{pager} set to 'less' and you will be able to browse the +results in unix 'less' and still have everything appended into a file +the same time. The difference between @code{UNIX tee} used with the +@code{pager} and the @code{mysql} client in-built @code{tee}, is that +the in-built @code{tee} works even if you don't have the @code{UNIX tee} +available. The in-built @code{tee} also logs everything that is printed +on the screen, where the @code{UNIX tee} used with @code{pager} doesn't +log quite that much. Last, but not least, the interactive @code{tee} is +more handy to switch on and off, when you want to log something into a +file, but want to be able to turn the feature off sometimes. @end itemize -If you are using only an update log, you only have to flush the logs and then -move away the old update log files to a backup. -If you are using the normal logging, you can do something like: + +@cindex administration, server +@cindex server administration +@cindex @code{mysladmn} +@node mysqladmin, mysqldump, mysql, Tools +@section Administering a MySQL Server + +A utility for performing administrative operations. The syntax is: @example -shell> cd mysql-data-directory -shell> mv mysql.log mysql.old -shell> mysqladmin flush-logs +shell> mysqladmin [OPTIONS] command [command-option] command ... @end example -and then take a backup and remove @file{mysql.old}. +You can get a list of the options your version of @code{mysqladmin} supports +by executing @code{mysqladmin --help}. -@cindex functions, new -@cindex adding, new functions -@cindex user-defined functions, adding -@cindex UDFs, defined -@cindex functions, user-defined -@node Adding functions, Adding procedures, Maintenance, Top -@chapter Adding New Functions to MySQL +The current @code{mysqladmin} supports the following commands: -There are two ways to add new functions to @strong{MySQL}: +@multitable @columnfractions .3 .7 +@item create databasename @tab Create a new database. +@item drop databasename @tab Delete a database and all its tables. +@item extended-status @tab Gives an extended status message from the server. +@item flush-hosts @tab Flush all cached hosts. +@item flush-logs @tab Flush all logs. +@item flush-tables @tab Flush all tables. +@item flush-privileges @tab Reload grant tables (same as reload). +@item kill id,id,... @tab Kill mysql threads. +@item password @tab New-password. Change old password to new-password. +@item ping @tab Check if @code{mysqld} is alive. +@item processlist @tab Show list of active threads in server. +@item reload @tab Reload grant tables. +@item refresh @tab Flush all tables and close and open logfiles. +@item shutdown @tab Take server down. +@item slave-start @tab Start slave replication thread. +@item slave-stop @tab Stop slave replication thread. +@item status @tab Gives a short status message from the server. +@item variables @tab Prints variables available. +@item version @tab Get version info from server. +@end multitable -@itemize @bullet -@item You can add the function through the user-definable function (UDF) -interface. User-definable functions are added and removed dynamically using -the @code{CREATE FUNCTION} and @code{DROP FUNCTION} statements. -@xref{CREATE FUNCTION, , @code{CREATE FUNCTION}}. +All commands can be shortened to their unique prefix. For example: -@item You can add the function as a native (built in) @strong{MySQL} function. -Native functions are compiled into the @code{mysqld} server and become -available on a permanent basis. -@end itemize - -Each method has advantages and disadvantages: +@example +shell> mysqladmin proc stat ++----+-------+-----------+----+-------------+------+-------+------+ +| Id | User | Host | db | Command | Time | State | Info | ++----+-------+-----------+----+-------------+------+-------+------+ +| 6 | monty | localhost | | Processlist | 0 | | | ++----+-------+-----------+----+-------------+------+-------+------+ +Uptime: 10077 Threads: 1 Questions: 9 Slow queries: 0 Opens: 6 Flush tables: 1 Open tables: 2 Memory in use: 1092K Max memory used: 1116K +@end example -@itemize @bullet -@item -If you write a user-definable function, you must install the object file -in addition to the server itself. If you compile your function into the -server, you don't need to do that. -@item -You can add UDFs to a binary @strong{MySQL} distribution. Native functions -require you to modify a source distribution. -@item -If you upgrade your @strong{MySQL} distribution, you can continue to use your -previously installed UDFs. For native functions, you must repeat your -modifications each time you upgrade. -@end itemize +@cindex status command, results +The @code{mysqladmin status} command result has the following columns: -Whichever method you use to add new functions, they may be used just like -native functions such as @code{ABS()} or @code{SOUNDEX()}. +@cindex uptime +@multitable @columnfractions .3 .7 +@item Uptime @tab Number of seconds the @strong{MySQL} server has been up. +@cindex threads +@item Threads @tab Number of active threads (clients). +@cindex questions +@item Questions @tab Number of questions from clients since @code{mysqld} was started. +@cindex slow queries +@item Slow queries @tab Queries that have taken more than @code{long_query_time} seconds. @xref{Slow query log}. +@cindex opens +@item Opens @tab How many tables @code{mysqld} has opened. +@cindex flush tables +@cindex tables, flush +@item Flush tables @tab Number of @code{flush ...}, @code{refresh}, and @code{reload} commands. +@cindex open tables +@item Open tables @tab Number of tables that are open now. +@cindex memory use +@item Memory in use @tab Memory allocated directly by the @code{mysqld} code (only available when @strong{MySQL} is compiled with --with-debug=full). +@cindex max memory used +@item Max memory used @tab Maximum memory allocated directly by the @code{mysqld} code (only available when @strong{MySQL} is compiled with --with-debug=full). +@end multitable -@menu -* Adding UDF:: Adding a new user-definable function -* Adding native function:: Adding a new native function -@end menu +If you do @code{myslqadmin shutdown} on a socket (in other words, on a +the computer where @code{mysqld} is running), @code{mysqladmin} will +wait until the @strong{MySQL} @code{pid-file} is removed to ensure that +the @code{mysqld} server has stopped properly. -@cindex adding, user-definable functions -@cindex user-defined functions, adding -@cindex functions, user-definable, adding -@node Adding UDF, Adding native function, Adding functions, Adding functions -@section Adding a New User-definable Function +@cindex dumping, databases +@cindex databases, dumping +@cindex tables, dumping +@cindex backing up, databases +@node mysqldump, mysqlhotcopy, mysqladmin, Tools +@section Dumping the Structure and Data from MySQL Databases and Tables -@menu -* UDF calling sequences:: UDF calling sequences -* UDF arguments:: Argument processing -* UDF return values:: Return values and error handling -* UDF compiling:: Compiling and installing user-definable functions -@end menu +@cindex @code{mysqldump} +Utility to dump a database or a collection of database for backup or for +transferring the data to another SQL server (not necessarily a @strong{MySQL} +server). The dump will contain SQL statements to create the table +and/or populate the table. -For the UDF mechanism to work, functions must be written in C or C++ and your -operating system must support dynamic loading. The @strong{MySQL} source -distribution includes a file @file{sql/udf_example.cc} that defines 5 new -functions. Consult this file to see how UDF calling conventions work. +If you are doing a backup on the server, you should consider using +the @code{mysqlhotcopy} instead. @xref{mysqlhotcopy, , @code{mysqlhotcopy}}. -For @code{mysqld} to be able to use UDF functions, you should configure MySQL -with @code{--with-mysqld-ldflags=-rdynamic} The reason is that to on -many platforms (including Linux) you can load a dynamic library (with -@code{dlopen()}) from a static linked program, which you would get if -you are using @code{--with-mysqld-ldflags=-all-static} If you want to -use an UDF that needs to access symbols from @code{mysqld} (like the -@code{methaphone} example in @file{sql/udf_example.cc} that uses -@code{default_charset_info}), you must link the program with -@code{-rdynamic}. (see @code{man dlopen}). +@example +shell> mysqldump [OPTIONS] database [tables] +OR mysqldump [OPTIONS] --databases [OPTIONS] DB1 [DB2 DB3...] +OR mysqldump [OPTIONS] --all-databases [OPTIONS] +@end example -For each function that you want to use in SQL statements, you should define -corresponding C (or C++) functions. In the discussion below, the name -``xxx'' is used for an example function name. To distinquish between SQL and -C/C++ usage, @code{XXX()} (uppercase) indicates a SQL function call, and -@code{xxx()} (lowercase) indicates a C/C++ function call. +If you don't give any tables or use the @code{--databases} or +@code{--all-databases}, the whole database(s) will be dumped. -The C/C++ functions that you write to implement the interface for -@code{XXX()} are: +You can get a list of the options your version of @code{mysqldump} supports +by executing @code{mysqldump --help}. -@table @asis -@item @code{xxx()} (required) -The main function. This is where the function result is computed. -The correspondence between the SQL type and return type of your C/C++ -function is shown below: +Note that if you run @code{mysqldump} without @code{--quick} or +@code{--opt}, @code{mysqldump} will load the whole result set into +memory before dumping the result. This will probably be a problem if +you are dumping a big database. -@multitable @columnfractions .2 .8 -@item @strong{SQL type} @tab @strong{C/C++ type} -@item @code{STRING} @tab @code{char *} -@item @code{INTEGER} @tab @code{long long} -@item @code{REAL} @tab @code{double} -@end multitable +Note that if you are using a new copy of the @code{mysqldump} program +and you are going to do a dump that will be read into a very old @strong{MySQL} +server, you should not use the @code{--opt} or @code{-e} options. -@item @code{xxx_init()} (optional) -The initialization function for @code{xxx()}. It can be used to: +@code{mysqldump} supports the following options: -@itemize @bullet -@item -Check the number of arguments to @code{XXX()}. -@item -Check that the arguments are of a required type or, alternatively, -tell @strong{MySQL} to coerce arguments to the types you want when -the main function is called. -@item -Allocate any memory required by the main function. -@item -Specify the maximum length of the result. -@item -Specify (for @code{REAL} functions) the maximum number of decimals. -@item -Specify whether or not the result can be @code{NULL}. -@end itemize +@table @code +@item --add-locks +Add @code{LOCK TABLES} before and @code{UNLOCK TABLE} after each table dump. +(To get faster inserts into @strong{MySQL}.) +@item --add-drop-table +Add a @code{drop table} before each create statement. +@item -A, --all-databases +Dump all the databases. This will be same as @code{--databases} with all +databases selected. +@item -a, --all +Include all @strong{MySQL}-specific create options. +@item --allow-keywords +Allow creation of column names that are keywords. This works by +prefixing each column name with the table name. +@item -c, --complete-insert +Use complete insert statements (with column names). +@item -C, --compress +Compress all information between the client and the server if both support +compression. +@item -B, --databases +To dump several databases. Note the difference in usage. In this case +no tables are given. All name arguments are regarded as database names. +@code{USE db_name;} will be included in the output before each new database. +@item --delayed +Insert rows with the @code{INSERT DELAYED} command. +@item -e, --extended-insert +Use the new multiline @code{INSERT} syntax. (Gives more compact and +faster inserts statements.) +@item -#, --debug[=option_string] +Trace usage of the program (for debugging). +@item --help +Display a help message and exit. +@item --fields-terminated-by=... +@itemx --fields-enclosed-by=... +@itemx --fields-optionally-enclosed-by=... +@itemx --fields-escaped-by=... +@itemx --lines-terminated-by=... +These options are used with the @code{-T} option and have the same +meaning as the corresponding clauses for @code{LOAD DATA INFILE}. +@xref{LOAD DATA, , @code{LOAD DATA}}. +@item -F, --flush-logs +Flush log file in the @strong{MySQL} server before starting the dump. +@item -f, --force, +Continue even if we get a SQL error during a table dump. +@item -h, --host=.. +Dump data from the @strong{MySQL} server on the named host. The default host +is @code{localhost}. +@item -l, --lock-tables. +Lock all tables before starting the dump. The tables are locked with +@code{READ LOCAL} to allow concurrent inserts in the case of @code{MyISAM} +tables. +@item -n, --no-create-db +'CREATE DATABASE /*!32312 IF NOT EXISTS*/ db_name;' will not be put in the +output. The above line will be added otherwise, if --databases or +--all-databases option was given. +@item -t, --no-create-info +Don't write table creation information (The @code{CREATE TABLE} statement.) +@item -d, --no-data +Don't write any row information for the table. This is very useful if you +just want to get a dump of the structure for a table! +@item --opt +Same as @code{--quick --add-drop-table --add-locks --extended-insert +--lock-tables}. Should give you the fastest possible dump for reading +into a @strong{MySQL} server. +@item -pyour_pass, --password[=your_pass] +The password to use when connecting to the server. If you specify +no @samp{=your_pass} part, +@code{mysqldump} you will be prompted for a password. +@item -P port_num, --port=port_num +The TCP/IP port number to use for connecting to a host. (This is used for +connections to hosts other than @code{localhost}, for which Unix sockets are +used.) +@item -q, --quick +Don't buffer query, dump directly to stdout. Uses @code{mysql_use_result()} +to do this. +@item -r, --result-file=... +Direct output to a given file. This option should be used in MSDOS, +because it prevents new line '\n' from being converted to '\n\r' (new +line + carriage return). +@item -S /path/to/socket, --socket=/path/to/socket +The socket file to use when connecting to @code{localhost} (which is the +default host). +@item --tables +Overrides option --databases (-B). +@item -T, --tab=path-to-some-directory +Creates a @code{table_name.sql} file, that contains the SQL CREATE commands, +and a @code{table_name.txt} file, that contains the data, for each give table. +@strong{NOTE}: This only works if @code{mysqldump} is run on the same +machine as the @code{mysqld} daemon. The format of the @code{.txt} file +is made according to the @code{--fields-xxx} and @code{--lines--xxx} options. +@item -u user_name, --user=user_name +The @strong{MySQL} user name to use when connecting to the server. The +default value is your Unix login name. +@item -O var=option, --set-variable var=option +Set the value of a variable. The possible variables are listed below. +@item -v, --verbose +Verbose mode. Print out more information on what the program does. +@item -V, --version +Print version information and exit. +@item -w, --where='where-condition' +Dump only selected records. Note that QUOTES are mandatory: -@item @code{xxx_deinit()} (optional) -The deinitialization function for @code{xxx()}. It should deallocate any -memory allocated by the initialization function. +@example +"--where=user='jimf'" "-wuserid>1" "-wuserid<1" +@end example +@item -O net_buffer_length=#, where # < 16M +When creating multi-row-insert statements (as with option +@code{--extended-insert} or @code{--opt}), @code{mysqldump} will create +rows up to @code{net_buffer_length} length. If you increase this +variable, you should also ensure that the @code{max_allowed_packet} +variable in the @strong{MySQL} server is bigger than the +@code{net_buffer_length}. @end table -When a SQL statement invokes @code{XXX()}, @strong{MySQL} calls the -initialization function @code{xxx_init()} to let it perform any required -setup, such as argument checking or memory allocation. If @code{xxx_init()} -returns an error, the SQL statement is aborted with an error message and the -main and deinitialization functions are not called. Otherwise, the main -function @code{xxx()} is called once for each row. After all rows have been -processed, the deinitialization function @code{xxx_deinit()} is called so it -can perform any required cleanup. - -All functions must be thread safe (not just the main function, -but the initialization and deinitialization functions as well). This means -that you are not allowed to allocate any global or static variables that -change! If you need memory, you should allocate it in @code{xxx_init()} -and free it in @code{xxx_deinit()}. - -@cindex calling sequences, UDF -@node UDF calling sequences, UDF arguments, Adding UDF, Adding UDF -@subsection UDF Calling Sequences +The most normal use of @code{mysqldump} is probably for making a backup of +whole databases. @xref{Backup}. -The main function should be declared as shown below. Note that the return -type and parameters differ, depending on whether you will declare the SQL -function @code{XXX()} to return @code{STRING}, @code{INTEGER}, or @code{REAL} -in the @code{CREATE FUNCTION} statement: +@example +mysqldump --opt database > backup-file.sql +@end example -@noindent -For @code{STRING} functions: +You can read this back into @strong{MySQL} with: @example -char *xxx(UDF_INIT *initid, UDF_ARGS *args, - char *result, unsigned long *length, - char *is_null, char *error); +mysql database < backup-file.sql @end example -@noindent -For @code{INTEGER} functions: +or @example -long long xxx(UDF_INIT *initid, UDF_ARGS *args, - char *is_null, char *error); +mysql -e "source /patch-to-backup/backup-file.sql" database @end example -@noindent -For @code{REAL} functions: +However, it's also very useful to populate another @strong{MySQL} server with +information from a database: @example -double xxx(UDF_INIT *initid, UDF_ARGS *args, - char *is_null, char *error); +mysqldump --opt database | mysql ---host=remote-host -C database @end example -The initialization and deinitialization functions are declared like this: +It is possible to dump several databases with one command: @example -my_bool xxx_init(UDF_INIT *initid, UDF_ARGS *args, char *message); - -void xxx_deinit(UDF_INIT *initid); +mysqldump --databases database1 [database2 database3...] > my_databases.sql @end example -The @code{initid} parameter is passed to all three functions. It points to a -@code{UDF_INIT} structure that is used to communicate information between -functions. The @code{UDF_INIT} structure members are listed below. The -initialization function should fill in any members that it wishes to change. -(To use the default for a member, leave it unchanged.): - -@table @code -@item my_bool maybe_null -@code{xxx_init()} should set @code{maybe_null} to @code{1} if @code{xxx()} -can return @code{NULL}. The default value is @code{1} if any of the -arguments are declared @code{maybe_null}. +If all the databases are wanted, one can use: -@item unsigned int decimals -Number of decimals. The default value is the maximum number of decimals in -the arguments passed to the main function. (For example, if the function is -passed @code{1.34}, @code{1.345}, and @code{1.3}, the default would be 3, -because @code{1.345} has 3 decimals. +@example +mysqldump --all-databases > all_databases.sql +@end example -@item unsigned int max_length -The maximum length of the string result. The default value differs depending -on the result type of the function. For string functions, the default is the -length of the longest argument. For integer functions, the default is 21 -digits. For real functions, the default is 13 plus the number of decimals -indicated by @code{initid->decimals}. (For numeric functions, the length -includes any sign or decimal point characters.) +@cindex dumping, databases +@cindex databases, dumping +@cindex tables, dumping +@cindex backing up, databases +@node mysqlhotcopy, mysqlimport, mysqldump, Tools +@section Copying MySQL Databases and Tables -@item char *ptr -A pointer that the function can use for its own purposes. For example, -functions can use @code{initid->ptr} to communicate allocated memory -between functions. In @code{xxx_init()}, allocate the memory and assign it -to this pointer: +@code{mysqlhotcopy} is a perl script that uses @code{LOCK TABLES}, +@code{FLUSH TABLES} and @code{cp} or @code{scp} to quickly make a backup +of a database. It's the fastest way to make a backup of the database, +of single tables but it can only be run on the same machine where the +database directories are. @example -initid->ptr = allocated_memory; -@end example +mysqlhotcopy db_name [/path/to/new_directory] -In @code{xxx()} and @code{xxx_deinit()}, refer to @code{initid->ptr} to use -or deallocate the memory. -@end table +mysqlhotcopy db_name_1 ... db_name_n /path/to/new_directory -@cindex argument processing -@cindex processing, arguments -@node UDF arguments, UDF return values, UDF calling sequences, Adding UDF -@subsection Argument Processing +mysqlhotcopy db_name./regex/ +@end example -The @code{args} parameter points to a @code{UDF_ARGS} structure that thas the -members listed below: +@code{mysqlhotcopy} supports the following options: @table @code -@item unsigned int arg_count -The number of arguments. Check this value in the initialization function -if you want your function to be called with a particular number of arguments. -For example: +@item -?, --help +Display a help screen and exit +@item -u, --user=# +User for database login +@item -p, --password=# +Password to use when connecting to server +@item -P, --port=# +Port to use when connecting to local server +@item -S, --socket=# +Socket to use when connecting to local server +@item --allowold +Don't abort if target already exists (rename it _old) +@item --keepold +Don't delete previous (now renamed) target when done +@item --noindices +Don't include full index files in copy to make the backup smaller and faster +The indexes can later be reconstructed with @code{myisamchk -rq.}. +@item --method=# +Method for copy (@code{cp} or @code{scp}). +@item -q, --quiet +Be silent except for errors +@item --debug +Enable debug +@item -n, --dryrun +Report actions without doing them +@item --regexp=# +Copy all databases with names matching regexp +@item --suffix=# +Suffix for names of copied databases +@item --checkpoint=# +Insert checkpoint entry into specified db.table +@item --flushlog +Flush logs once all tables are locked. +@item --tmpdir=# +Temporary directory (instead of /tmp). +@end table -@example -if (args->arg_count != 2) -@{ - strcpy(message,"XXX() requires two arguments"); - return 1; -@} -@end example +You can use @code{perldoc mysqlhotcopy} to get a more complete +documentation for @code{mysqlhotcopy}. +@code{mysqlhotcopy} reads the groups @code{[client]} and @code{[mysqlhotcopy]} +from the option files. -@item enum Item_result *arg_type -The types for each argument. The possible type values are -@code{STRING_RESULT}, @code{INT_RESULT}, and @code{REAL_RESULT}. +To be able to execute @code{mysqlhotcopy} you need write access to the +backup directory, @code{SELECT} privilege to the tables you are about to +copy and the @strong{MySQL} @code{Reload} privilege (to be able to +execute @code{FLUSH TABLES}). -To make sure that arguments are of a given type and return an -error if they are not, check the @code{arg_type} array in the initialization -function. For example: +@cindex importing, data +@cindex data, importing +@cindex files, text +@cindex text files, importing +@cindex @code{mysqlimport} +@node mysqlimport, perror, mysqlhotcopy, Tools +@section Importing Data from Text Files -@example -if (args->arg_type[0] != STRING_RESULT || - args->arg_type[1] != INT_RESULT) -@{ - strcpy(message,"XXX() requires a string and an integer"); - return 1; -@} -@end example +@code{mysqlimport} provides a command-line interface to the @code{LOAD DATA +INFILE} SQL statement. Most options to @code{mysqlimport} correspond +directly to the same options to @code{LOAD DATA INFILE}. +@xref{LOAD DATA, , @code{LOAD DATA}}. -As an alternative to requiring your function's arguments to be of particular -types, you can use the initialization function to set the @code{arg_type} -elements to the types you want. This causes @strong{MySQL} to coerce -arguments to those types for each call to @code{xxx()}. For example, to -specify coercion of the first two arguments to string and integer, do this in -@code{xxx_init()}: +@code{mysqlimport} is invoked like this: @example -args->arg_type[0] = STRING_RESULT; -args->arg_type[1] = INT_RESULT; +shell> mysqlimport [options] database textfile1 [textfile2....] @end example -@item char **args -@code{args->args} communicates information to the initialization function -about the general nature of the arguments your function was called with. For a -constant argument @code{i}, @code{args->args[i]} points to the argument -value. (See below for instructions on how to access the value properly.) -For a non-constant argument, @code{args->args[i]} is @code{0}. -A constant argument is an expression that uses only constants, such as -@code{3} or @code{4*7-2} or @code{SIN(3.14)}. A non-constant argument is an -expression that refers to values that may change from row to row, such as -column names or functions that are called with non-constant arguments. +For each text file named on the command line, +@code{mysqlimport} strips any extension from the filename and uses the result +to determine which table to import the file's contents into. For example, +files named @file{patient.txt}, @file{patient.text}, and @file{patient} would +all be imported into a table named @code{patient}. -For each invocation of the main function, @code{args->args} contains the -actual arguments that are passed for the row currently being processed. +@code{mysqlimport} supports the following options: -Functions can refer to an argument @code{i} as follows: +@table @code +@item -c, --columns=... +This option takes a comma-separated list of field names as an argument. +The field list is used to create a proper @code{LOAD DATA INFILE} command, +which is then passed to @strong{MySQL}. @xref{LOAD DATA, , @code{LOAD DATA}}. -@itemize @bullet -@item -An argument of type @code{STRING_RESULT} is given as a string pointer plus a -length, to allow handling of binary data or data of arbitrary length. The -string contents are available as @code{args->args[i]} and the string length -is @code{args->lengths[i]}. You should not assume that strings are -null-terminated. +@item -C, --compress +Compress all information between the client and the server if both support +compression. -@item -For an argument of type @code{INT_RESULT}, you must cast -@code{args->args[i]} to a @code{long long} value: +@item -#, --debug[=option_string] +Trace usage of the program (for debugging). -@example -long long int_val; -int_val = *((long long*) args->args[i]); -@end example +@item -d, --delete +Empty the table before importing the text file. -@item -For an argument of type @code{REAL_RESULT}, you must cast -@code{args->args[i]} to a @code{double} value: - -@example -double real_val; -real_val = *((double*) args->args[i]); -@end example -@end itemize - -@item unsigned long *lengths -For the initialization function, the @code{lengths} array indicates the -maximum string length for each argument. For each invocation of the main -function, @code{lengths} contains the actual lengths of any string arguments -that are passed for the row currently being processed. For arguments of -types @code{INT_RESULT} or @code{REAL_RESULT}, @code{lengths} still contains -the maximum length of the argument (as for the initialization function). -@end table - -@cindex UDFs, return values -@cindex return values, UDFs -@cindex errors, handling for UDFs -@cindex handling, errors -@node UDF return values, UDF compiling, UDF arguments, Adding UDF -@subsection Return Values and Error Handling - -The initialization function should return @code{0} if no error occurred and -@code{1} otherwise. If an error occurs, @code{xxx_init()} should store a -null-terminated error message in the @code{message} parameter. The message -will be returned to the client. The message buffer is -@code{MYSQL_ERRMSG_SIZE} characters long, but you should try to keep the -message to less than 80 characters so that it fits the width of a standard -terminal screen. +@item --fields-terminated-by=... +@itemx --fields-enclosed-by=... +@itemx --fields-optionally-enclosed-by=... +@itemx --fields-escaped-by=... +@itemx --lines-terminated-by=... +These options have the same meaning as the corresponding clauses for +@code{LOAD DATA INFILE}. @xref{LOAD DATA, , @code{LOAD DATA}}. -The return value of the main function @code{xxx()} is the function value, for -@code{long long} and @code{double} functions. A string functions should -return a pointer to the result and store the length of the string in the -@code{length} arguments. @code{result} is a buffer at least 255 bytes long. -Set these to the contents and length of the return value. For example: +@item -f, --force +Ignore errors. For example, if a table for a text file doesn't exist, +continue processing any remaining files. Without @code{--force}, +@code{mysqlimport} exits if a table doesn't exist. -@example -memcpy(result, "result string", 13); -*length = 13; -@end example +@item --help +Display a help message and exit. -If your string functions that needs to return a string longer than 255 -bytes, you must allocate the space for it with @code{malloc()} in your -@code{xxx_init()} function or your @code{xxx()} function and free it in -your @code{xxx_deinit()} function. You can store the allocated memory -in the @code{ptr} slot in the @code{UDF_INIT} structure for reuse by -future @code{xxx()} calls. @xref{UDF calling sequences}. +@item -h host_name, --host=host_name +Import data to the @strong{MySQL} server on the named host. The default host +is @code{localhost}. -To indicate a return value of @code{NULL} in the main function, set -@code{is_null} to @code{1}: +@item -i, --ignore +See the description for the @code{--replace} option. -@example -*is_null = 1; -@end example +@item -l, --lock-tables +Lock @strong{ALL} tables for writing before processing any text files. This +ensures that all tables are synchronized on the server. -To indicate an error return in the main function, set the @code{error} -parameter to @code{1}: +@item -L, --local +Read input files from the client. By default, text files are assumed to be on +the server if you connect to @code{localhost} (which is the default host). -@example -*error = 1; -@end example +@item -pyour_pass, --password[=your_pass] +The password to use when connecting to the server. If you specify +no @samp{=your_pass} part, +@code{mysqlimport} you will be prompted for a password. -If @code{xxx()} sets @code{*error} to @code{1} for any row, the function -value is @code{NULL} for the current row and for any subsequent rows -processed by the statement in which @code{XXX()} was invoked. (@code{xxx()} -will not even be called for subsequent rows.) @strong{NOTE:} In -@strong{MySQL} versions prior to 3.22.10, you should set both @code{*error} -and @code{*is_null}: +@item -P port_num, --port=port_num +The TCP/IP port number to use for connecting to a host. (This is used for +connections to hosts other than @code{localhost}, for which Unix sockets are +used.) -@example -*error = 1; -*is_null = 1; -@end example +@item -r, --replace +The @code{--replace} and @code{--ignore} options control handling of input +records that duplicate existing records on unique key values. If you specify +@code{--replace}, new rows replace existing rows that have the same unique key +value. If you specify @code{--ignore}, input rows that duplicate an existing +row on a unique key value are skipped. If you don't specify either option, an +error occurs when a duplicate key value is found, and the rest of the text +file is ignored. -@cindex compiling, user-defined functions -@cindex UDFs, compiling -@cindex installing, user-defined functions -@node UDF compiling, , UDF return values, Adding UDF -@subsection Compiling and Installing User-definable Functions +@item -s, --silent +Silent mode. Write output only when errors occur. -Files implementing UDFs must be compiled and installed on the host where the -server runs. This process is described below for the example UDF file -@file{udf_example.cc} that is included in the @strong{MySQL} source -distribution. This file contains the following functions: +@item -S /path/to/socket, --socket=/path/to/socket +The socket file to use when connecting to @code{localhost} (which is the +default host). -@itemize @bullet -@item -@code{metaphon()} returns a metaphon string of the string argument. -This is something like a soundex string, but it's more tuned for English. -@item -@code{myfunc_double()} returns the sum of the ASCII values of the -characters in its arguments, divided by the sum of the length of its arguments. -@item -@code{myfunc_int()} returns the sum of the length of its arguments. -@item -@code{sequence([const int])} returns an sequence starting from the given -number or 1 if no number has been given. -@item -@code{lookup()} returns the IP number for a hostname. -@item -@code{reverse_lookup()} returns the hostname for an IP number. -The function may be called with a string @code{"xxx.xxx.xxx.xxx"} or -four numbers. -@end itemize +@item -u user_name, --user=user_name +The @strong{MySQL} user name to use when connecting to the server. The +default value is your Unix login name. -A dynamically loadable file should be compiled as a sharable object file, -using a command something like this: +@item -v, --verbose +Verbose mode. Print out more information what the program does. -@example -shell> gcc -shared -o udf_example.so myfunc.cc -@end example +@item -V, --version +Print version information and exit. +@end table -You can easily find out the correct compiler options for your system by -running this command in the @file{sql} directory of your @strong{MySQL} -source tree: +Here is a sample run using @code{mysqlimport}: @example -shell> make udf_example.o +$ mysql --version +mysql Ver 9.33 Distrib 3.22.25, for pc-linux-gnu (i686) +$ uname -a +Linux xxx.com 2.2.5-15 #1 Mon Apr 19 22:21:09 EDT 1999 i586 unknown +$ mysql -e 'CREATE TABLE imptest(id INT, n VARCHAR(30))' test +$ ed +a +100 Max Sydow +101 Count Dracula +. +w imptest.txt +32 +q +$ od -c imptest.txt +0000000 1 0 0 \t M a x S y d o w \n 1 0 +0000020 1 \t C o u n t D r a c u l a \n +0000040 +$ mysqlimport --local test imptest.txt +test.imptest: Records: 2 Deleted: 0 Skipped: 0 Warnings: 0 +$ mysql -e 'SELECT * FROM imptest' test ++------+---------------+ +| id | n | ++------+---------------+ +| 100 | Max Sydow | +| 101 | Count Dracula | ++------+---------------+ @end example -You should run a compile command similar to the one that @code{make} displays, -except that you should remove the @code{-c} option near the end of the line -and add @code{-o udf_example.so} to the end of the line. (On some systems, -you may need to leave the @code{-c} on the command.) - -Once you compile a shared object containing UDFs, you must install it -and tell @strong{MySQL} about it. Compiling a shared object from -@file{udf_example.cc} produces a file named something like -@file{udf_example.so} (the exact name may vary from platform to platform). -Copy this file to some directory searched by @code{ld}, such as -@file{/usr/lib}. On many systems, you can set the @code{LD_LIBRARY} or -@code{LD_LIBRARY_PATH} environment variable to point at the directory where -you have your UDF function files. The @code{dlopen} manual page tells you -which variable you should use on your system. You should set this in -@code{mysql.server} or @code{safe_mysqld} and restart @code{mysqld}. +@cindex error messages, displaying +@cindex perror +@node perror, mysqlshow, mysqlimport, Tools +@section Converting an error code to the corresponding error message -After the library is installed, notify @code{mysqld} about the new -functions with these commands: +@code{perror} can be used to print error message(s). @code{perror} can +be invoked like this: @example -mysql> CREATE FUNCTION metaphon RETURNS STRING SONAME "udf_example.so"; -mysql> CREATE FUNCTION myfunc_double RETURNS REAL SONAME "udf_example.so"; -mysql> CREATE FUNCTION myfunc_int RETURNS INTEGER SONAME "udf_example.so"; -mysql> CREATE FUNCTION lookup RETURNS STRING SONAME "udf_example.so"; -mysql> CREATE FUNCTION reverse_lookup RETURNS STRING SONAME "udf_example.so"; -@end example +shell> perror [OPTIONS] [ERRORCODE [ERRORCODE...]] -Functions can be deleted using @code{DROP FUNCTION}: +For example: -@example -mysql> DROP FUNCTION metaphon; -mysql> DROP FUNCTION myfunc_double; -mysql> DROP FUNCTION myfunc_int; -mysql> DROP FUNCTION lookup; -mysql> DROP FUNCTION reverse_lookup; +shell> perror 64 79 +Error code 64: Machine is not on the network +Error code 79: Can not access a needed shared library @end example -The @code{CREATE FUNCTION} and @code{DROP FUNCTION} statements update the -system table @code{func} in the @code{mysql} database. The function's name, -type and shared library name are saved in the table. You must have the -@strong{insert} and @strong{delete} privileges for the @code{mysql} database -to create and drop functions. - -You should not use @code{CREATE FUNCTION} to add a function that has already -been created. If you need to reinstall a function, you should remove it with -@code{DROP FUNCTION} and then reinstall it with @code{CREATE FUNCTION}. You -would need to do this, for example, if you recompile a new version of your -function, so that @code{mysqld} gets the new version. Otherwise the server -will continue to use the old version. - -Active functions are reloaded each time the server starts, unless you start -@code{mysqld} with the @code{--skip-grant-tables} option. In this case, UDF -initialization is skipped and UDFs are unavailable. (An active function is -one that has been loaded with @code{CREATE FUNCTION} and not removed with -@code{DROP FUNCTION}.) - -@cindex adding, native functions -@cindex native functions, adding -@cindex functions, native, adding -@node Adding native function, , Adding UDF, Adding functions -@section Adding a New Native Function - -The procedure for adding a new native function is described below. Note -that you cannot add native functions to a binary distribution because -the procedure involves modifying @strong{MySQL} source code. You must -compile @strong{MySQL} yourself from a source distribution. Also note -that if you migrate to another version of @strong{MySQL} (for example, -when a new version is released), you will need to repeat the procedure -with the new version. +@code{perror} can be used to display a description for a system error +code, or an MyISAM/ISAM table handler error code. The error messages +are mostly system dependent. -To add a new native @strong{MySQL} function, follow these steps: +@cindex databases, displaying +@cindex displaying, database information +@cindex tables, displaying +@cindex columns, displaying +@cindex showing, database information +@node mysqlshow, myisampack, perror, Tools +@section Showing Databases, Tables, and Columns -@enumerate -@item -Add one line to @file{lex.h} that defines the function name in the -@code{sql_functions[]} array. -@item -If the function prototype is simple (just takes zero, one, two or three -arguments), you should in lex.h specify SYM(FUNC_ARG#) (where # is the -number of arguments) as the second argument in the -@code{sql_functions[]} array and add a function that creates a function -object in @file{item_create.cc}. Take a look at @code{"ABS"} and -@code{create_funcs_abs()} for an example of this. +@code{mysqlshow} can be used to quickly look at which databases exist, +their tables, and the table's columns. -If the function prototype is complicated (for example takes a variable number -of arguments), you should add two lines to @file{sql_yacc.yy}. One -indicates the preprocessor symbol that @code{yacc} should define (this -should be added at the beginning of the file). Then define the function -parameters and add an ``item'' with these parameters to the -@code{simple_expr} parsing rule. For an example, check all occurrences -of @code{ATAN} in @file{sql_yacc.yy} to see how this is done. -@item -In @file{item_func.h}, declare a class inheriting from @code{Item_num_func} or -@code{Item_str_func}, depending on whether your function returns a number or a -string. -@item -In @file{item_func.cc}, add one of the following declarations, depending -on whether you are defining a numeric or string function: -@example -double Item_func_newname::val() -longlong Item_func_newname::val_int() -String *Item_func_newname::Str(String *str) -@end example +With the @code{mysql} program you can get the same information with the +@code{SHOW} commands. @xref{SHOW}. -If you inherit your object from any of the standard items (like -@code{Item_num_func} you probably only have to define one of the above -functions and let the parent object take care of the other functions. -For example, the @code{Item_str_func} class defines a @code{val()} function -that executes @code{atof()} on the value returned by @code{::str()}. +@code{mysqlshow} is invoked like this: -@item -You should probably also define the following object function: @example -void Item_func_newname::fix_length_and_dec() +shell> mysqlshow [OPTIONS] [database [table [column]]] @end example -This function should at least calculate @code{max_length} based on the -given arguments. @code{max_length} is the maximum number of characters -the function may return. This function should also set @code{maybe_null -= 0} if the main function can't return a @code{NULL} value. The -function can check if any of the function arguments can return -@code{NULL} by checking the arguments @code{maybe_null} variable. You -can take a look at @code{Item_func_mod::fix_length_and_dec} for a -typical example of how to do this. -@end enumerate - -All functions must be thread safe (In other words, don't use any global or -static variables in the functions without protecting them with mutexes). - -If you want to return @code{NULL}, from @code{::val()}, @code{::val_int()} -or @code{::str()} you should set @code{null_value} to 1 and return 0. - -For @code{::str()} object functions, there are some additional -considerations to be aware of: @itemize @bullet @item -The @code{String *str} argument provides a string buffer that may be -used to hold the result. (For more information about the @code{String} type, -take a look at the @file{sql_string.h} file.) +If no database is given, all matching databases are shown. @item -The @code{::str()} function should return the string that holds the result or -@code{(char*) 0} if the result is @code{NULL}. +If no table is given, all matching tables in the database are shown. @item -All current string functions try to avoid allocating any memory unless -absolutely necessary! +If no column is given, all matching columns and column types in the table +are shown. @end itemize -@cindex procedures, adding -@cindex adding, procedures -@cindex new procedures, adding -@node Adding procedures, ODBC, Adding functions, Top -@chapter Adding New Procedures to MySQL - -In @strong{MySQL}, you can define a procedure in C++ that can access and -modify the data in a query before it is sent to the client. The modification -can be done on row-by-row or @code{GROUP BY} level. - -We have created an example procedure in @strong{MySQL} Version 3.23 to -show you what can be done. +Note that in newer @strong{MySQL} versions, you only see those +database/tables/columns for which you have some privileges. -Additionally we recommend you to take a look at 'mylua', which you can find in the Contrib directory. @xref{Contrib}. Which this you can use the LUA -language to load a procedure at runtime into @code{mysqld}. +If the last argument contains a shell or SQL wild-card (@code{*}, @code{?}, +@code{%} or @code{_}) then only what's matched by the wild card is shown. +This may cause some confusion when you try to display the columns for a +table with a @code{_} as in this case @code{mysqlshow} only shows you +the table names that match the pattern. This is easily fixed by +adding an extra @code{%} last on the command line (as a separate +argument). -@menu -* procedure analyse:: Procedure analyse -* Writing a procedure:: Writing a procedure. -@end menu +@cindex compressed tables +@cindex tables, compressed +@cindex MyISAM, compressed tables +@cindex @code{myisampack} +@cindex @code{pack_isam} +@node myisampack, , mysqlshow, Tools +@section The MySQL Compressed Read-only Table Generator -@node procedure analyse, Writing a procedure, Adding procedures, Adding procedures -@section Procedure Analyse +@code{myisampack} is used to compress MyISAM tables, and @code{pack_isam} +is used to compress ISAM tables. Because ISAM tables are deprecated, we +will only discuss @code{myisampack} here, but everything said about +@code{myisampack} should also be true for @code{pack_isam}. -@code{analyse([max elements,[max memory]])} +@code{myisampack} works by compressing each column in the table separately. +The information needed to decompress columns is read into memory when the +table is opened. This results in much better performance when accessing +individual records, because you only have to uncompress exactly one record, not +a much larger disk block as when using Stacker on MS-DOS. +Usually, @code{myisampack} packs the data file 40%-70%. -This procedure is defined in the @file{sql/sql_analyse.cc}. This -examines the result from your query and returns an analysis of the -results: +@strong{MySQL} uses memory mapping (@code{mmap()}) on compressed tables and +falls back to normal read/write file usage if @code{mmap()} doesn't work. +There are currently two limitations with @code{myisampack}: @itemize @bullet @item -@code{max elements} (default 256) is the maximum number of distinct values -@code{analyse} will notice per column. This is used by @code{analyse} to check if -the optimal column type should be of type @code{ENUM}. +After packing, the table is read-only. @item -@code{max memory} (default 8192) is the maximum memory @code{analyse} should -allocate per column while trying to find all distinct values. +@code{myisampack} can also pack @code{BLOB} or @code{TEXT} columns. The +older @code{pack_isam} could not do this. @end itemize +Fixing these limitations is on our TODO list but with low priority. + +@code{myisampack} is invoked like this: + @example -SELECT ... FROM ... WHERE ... PROCEDURE ANALYSE([max elements,[max memory]]) +shell> myisampack [options] filename ... @end example -@node Writing a procedure, , procedure analyse, Adding procedures -@section Writing a Procedure +Each filename should be the name of an index (@file{.MYI}) file. If you +are not in the database directory, you should specify the pathname to the +file. It is permissible to omit the @file{.MYI} extension. -For the moment, the only documentation for this is the source. +@code{myisampack} supports the following options: -You can find all information about procedures by examining the following files: +@table @code +@item -b, --backup +Make a backup of the table as @code{tbl_name.OLD}. -@itemize @bullet -@item @file{sql/sql_analyse.cc} -@item @file{sql/procedure.h} -@item @file{sql/procedure.cc} -@item @file{sql/sql_select.cc} -@end itemize +@item -#, --debug=debug_options +Output debug log. The @code{debug_options} string often is +@code{'d:t:o,filename'}. -@cindex ODBC -@cindex Windows -@cindex MyODBC -@node ODBC, Common programs, Adding procedures, Top -@chapter MySQL ODBC Support +@item -f, --force +Force packing of the table even if it becomes bigger or if the temporary file +exists. @code{myisampack} creates a temporary file named @file{tbl_name.TMD} +while it compresses the table. If you kill @code{myisampack}, the @file{.TMD} +file may not be deleted. Normally, @code{myisampack} exits with an error if +it finds that @file{tbl_name.TMD} exists. With @code{--force}, +@code{myisampack} packs the table anyway. -@menu -* Installing MyODBC:: How to install MyODBC -* ODBC administrator:: How to fill in the various fields in the ODBC administrator program -* MyODBC connect parameters:: -* ODBC Problems:: How to report problems with @strong{MySQL} ODBC -* MyODBC clients:: Programs known to work with @strong{MyODBC} -* ODBC and last_insert_id:: How to get the value of an @code{AUTO_INCREMENT} column in ODBC -* MyODBC bug report:: Reporting problems with MyODBC -@end menu +@item -?, --help +Display a help message and exit. -@strong{MySQL} provides support for ODBC by means of the @strong{MyODBC} -program. This chapter will teach you how to install @strong{MyODBC}, -and how to use it. Here, you will also find a list of common programs that -are known to work with @strong{MyODBC}. +@item -j big_tbl_name, --join=big_tbl_name +Join all tables named on the command line into a single table +@code{big_tbl_name}. All tables that are to be combined +MUST be identical (same column names and types, same indexes, etc.). -@node Installing MyODBC, ODBC administrator, ODBC, ODBC -@section How To Install MyODBC +@item -p #, --packlength=# +Specify the record length storage size, in bytes. The value should be 1, 2, +or 3. (@code{myisampack} stores all rows with length pointers of 1, 2, or 3 +bytes. In most normal cases, @code{myisampack} can determine the right length +value before it begins packing the file, but it may notice during the packing +process that it could have used a shorter length. In this case, +@code{myisampack} will print a note that the next time you pack the same file, +you could use a shorter record length.) -@strong{MyODBC} is a 32-bit ODBC (2.50) level 0 (with level 1 and level -2 features) driver for connecting an ODBC-aware application to -@strong{MySQL}. @strong{MyODBC} works on Windows95, Windows98, NT, and -on most Unix platforms. +@item -s, --silent +Silent mode. Write output only when errors occur. -@strong{MyODBC} is in public domain, and you can find the newest version -at @uref{http://www.mysql.com/downloads/api-myodbc.html}. +@item -t, --test +Don't actually pack table, just test packing it. -If you have problem with @strong{MyODBC} and your program also works -with OLEDB, you should try the OLEDB driver that you can find in the -Contrib section. @xref{Contrib}. +@item -T dir_name, --tmp_dir=dir_name +Use the named directory as the location in which to write the temporary table. -Normally you only need to install @strong{MyODBC} on Windows machines. -You only need @strong{MyODBC} for Unix if you have a program like -ColdFusion that is running on the Unix machine and uses ODBC to connect -to the databases. +@item -v, --verbose +Verbose mode. Write information about progress and packing result. -If you want to install @strong{MyODBC} on a Unix box, you will also need -an @strong{ODBC} manager. @strong{MyODBC} is known to work with -most of the Unix ODBC managers. You can find a list at these in the -@strong{ODBC}-related links section on the @strong{MySQL} useful links page. -@xref{Useful Links}. +@item -V, --version +Display version information and exit. -To install @strong{MyODBC} on Windows, you should download the -appropriate @strong{MyODBC} .zip file (for Windows or NT/Win2000), -unpack it with @code{WINZIP}, or some similar program, and execute the -@code{SETUP.EXE} file. +@item -w, --wait -On Windows/NT you may get the following error when trying to install -@strong{MyODBC}: +Wait and retry if table is in use. If the @code{mysqld} server was +invoked with the @code{--skip-locking} option, it is not a good idea to +invoke @code{myisampack} if the table might be updated during the +packing process. +@end table + +@cindex examples, compressed tables +The sequence of commands shown below illustrates a typical table compression +session: @example -An error occurred while copying C:\WINDOWS\SYSTEM\MFC30.DLL. Restart -Windows and try installing again (before running any applications which -use ODBC) -@end example - -The problem in this case is that some other program is using ODBC and -because of how Windows is designed, you may not in this case be able to -install a new ODBC drivers with Microsoft's ODBC setup program. In most -cases you can continue by just pressing @code{Ignore} to copy the rest -of the MyODBC files and the final installation should still work. If -this doesn't work, the solution is to reboot your computer in ``safe -mode`` (Choose this by pressing F8 just before your machine starts -Windows during rebooting), install @strong{MyODBC}, and reboot to normal -mode. - -@itemize @bullet -@item -To make a connection to a Unix box from a Windows box, with an ODBC -application (one that doesn't support @strong{MySQL} natively), you must -first install @strong{MyODBC} on the Windows machine. -@item -The user and Windows machine must have the access privileges to the -@strong{MySQL} server on the Unix machine. This is set up with the -@code{GRANT} command. @xref{GRANT,,@code{GRANT}}. -@item -You must create an ODBC DSN entry as follows: +shell> ls -l station.* +-rw-rw-r-- 1 monty my 994128 Apr 17 19:00 station.MYD +-rw-rw-r-- 1 monty my 53248 Apr 17 19:00 station.MYI +-rw-rw-r-- 1 monty my 5767 Apr 17 19:00 station.frm -@itemize @minus -@item -Open the Control Panel on the Windows machine. -@item -Double-click the ODBC Data Sources 32 bits icon. -@item -Click the tab User DSN. -@item -Click the button Add. -@item -Select @strong{MySQL} in the screen Create New Data Source and click -the Finish button. -@item -The @strong{MySQL} Driver default configuration screen is shown. -@xref{ODBC administrator}. -@end itemize +shell> myisamchk -dvv station -@item -Now start your application and select the ODBC driver with the DSN you -specified in the ODBC administrator. -@end itemize +MyISAM file: station +Isam-version: 2 +Creation time: 1996-03-13 10:08:58 +Recover time: 1997-02-02 3:06:43 +Data records: 1192 Deleted blocks: 0 +Datafile: Parts: 1192 Deleted data: 0 +Datafile pointer (bytes): 2 Keyfile pointer (bytes): 2 +Max datafile length: 54657023 Max keyfile length: 33554431 +Recordlength: 834 +Record format: Fixed length -Notice that there are other configuration options on the screen of -@strong{MySQL} (trace, don't prompt on connect, etc) that you can try if -you run into problems. +table description: +Key Start Len Index Type Root Blocksize Rec/key +1 2 4 unique unsigned long 1024 1024 1 +2 32 30 multip. text 10240 1024 1 -@cindex ODBC, administrator -@node ODBC administrator, MyODBC connect parameters, Installing MyODBC, ODBC -@section How to Fill in the Various Fields in the ODBC Administrator Program +Field Start Length Type +1 1 1 +2 2 4 +3 6 4 +4 10 1 +5 11 20 +6 31 1 +7 32 30 +8 62 35 +9 97 35 +10 132 35 +11 167 4 +12 171 16 +13 187 35 +14 222 4 +15 226 16 +16 242 20 +17 262 20 +18 282 20 +19 302 30 +20 332 4 +21 336 4 +22 340 1 +23 341 8 +24 349 8 +25 357 8 +26 365 2 +27 367 2 +28 369 4 +29 373 4 +30 377 1 +31 378 2 +32 380 8 +33 388 4 +34 392 4 +35 396 4 +36 400 4 +37 404 1 +38 405 4 +39 409 4 +40 413 4 +41 417 4 +42 421 4 +43 425 4 +44 429 20 +45 449 30 +46 479 1 +47 480 1 +48 481 79 +49 560 79 +50 639 79 +51 718 79 +52 797 8 +53 805 1 +54 806 1 +55 807 20 +56 827 4 +57 831 4 -There are three possibilities for specifying the server name on -Windows95: +shell> myisampack station.MYI +Compressing station.MYI: (1192 records) +- Calculating statistics -@itemize @bullet -@item -Use the IP address of the server. -@item -Add a file @file{\windows\lmhosts} with the following information: +normal: 20 empty-space: 16 empty-zero: 12 empty-fill: 11 +pre-space: 0 end-space: 12 table-lookups: 5 zero: 7 +Original trees: 57 After join: 17 +- Compressing file +87.14% -@example -ip hostname -@end example +shell> ls -l station.* +-rw-rw-r-- 1 monty my 127874 Apr 17 19:00 station.MYD +-rw-rw-r-- 1 monty my 55296 Apr 17 19:04 station.MYI +-rw-rw-r-- 1 monty my 5767 Apr 17 19:00 station.frm -For example: +shell> myisamchk -dvv station -@example -194.216.84.21 my_hostname -@end example +MyISAM file: station +Isam-version: 2 +Creation time: 1996-03-13 10:08:58 +Recover time: 1997-04-17 19:04:26 +Data records: 1192 Deleted blocks: 0 +Datafile: Parts: 1192 Deleted data: 0 +Datafilepointer (bytes): 3 Keyfile pointer (bytes): 1 +Max datafile length: 16777215 Max keyfile length: 131071 +Recordlength: 834 +Record format: Compressed -@item -Configure the PC to use DNS. -@end itemize +table description: +Key Start Len Index Type Root Blocksize Rec/key +1 2 4 unique unsigned long 10240 1024 1 +2 32 30 multip. text 54272 1024 1 -Example of how to fill in the @code{ODBC setup}: -@example -Windows DSN name: test -Description: This is my test database -MySql Database: test -Server: 194.216.84.21 -User: monty -Password: my_password -Port: +Field Start Length Type Huff tree Bits +1 1 1 constant 1 0 +2 2 4 zerofill(1) 2 9 +3 6 4 no zeros, zerofill(1) 2 9 +4 10 1 3 9 +5 11 20 table-lookup 4 0 +6 31 1 3 9 +7 32 30 no endspace, not_always 5 9 +8 62 35 no endspace, not_always, no empty 6 9 +9 97 35 no empty 7 9 +10 132 35 no endspace, not_always, no empty 6 9 +11 167 4 zerofill(1) 2 9 +12 171 16 no endspace, not_always, no empty 5 9 +13 187 35 no endspace, not_always, no empty 6 9 +14 222 4 zerofill(1) 2 9 +15 226 16 no endspace, not_always, no empty 5 9 +16 242 20 no endspace, not_always 8 9 +17 262 20 no endspace, no empty 8 9 +18 282 20 no endspace, no empty 5 9 +19 302 30 no endspace, no empty 6 9 +20 332 4 always zero 2 9 +21 336 4 always zero 2 9 +22 340 1 3 9 +23 341 8 table-lookup 9 0 +24 349 8 table-lookup 10 0 +25 357 8 always zero 2 9 +26 365 2 2 9 +27 367 2 no zeros, zerofill(1) 2 9 +28 369 4 no zeros, zerofill(1) 2 9 +29 373 4 table-lookup 11 0 +30 377 1 3 9 +31 378 2 no zeros, zerofill(1) 2 9 +32 380 8 no zeros 2 9 +33 388 4 always zero 2 9 +34 392 4 table-lookup 12 0 +35 396 4 no zeros, zerofill(1) 13 9 +36 400 4 no zeros, zerofill(1) 2 9 +37 404 1 2 9 +38 405 4 no zeros 2 9 +39 409 4 always zero 2 9 +40 413 4 no zeros 2 9 +41 417 4 always zero 2 9 +42 421 4 no zeros 2 9 +43 425 4 always zero 2 9 +44 429 20 no empty 3 9 +45 449 30 no empty 3 9 +46 479 1 14 4 +47 480 1 14 4 +48 481 79 no endspace, no empty 15 9 +49 560 79 no empty 2 9 +50 639 79 no empty 2 9 +51 718 79 no endspace 16 9 +52 797 8 no empty 2 9 +53 805 1 17 1 +54 806 1 3 9 +55 807 20 no empty 3 9 +56 827 4 no zeros, zerofill(2) 2 9 +57 831 4 no zeros, zerofill(1) 2 9 @end example -The value for the @code{Windows DSN name} field is any name that is unique -in your Windows ODBC setup. +The information printed by @code{myisampack} is described below: -You don't have to specify values for the @code{Server}, @code{User}, -@code{Password}, or @code{Port} fields in the ODBC setup screen. -However, if you do, the values will be used as the defaults later when -you attempt to make a connection. You have the option of changing the -values at that time. +@table @code +@item normal +The number of columns for which no extra packing is used. -If the port number is not given, the default port (@value{default_port}) -is used. +@item empty-space +The number of columns containing +values that are only spaces; these will occupy 1 bit. -If you specify the option @code{Read options from C:\my.cnf}, the groups -@code{client} and @code{odbc} will be read from the @file{C:\my.cnf} file. -You can use all options that are usable by @code{mysql_options()}. -@xref{mysql_options, , @code{mysql_options}}. +@item empty-zero +The number of columns containing +values that are only binary 0's; these will occupy 1 bit. -@node MyODBC connect parameters, ODBC Problems, ODBC administrator, ODBC -@section Connect parameters for MyODBC +@item empty-fill +The number of integer columns that don't occupy the full byte range of their +type; these are changed to a smaller type (for example, an @code{INTEGER} +column may be changed to @code{MEDIUMINT}). -One can specify the following parameters for @strong{MyODBC} on -the @code{[Servername]} section of an @code{ODBC.INI} file or -through the @code{InConnectionString} argument in the -@code{SQLDriverConnect()} call. +@item pre-space +The number of decimal columns that are stored with leading spaces. In this +case, each value will contain a count for the number of leading spaces. -@multitable @columnfractions .2 .2 .6 -@item @strong{Parameter} @tab @strong{Default value} @tab @strong{Comment} -@item user @tab ODBC (on Windows) @tab The username used to connect to @strong{MySQL}. -@item server @tab localhost @tab The hostname of the @strong{MySQL} server. -@item database @tab @tab The default database -@item option @tab 0 @tab A integer by which you can specify how @strong{MyODBC} should work. See below. -@item port @tab 3306 @tab The TCP/IP port to use if @code{server} is not @code{localhost}. -@item stmt @tab @tab A statement that will be executed when connection to @code{MySQL}. -@item password @tab @tab The password for the @code{server} @code{user} combination. -@item socket @tab @tab The socket or Windows pipe to connect to. -@end multitable +@item end-space +The number of columns that have a lot of trailing spaces. In this case, each +value will contain a count for the number of trailing spaces. -The option argument is used to tell @strong{MyODBC} that the client isn't 100% -ODBC compliant. On Windows, one normally sets the option flag by -toggling the different options on the connection screen but one can also -set this in the opton argument. The following options are listed in the -same order as they appear in the @strong{MyODBC} connect screen: +@item table-lookup +The column had only a small number of different values, which were +converted to an @code{ENUM} before Huffman compression. -@multitable @columnfractions .1 .9 -@item @strong{Bit} @tab @strong{Meaning} -@item 1 @tab The client can't handle that @strong{MyODBC} returns the real width of a column. -@item 2 @tab The client can't handle that MySQL returns the true value of affected rows. If this flag is set then MySQL returns 'found rows' instead. One must have MySQL 3.21.14 or newer to get this to work. -@item 4 @tab Make a debug log in c:\myodbc.log. This is the same as putting @code{MYSQL_DEBUG=d:t:O,c::\myodbc.log} in @file{AUTOEXEC.BAT} -@item 8 @tab Don't set any packet limit for results and parameters. -@item 16 @tab Don't prompt for questions even if driver would like to prompt -@item 32 @tab Simulate a ODBC 1.0 driver in some context. -@item 64 @tab Ignore use of database name in 'database.table.column'. -@item 128 @tab Force use of ODBC manager cursors (experimental). -@item 256 @tab Disable the use of extended fetch (experimental) -@item 512 @tab Pad CHAR fields to full column length. -@item 1024 @tab SQLDescribeCol() will return fully qualifed column names -@item 2048 @tab Use the compressed server/client protocol -@item 4096 @tab Tell server to ignore space after function name and before @code{'('} (needed by PowerBuilder). This will make all function names keywords! -@item 8192 @tab Connect with named pipes to a @code{mysqld} server running on NT. -@item 16384 @tab Change LONGLONG columns to INT columns (Some applications can't handle LONGLONG). -@item 32768 @tab Return 'user' as Table_qualifier and Table_owner from SQLTables (experimental) -@item 65536 @tab Read parameters from the @code{client} and @code{odbc} groups from @file{my.cnf} -@item 131072 @tab Add some extra safety checks (should not bee needed but...) -@end multitable +@item zero +The number of columns for which all values are zero. -If you want to have many options, you should add the above flags! For -example setting option to 12 (4+8) gives you debugging without package -limits! +@item Original trees +The initial number of Huffman trees. -The default @file{MYODBC.DLL} is compiled for optimal performance. If -you want to to debug @strong{MyODBC} (for example to enable tracing), -you should instead use @code{MYODBCD.DLL}. To install this file, copy -@file{MYODBCD.DLL} over the installed @code{MYODBC.DLL} file. +@item After join +The number of distinct Huffman trees left after joining +trees to save some header space. +@end table -@node ODBC Problems, MyODBC clients, MyODBC connect parameters, ODBC -@section How to Report Problems with MyODBC +After a table has been compressed, @code{myisamchk -dvv} prints additional +information about each field: -@strong{MyODBC} has been tested with Access, Admndemo.exe, C++-Builder, -Borland Builder 4, Centura Team Developer (formerly Gupta SQL/Windows), -ColdFusion (on Solaris and NT with svc pack 5), Crystal Reports, -DataJunction, Delphi, ERwin, Excel, iHTML, FileMaker Pro, FoxPro, Notes -4.5/4.6, SBSS, Perl DBD-ODBC, Paradox, Powerbuilder, Powerdesigner 32 -bit, VC++, and Visual Basic. +@table @code +@item Type +The field type may contain the following descriptors: -If you know of any other applications that work with @strong{MyODBC}, please -send mail to @email{myodbc@@lists.mysql.com} about this! +@table @code +@item constant +All rows have the same value. -With some programs you may get an error like: -@code{Another user has modifies the record that you have modified}. In most -cases this can be solved by doing one of the following things: +@item no endspace +Don't store endspace. -@itemize @bullet -@item -Add a primary key for the table if there isn't one already. -@item -Add a timestamp column if there isn't one already. -@item -Only use double float fields. Some programs may fail when they compare -single floats. -@end itemize +@item no endspace, not_always +Don't store endspace and don't do end space compression for all values. -If the above doesn't help, you should do a @code{MyODBC} trace file and -try to figure out why things go wrong. +@item no endspace, no empty +Don't store endspace. Don't store empty values. -@node MyODBC clients, ODBC and last_insert_id, ODBC Problems, ODBC -@section Programs Known to Work with MyODBC +@item table-lookup +The column was converted to an @code{ENUM}. -Most programs should work with @strong{MyODBC}, but for each of those -listed below, we have tested it ourselves or received confirmation from -some user that it works: +@item zerofill(n) +The most significant @code{n} bytes in the value are always 0 and are not +stored. -@table @asis -@item @strong{Program} -@strong{Comment} -@cindex Access program +@item no zeros +Don't store zeros. -@item Access -To make Access work: -@itemize @bullet -@item -If you are using Access 2000, you should get and install the newest -Microsoft MDAC (@code{Microsoft Data Access Components}) from -@uref{http://www.microsoft.com/data}. This will fix the following bug -in Access: when you export data to @strong{MySQL}, the table and column -names aren't specified. Another way to around this bug is to upgrade to -MyODBC Version 2.50.33 and @strong{MySQL} Version 3.23.x, which together -provide a workaround for this bug! +@item always zero +0 values are stored in 1 bit. +@end table -Note that if you are using @strong{MySQL} Version 3.22, you must to apply the -MDAC patch and use MyODBC 2.50.32 or 2.50.34 and above to go around -this problem. -@item -Set the ``Return matching rows'' MyODBC option field when connecting to -@strong{MySQL}. -@item -You should have a primary key in the table. If not, new or updated rows -may show up as @code{#Deleted#}. -@item -You should have a timestamp in all tables you want to be able to update. -For maximum portability @code{TIMESTAMP(14)} or simple @code{TIMESTAMP} -is recommended instead of other @code{TIMESTAMP(X)} variations. -@item -Only use double float fields. Access fails when comparing with single floats. -The symptom usually is that new or updated rows may show up as @code{#Deleted#} -or that you can't find or update rows. -@item -If you still get the error @code{Another user has changed your data} after -adding a @code{TIMESTAMP} column, the following trick may help you: +@item Huff tree +The Huffman tree associated with the field. -Don't use @code{table} data sheet view. Create instead a form with the -fields you want, and use that @code{form} data sheet view. You should -set the @code{DefaultValue} property for the @code{TIMESTAMP} column to -@code{NOW()}. It may be a good idea to hide the @code{TIMESTAMP} column -from view so your users are not confused. -@item -Access on NT will report @code{BLOB} columns as @code{OLE OBJECTS}. If -you want to have @code{MEMO} columns instead, you should change the -column to @code{TEXT} with @code{ALTER TABLE}. -@item -Access can't always handle @code{DATE} columns properly. If you have a problem -with these, change the columns to @code{DATETIME}. -@item -In some cases, Access may generate illegal SQL queries that -@strong{MySQL} can't understand. You can fix this by selecting -@code{"Query|SQLSpecific|Pass-Through"} from the Access menu. -@item -If you have in Access a column defined as BYTE, Access will try to export this -as @code{TINYINT} instead of @code{TINYINT UNSIGNED}. This will give you -problems if you have values > 127 in the column! -@item -If you are using Access 7.0, You should use the option flag @code{Return -matching rows}. -@item -If you are using Access 2.0, You should use the option flags @code{Return -matching rows} and @code{Simulate ODBC 1.0}. -@end itemize +@item Bits +The number of bits used in the Huffman tree. +@end table -@cindex ADO program -@item ADO -When you are coding with the ADO API and @strong{MyODBC} you need to put -attention in some default properties that aren't supported by the -@strong{MySQL} server. For example, using the @code{CursorLocation -Property} as @code{adUseServer} will return for the @code{RecordCount -Property} a result of -1. To have the right value, you need to set this -property to @code{adUseClient}, like is showing in the VB code below: +After you have run @code{pack_isam}/@code{myisampack} you must run +@code{isamchk}/@code{myisamchk} to re-create the index. At this time you +can also sort the index blocks and create statistics needed for +the @strong{MySQL} optimizer to work more efficiently: @example -Dim myconn As New ADODB.Connection -Dim myrs As New Recordset -Dim mySQL As String -Dim myrows As Long - -myconn.Open "DSN=MyODBCsample" -mySQL = "SELECT * from user" -myrs.Source = mySQL -Set myrs.ActiveConnection = myconn -myrs.CursorLocation = adUseClient -myrs.Open -myrows = myrs.RecordCount - -myrs.Close -myconn.Close +myisamchk -rq --analyze --sort-index table_name.MYI +isamchk -rq --analyze --sort-index table_name.ISM @end example -Another workaround is to use a @code{SELECT COUNT(*)} statement -for a similar query to get the correct row count. - -@item Active server pages (ASP) -You should use the option flag @code{Return matching rows}. - -@item BDE applications -To get these to work, you should set the option flags -@code{Don't optimize column widths} and @code{Return matching rows}. +After you have installed the packed table into the @strong{MySQL} database +directory you should do @code{mysqladmin flush-tables} to force @code{mysqld} +to start using the new table. -@cindex Borland Buidler 4 program -@item Borland Builder 4 -When you start a query you can use the property @code{Active} or use the -method @code{Open}. Note that @code{Active} will start by automatically -issuing a @code{SELECT * FROM ...} query that may not be a good thing if -your tables are big! -@item ColdFusion (On Unix) -The following information is taken from the ColdFusion documentation: +If you want to unpack a packed table, you can do this with the +@code{--unpack} option to @code{isamchk} or @code{myisamchk}. -Use the following information to configure ColdFusion Server for Linux -to use the unixODBC driver with @strong{MyODBC} for @strong{MySQL} data -sources. Allaire has verified that @strong{MyODBC} Version 2.50.26 -works with @strong{MySQL} Version 3.22.27 and ColdFusion for Linux. (Any -newer version should also work.) You can download @strong{MyODBC} at -@uref{http://www.mysql.com/downloads/api-myodbc.html} -@cindex ColdFusion program -ColdFusion Version 4.5.1 allows you to us the ColdFusion Administrator -to add the @strong{MySQL} data source. However, the driver is not -included with ColdFusion Version 4.5.1. Before the @strong{MySQL} driver -will appear in the ODBC datasources drop-down list, you must build and -copy the @strong{MyODBC} driver to -@file{/opt/coldfusion/lib/libmyodbc.so}. - -The Contrib directory contains the program mydsn-xxx.zip which allows -you to build and remove the DSN registry file for the MyODBC driver -on Coldfusion applications. - -@cindex DataJunction -@item DataJunction -You have to change it to output @code{VARCHAR} rather than @code{ENUM}, as -it exports the latter in a manner that causes @strong{MySQL} grief. -@cindex Excel -@item Excel -Works. Some tips: -@itemize @bullet -@item -If you have problems with dates, try to select them as strings using the -@code{CONCAT()} function. For example: -@example -select CONCAT(rise_time), CONCAT(set_time) - from sunrise_sunset; -@end example -Values retrieved as strings this way should be correctly recognized -as time values by Excel97. - -The purpose of @code{CONCAT()} in this example is to fool ODBC into thinking -the column is of ``string type''. Without the @code{CONCAT()}, ODBC knows the -column is of time type, and Excel does not understand that. - -Note that this is a bug in Excel, because it automatically converts a -string to a time. This would be great if the source was a text file, but -is plain stupid when the source is an ODBC connection that reports -exact types for each column. -@end itemize -@cindex Word program -@item Word +@cindex installation maintenance +@cindex maintaining, tables +@cindex tables, maintaining +@cindex databases, maintaining +@cindex @code{myisamchk} +@cindex @code{mysqlcheck} +@cindex crash, recovery +@cindex recovery, from crash +@node Maintenance, Adding functions, Tools, Top +@chapter Maintaining a MySQL Installation -To retrieve data from @strong{MySQL} to Word/Excel documents, you need to -use the @code{MyODBC} driver and the Add-in Microsoft Query help. +@menu +* Table maintenance:: Table maintenance and crash recovery +* Using mysqlcheck:: Using mysqlcheck for maintenance and recovery +* Maintenance regimen:: Setting up a table maintenance regimen +* Table-info:: Getting information about a table +* Crash recovery:: Using @code{myisamchk} for crash recovery +* Log file maintenance:: Log file maintenance +@end menu -For example, create a db with a table containing 2 columns of text: +This chapter covers what you should know about maintaining a @strong{MySQL} +distribution. You will learn how to care for your tables on a regular +basis, and what to do when disaster strikes. -@itemize @bullet -@item -Insert rows using the @code{mysql} client command-line tool. -@item -Create a DSN file using the MyODBC driver, for example, my for the db above. -@item -Open the Word application. -@item -Create a blank new documentation. -@item -Using the tool bar called Database, press the button insert database. -@item -Press the button Get Data. -@item -At the right hand of the screen Get Data, press the button Ms Query. -@item -In the Ms Query create a New Data Source using the DSN file my. -@item -Select the new query. -@item -Select the columns that you want. -@item -Make a filter if you want. -@item -Make a Sort if you want. -@item -Select Return Data to Microsoft Word. -@item -Click Finish. -@item -Click Insert data and select the records. -@item -Click OK and you see the rows in your Word document. -@end itemize +@node Table maintenance, Using mysqlcheck, Maintenance, Maintenance +@section Using @code{myisamchk} for Table Maintenance and Crash Recovery -@cindex odbcadmin program -@item odbcadmin -Test program for ODBC. -@cindex Delphi program -@item Delphi -You must use BDE Version 3.2 or newer. Set the `Don't optimize column width' -option field when connecting to @strong{MySQL}. +Starting with @strong{MySQL} Version 3.23.13, you can check MyISAM +tables with the @code{CHECK TABLE} command. @xref{CHECK TABLE}. You can +repair tables with the @code{REPAIR TABLE} command. @xref{REPAIR TABLE}. -Also, here is some potentially useful Delphi code that sets up both an -ODBC entry and a BDE entry for @strong{MyODBC} (the BDE entry requires a BDE -Alias Editor that is free at a Delphi Super Page near -you. (Thanks to Bryan Brunton @email{bryan@@flesherfab.com} for this): +To check/repair MyISAM tables (@code{.MYI} and @code{.MYD}) you should +use the @code{myisamchk} utility. To check/repair ISAM tables +(@code{.ISM} and @code{.ISD}) you should use the @code{isamchk} +utility. @xref{Table types}. -@example -fReg:= TRegistry.Create; - fReg.OpenKey('\Software\ODBC\ODBC.INI\DocumentsFab', True); - fReg.WriteString('Database', 'Documents'); - fReg.WriteString('Description', ' '); - fReg.WriteString('Driver', 'C:\WINNT\System32\myodbc.dll'); - fReg.WriteString('Flag', '1'); - fReg.WriteString('Password', ''); - fReg.WriteString('Port', ' '); - fReg.WriteString('Server', 'xmark'); - fReg.WriteString('User', 'winuser'); - fReg.OpenKey('\Software\ODBC\ODBC.INI\ODBC Data Sources', True); - fReg.WriteString('DocumentsFab', 'MySQL'); - fReg.CloseKey; - fReg.Free; +In the following text we will talk about @code{myisamchk}, but everything +also applies to the old @code{isamchk}. - Memo1.Lines.Add('DATABASE NAME='); - Memo1.Lines.Add('USER NAME='); - Memo1.Lines.Add('ODBC DSN=DocumentsFab'); - Memo1.Lines.Add('OPEN MODE=READ/WRITE'); - Memo1.Lines.Add('BATCH COUNT=200'); - Memo1.Lines.Add('LANGDRIVER='); - Memo1.Lines.Add('MAX ROWS=-1'); - Memo1.Lines.Add('SCHEMA CACHE DIR='); - Memo1.Lines.Add('SCHEMA CACHE SIZE=8'); - Memo1.Lines.Add('SCHEMA CACHE TIME=-1'); - Memo1.Lines.Add('SQLPASSTHRU MODE=SHARED AUTOCOMMIT'); - Memo1.Lines.Add('SQLQRYMODE='); - Memo1.Lines.Add('ENABLE SCHEMA CACHE=FALSE'); - Memo1.Lines.Add('ENABLE BCD=FALSE'); - Memo1.Lines.Add('ROWSET SIZE=20'); - Memo1.Lines.Add('BLOBS TO CACHE=64'); - Memo1.Lines.Add('BLOB SIZE=32'); +You can use the @code{myisamchk} utility to get information about your +database tables, check and repair them, or optimize them. The following +sections describe how to invoke @code{myisamchk} (including a +description of its options), how to set up a table maintenance schedule, +and how to use @code{myisamchk} to perform its various functions. - AliasEditor.Add('DocumentsFab','MySQL',Memo1.Lines); -@end example +You can, in most cases, also use the command @code{OPTIMIZE TABLES} to +optimize and repair tables, but this is not as fast or reliable (in case +of real fatal errors) as @code{myisamchk}. On the other hand, +@code{OPTIMIZE TABLE} is easier to use and you don't have to worry about +flushing tables. +@xref{OPTIMIZE TABLE, , @code{OPTIMIZE TABLE}}. -@cindex C++Builder -@item C++Builder -Tested with BDE Version 3.0. The only known problem is that when the table -schema changes, query fields are not updated. BDE, however, does not seem -to recognize primary keys, only the index PRIMARY, though this has not -been a problem. +Even that the repair in @code{myisamchk} is quite secure, it's always a +good idea to make a backup BEFORE doing a repair (or anything that could +make a lot of changes to a table) -@item Vision -You should use the option flag @code{Return matching rows}. +@menu +* myisamchk syntax:: @code{myisamchk} invocation syntax +* myisamchk memory:: @code{myisamchk} memory usage +@end menu -@cindex Visual Basic -@item Visual Basic -To be able to update a table, you must define a primary key for the table. +@node myisamchk syntax, myisamchk memory, Table maintenance, Table maintenance +@subsection @code{myisamchk} Invocation Syntax -Visual Basic with ADO can't handle big integers. This means that some queries -like @code{SHOW PROCESSLIST} will not work properly. The fix is to set -add the option @code{OPTION=16834} in the ODBC connect string or set -the @code{Change BIGINT columns to INT} option in the MyODBC connect screen. -You may also want to set the @code{Return matching rows} option. +@code{myisamchk} is invoked like this: -@item VisualInterDev -If you get the error @code{[Microsoft][ODBC Driver Manager] Driver does -not support this parameter} the reason may be that you have a -@code{BIGINT} in your result. Try setting the @code{Change BIGINT -columns to INT} option in the MyODBC connect screen. +@example +shell> myisamchk [options] tbl_name +@end example -@item Visual Objects -You should use the option flag @code{Don't optimize column widths}. -@end table +The @code{options} specify what you want @code{myisamchk} to do. They are +described below. (You can also get a list of options by invoking +@code{myisamchk --help}.) With no options, @code{myisamchk} simply checks your +table. To get more information or to tell @code{myisamchk} to take corrective +action, specify options as described below and in the following sections. -@cindex AUTO-INCREMENT, ODBC -@node ODBC and last_insert_id, MyODBC bug report, MyODBC clients, ODBC -@section How to Get the Value of an @code{AUTO_INCREMENT} Column in ODBC +@code{tbl_name} is the database table you want to check/repair. If you run +@code{myisamchk} somewhere other than in the database directory, you must +specify the path to the file, because @code{myisamchk} has no idea where your +database is located. Actually, @code{myisamchk} doesn't care whether or not +the files you are working on are located in a database directory; you can +copy the files that correspond to a database table into another location and +perform recovery operations on them there. -A common problem is how to get the value of an automatically generated ID -from an @code{INSERT}. With ODBC, you can do something like this (assuming -that @code{auto} is an @code{AUTO_INCREMENT} field): +You can name several tables on the @code{myisamchk} command line if you +wish. You can also specify a name as an index file +name (with the @file{.MYI} suffix), which allows you to specify all +tables in a directory by using the pattern @file{*.MYI}. +For example, if you are in a database directory, you can check all the +tables in the directory like this: @example -INSERT INTO foo (auto,text) VALUES(NULL,'text'); -SELECT LAST_INSERT_ID(); +shell> myisamchk *.MYI @end example -Or, if you are just going to insert the ID into another table, you can do this: +If you are not in the database directory, you can check all the tables there +by specifying the path to the directory: @example -INSERT INTO foo (auto,text) VALUES(NULL,'text'); -INSERT INTO foo2 (id,text) VALUES(LAST_INSERT_ID(),'text'); +shell> myisamchk /path/to/database_dir/*.MYI @end example -@xref{Getting unique ID}. +You can even check all tables in all databases by specifying a wild card +with the path to the @strong{MySQL} data directory: -For the benefit of some ODBC applications (at least Delphi and Access), -the following query can be used to find a newly inserted row: @example -SELECT * FROM tbl_name WHERE auto IS NULL; +shell> myisamchk /path/to/datadir/*/*.MYI @end example -@cindex reporting, MyODBC problems -@cindex problems, ODBC -@cindex MyODBC, reporting problems -@node MyODBC bug report, , ODBC and last_insert_id, ODBC -@section Reporting Problems with MyODBC - -If you encounter difficulties with @strong{MyODBC}, you should start by -making a log file from the ODBC manager (the log you get when requesting -logs from ODBCADMIN) and a @strong{MyODBC} log. - -To get a @strong{MyODBC} log, you need to do the following: +The recommended way to quickly check all tables is: -@enumerate -@item -Ensure that you are using @code{myodbcd.dll} and not @code{myodbc.dll}. -The easiest way to do this is to get @code{myodbcd.dll} from the MyODBC -distribution and copy it over the @code{myodbc.dll}, which is probably -in your @code{C:\windows\system32} or @code{C:\winnt\system32} directory. +@example +myisamchk --silent --fast /path/to/datadir/*/*.MYI +isamchk --silent /path/to/datadir/*/*.ISM +@end example -Note that you probably want to restore the old myodbc.dll file when you -have finished testing, as this is a lot faster than @code{myodbcd.dll}. -@item -Tag the `Trace MyODBC' option flag in the @strong{MyODBC} connect/configure -screen. The log will be written to file @file{C:\myodbc.log}. +If you want to check all tables and repair all tables that are corrupted, +you can use the following line: -If the trace option is not remembered when you are going back to the -above screen, it means that you are not using the @code{myodbcd.dll} -driver (see above). -@item -Start your application and try to get it to fail. -@end enumerate +@example +myisamchk --silent --force --fast --update-state -O key_buffer=64M -O sort_buffer=64M -O read_buffer=1M -O write_buffer=1M /path/to/datadir/*/*.MYI +isamchk --silent --force -O key_buffer=64M -O sort_buffer=64M -O read_buffer=1M -O write_buffer=1M /path/to/datadir/*/*.ISM +@end example -Check the @code{MyODBC trace file}, to find out what could be wrong. -You should be able to find out the issued queries by searching after -the string @code{>mysql_real_query} in the @file{myodbc.log} file. +The above assumes that you have more than 64 M free. -You should also try duplicating the queries in the @code{mysql} monitor -or @code{admndemo} to find out if the error is MyODBC or @strong{MySQL}. +Note that if you get an error like: -If you find out something is wrong, please only send the relevant rows -(max 40 rows) to @email{myodbc@@lists.mysql.com}. Please never -send the whole MyODBC or ODBC log file! +@example +myisamchk: warning: 1 clients is using or hasn't closed the table properly +@end example -If you are unable to find out what's wrong, the last option is to -make an archive (tar or zip) that contains a MyODBC trace file, the ODBC -log file, and a README file that explains the problem. You can send this -to @uref{ftp://support.mysql.com/pub/mysql/secret}. Only we at -@strong{MySQL AB} will have access to the files you upload, and we will -be very discrete with the data! +This means that you are trying to check a table that has been updated by +the another program (like the @code{mysqld} server) that hasn't yet closed +the file or that has died without closing the file properly. -If you can create a program that also shows this problem, please -upload this too! +If you @code{mysqld} is running, you must force a sync/close of all +tables with @code{FLUSH TABLES} and ensure that no one is using the +tables while you are running @code{myisamchk}. In @strong{MySQL} Version 3.23 +the easiest way to avoid this problem is to use @code{CHECK TABLE} +instead of @code{myisamchk} to check tables. -If the program works with some other SQL server, you should make an ODBC log -file where you do exactly the same thing in the other SQL server. +@menu +* myisamchk general options:: +* myisamchk check options:: +* myisamchk repair options:: +* myisamchk other options:: +@end menu -Remember that the more information you can supply to us, the more -likely it is that we can fix the problem! +@cindex options, @code{myisamchk} +@cindex @code{myisamchk}, options +@node myisamchk general options, myisamchk check options, myisamchk syntax, myisamchk syntax +@subsubsection General Options for @code{myisamchk} -@node Common programs, Problems, ODBC, Top -@chapter Using MySQL with Some Common Programs +@code{myisamchk} supports the following options. -@menu -* Apache:: Using @strong{MySQL} with Apache -* Borland C++:: -@end menu +@table @code +@item -# or --debug=debug_options +Output debug log. The @code{debug_options} string often is +@code{'d:t:o,filename'}. +@item -? or --help +Display a help message and exit. +@item -O var=option, --set-variable var=option +Set the value of a variable. The possible variables and their default values +for myisamchk can be examined with @code{myisamchk --help}: +@multitable @columnfractions .3 .7 +@item key_buffer_size @tab 523264 +@item read_buffer_size @tab 262136 +@item write_buffer_size @tab 262136 +@item sort_buffer_size @tab 2097144 +@item sort_key_blocks @tab 16 +@item decode_bits @tab 9 +@end multitable -This chapter describes how to use @strong{MySQL} with some common programs. +@code{sort_buffer_size} is used when the keys are repaired by sorting +keys, which is the normal case when you use @code{--recover}. -In this chapter you will: +@code{key_buffer_size} is used when you are checking the table with +@code{--extended-check} or when the keys are repaired by inserting key +row by row in to the table (like when doing normal inserts). Repairing +through the key buffer is used in the following cases: @itemize @bullet @item -Learn how to easily store your Apache log files in a @strong{MySQL} -database. +If you use @code{--safe-recover}. @item -Find some tips on how to compile @strong{MySQL} and @strong{MySQL}-based -programs using Borland C++. +If you are using a @code{FULLTEXT} index. +@item +If the temporary files needed to sort the keys would be more than twice +as big as when creating the key file directly. This is often the case +when you have big @code{CHAR}, @code{VARCHAR} or @code{TEXT} keys as the +sort needs to store the whole keys during sorting. If you have lots +of temporary space and you can force @code{myisamchk} to repair by sorting +you can use the @code{--sort-recover} option. @end itemize -@cindex Apache -@node Apache, Borland C++, Common programs, Common programs -@section Using MySQL with Apache +Reparing through the key buffer takes much less disk space than using +sorting, but is also much slower. -The Contrib section includes programs that let you authenticate your -users from a @strong{MySQL} database and also let you log your log files -into a @strong{MySQL} table. @xref{Contrib}. +If you want a faster repair, set the above variables to about 1/4 of your +available memory. You can set both variables to big values, as only one +of the above buffers will be used at a time. -You can change the Apache logging format to be easily readable by -@strong{MySQL} by putting the following into the Apache configuration file: +@item -s or --silent +Silent mode. Write output only when errors occur. You can use @code{-s} +twice (@code{-ss}) to make @code{myisamchk} very silent. +@item -v or --verbose +Verbose mode. Print more information. This can be used with @code{-d} and +@code{-e}. Use @code{-v} multiple times (@code{-vv}, @code{-vvv}) for more +verbosity! +@item -V or --version +Print the @code{myisamchk} version and exit. +@item -w or, --wait +Instead of giving an error if the table is locked, wait until the table +is unlocked before continuing. Note that if you are running @code{mysqld} +on the table with @code{--skip-locking}, the table can only be locked +by another @code{myisamchk} command. +@end table -@example -LogFormat \ - "\"%h\",%@{%Y%m%d%H%M%S@}t,%>s,\"%b\",\"%@{Content-Type@}o\", \ - \"%U\",\"%@{Referer@}i\",\"%@{User-Agent@}i\"" -@end example +@cindex check options, myisamchk +@cindex tables, checking +@node myisamchk check options, myisamchk repair options, myisamchk general options, myisamchk syntax +@subsubsection Check Options for @code{myisamchk} -In @strong{MySQL} you can do something like this: +@table @code +@item -c or --check +Check table for errors. This is the default operation if you are not +giving @code{myisamchk} any options that override this. -@example -LOAD DATA INFILE '/local/access_log' INTO TABLE table_name -FIELDS TERMINATED BY ',' OPTIONALLY ENCLOSED BY '"' ESCAPED BY '\\' -@end example +@item -e or --extend-check +Check the table VERY thoroughly (which is quite slow if you have many +indexes). This option should only be used in extreme cases. Normally, +@code{myisamchk} or @code{myisamchk --medium-check} should, in most +cases, be able to find out if there are any errors in the table. -@cindex Borland C++ compiler -@node Borland C++, , Apache, Common programs -@section Borland C++ +If you are using @code{--extended-check} and have much memory, you should +increase the value of @code{key_buffer_size} a lot! -You can compile the @strong{MySQL} Windows source with Borland C++ 5.02. -(The Windows source includes only projects for Microsoft VC++, for -Borland C++ you have to do the project files yourself). +@item -F or --fast +Check only tables that haven't been closed properly. +@item -C or --check-only-changed +Check only tables that have changed since the last check. +@item -f or --force +Restart @code{myisamchk} with @code{-r} (repair) on the table, if +@code{myisamchk} finds any errors in the table. +@item -i or --information +Print informational statistics about the table that is checked. +@item -m or --medium-check +Faster than extended-check, but only finds 99.99% of all errors. +Should, however, be good enough for most cases. +@item -U or --update-state +Store in the @file{.MYI} file when the table was checked and if the table crashed. This should be used to get full benefit of the +@code{--check-only-changed} option, but you shouldn't use this +option if the @code{mysqld} server is using the table and you are +running @code{mysqld} with @code{--skip-locking}. +@item -T or --read-only +Don't mark table as checked. This is useful if you use @code{myisamchk} +to check a table that is in use by some other application that doesn't +use locking (like @code{mysqld --skip-locking}). +@end table -One known problem with Borland C++ is that it uses a different structure -alignment than VC++. This means that you will run into problems if you -try to use the default @code{libmysql.dll} libraries (that was compiled -with VC++) with Borland C++. You can do one of the following to avoid -this problem. +@cindex repair options, myisamchk +@cindex files, repairing +@node myisamchk repair options, myisamchk other options, myisamchk check options, myisamchk syntax +@subsubsection Repair Options for myisamchk -@itemize @bullet -@item -You can use the static @strong{MySQL} libraries for Borland C++ that you -can find on @uref{http://www.mysql.com/downloads/os-win32.html}. -@item -Only call @code{mysql_init()} with @code{NULL} as an argument, not a -pre-allocated MYSQL struct. -@end itemize +The following options are used if you start @code{myisamchk} with +@code{-r} or @code{-o}: -@cindex problems, common errors -@cindex errors, common -@node Problems, Common problems, Common programs, Top -@chapter Problems and Common Errors +@table @code +@item -D # or --data-file-length=# +Max length of data file (when re-creating data file when it's 'full'). +@item -e or --extend-check +Try to recover every possible row from the data file. +Normally this will also find a lot of garbage rows. Don't use this option +if you are not totally desperate. +@item -f or --force +Overwrite old temporary files (@code{table_name.TMD}) instead of aborting. +@item -k # or keys-used=# +If you are using ISAM, tells the ISAM table handler to update only the +first @code{#} indexes. If you are using @code{MyISAM}, tells which keys +to use, where each binary bit stands for one key (first key is bit 0). +This can be used to get faster inserts! Deactivated indexes can be +reactivated by using @code{myisamchk -r}. keys. +@item -l or --no-symlinks +Do not follow symbolic links. Normally @code{myisamchk} repairs the +table a symlink points at. This option doesn't exist in MySQL 4.0, +as MySQL 4.0 will not remove symlinks during repair. +@item -r or --recover +Can fix almost anything except unique keys that aren't unique +(which is an extremely unlikely error with ISAM/MyISAM tables). +If you want to recover a table, this is the option to try first. Only if +myisamchk reports that the table can't be recovered by @code{-r}, you +should then try @code{-o}. (Note that in the unlikely case that @code{-r} +fails, the data file is still intact.) +If you have lots of memory, you should increase the size of +@code{sort_buffer_size}! +@item -o or --safe-recover +Uses an old recovery method (reads through all rows in order and updates +all index trees based on the found rows); this is a magnitude slower +than @code{-r}, but can handle a couple of very unlikely cases that +@code{-r} cannot handle. This recovery method also uses much less disk +space than @code{-r}. Normally one should always first repair with +@code{-r}, and only if this fails use @code{-o}. -@menu -* What is crashing:: How to determine what is causing problems -* Crashing:: What to do if @strong{MySQL} keeps crashing -* Link errors:: Problems when linking with the @strong{MySQL} client library -* Common errors:: Some common errors when using @strong{MySQL} -* Full disk:: How @strong{MySQL} handles a full disk -* Multiple sql commands:: How to run SQL commands from a text file -* Temporary files:: Where @strong{MySQL} stores temporary files -* Problems with mysql.sock:: How to protect @file{/tmp/mysql.sock} -* Changing MySQL user:: How to run @strong{MySQL} as a normal user -* Resetting permissions:: How to reset a forgotten password. -* File permissions :: Problems with file permissions -* Not enough file handles:: File not found -* Using DATE:: Problems using @code{DATE} columns -* Timezone problems:: Timezone problems -* Case sensitivity:: Case sensitivity in searches -* Problems with NULL:: Problems with @code{NULL} values -* Problems with alias:: Problems with @code{alias} -* Deleting from related tables:: Deleting rows from related tables -* No matching rows:: Solving problems with no matching rows -* ALTER TABLE problems:: Problems with @code{ALTER TABLE}. -* Change column order:: How to change the order of columns in a table -* Temporary table problems:: -@end menu - -This chapter lists some common problems and error messages that users have -run into. You will learn how to figure out what the problem is, and what -to do to solve it. You will also find proper solutions to some common -problems. - -@node What is crashing, Crashing, Problems, Problems -@section How to Determine What Is Causing Problems - -When you run into problems, the first thing you should do is to find out -which program / piece of equipment is causing problems: - -@itemize @bullet -@item -If you have one of the following symptoms, then it is probably a hardware -(like memory, motherboard, CPU, or hard disk) or kernel problem: -@itemize @minus -@item -The keyboard doesn't work. This can normally be checked by pressing -Caps Lock. If the Caps Lock light doesn't change you have to replace -your keyboard. (Before doing this, you should try to reboot -your computer and check all cables to the keyboard.) -@item -The mouse pointer doesn't move. -@item -The machine doesn't answer to a remote machine's pings. -@item -Different, unrelated programs don't behave correctly. -@item -If your system rebooted unexpectedly (a faulty user level program should -NEVER be able to take down your system). -@end itemize +If you have lots of memory, you should increase the size of +@code{key_buffer_size}! +@item -n or --sort-recover +Force @code{myisamchk} to use sorting to resolve the keys even if the +temporary files should be very big. This will not have any effect if you have +fulltext keys in the table. -In this case you should start by checking all your cables and run some -diagnostic tool to check your hardware! -You should also check if there are any patches, updates, or service -packs for your operating system that could likely solve your problems. -Check also that all your libraries (like glibc) are up to date. +@item --character-sets-dir=... +Directory where character sets are stored. +@item --set-character-set=name +Change the character set used by the index +@item .t or --tmpdir=path +Path for storing temporary files. If this is not set, @code{myisamchk} will +use the environment variable @code{TMPDIR} for this. +@item -q or --quick +Faster repair by not modifying the data file. One can give a second +@code{-q} to force @code{myisamchk} to modify the original datafile in case +of duplicate keys +@item -u or --unpack +Unpack file packed with myisampack. +@end table -It's always good to use a machine with ECC memory to discover -memory problems early! -@item -If your keyboard is locked up, you may be able to fix this by -logging into your machine from another machine and execute -@code{kbd_mode -a} on it. +@node myisamchk other options, , myisamchk repair options, myisamchk syntax +@subsubsection Other Options for @code{myisamchk} -@item -Please examine your system log file (/var/log/messages or similar) for -reasons for your problems. If you think the problem is in @strong{MySQL} -then you should also examine @strong{MySQL}'s log files. @xref{Update log}. +Other actions that @code{myisamchk} can do, besides repair and check tables: -@item -If you don't think you have hardware problems, you should try to find -out which program is causing problems. +@table @code +@item -a or --analyze +Analyze the distribution of keys. This improves join performance by +enabling the join optimizer to better choose in which order it should +join the tables and which keys it should use: +@code{myisamchk --describe --verbose table_name'} or using @code{SHOW KEYS} in +@strong{MySQL}. +@item -d or --description +Prints some information about table. +@item -A or --set-auto-increment[=value] +Force auto_increment to start at this or higher value. If no value is +given, then sets the next auto_increment value to the highest used value +for the auto key + 1. +@item -S or --sort-index +Sort the index tree blocks in high-low order. +This will optimize seeks and will make table scanning by key faster. +@item -R or --sort-records=# +Sorts records according to an index. This makes your data much more localized +and may speed up ranged @code{SELECT} and @code{ORDER BY} operations on +this index. (It may be VERY slow to do a sort the first time!) +To find out a table's index numbers, use @code{SHOW INDEX}, which shows a +table's indexes in the same order that @code{myisamchk} sees them. Indexes are +numbered beginning with 1. +@end table -Try using @code{top}, @code{ps}, @code{taskmanager}, or some similar program, -to check which program is taking all CPU or is locking the machine. +@cindex memory usage, myisamchk +@node myisamchk memory, , myisamchk syntax, Table maintenance +@subsection @code{myisamchk} Memory Usage -@item -Check with @code{top}, @code{df}, or a similar program if you are out of -memory, disk space, open files, or some other critical resource. +Memory allocation is important when you run @code{myisamchk}. +@code{myisamchk} uses no more memory than you specify with the @code{-O} +options. If you are going to use @code{myisamchk} on very large files, +you should first decide how much memory you want it to use. The default +is to use only about 3M to fix things. By using larger values, you can +get @code{myisamchk} to operate faster. For example, if you have more +than 32M RAM, you could use options such as these (in addition to any +other options you might specify): -@item -If the problem is some runaway process, you can always try to kill it. If it -doesn't want to die, there is probably a bug in the operating system. -@end itemize +@example +shell> myisamchk -O sort=16M -O key=16M -O read=1M -O write=1M ... +@end example -If after you have examined all other possibilities and you have -concluded that it's the @strong{MySQL} server or a @strong{MySQL} client -that is causing the problem, it's time to do a bug report for our -mailing list or our support team. In the bug report, try to give a -very detailed description of how the system is behaving and what you think is -happening. You should also state why you think it's @strong{MySQL} that -is causing the problems. Take into consideration all the situations in -this chapter. State any problems exactly how they appear when you -examine your system. Use the 'cut and paste' method for any output -and/or error messages from programs and/or log files! +Using @code{-O sort=16M} should probably be enough for most cases. -Try to describe in detail which program is not working and all -symptoms you see! We have in the past received many bug reports that just -state "the system doesn't work". This doesn't provide us with any -information about what could be the problem. +Be aware that @code{myisamchk} uses temporary files in @code{TMPDIR}. If +@code{TMPDIR} points to a memory file system, you may easily get out of +memory errors. If this happens, set @code{TMPDIR} to point at some directory +with more space and restart @code{myisamchk}. -If a program fails, it's always useful to know: +When repairing, @code{myisamchk} will also need a lot of disk space: @itemize @bullet @item -Has the program in question made a segmentation fault (core dumped)? -@item -Is the program taking the whole CPU? Check with @code{top}. Let the -program run for a while, it may be evaluating something heavy. +Double the size of the record file (the original one and a copy). This +space is not needed if one does a repair with @code{--quick}, as in this +case only the index file will be re-created. This space is needed on the +same disk as the original record file! @item -If it's the @code{mysqld} server that is causing problems, can you -do @code{mysqladmin -u root ping} or @code{mysqladmin -u root processlist}? +Space for the new index file that replaces the old one. The old +index file is truncated at start, so one usually ignore this space. +This space is needed on the same disk as the original index file! @item -What does a client program say (try with @code{mysql}, for example) -when you try to connect to the @strong{MySQL} server? -Does the client jam? Do you get any output from the program? +When using @code{--recover} or @code{--sort-recover} +(but not when using @code{--safe-recover}, you will need space for a +sort buffer for: +@code{(largest_key + row_pointer_length)*number_of_rows * 2}. +You can check the length of the keys and the row_pointer_length with +@code{myisamchk -dv table}. +This space is allocated on the temporary disk (specified by @code{TMPDIR} or +@code{--tmpdir=#}). @end itemize -When sending a bug report, you should of follow the outlines -described in this manual. @xref{Asking questions}. - -@cindex crash, repeated -@node Crashing, Link errors, What is crashing, Problems -@section What to Do if MySQL Keeps Crashing - -All @strong{MySQL} versions are tested on many platforms before they are -released. This doesn't mean that there aren't any bugs in -@strong{MySQL}, but it means if there are bugs, they are very few and can be -hard to find. If you have a problem, it will always help if you try to -find out exactly what crashes your system, as you will have a much better -chance of getting this fixed quickly. +If you have a problem with disk space during repair, you can try to use +@code{--safe-recover} instead of @code{--recover}. -First, you should try to find out whether the problem is that the -@code{mysqld} daemon dies or whether your problem has to do with your -client. You can check how long your @code{mysqld} server has been up by -executing @code{mysqladmin version}. If @code{mysqld} has died, you may -find the reason for this in the file -@file{mysql-data-directory/`hostname`.err}. @xref{Error log}. +@node Using mysqlcheck, Maintenance regimen, Table maintenance, Maintenance +@section Using @code{mysqlcheck} for Table Maintenance and Crash Recovery -Many crashes of @strong{MySQL} are caused by corrupted index / data -files. @strong{MySQL} will update the data on disk, with the -@code{write()} system call, after every SQL statement and before the -client is notified about the result. (This is not true if you are running -with @code{delayed_key_writes}, in which case only the data is written.) -This means that the data is safe even if @code{mysqld} crashes, as the OS will -ensure that the not flushed data is written to disk. You can force -@strong{MySQL} to sync everything to disk after every SQL command by -starting @code{mysqld} with @code{--flush}. +Since @strong{MySQL} version 3.23.38 you will be able to use a new +checking and repairing tool for @code{MyISAM} tables. The difference to +@code{myisamchk} is that @code{mysqlcheck} should be used when the +@code{mysqld} server is running, where as @code{myisamchk} should be used +when it is not. The benefit is that you no longer have to take the +server down for checking or repairing your tables. -The above means that normally you shouldn't get corrupted tables unless: +@code{mysqlcheck} uses @strong{MySQL} server commands @code{CHECK}, +@code{REPAIR}, @code{ANALYZE} and @code{OPTIMIZE} in a convenient way +for the user. -@itemize @bullet -@item -Someone/something killed @code{mysqld} or the machine in the middle -of an update. -@item -You have found a bug in @code{mysqld} that caused it to die in the -middle of an update. -@item -Someone is manipulating the data/index files outside of @strong{mysqld} -without locking the table properly. -@item -If you are running many @code{mysqld} servers on the same data on a -system that doesn't support good file system locks (normally handled by -the @code{lockd} daemon ) or if you are running -multiple servers with @code{--skip-locking} -@item -You have a crashed index/data file that contains very wrong data that -got @code{mysqld} confused. -@item -You have found a bug in the data storage code. This isn't that likely, -but it's at least possible. In this case you can try to change the file -type to another database handler by using @code{ALTER TABLE} on a -repaired copy of the table! -@end itemize +There are three alternative ways to invoke @code{mysqlcheck}: -Because it is very difficult to know why something is crashing, first try to -check whether or not things that work for others crash for you. Please try -the following things: +@example +shell> mysqlcheck [OPTIONS] database [tables] +shell> mysqlcheck [OPTIONS] --databases DB1 [DB2 DB3...] +shell> mysqlcheck [OPTIONS] --all-databases +@end example -@itemize @bullet -@item -Take down the @code{mysqld} daemon with @code{mysqladmin shutdown}, run -@code{myisamchk --silent --force */*.MYI} on all tables, and restart the -@code{mysqld} daemon. This will ensure that you are running from a clean -state. @xref{Maintenance}. +So it can be used in a similar way as @code{mysqldump} when it +comes to what databases and tables you want to choose. -@item -Use @code{mysqld --log} and try to determine from the information in the log -whether or not some specific query kills the server. About 95% of all bugs are -related to a particular query! Normally this is one of the last queries in -the log file just before @strong{MySQL} restarted. @xref{Query log}. -If you can repeatadly kill @strong{MySQL} with one of the queries, even -when you have checked all tables just before doing the query, then you -have been able to locate the bug and should do a bug report for this! -@xref{Bug reports}. +@code{mysqlcheck} does have a special feature compared to the other +clients; the default behavior, checking tables (-c), can be changed by +renaming the binary. So if you want to have a tool that repairs tables +by default, you should just copy @code{mysqlcheck} to your harddrive +with a new name, @code{mysqlrepair}, or alternatively make a symbolic +link to @code{mysqlrepair} and name the symbolic link as +@code{mysqlrepair}. If you invoke @code{mysqlrepair} now, it will repair +tables by default. -@item -Try to make a test case that we can use to reproduce the problem. -@xref{Reproduceable test case}. +The names that you can use to change @code{mysqlcheck} default behavior +are here: -@item -Try running the included mysql-test test and the @strong{MySQL} -benchmarks. @xref{MySQL test suite}. They should test @strong{MySQL} -rather well. You can also add code that to the benchmarks to simulates -your application! The benchmarks can be found in the @file{bench} -directory in the source distribution or, for a binary distribution, in -the @file{sql-bench} directory under your @strong{MySQL} installation -directory. +@example +mysqlrepair: The default option will be -r +mysqlanalyze: The default option will be -a +mysqloptimize: The default option will be -o +@end example -@item -Try @code{fork_test.pl} and @code{fork2_test.pl}. +The options available for @code{mysqlcheck} are listed here, please +check what your version supports with @code{mysqlcheck --help}. -@item -If you configure @strong{MySQL} for debugging, it will be much easier to -gather information about possible errors if something goes wrong. -Reconfigure @strong{MySQL} with the @code{--with-debug} option or -@code{--with-debug=full} to @code{configure} and then recompile. -@xref{Debugging server}. +@table @code +@item -A, --all-databases +Check all the databases. This will be same as --databases with all +databases selected +@item -1, --all-in-1 +Instead of making one query for each table, execute all queries in 1 +query separately for each database. Table names will be in a comma +separated list. +@item -a, --analyze +Analyze given tables. +@item --auto-repair +If a checked table is corrupted, automatically fix it. Repairing will be +done after all tables have been checked, if corrupted ones were found. +@item -#, --debug=... +Output debug log. Often this is 'd:t:o,filename' +@item --character-sets-dir=... +Directory where character sets are +@item -c, --check +Check table for errors +@item -C, --check-only-changed +Check only tables that have changed since last check or haven't been +closed properly. +@item --compress +Use compression in server/client protocol. +@item -?, --help +Display this help message and exit. +@item -B, --databases +To check several databases. Note the difference in usage; In this case +no tables are given. All name arguments are regarded as database names. +@item --default-character-set=... +Set the default character set +@item -F, --fast +Check only tables that hasn't been closed properly +@item -f, --force +Continue even if we get an sql-error. +@item -e, --extended +If you are using this option with CHECK TABLE, it will ensure that the +table is 100 percent consistent, but will take a long time. -@item -Configuring @strong{MySQL} for debugging causes a safe memory allocator to be -included that can find some errors. It also provides a lot of output about -what is happening. +If you are using this option with REPAIR TABLE, it will run an extended +repair on the table, which may not only take a long time to execute, but +may produce a lot of garbage rows also! +@item -h, --host=... +Connect to host. +@item -m, --medium-check +Faster than extended-check, but only finds 99.99 percent of all +errors. Should be good enough for most cases. +@item -o, --optimize +Optimize table +@item -p, --password[=...] +Password to use when connecting to server. If password is not given +it's solicited on the tty. +@item -P, --port=... +Port number to use for connection. +@item -q, --quick +If you are using this option with CHECK TABLE, it prevents the check +from scanning the rows to check for wrong links. This is the fastest +check. -@item -Have you applied the latest patches for your operating system? +If you are using this option with REPAIR TABLE, it will try to repair +only the index tree. This is the fastest repair method for a table. +@item -r, --repair +Can fix almost anything except unique keys that aren't unique. +@item -s, --silent +Print only error messages. +@item -S, --socket=... +Socket file to use for connection. +@item --tables +Overrides option --databases (-B). +@item -u, --user=# +User for login if not current user. +@item -v, --verbose +Print info about the various stages. +@item -V, --version +Output version information and exit. +@end table -@item -Use the @code{--skip-locking} option to @code{mysqld}. On some systems, the -@code{lockd} lock manager does not work properly; the @code{--skip-locking} -option tells @code{mysqld} not to use external locking. (This means that you -cannot run 2 @code{mysqld} servers on the same data and that you must be -careful if you use @code{myisamchk}, but it may be instructive to try the -option as a test.) +@cindex maintaining, tables +@cindex tables, maintenance regimen +@node Maintenance regimen, Table-info, Using mysqlcheck, Maintenance +@section Setting Up a Table Maintenance Regimen -@item -Have you tried @code{mysqladmin -u root processlist} when @code{mysqld} -appears to be running but not responding? Sometimes @code{mysqld} is not -comatose even though you might think so. The problem may be that all -connections are in use, or there may be some internal lock problem. -@code{mysqladmin processlist} will usually be able to make a connection even -in these cases, and can provide useful information about the current number -of connections and their status. +Starting with @strong{MySQL} Version 3.23.13, you can check MyISAM +tables with the @code{CHECK TABLE} command. @xref{CHECK TABLE}. You can +repair tables with the @code{REPAIR TABLE} command. @xref{REPAIR TABLE}. -@item -Run the command @code{mysqladmin -i 5 status} or @code{mysqladmin -i 5 --r status} or in a separate window to produce statistics while you run -your other queries. +It is a good idea to perform table checks on a regular basis rather than +waiting for problems to occur. For maintenance purposes, you can use +@code{myisamchk -s} to check tables. The @code{-s} option (short for +@code{--silent}) causes @code{myisamchk} to run in silent mode, printing +messages only when errors occur. -@item -Try the following: -@enumerate -@item -Start @code{mysqld} from @code{gdb} (or in another debugger). -@xref{Using gdb on mysqld}. +@tindex .pid (process ID) file +It's also a good idea to check tables when the server starts up. +For example, whenever the machine has done a reboot in the middle of an +update, you usually need to check all the tables that could have been +affected. (This is an ``expected crashed table''.) You could add a test to +@code{safe_mysqld} that runs @code{myisamchk} to check all tables that have +been modified during the last 24 hours if there is an old @file{.pid} +(process ID) file left after a reboot. (The @file{.pid} file is created by +@code{mysqld} when it starts up and removed when it terminates normally. The +presence of a @file{.pid} file at system startup time indicates that +@code{mysqld} terminated abnormally.) -@item -Run your test scripts. +An even better test would be to check any table whose last-modified time +is more recent than that of the @file{.pid} file. -@item -Print the backtrace and the local variables at the 3 lowest levels. In gdb you -can do this with the following commands when @code{mysqld} has crashed inside -gdb: +You should also check your tables regularly during normal system +operation. At @strong{MySQL AB}, we run a @code{cron} job to check all +our important tables once a week, using a line like this in a @file{crontab} +file: @example -backtrace -info local -up -info local -up -info local +35 0 * * 0 /path/to/myisamchk --fast --silent /path/to/datadir/*/*.MYI @end example -With gdb you can also examine which threads exist with @code{info -threads} and switch to a specific thread with @code{thread #}, where -@code{#} is the thread id. -@end enumerate - -@item -Try to simulate your application with a Perl script to force -@strong{MySQL} to crash or misbehave. +This prints out information about crashed tables so we can examine and repair +them when needed. -@item -Send a normal bug report. @xref{Bug reports}. Be even more detailed -than usual. Because @strong{MySQL} works for many people, it may be that the -crash results from something that exists only on your computer (for example, -an error that is related to your particular system libraries). -@item -If you have a problem with tables with dynamic-length rows and you are -not using @code{BLOB/TEXT} columns (but only @code{VARCHAR} columns), you -can try to change all @code{VARCHAR} to @code{CHAR} with @code{ALTER -TABLE}. This will force @strong{MySQL} to use fixed-size rows. -Fixed-size rows take a little extra space, but are much more tolerant to -corruption! +As we haven't had any unexpectedly crashed tables (tables that become +corrupted for reasons other than hardware trouble) +for a couple of years now (this is really true), once a week is +more than enough for us. -The current dynamic row code has been in use at @strong{MySQL AB} for at -least 3 years without any problems, but by nature dynamic-length rows are -more prone to errors, so it may be a good idea to try the above to see if -it helps! -@end itemize +We recommend that to start with, you execute @code{myisamchk -s} each +night on all tables that have been updated during the last 24 hours, +until you come to trust @strong{MySQL} as much as we do. -@cindex linking, errors -@cindex errors, linking -@cindex problems, linking -@node Link errors, Common errors, Crashing, Problems -@section Problems When Linking with the MySQL Client Library +@cindex tables, defragment +Normally you don't need to maintain @strong{MySQL} tables that much. If +you are changing tables with dynamic size rows (tables with @code{VARCHAR}, +@code{BLOB} or @code{TEXT} columns) or have tables with many deleted rows +you may want to from time to time (once a month?) defragment/reclaim space +from the tables. -If you are linking your program and you get errors for unreferenced -symbols that start with @code{mysql_}, like the following: +You can do this by using @code{OPTIMIZE TABLE} on the tables in question or +if you can take the @code{mysqld} server down for a while do: @example -/tmp/ccFKsdPa.o: In function `main': -/tmp/ccFKsdPa.o(.text+0xb): undefined reference to `mysql_init' -/tmp/ccFKsdPa.o(.text+0x31): undefined reference to `mysql_real_connect' -/tmp/ccFKsdPa.o(.text+0x57): undefined reference to `mysql_real_connect' -/tmp/ccFKsdPa.o(.text+0x69): undefined reference to `mysql_error' -/tmp/ccFKsdPa.o(.text+0x9a): undefined reference to `mysql_close' +isamchk -r --silent --sort-index -O sort_buffer_size=16M */*.ISM +myisamchk -r --silent --sort-index -O sort_buffer_size=16M */*.MYI @end example -you should be able to solve this by adding @code{-Lpath-to-the-mysql-library --lmysqlclient} @strong{LAST} on your link line. +@cindex tables, information +@node Table-info, Crash recovery, Maintenance regimen, Maintenance +@section Getting Information About a Table -If you get @code{undefined reference} errors for the @code{uncompress} -or @code{compress} function, add @code{-lgz} @strong{LAST} on your link -line and try again! +To get a description of a table or statistics about it, use the commands shown +below. We explain some of the information in more detail later: -If you get @code{undefined reference} errors for functions that should -exist on your system, like @code{connect}, check the man page for the -function in question, for which libraries you should add to the link -line! +@table @code +@item myisamchk -d tbl_name +Runs @code{myisamchk} in ``describe mode'' to produce a description of +your table. If you start the @strong{MySQL} server using the +@code{--skip-locking} option, @code{myisamchk} may report an error for a +table that is updated while it runs. However, because @code{myisamchk} +doesn't change the table in describe mode, there isn't any risk of +destroying data. -If you get @code{undefined reference} errors for functions that don't -exist on your system, like the following: +@item myisamchk -d -v tbl_name +To produce more information about what @code{myisamchk} is doing, add @code{-v} +to tell it to run in verbose mode. +@item myisamchk -eis tbl_name +Shows only the most important information from a table. It is slow because it +must read the whole table. + +@item myisamchk -eiv tbl_name +This is like @code{-eis}, but tells you what is being done. +@end table + +@cindex examples, @code{myisamchk} output +@cindex @code{myisamchk}, example output +Example of @code{myisamchk -d} output: @example -mf_format.o(.text+0x201): undefined reference to `__lxstat' -@end example +MyISAM file: company.MYI +Record format: Fixed length +Data records: 1403698 Deleted blocks: 0 +Recordlength: 226 -it usually means that your library is compiled on a system that is not -100 % compatible with yours. In this case you should download the -latest @strong{MySQL} source distribution and compile this yourself. -@xref{Installing source}. +table description: +Key Start Len Index Type +1 2 8 unique double +2 15 10 multip. text packed stripped +3 219 8 multip. double +4 63 10 multip. text packed stripped +5 167 2 multip. unsigned short +6 177 4 multip. unsigned long +7 155 4 multip. text +8 138 4 multip. unsigned long +9 177 4 multip. unsigned long + 193 1 text +@end example -If you are trying to run a program and you then get errors for -unreferenced symbols that start with @code{mysql_} or that the -@code{mysqlclient} library can't be found, this means that your system -can't find the share @code{libmysqlclient.so} library. +Example of @code{myisamchk -d -v} output: +@example +MyISAM file: company +Record format: Fixed length +File-version: 1 +Creation time: 1999-10-30 12:12:51 +Recover time: 1999-10-31 19:13:01 +Status: checked +Data records: 1403698 Deleted blocks: 0 +Datafile parts: 1403698 Deleted data: 0 +Datafilepointer (bytes): 3 Keyfile pointer (bytes): 3 +Max datafile length: 3791650815 Max keyfile length: 4294967294 +Recordlength: 226 -The fix for this is to tell your system to search after shared -libraries where the library is located by one of the following methods: +table description: +Key Start Len Index Type Rec/key Root Blocksize +1 2 8 unique double 1 15845376 1024 +2 15 10 multip. text packed stripped 2 25062400 1024 +3 219 8 multip. double 73 40907776 1024 +4 63 10 multip. text packed stripped 5 48097280 1024 +5 167 2 multip. unsigned short 4840 55200768 1024 +6 177 4 multip. unsigned long 1346 65145856 1024 +7 155 4 multip. text 4995 75090944 1024 +8 138 4 multip. unsigned long 87 85036032 1024 +9 177 4 multip. unsigned long 178 96481280 1024 + 193 1 text +@end example -@itemize @bullet -@item -Add the path to the directory where you have @code{libmysqlclient.so} the -@code{LD_LIBRARY_PATH} environment variable. -@item -Add the path to the directory where you have @code{libmysqlclient.so} the -@code{LD_LIBRARY} environment variable. -@item -Copy @code{libmysqlclient.so} to some place that is searched by your system, -like @file{/lib}, and update the shared library information by executing -@code{ldconfig}. -@end itemize +Example of @code{myisamchk -eis} output: +@example +Checking MyISAM file: company +Key: 1: Keyblocks used: 97% Packed: 0% Max levels: 4 +Key: 2: Keyblocks used: 98% Packed: 50% Max levels: 4 +Key: 3: Keyblocks used: 97% Packed: 0% Max levels: 4 +Key: 4: Keyblocks used: 99% Packed: 60% Max levels: 3 +Key: 5: Keyblocks used: 99% Packed: 0% Max levels: 3 +Key: 6: Keyblocks used: 99% Packed: 0% Max levels: 3 +Key: 7: Keyblocks used: 99% Packed: 0% Max levels: 3 +Key: 8: Keyblocks used: 99% Packed: 0% Max levels: 3 +Key: 9: Keyblocks used: 98% Packed: 0% Max levels: 4 +Total: Keyblocks used: 98% Packed: 17% -Another way to solve this problem is to link your program statically, with -@code{-static}, or by removing the dynamic @strong{MySQL} libraries -before linking your code. In the second case you should be -sure that no other programs are using the dynamic libraries! +Records: 1403698 M.recordlength: 226 Packed: 0% +Recordspace used: 100% Empty space: 0% Blocks/Record: 1.00 +Record blocks: 1403698 Delete blocks: 0 +Recorddata: 317235748 Deleted data: 0 +Lost space: 0 Linkdata: 0 -@cindex errors, list of -@node Common errors, Full disk, Link errors, Problems -@section Some Common Errors When Using MySQL +User time 1626.51, System time 232.36 +Maximum resident set size 0, Integral resident set size 0 +Non physical pagefaults 0, Physical pagefaults 627, Swaps 0 +Blocks in 0 out 0, Messages in 0 out 0, Signals 0 +Voluntary context switches 639, Involuntary context switches 28966 +@end example -@menu -* Error Access denied:: @code{Access denied} Error -* Gone away:: @code{MySQL server has gone away} error -* Can not connect to server:: @code{Can't connect to [local] MySQL server} error -* Blocked host:: @code{Host '...' is blocked} error -* Too many connections:: @code{Too many connections} error -* Non-transactional tables:: @code{Some non-transactional changed tables couldn't be rolled back} Error -* Out of memory:: @code{Out of memory} error -* Packet too large:: @code{Packet too large} error -* Communication errors:: Communication errors / Aborted connection -* Full table:: @code{The table is full} error -* Cannot create:: @code{Can't create/write to file} Error -* Commands out of sync:: @code{Commands out of sync} error in client -* Ignoring user:: @code{Ignoring user} error -* Cannot find table:: @code{Table 'xxx' doesn't exist} error -* Cannot initialize character set:: -@end menu +Example of @code{myisamchk -eiv} output: +@example +Checking MyISAM file: company +Data records: 1403698 Deleted blocks: 0 +- check file-size +- check delete-chain +block_size 1024: +index 1: +index 2: +index 3: +index 4: +index 5: +index 6: +index 7: +index 8: +index 9: +No recordlinks +- check index reference +- check data record references index: 1 +Key: 1: Keyblocks used: 97% Packed: 0% Max levels: 4 +- check data record references index: 2 +Key: 2: Keyblocks used: 98% Packed: 50% Max levels: 4 +- check data record references index: 3 +Key: 3: Keyblocks used: 97% Packed: 0% Max levels: 4 +- check data record references index: 4 +Key: 4: Keyblocks used: 99% Packed: 60% Max levels: 3 +- check data record references index: 5 +Key: 5: Keyblocks used: 99% Packed: 0% Max levels: 3 +- check data record references index: 6 +Key: 6: Keyblocks used: 99% Packed: 0% Max levels: 3 +- check data record references index: 7 +Key: 7: Keyblocks used: 99% Packed: 0% Max levels: 3 +- check data record references index: 8 +Key: 8: Keyblocks used: 99% Packed: 0% Max levels: 3 +- check data record references index: 9 +Key: 9: Keyblocks used: 98% Packed: 0% Max levels: 4 +Total: Keyblocks used: 9% Packed: 17% -This section lists some errors that users frequently get. You will find -descriptions of the errors, and how to solve the problem here. +- check records and index references +[LOTS OF ROW NUMBERS DELETED] -@cindex errors, access denied -@cindex problems, access denied errors -@cindex access denied errors -@node Error Access denied, Gone away, Common errors, Common errors -@subsection @code{Access denied} Error +Records: 1403698 M.recordlength: 226 Packed: 0% +Recordspace used: 100% Empty space: 0% Blocks/Record: 1.00 +Record blocks: 1403698 Delete blocks: 0 +Recorddata: 317235748 Deleted data: 0 +Lost space: 0 Linkdata: 0 -@xref{Privileges}, and especially. @xref{Access denied}. +User time 1639.63, System time 251.61 +Maximum resident set size 0, Integral resident set size 0 +Non physical pagefaults 0, Physical pagefaults 10580, Swaps 0 +Blocks in 4 out 0, Messages in 0 out 0, Signals 0 +Voluntary context switches 10604, Involuntary context switches 122798 +@end example -@node Gone away, Can not connect to server, Error Access denied, Common errors -@subsection @code{MySQL server has gone away} Error +Here are the sizes of the data and index files for the table used in the +preceding examples: -This section also covers the related @code{Lost connection to server -during query} error. +@example +-rw-rw-r-- 1 monty tcx 317235748 Jan 12 17:30 company.MYD +-rw-rw-r-- 1 davida tcx 96482304 Jan 12 18:35 company.MYM +@end example -The most common reason for the @code{MySQL server has gone away} error -is that the server timed out and closed the connection. By default, the -server closes the connection after 8 hours if nothing has happened. You -can change the time limit by setting the @code{wait_timeout} variable when -you start @code{mysqld}. +Explanations for the types of information @code{myisamchk} produces are +given below. The ``keyfile'' is the index file. ``Record'' and ``row'' +are synonymous: -Another common reason to receive the @code{MySQL server has gone away} error -is because you have issued a ``close'' on your @strong{MySQL} connection -and then tried to run a query on the closed connection. +@table @code +@item ISAM file +Name of the ISAM (index) file. -You can check that the @strong{MySQL} hasn't died by executing -@code{mysqladmin version} and examining the uptime. +@item Isam-version +Version of ISAM format. Currently always 2. -If you have a script, you just have to issue the query again for the client -to do an automatic reconnection. +@item Creation time +When the data file was created. -You normally can get the following error codes in this case -(which one you get is OS-dependent): +@item Recover time +When the index/data file was last reconstructed. -@multitable @columnfractions .3 .7 -@item @code{CR_SERVER_GONE_ERROR} @tab The client couldn't send a question to the -server. -@item @code{CR_SERVER_LOST} @tab The client didn't get an error when writing -to the server, but it didn't get a full answer (or any answer) to the question. -@end multitable +@item Data records +How many records are in the table. -You can also get these errors if you send a query to the server that is -incorrect or too large. If @code{mysqld} gets a packet that is too large -or out of order, it assumes that something has gone wrong with the client and -closes the connection. If you need big queries (for example, if you are -working with big @code{BLOB} columns), you can increase the query limit by -starting @code{mysqld} with the @code{-O max_allowed_packet=#} option -(default 1M). The extra memory is allocated on demand, so @code{mysqld} will -use more memory only when you issue a big query or when @code{mysqld} must -return a big result row! +@item Deleted blocks +How many deleted blocks still have reserved space. +You can optimize your table to minimize this space. +@xref{Optimization}. -@node Can not connect to server, Blocked host, Gone away, Common errors -@subsection @code{Can't connect to [local] MySQL server} error +@item Datafile: Parts +For dynamic record format, this indicates how many data blocks there are. For +an optimized table without fragmented records, this is the same as @code{Data +records}. -A @strong{MySQL} client on Unix can connect to the @code{mysqld} server in two -different ways: Unix sockets, which connect through a file in the file -system (default @file{/tmp/mysqld.sock}) or TCP/IP, which connects -through a port number. Unix sockets are faster than TCP/IP but can only -be used when connecting to a server on the same computer. Unix sockets -are used if you don't specify a hostname or if you specify the special -hostname @code{localhost}. +@item Deleted data +How many bytes of non-reclaimed deleted data there are. +You can optimize your table to minimize this space. +@xref{Optimization}. -On Windows you can connect only with TCP/IP if the @code{mysqld} server -is running on Win95/Win98. If it's running on NT, you can also connect -with named pipes. The name of the named pipe is @strong{MySQL}. If you -don't give a hostname when connecting to @code{mysqld}, a @strong{MySQL} client -will first try to connect to the named pipe, and if this doesn't work it -will connect to the TCP/IP port. You can force the use of named pipes -on Windows by using @code{.} as the hostname. +@item Datafile pointer +The size of the data file pointer, in bytes. It is usually 2, 3, 4, or 5 +bytes. Most tables manage with 2 bytes, but this cannot be controlled +from @strong{MySQL} yet. For fixed tables, this is a record address. For +dynamic tables, this is a byte address. -The error (2002) @code{Can't connect to ...} normally means that there -isn't a @strong{MySQL} server running on the system or that you are -using a wrong socket file or TCP/IP port when trying to connect to the -@code{mysqld} server. +@item Keyfile pointer +The size of the index file pointer, in bytes. It is usually 1, 2, or 3 +bytes. Most tables manage with 2 bytes, but this is calculated +automatically by @strong{MySQL}. It is always a block address. -Start by checking (using @code{ps} or the task manager on Windows) that -there is a process running named @code{mysqld} on your server! If there -isn't any @code{mysqld} process, you should start one. @xref{Starting -server}. +@item Max datafile length +How long the table's data file (@code{.MYD} file) can become, in bytes. -If a @code{mysqld} process is running, you can check the server by -trying these different connections (the port number and socket pathname -might be different in your setup, of course): +@item Max keyfile length +How long the table's key file (@code{.MYI} file) can become, in bytes. -@example -shell> mysqladmin version -shell> mysqladmin variables -shell> mysqladmin -h `hostname` version variables -shell> mysqladmin -h `hostname` --port=3306 version -shell> mysqladmin -h 'ip for your host' version -shell> mysqladmin --socket=/tmp/mysql.sock version -@end example +@item Recordlength +How much space each record takes, in bytes. -Note the use of backquotes rather than forward quotes with the @code{hostname} -command; these cause the output of @code{hostname} (that is, the current -hostname) to be substituted into the @code{mysqladmin} command. +@item Record format +The format used to store table rows. +The examples shown above use @code{Fixed length}. +Other possible values are @code{Compressed} and @code{Packed}. -Here are some reasons the @code{Can't connect to local MySQL server} -error might occur: +@item table description +A list of all keys in the table. For each key, some low-level information +is presented: -@itemize @bullet -@item -@code{mysqld} is not running. -@item -You are running on a system that uses MIT-pthreads. -If you are running on a system that doesn't have native threads, -@code{mysqld} uses the MIT-pthreads package. @xref{Which OS}. However, -all MIT-pthreads versions doesn't support Unix sockets. On a system -without sockets support you must always specify the hostname explicitly -when connecting to the server. Try using this command to check the -connection to the server: -@example -shell> mysqladmin -h `hostname` version -@end example -@item -Someone has removed the Unix socket that @code{mysqld} uses (default -@file{/tmp/mysqld.sock}). You might have a @code{cron} job that removes -the @strong{MySQL} socket (for example, a job that removes old files -from the @file{/tmp} directory). You can always run @code{mysqladmin -version} and check that the socket @code{mysqladmin} is trying to use -really exists. The fix in this case is to change the @code{cron} job to -not remove @file{mysqld.sock} or to place the socket somewhere else. You -can specify a different socket location at @strong{MySQL} configuration -time with this command: -@example -shell> ./configure --with-unix-socket-path=/path/to/socket -@end example -You can also start @code{safe_mysqld} with the -@code{--socket=/path/to/socket} option and set the environment variable -@code{MYSQL_UNIX_PORT} to the socket pathname before starting your -@strong{MySQL} clients. -@item -You have started the @code{mysqld} server with -the @code{--socket=/path/to/socket} option. If you change the socket -pathname for the server, you must also notify the @strong{MySQL} clients -about the new path. You can do this by setting the environment variable -@code{MYSQL_UNIX_PORT} to the socket pathname or by providing the socket path -as an argument to the clients. You can test the socket with this command: +@table @code +@item Key +This key's number. -@example -shell> mysqladmin --socket=/path/to/socket version -@end example -@item -You are using Linux and one thread has died (core dumped). In this case -you must kill the other @code{mysqld} threads (for example, with the -@code{mysql_zap} script before you can start a new @strong{MySQL} -server. @xref{Crashing}. -@item -You may not have read and write privilege to either the directory that holds -the socket file or privilege to the socket file itself. In this case you -have to either change the privilege for the directory / file or restart -@code{mysqld} so that it uses a directory that you can access. -@end itemize +@item Start +Where in the record this index part starts. -If you get the error message @code{Can't connect to MySQL server on -some_hostname}, you can try the following things to find out what the -problem is : +@item Len +How long this index part is. For packed numbers, this should always be +the full length of the column. For strings, it may be shorter than the full +length of the indexed column, because you can index a prefix of a string +column. -@itemize @bullet -@item -Check if the server is up by doing @code{telnet your-host-name -tcp-ip-port-number} and press @code{RETURN} a couple of times. If there -is a @strong{MySQL} server running on this port you should get a -responses that includes the version number of the running @strong{MySQL} -server. If you get an error like @code{telnet: Unable to connect to -remote host: Connection refused}, then there is no server running on the -given port. -@item -Try connecting to the @code{mysqld} daemon on the local machine and check -the TCP/IP port that @code{mysqld} it's configured to use (variable @code{port}) with -@code{mysqladmin variables}. -@item -Check that your @code{mysqld} server is not started with the -@code{--skip-networking} option. -@end itemize +@item Index +@code{unique} or @code{multip.} (multiple). Indicates whether or not one value +can exist multiple times in this index. -@node Blocked host, Too many connections, Can not connect to server, Common errors -@subsection @code{Host '...' is blocked} Error +@item Type +What data-type this index part has. This is an ISAM data-type +with the options @code{packed}, @code{stripped} or @code{empty}. -If you get an error like this: +@item Root +Address of the root index block. -@example -Host 'hostname' is blocked because of many connection errors. -Unblock with 'mysqladmin flush-hosts' -@end example +@item Blocksize +The size of each index block. By default this is 1024, but the value may be +changed at compile time. -this means that @code{mysqld} has gotten a lot (@code{max_connect_errors}) -of connect requests from the host @code{'hostname'} that have been interrupted -in the middle. After @code{max_connect_errors} failed requests, @code{mysqld} -assumes that something is wrong (like an attack from a cracker), and -blocks the site from further connections until someone executes the command -@code{mysqladmin flush-hosts}. +@item Rec/key +This is a statistical value used by the optimizer. It tells how many +records there are per value for this key. A unique key always has a +value of 1. This may be updated after a table is loaded (or greatly +changed) with @code{myisamchk -a}. If this is not updated at all, a default +value of 30 is given. +@end table -By default, @code{mysqld} blocks a host after 10 connection errors. -You can easily adjust this by starting the server like this: +@item +In the first example above, the 9th key is a multi-part key with two parts. -@example -shell> safe_mysqld -O max_connect_errors=10000 & -@end example +@item Keyblocks used +What percentage of the keyblocks are used. Because the table used in the +examples had just been reorganized with @code{myisamchk}, the values are very +high (very near the theoretical maximum). -Note that if you get this error message for a given host, you should first -check that there isn't anything wrong with TCP/IP connections from that -host. If your TCP/IP connections aren't working, it won't do you any good to -increase the value of the @code{max_connect_errors} variable! +@item Packed +@strong{MySQL} tries to pack keys with a common suffix. This can only be used +for @code{CHAR}/@code{VARCHAR}/@code{DECIMAL} keys. For long strings like +names, this can significantly reduce the space used. In the third example +above, the 4th key is 10 characters long and a 60% reduction in space is +achieved. -@node Too many connections, Non-transactional tables, Blocked host, Common errors -@subsection @code{Too many connections} Error +@item Max levels +How deep the B-tree for this key is. Large tables with long keys get high +values. -If you get the error @code{Too many connections} when you try to connect -to @strong{MySQL}, this means that there is already @code{max_connections} -clients connected to the @code{mysqld} server. +@item Records +How many rows are in the table. -If you need more connections than the default (100), then you should restart -@code{mysqld} with a bigger value for the @code{max_connections} variable. +@item M.recordlength +The average record length. For tables with fixed-length records, this is the +exact record length. -Note that @code{mysqld} actually allows (@code{max_connections}+1) -clients to connect. The last connection is reserved for a user with the -@strong{process} privilege. By not giving this privilege to normal -users (they shouldn't need this), an administrator with this privilege -can log in and use @code{SHOW PROCESSLIST} to find out what could be -wrong. @xref{SHOW}. +@item Packed +@strong{MySQL} strips spaces from the end of strings. The @code{Packed} +value indicates the percentage of savings achieved by doing this. -The maximum number of connects @strong{MySQL} is depending on how good -the thread library is on a given platform. Linux or Solaris should be -able to support 500-1000 simultaneous connections, depending on how much -RAM you have and what your clients are doing. +@item Recordspace used +What percentage of the data file is used. -@cindex Non-transactional tables -@node Non-transactional tables, Out of memory, Too many connections, Common errors -@subsection @code{Some non-transactional changed tables couldn't be rolled back} Error +@item Empty space +What percentage of the data file is unused. -If you get the error/warning: @code{Warning: Some non-transactional -changed tables couldn't be rolled back} when trying to do a -@code{ROLLBACK}, this means that some of the tables you used in the -transaction didn't support transactions. These non-transactional tables -will not be affected by the @code{ROLLBACK} statement. +@item Blocks/Record +Average number of blocks per record (that is, how many links a fragmented +record is composed of). This is always 1 for fixed-format tables. This value +should stay as close to 1.0 as possible. If it gets too big, you can +reorganize the table with @code{myisamchk}. +@xref{Optimization}. -The most typical case when this happens is when you have tried to create -a table of a type that is not supported by your @code{mysqld} binary. -If @code{mysqld} doesn't support a table type (or if the table type is -disabled by a startup option) , it will instead create the table type -with the table type that is most resembles to the one you requested, -probably @code{MyISAM}. +@item Recordblocks +How many blocks (links) are used. For fixed format, this is the same as the number +of records. -You can check the table type for a table by doing: +@item Deleteblocks +How many blocks (links) are deleted. -@code{SHOW TABLE STATUS LIKE 'table_name'}. @xref{SHOW TABLE STATUS}. +@item Recorddata +How many bytes in the data file are used. -You can check the extensions your @code{mysqld} binary supports by doing: +@item Deleted data +How many bytes in the data file are deleted (unused). -@code{show variables like 'have_%'}. @xref{SHOW VARIABLES}. +@item Lost space +If a record is updated to a shorter length, some space is lost. This is +the sum of all such losses, in bytes. -@node Out of memory, Packet too large, Non-transactional tables, Common errors -@subsection @code{Out of memory} Error +@item Linkdata +When the dynamic table format is used, record fragments are linked with +pointers (4 to 7 bytes each). @code{Linkdata} is the sum of the amount of +storage used by all such pointers. +@end table -If you issue a query and get something like the following error: +If a table has been compressed with @code{myisampack}, @code{myisamchk +-d} prints additional information about each table column. See +@ref{myisampack, , @code{myisampack}}, for an example of this +information and a description of what it means. -@example -mysql: Out of memory at line 42, 'malloc.c' -mysql: needed 8136 byte (8k), memory in use: 12481367 bytes (12189k) -ERROR 2008: MySQL client ran out of memory -@end example +@cindex crash, recovery +@cindex recovery, from crash +@node Crash recovery, Log file maintenance, Table-info, Maintenance +@section Using @code{myisamchk} for Crash Recovery -note that the error refers to the @strong{MySQL} client @code{mysql}. The -reason for this error is simply that the client does not have enough memory to -store the whole result. +If you run @code{mysqld} with @code{--skip-locking} (which is the default on +some systems, like Linux), you can't reliably use @code{myisamchk} to +check a table when @code{mysqld} is using the same table. If you +can be sure that no one is accessing the tables through @code{mysqld} +while you run @code{myisamchk}, you only have to do @code{mysqladmin +flush-tables} before you start checking the tables. If you can't +guarantee the above, then you must take down @code{mysqld} while you +check the tables. If you run @code{myisamchk} while @code{mysqld} is updating +the tables, you may get a warning that a table is corrupt even if it +isn't. -To remedy the problem, first check that your query is correct. Is it -reasonable that it should return so many rows? If so, -you can use @code{mysql --quick}, which uses @code{mysql_use_result()} -to retrieve the result set. This places less of a load on the client (but -more on the server). +If you are not using @code{--skip-locking}, you can use @code{myisamchk} +to check tables at any time. While you do this, all clients that try +to update the table will wait until @code{myisamchk} is ready before +continuing. -@node Packet too large, Communication errors, Out of memory, Common errors -@subsection @code{Packet too large} Error +If you use @code{myisamchk} to repair or optimize tables, you +@strong{MUST} always ensure that the @code{mysqld} server is not using +the table (this also applies if you are using @code{--skip-locking}). +If you don't take down @code{mysqld} you should at least do a +@code{mysqladmin flush-tables} before you run @code{myisamchk}. -When a @strong{MySQL} client or the @code{mysqld} server gets a packet bigger -than @code{max_allowed_packet} bytes, it issues a @code{Packet too large} -error and closes the connection. +This chapter describes how to check for and deal with data corruption +in @strong{MySQL} databases. If your tables get corrupted a lot you should +try to find the reason for this! @xref{Crashing}. -If you are using the @code{mysql} client, you may specify a bigger buffer by -starting the client with @code{mysql --set-variable=max_allowed_packet=8M}. +The @code{MyISAM} table section contains reason for why a table could be +corrupted. @xref{MyISAM table problems}. -If you are using other clients that do not allow you to specify the maximum -packet size (such as @code{DBI}), you need to set the packet size when you -start the server. You cau use a command-line option to @code{mysqld} to set -@code{max_allowed_packet} to a larger size. For example, if you are -expecting to store the full length of a @code{BLOB} into a table, you'll need -to start the server with the @code{--set-variable=max_allowed_packet=16M} -option. +When performing crash recovery, it is important to understand that each table +@code{tbl_name} in a database corresponds to three files in the database +directory: -@cindex aborted clients -@cindex aborted connection -@cindex connection, aborted -@node Communication errors, Full table, Packet too large, Common errors -@subsection Communication Errors / Aborted Connection +@multitable @columnfractions .2 .8 +@item @strong{File} @tab @strong{Purpose} +@item @file{tbl_name.frm} @tab Table definition (form) file +@item @file{tbl_name.MYD} @tab Data file +@item @file{tbl_name.MYI} @tab Index file +@end multitable -Starting with @code{MySQL 3.23.40} you only get the @code{Aborted -connection} error of you start @code{mysqld} with @code{--warnings}. +Each of these three file types is subject to corruption in various ways, but +problems occur most often in data files and index files. -If you find errors like the following in your error log. +@code{myisamchk} works by creating a copy of the @file{.MYD} (data) file +row by row. It ends the repair stage by removing the old @file{.MYD} +file and renaming the new file to the original file name. If you use +@code{--quick}, @code{myisamchk} does not create a temporary @file{.MYD} +file, but instead assumes that the @file{.MYD} file is correct and only +generates a new index file without touching the @file{.MYD} file. This +is safe, because @code{myisamchk} automatically detects if the +@file{.MYD} file is corrupt and aborts the repair in this case. You can +also give two @code{--quick} options to @code{myisamchk}. In this case, +@code{myisamchk} does not abort on some errors (like duplicate key) but +instead tries to resolve them by modifying the @file{.MYD} +file. Normally the use of two @code{--quick} options is useful only if +you have too little free disk space to perform a normal repair. In this +case you should at least make a backup before running @code{myisamchk}. -@example -010301 14:38:23 Aborted connection 854 to db: 'users' user: 'josh' -@end example +@menu +* Check:: How to check tables for errors +* Repair:: How to repair tables +* Optimization:: Table optimization +@end menu -@xref{Error log}. +@cindex checking, tables for errors +@cindex tables, error checking +@cindex errors, checking tables for +@node Check, Repair, Crash recovery, Crash recovery +@subsection How to Check Tables for Errors -This means that something of the following has happened: +To check a MyISAM table, use the following commands: -@itemize @bullet -@item -The client program did not call @code{mysql_close()} before exit. -@item -The client had been sleeping more than @code{wait_timeout} or -@code{interactive_timeout} without doing any requests. @xref{SHOW -VARIABLES}. -@item -The client program ended abruptly in the middle of the transfer. -@end itemize +@table @code +@item myisamchk tbl_name +This finds 99.99% of all errors. What it can't find is corruption that +involves @strong{ONLY} the data file (which is very unusual). If you want +to check a table, you should normally run @code{myisamchk} without options or +with either the @code{-s} or @code{--silent} option. -When the above happens, the server variable @code{Aborted_clients} is -incremented. +@item myisamchk -m tbl_name +This finds 99.999% of all errors. It checks first all index entries for errors and +then it reads through all rows. It calculates a checksum for all keys in +the rows and verifies that they checksum matches the checksum for the keys +in the index tree. -The server variable @code{Aborted_connects} is incremented when: +@item myisamchk -e tbl_name +This does a complete and thorough check of all data (@code{-e} means +``extended check''). It does a check-read of every key for each row to verify +that they indeed point to the correct row. This may take a LONG time on a +big table with many keys. @code{myisamchk} will normally stop after the first +error it finds. If you want to obtain more information, you can add the +@code{--verbose} (@code{-v}) option. This causes @code{myisamchk} to keep +going, up through a maximum of 20 errors. In normal usage, a simple +@code{myisamchk} (with no arguments other than the table name) is sufficient. -@itemize @bullet -@item -When a connection packet doesn't contain the right information. -@item -When the user didn't have privileges to connect to a database. -@item -When a user uses a wrong password. -@item -When it takes more than @code{connect_timeout} seconds to get -a connect package. -@end itemize +@item myisamchk -e -i tbl_name +Like the previous command, but the @code{-i} option tells @code{myisamchk} to +print some informational statistics, too. +@end table -Note that the above could indicate that someone is trying to break into -your database! +@cindex tables, repairing +@cindex repairing, tables +@node Repair, Optimization, Check, Crash recovery +@subsection How to Repair Tables -@xref{SHOW VARIABLES}. +In the following section we only talk about using @code{myisamchk} on +@code{MyISAM} tables (extensions @code{.MYI} and @code{.MYD}). If you +are using @code{ISAM} tables (extensions @code{.ISM} and @code{.ISD}), +you should use @code{isamchk} instead. + +Starting with @strong{MySQL} Version 3.23.14, you can repair MyISAM +tables with the @code{REPAIR TABLE} command. @xref{REPAIR TABLE}. + +The symptoms of a corrupted table include queries that abort unexpectedly +and observable errors such as these: -Other reasons for problems with Aborted clients / Aborted connections. @itemize @bullet @item -Usage of duplex Ethernet protocol, both half and full with -Linux. Many Linux Ethernet drivers have this bug. You should test -for this bug by transferring a huge file via ftp between these two -machines. If a transfer goes in burst-pause-burst-pause ... mode then -you are experiencing a Linux duplex syndrome. The only solution to -this problem is switching of both half and full duplexing on hubs -and switches. +@file{tbl_name.frm} is locked against change @item -Some problem with the thread library that causes interrupts on reads. +Can't find file @file{tbl_name.MYI} (Errcode: ###) @item -Badly configured TCP/IP. +Unexpected end of file @item -Faulty Ethernets or hubs or switches, cables ... This can be diagnosed -properly only by replacing hardware. -@end itemize +Record file is crashed +@item +Got error ### from table handler +To get more information about the error you can run @code{perror ###}. Here +is the most common errors that indicates a problem with the table: -@cindex table is full -@node Full table, Cannot create, Communication errors, Common errors -@subsection @code{The table is full} Error +@example +shell> perror 126 127 132 134 135 136 141 144 145 +126 = Index file is crashed / Wrong file format +127 = Record-file is crashed +132 = Old database file +134 = Record was already deleted (or record file crashed) +135 = No more room in record file +136 = No more room in index file +141 = Duplicate unique key or constraint on write or update +144 = Table is crashed and last repair failed +145 = Table was marked as crashed and should be repaired +@end example -This error occurs in older @strong{MySQL} versions when an in-memory temporary -table becomes larger than @code{tmp_table_size} bytes. To avoid this -problem, you can use the @code{-O tmp_table_size=#} option to -@code{mysqld} to increase the temporary table size or use the SQL -option @code{SQL_BIG_TABLES} before you issue the problematic -query. @xref{SET OPTION, , @code{SET OPTION}}. +Note that error 135, no more room in record file, is not an error that +can be fixed by a simple repair. In this case you have to do: -You can also start @code{mysqld} with the @code{--big-tables} option. -This is exactly the same as using @code{SQL_BIG_TABLES} for all queries. +@example +ALTER TABLE table MAX_ROWS=xxx AVG_ROW_LENGTH=yyy; +@end example -In @strong{MySQL} Version 3.23, in-memory temporary tables will automatically be -converted to a disk-based @code{MyISAM} table after the table size gets -bigger than @code{tmp_table_size}. +@end itemize -@cindex can't create/write to file -@node Cannot create, Commands out of sync, Full table, Common errors -@subsection @code{Can't create/write to file} Error +In the other cases, you must repair your tables. @code{myisamchk} +can usually detect and fix most things that go wrong. -If you get an error for some queries of type: +The repair process involves up to four stages, described below. Before you +begin, you should @code{cd} to the database directory and check the +permissions of the table files. Make sure they are readable by the Unix user +that @code{mysqld} runs as (and to you, because you need to access the files +you are checking). If it turns out you need to modify files, they must also +be writable by you. -@example -Can't create/write to file '\\sqla3fe_0.ism'. -@end example +If you are using @strong{MySQL} Version 3.23.16 and above, you can (and +should) use the @code{CHECK} and @code{REPAIR} commands to check and repair +@code{MyISAM} tables. @xref{CHECK TABLE}. @xref{REPAIR TABLE}. -this means that @strong{MySQL} can't create a temporary file for the -result set in the given temporary directory. (The above error is a -typical error message on Windows, and the Unix error message is similar.) -The fix is to start @code{mysqld} with @code{--tmpdir=path} or to add to your option -file: +The manual section about table maintenance includes the options to +@code{isamchk}/@code{myisamchk}. @xref{Table maintenance}. -@example -[mysqld] -tmpdir=C:/temp -@end example +The following section is for the cases where the above command fails or +if you want to use the extended features that @code{isamchk}/@code{myisamchk} provides. -assuming that the @file{c:\\temp} directory exists. @xref{Option files}. +If you are going to repair a table from the command line, you must first +take down the @code{mysqld} server. Note that when you do +@code{mysqladmin shutdown} on a remote server, the @code{mysqld} server +will still be alive for a while after @code{mysqladmin} returns, until +all queries are stopped and all keys have been flushed to disk. -Check also the error code that you get with @code{perror}. One reason -may also be a disk full error; +@noindent +@strong{Stage 1: Checking your tables} -@example -shell> perror 28 -Error code 28: No space left on device -@end example +Run @code{myisamchk *.MYI} or @code{myisamchk -e *.MYI} if you have +more time. Use the @code{-s} (silent) option to suppress unnecessary +information. -@cindex commands out of sync -@node Commands out of sync, Ignoring user, Cannot create, Common errors -@subsection @code{Commands out of sync} Error in Client +If the @code{mysqld} server is done you should use the --update option to tell +@code{myisamchk} to mark the table as 'checked'. -If you get @code{Commands out of sync; You can't run this command now} -in your client code, you are calling client functions in the wrong order! +You have to repair only those tables for which @code{myisamchk} announces an +error. For such tables, proceed to Stage 2. -This can happen, for example, if you are using @code{mysql_use_result()} and -try to execute a new query before you have called @code{mysql_free_result()}. -It can also happen if you try to execute two queries that return data without -a @code{mysql_use_result()} or @code{mysql_store_result()} in between. +If you get weird errors when checking (such as @code{out of +memory} errors), or if @code{myisamchk} crashes, go to Stage 3. -@node Ignoring user, Cannot find table, Commands out of sync, Common errors -@subsection @code{Ignoring user} Error +@noindent +@strong{Stage 2: Easy safe repair} -If you get the following error: +NOTE: If you want repairing to go much faster, you should add: @code{-O +sort_buffer=# -O key_buffer=#} (where # is about 1/4 of the available +memory) to all @code{isamchk/myisamchk} commands. -@code{Found wrong password for user: 'some_user@@some_host'; Ignoring user} +First, try @code{myisamchk -r -q tbl_name} (@code{-r -q} means ``quick +recovery mode''). This will attempt to repair the index file without +touching the data file. If the data file contains everything that it +should and the delete links point at the correct locations within the +data file, this should work, and the table is fixed. Start repairing the +next table. Otherwise, use the following procedure: -this means that when @code{mysqld} was started or when it reloaded the -permissions tables, it found an entry in the @code{user} table with -an invalid password. As a result, the entry is simply ignored by the -permission system. +@enumerate +@item +Make a backup of the data file before continuing. -Possible causes of and fixes for this problem: +@item +Use @code{myisamchk -r tbl_name} (@code{-r} means ``recovery mode''). This will +remove incorrect records and deleted records from the data file and +reconstruct the index file. -@itemize @bullet @item -You may be running a new version of @code{mysqld} with an old -@code{user} table. -You can check this by executing @code{mysqlshow mysql user} to see if -the password field is shorter than 16 characters. If so, you can correct this -condition by running the @code{scripts/add_long_password} script. +If the preceding step fails, use @code{myisamchk --safe-recover tbl_name}. +Safe recovery mode uses an old recovery method that handles a few cases that +regular recovery mode doesn't (but is slower). +@end enumerate + +If you get weird errors when repairing (such as @code{out of +memory} errors), or if @code{myisamchk} crashes, go to Stage 3. + +@noindent +@strong{Stage 3: Difficult repair} + +You should only reach this stage if the first 16K block in the index file is +destroyed or contains incorrect information, or if the index file is +missing. In this case, it's necessary to create a new index file. Do so as +follows: +@enumerate @item -The user has an old password (8 characters long) and you didn't start -@code{mysqld} with the @code{--old-protocol} option. -Update the user in the @code{user} table with a new password or -restart @code{mysqld} with @code{--old-protocol}. +Move the data file to some safe place. @item -@findex PASSWORD() -You have specified a password in the @code{user} table without using the -@code{PASSWORD()} function. Use @code{mysql} to update the user in the -@code{user} table with a new password. Make sure to use the @code{PASSWORD()} -function: +Use the table description file to create new (empty) data and index files: @example -mysql> update user set password=PASSWORD('your password') - where user='XXX'; +shell> mysql db_name +mysql> SET AUTOCOMMIT=1; +mysql> TRUNCATE TABLE table_name; +mysql> quit @end example -@end itemize -@node Cannot find table, Cannot initialize character set, Ignoring user, Common errors -@subsection @code{Table 'xxx' doesn't exist} Error +If your SQL version doesn't have @code{TRUNCATE TABLE}, use @code{DELETE FROM +table_name} instead. -If you get the error @code{Table 'xxx' doesn't exist} or @code{Can't -find file: 'xxx' (errno: 2)}, this means that no table exists -in the current database with the name @code{xxx}. +@item +Copy the old data file back onto the newly created data file. +(Don't just move the old file back onto the new file; you want to retain +a copy in case something goes wrong.) +@end enumerate -Note that as @strong{MySQL} uses directories and files to store databases and -tables, the database and table names are @strong{case sensitive}! -(On Windows the databases and tables names are not case sensitive, but all -references to a given table within a query must use the same case!) +Go back to Stage 2. @code{myisamchk -r -q} should work now. (This shouldn't +be an endless loop.) -You can check which tables you have in the current database with -@code{SHOW TABLES}. @xref{SHOW, , @code{SHOW}}. +@noindent +@strong{Stage 4: Very difficult repair} -@cindex multibyte character sets -@node Cannot initialize character set, , Cannot find table, Common errors -@subsection @code{Can@'t initialize character set xxx} error. +You should reach this stage only if the description file has also +crashed. That should never happen, because the description file isn't changed +after the table is created: -If you get an error like: +@enumerate +@item +Restore the description file from a backup and go back to Stage 3. You can +also restore the index file and go back to Stage 2. In the latter case, you +should start with @code{myisamchk -r}. + +@item +If you don't have a backup but know exactly how the table was created, create +a copy of the table in another database. Remove the new data file, then move +the description and index files from the other database to your crashed +database. This gives you new description and index files, but leaves +the data file alone. Go back to Stage 2 and attempt to reconstruct +the index file. +@end enumerate + +@cindex tables, optimizing +@cindex optimizing, tables +@node Optimization, , Repair, Crash recovery +@subsection Table Optimization + +To coalesce fragmented records and eliminate wasted space resulting from +deleting or updating records, run @code{myisamchk} in recovery mode: @example -MySQL Connection Failed: Can't initialize character set xxx +shell> myisamchk -r tbl_name @end example -This means one of the following things: +You can optimize a table in the same way using the SQL @code{OPTIMIZE TABLE} +statement. @code{OPTIMIZE TABLE} does a repair of the table, a key +analyzes and also sorts the index tree to give faster key lookups. +There is also no possibility of unwanted interaction between a utility +and the server, because the server does all the work when you use +@code{OPTIMIZE TABLE}. @xref{OPTIMIZE TABLE}. -@itemize @bullet -@item -The character set is a multi-byte character set and you have not support -for the character set in the client. +@code{myisamchk} also has a number of other options you can use to improve +the performance of a table: -In this case you need to recompile the client with -@code{--with-charset=xxx} or with @code{--with-extra-charsets=xxx}. -@xref{configure options}. +@table @code +@item -S, --sort-index +@item -R index_num, --sort-records=index_num +@item -a, --analyze +@end table -All standard @strong{MySQL} binaries are compiled with -@code{--with-extra-character-sets=complex} which will enable support for -all multi-byte character sets. @xref{Character sets}. +For a full description of the option. @xref{myisamchk syntax}. -@item -The character set is a simple character set which is not compiled into -@code{mysqld} and the character set definition files is not in the place -where the client expect to find them. +@cindex files, log +@cindex maintaining, log files +@cindex log files, maintaining +@node Log file maintenance, , Crash recovery, Maintenance +@section Log file Maintenance -In this case you need to: +@strong{MySQL} has a lot of log files which make it easy to see what is +going. @xref{Log files}. One must however from time to time clean up +after @code{MysQL} to ensure that the logs don't take up too much disk +space. + +When using @strong{MySQL} with log files, you will, from time to time, +want to remove/backup old log files and tell @strong{MySQL} to start +logging on new files. @xref{Backup}. + +On a Linux (@code{Redhat}) installation, you can use the +@code{mysql-log-rotate} script for this. If you installed @strong{MySQL} +from an RPM distribution, the script should have been installed +automatically. Note that you should be careful with this if you are using +the log for replication! + +On other systems you must install a short script yourself that you +start from @code{cron} to handle log files. + +You can force @strong{MySQL} to start using new log files by using +@code{mysqladmin flush-logs} or by using the SQL command @code{FLUSH LOGS}. +If you are using @strong{MySQL} Version 3.21 you must use @code{mysqladmin refresh}. + +The above command does the following: @itemize @bullet @item -Recompile the client with support for the character set. -@xref{configure options}. -@item -Specify to the client where the character set definition files are. For many -client you can do this with the -@code{--character-sets-dir=path-to-charset-dir} option. +If standard logging (@code{--log}) or slow query logging +(@code{--log-slow-queries}) is used, closes and reopens the log file. +(@file{mysql.log} and @file{`hostname`-slow.log} as default). @item -Copy the character definition files to the path where the client expect them -to be. -@end itemize +If update logging (@code{--log-update}) is used, closes the update log and +opens a new log file with a higher sequence number. @end itemize -@cindex full disk -@cindex disk full -@node Full disk, Multiple sql commands, Common errors, Problems -@section How MySQL Handles a Full Disk +If you are using only an update log, you only have to flush the logs and then +move away the old update log files to a backup. +If you are using the normal logging, you can do something like: -@noindent -When a disk-full condition occurs, @strong{MySQL} does the following: +@example +shell> cd mysql-data-directory +shell> mv mysql.log mysql.old +shell> mysqladmin flush-logs +@end example + +and then take a backup and remove @file{mysql.old}. + +@cindex functions, new +@cindex adding, new functions +@cindex user-defined functions, adding +@cindex UDFs, defined +@cindex functions, user-defined +@node Adding functions, Adding procedures, Maintenance, Top +@chapter Adding New Functions to MySQL + +There are two ways to add new functions to @strong{MySQL}: + +@itemize @bullet +@item You can add the function through the user-definable function (UDF) +interface. User-definable functions are added and removed dynamically using +the @code{CREATE FUNCTION} and @code{DROP FUNCTION} statements. +@xref{CREATE FUNCTION, , @code{CREATE FUNCTION}}. + +@item You can add the function as a native (built in) @strong{MySQL} function. +Native functions are compiled into the @code{mysqld} server and become +available on a permanent basis. +@end itemize + +Each method has advantages and disadvantages: @itemize @bullet @item -It checks once every minute to see whether or not there is enough space to -write the current row. If there is enough space, it continues as if nothing had -happened. +If you write a user-definable function, you must install the object file +in addition to the server itself. If you compile your function into the +server, you don't need to do that. @item -Every 6 minutes it writes an entry to the log file warning about the disk -full condition. +You can add UDFs to a binary @strong{MySQL} distribution. Native functions +require you to modify a source distribution. +@item +If you upgrade your @strong{MySQL} distribution, you can continue to use your +previously installed UDFs. For native functions, you must repeat your +modifications each time you upgrade. @end itemize -@noindent -To alleviate the problem, you can take the following actions: +Whichever method you use to add new functions, they may be used just like +native functions such as @code{ABS()} or @code{SOUNDEX()}. + +@menu +* Adding UDF:: Adding a new user-definable function +* Adding native function:: Adding a new native function +@end menu + +@cindex adding, user-definable functions +@cindex user-defined functions, adding +@cindex functions, user-definable, adding +@node Adding UDF, Adding native function, Adding functions, Adding functions +@section Adding a New User-definable Function + +@menu +* UDF calling sequences:: UDF calling sequences +* UDF arguments:: Argument processing +* UDF return values:: Return values and error handling +* UDF compiling:: Compiling and installing user-definable functions +@end menu + +For the UDF mechanism to work, functions must be written in C or C++ and your +operating system must support dynamic loading. The @strong{MySQL} source +distribution includes a file @file{sql/udf_example.cc} that defines 5 new +functions. Consult this file to see how UDF calling conventions work. + +For @code{mysqld} to be able to use UDF functions, you should configure MySQL +with @code{--with-mysqld-ldflags=-rdynamic} The reason is that to on +many platforms (including Linux) you can load a dynamic library (with +@code{dlopen()}) from a static linked program, which you would get if +you are using @code{--with-mysqld-ldflags=-all-static} If you want to +use an UDF that needs to access symbols from @code{mysqld} (like the +@code{methaphone} example in @file{sql/udf_example.cc} that uses +@code{default_charset_info}), you must link the program with +@code{-rdynamic}. (see @code{man dlopen}). + +For each function that you want to use in SQL statements, you should define +corresponding C (or C++) functions. In the discussion below, the name +``xxx'' is used for an example function name. To distinquish between SQL and +C/C++ usage, @code{XXX()} (uppercase) indicates a SQL function call, and +@code{xxx()} (lowercase) indicates a C/C++ function call. + +The C/C++ functions that you write to implement the interface for +@code{XXX()} are: + +@table @asis +@item @code{xxx()} (required) +The main function. This is where the function result is computed. +The correspondence between the SQL type and return type of your C/C++ +function is shown below: + +@multitable @columnfractions .2 .8 +@item @strong{SQL type} @tab @strong{C/C++ type} +@item @code{STRING} @tab @code{char *} +@item @code{INTEGER} @tab @code{long long} +@item @code{REAL} @tab @code{double} +@end multitable + +@item @code{xxx_init()} (optional) +The initialization function for @code{xxx()}. It can be used to: @itemize @bullet @item -To continue, you only have to free enough disk space to insert all records. +Check the number of arguments to @code{XXX()}. @item -To abort the thread, you must send a @code{mysqladmin kill} to the thread. -The thread will be aborted the next time it checks the disk (in 1 minute). +Check that the arguments are of a required type or, alternatively, +tell @strong{MySQL} to coerce arguments to the types you want when +the main function is called. @item -Note that other threads may be waiting for the table that caused the disk -full condition. If you have several ``locked'' threads, killing the one -thread that is waiting on the disk-full condition will allow the other -threads to continue. +Allocate any memory required by the main function. +@item +Specify the maximum length of the result. +@item +Specify (for @code{REAL} functions) the maximum number of decimals. +@item +Specify whether or not the result can be @code{NULL}. @end itemize -Exceptions to the above behaveour is when you use @code{REPAIR} or -@code{OPTIMIZE} or when the indexes are created in a batch after an -@code{LOAD DATA INFILE} or after an @code{ALTER TABLE} statement. +@item @code{xxx_deinit()} (optional) +The deinitialization function for @code{xxx()}. It should deallocate any +memory allocated by the initialization function. +@end table -All of the above commands may use big temporary files that left to -themself would cause big problems for the rest of the system. If -@strong{MySQL} gets disk full while doing any of the above operations, -it will remove the big temporary files and mark the table as crashed -(except for @code{ALTER TABLE}, in which the old table will be left -unchanged). +When a SQL statement invokes @code{XXX()}, @strong{MySQL} calls the +initialization function @code{xxx_init()} to let it perform any required +setup, such as argument checking or memory allocation. If @code{xxx_init()} +returns an error, the SQL statement is aborted with an error message and the +main and deinitialization functions are not called. Otherwise, the main +function @code{xxx()} is called once for each row. After all rows have been +processed, the deinitialization function @code{xxx_deinit()} is called so it +can perform any required cleanup. -@node Multiple sql commands, Temporary files, Full disk, Problems -@section How to Run SQL Commands from a Text File +All functions must be thread safe (not just the main function, +but the initialization and deinitialization functions as well). This means +that you are not allowed to allocate any global or static variables that +change! If you need memory, you should allocate it in @code{xxx_init()} +and free it in @code{xxx_deinit()}. -The @code{mysql} client typically is used interactively, like this: +@cindex calling sequences, UDF +@node UDF calling sequences, UDF arguments, Adding UDF, Adding UDF +@subsection UDF Calling Sequences -@example -shell> mysql database -@end example +The main function should be declared as shown below. Note that the return +type and parameters differ, depending on whether you will declare the SQL +function @code{XXX()} to return @code{STRING}, @code{INTEGER}, or @code{REAL} +in the @code{CREATE FUNCTION} statement: -However, it's also possible to put your SQL commands in a file and tell -@code{mysql} to read its input from that file. To do so, create a text -file @file{text_file} that contains the commands you wish to execute. -Then invoke @code{mysql} as shown below: +@noindent +For @code{STRING} functions: @example -shell> mysql database < text_file +char *xxx(UDF_INIT *initid, UDF_ARGS *args, + char *result, unsigned long *length, + char *is_null, char *error); @end example -You can also start your text file with a @code{USE db_name} statement. In -this case, it is unnecessary to specify the database name on the command -line: +@noindent +For @code{INTEGER} functions: @example -shell> mysql < text_file +long long xxx(UDF_INIT *initid, UDF_ARGS *args, + char *is_null, char *error); @end example -@xref{Programs}. - -@node Temporary files, Problems with mysql.sock, Multiple sql commands, Problems -@section Where MySQL Stores Temporary Files - -@strong{MySQL} uses the value of the @code{TMPDIR} environment variable as -the pathname of the directory in which to store temporary files. If you don't -have @code{TMPDIR} set, @strong{MySQL} uses the system default, which is -normally @file{/tmp} or @file{/usr/tmp}. If the file system containing your -temporary file directory is too small, you should edit @code{safe_mysqld} to -set @code{TMPDIR} to point to a directory in a file system where you have -enough space! You can also set the temporary directory using the -@code{--tmpdir} option to @code{mysqld}. +@noindent +For @code{REAL} functions: -@strong{MySQL} creates all temporary files as hidden files. This ensures -that the temporary files will be removed if @code{mysqld} is terminated. The -disadvantage of using hidden files is that you will not see a big temporary -file that fills up the file system in which the temporary file directory is -located. +@example +double xxx(UDF_INIT *initid, UDF_ARGS *args, + char *is_null, char *error); +@end example -When sorting (@code{ORDER BY} or @code{GROUP BY}), @strong{MySQL} normally -uses one or two temporary files. The maximum disk-space needed is: +The initialization and deinitialization functions are declared like this: @example -(length of what is sorted + sizeof(database pointer)) -* number of matched rows -* 2 +my_bool xxx_init(UDF_INIT *initid, UDF_ARGS *args, char *message); + +void xxx_deinit(UDF_INIT *initid); @end example -@code{sizeof(database pointer)} is usually 4, but may grow in the future for -really big tables. +The @code{initid} parameter is passed to all three functions. It points to a +@code{UDF_INIT} structure that is used to communicate information between +functions. The @code{UDF_INIT} structure members are listed below. The +initialization function should fill in any members that it wishes to change. +(To use the default for a member, leave it unchanged.): -For some @code{SELECT} queries, @strong{MySQL} also creates temporary SQL -tables. These are not hidden and have names of the form @file{SQL_*}. +@table @code +@item my_bool maybe_null +@code{xxx_init()} should set @code{maybe_null} to @code{1} if @code{xxx()} +can return @code{NULL}. The default value is @code{1} if any of the +arguments are declared @code{maybe_null}. -@code{ALTER TABLE} creates a temporary table in the same directory as -the original table. +@item unsigned int decimals +Number of decimals. The default value is the maximum number of decimals in +the arguments passed to the main function. (For example, if the function is +passed @code{1.34}, @code{1.345}, and @code{1.3}, the default would be 3, +because @code{1.345} has 3 decimals. -@cindex @code{mysql.sock}, protection -@cindex deletion, @code{mysql.sock} -@node Problems with mysql.sock, Changing MySQL user, Temporary files, Problems -@section How to Protect @file{/tmp/mysql.sock} from Being Deleted +@item unsigned int max_length +The maximum length of the string result. The default value differs depending +on the result type of the function. For string functions, the default is the +length of the longest argument. For integer functions, the default is 21 +digits. For real functions, the default is 13 plus the number of decimals +indicated by @code{initid->decimals}. (For numeric functions, the length +includes any sign or decimal point characters.) -If you have problems with the fact that anyone can delete the -@strong{MySQL} communication socket @file{/tmp/mysql.sock}, you can, -on most versions of Unix, protect your @file{/tmp} file system by setting -the @code{sticky} bit on it. Log in as @code{root} and do the following: +@item char *ptr +A pointer that the function can use for its own purposes. For example, +functions can use @code{initid->ptr} to communicate allocated memory +between functions. In @code{xxx_init()}, allocate the memory and assign it +to this pointer: @example -shell> chmod +t /tmp +initid->ptr = allocated_memory; @end example -This will protect your @file{/tmp} file system so that files can be deleted -only by their owners or the superuser (@code{root}). - -You can check if the @code{sticky} bit is set by executing @code{ls -ld /tmp}. -If the last permission bit is @code{t}, the bit is set. - -@cindex starting, @code{mysqld} -@cindex @code{mysqld}, starting -@node Changing MySQL user, Resetting permissions, Problems with mysql.sock, Problems -@section How to Run MySQL As a Normal User +In @code{xxx()} and @code{xxx_deinit()}, refer to @code{initid->ptr} to use +or deallocate the memory. +@end table -The @strong{MySQL} server @code{mysqld} can be started and run by any user. -In order to change @code{mysqld} to run as a Unix user @code{user_name}, you must -do the following: +@cindex argument processing +@cindex processing, arguments +@node UDF arguments, UDF return values, UDF calling sequences, Adding UDF +@subsection Argument Processing -@enumerate -@item -Stop the server if it's running (use @code{mysqladmin shutdown}). +The @code{args} parameter points to a @code{UDF_ARGS} structure that thas the +members listed below: -@item -Change the database directories and files so that @code{user_name} has -privileges to read and write files in them (you may need to do this as -the Unix @code{root} user): +@table @code +@item unsigned int arg_count +The number of arguments. Check this value in the initialization function +if you want your function to be called with a particular number of arguments. +For example: @example -shell> chown -R user_name /path/to/mysql/datadir +if (args->arg_count != 2) +@{ + strcpy(message,"XXX() requires two arguments"); + return 1; +@} @end example -If directories or files within the @strong{MySQL} data directory are -symlinks, you'll also need to follow those links and change the directories -and files they point to. @code{chown -R} may not follow symlinks for -you. -@item -Start the server as user @code{user_name}, or, if you are using -@strong{MySQL} Version 3.22 or later, start @code{mysqld} as the Unix @code{root} -user and use the @code{--user=user_name} option. @code{mysqld} will switch -to run as the Unix user @code{user_name} before accepting any connections. +@item enum Item_result *arg_type +The types for each argument. The possible type values are +@code{STRING_RESULT}, @code{INT_RESULT}, and @code{REAL_RESULT}. -@item -To start the server as the given user name automatically at system -startup time, add a @code{user} line that specifies the user name to -the @code{[mysqld]} group of the @file{/etc/my.cnf} option file or the -@file{my.cnf} option file in the server's data directory. For example: +To make sure that arguments are of a given type and return an +error if they are not, check the @code{arg_type} array in the initialization +function. For example: @example -[mysqld] -user=user_name +if (args->arg_type[0] != STRING_RESULT || + args->arg_type[1] != INT_RESULT) +@{ + strcpy(message,"XXX() requires a string and an integer"); + return 1; +@} @end example -@end enumerate -At this point, your @code{mysqld} process should be running fine and dandy as -the Unix user @code{user_name}. One thing hasn't changed, though: the -contents of the permissions tables. By default (right after running the -permissions table install script @code{mysql_install_db}), the @strong{MySQL} -user @code{root} is the only user with permission to access the @code{mysql} -database or to create or drop databases. Unless you have changed those -permissions, they still hold. This shouldn't stop you from accessing -@strong{MySQL} as the @strong{MySQL} @code{root} user when you're logged in -as a Unix user other than @code{root}; just specify the @code{-u root} option -to the client program. +As an alternative to requiring your function's arguments to be of particular +types, you can use the initialization function to set the @code{arg_type} +elements to the types you want. This causes @strong{MySQL} to coerce +arguments to those types for each call to @code{xxx()}. For example, to +specify coercion of the first two arguments to string and integer, do this in +@code{xxx_init()}: -Note that accessing @strong{MySQL} as @code{root}, by supplying @code{-u -root} on the command line, has @emph{nothing} to do with @strong{MySQL} running -as the Unix @code{root} user, or, indeed, as another Unix user. The access -permissions and user names of @strong{MySQL} are completely separate from -Unix user names. The only connection with Unix user names is that if you -don't provide a @code{-u} option when you invoke a client program, the client -will try to connect using your Unix login name as your @strong{MySQL} user -name. +@example +args->arg_type[0] = STRING_RESULT; +args->arg_type[1] = INT_RESULT; +@end example -If your Unix box itself isn't secured, you should probably at least put a -password on the @strong{MySQL} @code{root} users in the access tables. -Otherwise, any user with an account on that machine can run @code{mysql -u -root db_name} and do whatever he likes. +@item char **args +@code{args->args} communicates information to the initialization function +about the general nature of the arguments your function was called with. For a +constant argument @code{i}, @code{args->args[i]} points to the argument +value. (See below for instructions on how to access the value properly.) +For a non-constant argument, @code{args->args[i]} is @code{0}. +A constant argument is an expression that uses only constants, such as +@code{3} or @code{4*7-2} or @code{SIN(3.14)}. A non-constant argument is an +expression that refers to values that may change from row to row, such as +column names or functions that are called with non-constant arguments. -@cindex passwords, forgotten -@cindex passwords, resetting -@cindex root user, password resetting -@node Resetting permissions, File permissions , Changing MySQL user, Problems -@section How to Reset a Forgotten Password +For each invocation of the main function, @code{args->args} contains the +actual arguments that are passed for the row currently being processed. -If you have forgotten the @code{root} user password for @strong{MySQL}, you -can restore it with the following procedure: +Functions can refer to an argument @code{i} as follows: -@enumerate +@itemize @bullet @item -Take down the @code{mysqld} server by sending a @code{kill} (not @code{kill --9}) to the @code{mysqld} server. The pid is stored in a @code{.pid} -file, which is normally in the @strong{MySQL} database directory: +An argument of type @code{STRING_RESULT} is given as a string pointer plus a +length, to allow handling of binary data or data of arbitrary length. The +string contents are available as @code{args->args[i]} and the string length +is @code{args->lengths[i]}. You should not assume that strings are +null-terminated. + +@item +For an argument of type @code{INT_RESULT}, you must cast +@code{args->args[i]} to a @code{long long} value: @example -kill `cat /mysql-data-directory/hostname.pid` +long long int_val; +int_val = *((long long*) args->args[i]); @end example -You must be either the Unix @code{root} user or the same user the server -runs as to do this. - -@item -Restart @code{mysqld} with the @code{--skip-grant-tables} option. -@item -Connect to the @code{mysqld} server with @code{mysql -h hostname mysql} and change -the password with a @code{GRANT} command. @xref{GRANT,,@code{GRANT}}. -You can also do this with -@code{mysqladmin -h hostname -u user password 'new password'} @item -Load the privilege tables with: @code{mysqladmin -h hostname -flush-privileges} or with the SQL command @code{FLUSH PRIVILEGES}. -@end enumerate +For an argument of type @code{REAL_RESULT}, you must cast +@code{args->args[i]} to a @code{double} value: -Note that after you started @code{mysqld} with @code{--skip-grant-tables}, -any usage of @code{GRANT} commands will give you an @code{Unknown command} -error until you have executed @code{FLUSH PRIVILEGES}. +@example +double real_val; +real_val = *((double*) args->args[i]); +@end example +@end itemize -@cindex files, permissions -@cindex error mesaages, can't find file -@cindex files, not found message -@node File permissions , Not enough file handles, Resetting permissions, Problems -@section Problems with File Permissions +@item unsigned long *lengths +For the initialization function, the @code{lengths} array indicates the +maximum string length for each argument. For each invocation of the main +function, @code{lengths} contains the actual lengths of any string arguments +that are passed for the row currently being processed. For arguments of +types @code{INT_RESULT} or @code{REAL_RESULT}, @code{lengths} still contains +the maximum length of the argument (as for the initialization function). +@end table -If you have problems with file permissions, for example, if @code{mysql} -issues the following error message when you create a table: +@cindex UDFs, return values +@cindex return values, UDFs +@cindex errors, handling for UDFs +@cindex handling, errors +@node UDF return values, UDF compiling, UDF arguments, Adding UDF +@subsection Return Values and Error Handling + +The initialization function should return @code{0} if no error occurred and +@code{1} otherwise. If an error occurs, @code{xxx_init()} should store a +null-terminated error message in the @code{message} parameter. The message +will be returned to the client. The message buffer is +@code{MYSQL_ERRMSG_SIZE} characters long, but you should try to keep the +message to less than 80 characters so that it fits the width of a standard +terminal screen. + +The return value of the main function @code{xxx()} is the function value, for +@code{long long} and @code{double} functions. A string functions should +return a pointer to the result and store the length of the string in the +@code{length} arguments. @code{result} is a buffer at least 255 bytes long. +Set these to the contents and length of the return value. For example: @example -ERROR: Can't find file: 'path/with/filename.frm' (Errcode: 13) +memcpy(result, "result string", 13); +*length = 13; @end example -@tindex UMASK environment variable -@tindex Environment variable, UMASK -then the environment variable @code{UMASK} might be set incorrectly when -@code{mysqld} starts up. The default umask value is @code{0660}. You can -change this behavior by starting @code{safe_mysqld} as follows: +If your string functions that needs to return a string longer than 255 +bytes, you must allocate the space for it with @code{malloc()} in your +@code{xxx_init()} function or your @code{xxx()} function and free it in +your @code{xxx_deinit()} function. You can store the allocated memory +in the @code{ptr} slot in the @code{UDF_INIT} structure for reuse by +future @code{xxx()} calls. @xref{UDF calling sequences}. + +To indicate a return value of @code{NULL} in the main function, set +@code{is_null} to @code{1}: @example -shell> UMASK=384 # = 600 in octal -shell> export UMASK -shell> /path/to/safe_mysqld & +*is_null = 1; @end example -@tindex UMASK_DIR environment variable -@tindex Environment variable, UMASK_DIR -By default @strong{MySQL} will create database and @code{RAID} -directories with permission type 0700. You can modify this behavior by -setting the @code{UMASK_DIR} variable. If you set this, new -directories are created with the combined @code{UMASK} and -@code{UMASK_DIR}. For example, if you want to give group access to -all new directories, you can do: +To indicate an error return in the main function, set the @code{error} +parameter to @code{1}: @example -shell> UMASK_DIR=504 # = 770 in octal -shell> export UMASK_DIR -shell> /path/to/safe_mysqld & +*error = 1; @end example -In @strong{MySQL} Version 3.23.25 and above, @strong{MySQL} assumes that the -value for @code{UMASK} and @code{UMASK_DIR} is in octal if it starts -with a zero. - -@xref{Environment variables}. - -@node Not enough file handles, Using DATE, File permissions , Problems -@section File Not Found - -If you get @code{ERROR '...' not found (errno: 23)}, @code{Can't open -file: ... (errno: 24)}, or any other error with @code{errno 23} or -@code{errno 24} from @strong{MySQL}, it means that you haven't allocated -enough file descriptors for @strong{MySQL}. You can use the -@code{perror} utility to get a description of what the error number -means: +If @code{xxx()} sets @code{*error} to @code{1} for any row, the function +value is @code{NULL} for the current row and for any subsequent rows +processed by the statement in which @code{XXX()} was invoked. (@code{xxx()} +will not even be called for subsequent rows.) @strong{NOTE:} In +@strong{MySQL} versions prior to 3.22.10, you should set both @code{*error} +and @code{*is_null}: @example -shell> perror 23 -File table overflow -shell> perror 24 -Too many open files -shell> perror 11 -Resource temporarily unavailable +*error = 1; +*is_null = 1; @end example -The problem here is that @code{mysqld} is trying to keep open too many -files simultaneously. You can either tell @code{mysqld} not to open so -many files at once or increase the number of file descriptors -available to @code{mysqld}. - -To tell @code{mysqld} to keep open fewer files at a time, you can make -the table cache smaller by using the @code{-O table_cache=32} option to -@code{safe_mysqld} (the default value is 64). Reducing the value of -@code{max_connections} will also reduce the number of open files (the -default value is 90). +@cindex compiling, user-defined functions +@cindex UDFs, compiling +@cindex installing, user-defined functions +@node UDF compiling, , UDF return values, Adding UDF +@subsection Compiling and Installing User-definable Functions -@tindex ulimit -To change the number of file descriptors available to @code{mysqld}, you -can use the option @code{--open-files-limit=#} to @code{safe_mysqld} or -@code{-O open-files-limit=#} to @code{mysqld}. @xref{SHOW VARIABLES}. -The easiest way to do that is to add the option to your option file. -@xref{Option files}. If you have an old @code{mysqld} version that -doesn't support this, you can edit the @code{safe_mysqld} script. There -is a commented-out line @code{ulimit -n 256} in the script. You can -remove the @code{'#'} character to uncomment this line, and change the -number 256 to affect the number of file descriptors available to -@code{mysqld}. +Files implementing UDFs must be compiled and installed on the host where the +server runs. This process is described below for the example UDF file +@file{udf_example.cc} that is included in the @strong{MySQL} source +distribution. This file contains the following functions: -@code{ulimit} (and @code{open-files-limit}) can increase the number of -file descriptors, but only up to the limit imposed by the operating -system. There is also a 'hard' limit that can only be overrided if you -start @code{safe_mysqld} or @code{mysqld} as root (Just remember that -you need to also use the @code{--user=..} option in this case). If you -need to increase the OS limit on the number of file descriptors -available to each process, consult the documentation for your operating -system. +@itemize @bullet +@item +@code{metaphon()} returns a metaphon string of the string argument. +This is something like a soundex string, but it's more tuned for English. +@item +@code{myfunc_double()} returns the sum of the ASCII values of the +characters in its arguments, divided by the sum of the length of its arguments. +@item +@code{myfunc_int()} returns the sum of the length of its arguments. +@item +@code{sequence([const int])} returns an sequence starting from the given +number or 1 if no number has been given. +@item +@code{lookup()} returns the IP number for a hostname. +@item +@code{reverse_lookup()} returns the hostname for an IP number. +The function may be called with a string @code{"xxx.xxx.xxx.xxx"} or +four numbers. +@end itemize -Note that if you run the @code{tcsh} shell, @code{ulimit} will not work! -@code{tcsh} will also report incorrect values when you ask for the current -limits! In this case you should start @code{safe_mysqld} with @code{sh}! +A dynamically loadable file should be compiled as a sharable object file, +using a command something like this: -@findex DATE -@cindex DATE columns, problems -@cindex problems, @code{DATE} columns -@node Using DATE, Timezone problems, Not enough file handles, Problems -@section Problems Using @code{DATE} Columns +@example +shell> gcc -shared -o udf_example.so myfunc.cc +@end example -The format of a @code{DATE} value is @code{'YYYY-MM-DD'}. According to ANSI -SQL, no other format is allowed. You should use this format in @code{UPDATE} -expressions and in the WHERE clause of @code{SELECT} statements. For -example: +You can easily find out the correct compiler options for your system by +running this command in the @file{sql} directory of your @strong{MySQL} +source tree: @example -mysql> SELECT * FROM tbl_name WHERE date >= '1997-05-05'; +shell> make udf_example.o @end example -As a convenience, @strong{MySQL} automatically converts a date to a number if -the date is used in a numeric context (and vice versa). It is also smart -enough to allow a ``relaxed'' string form when updating and in a @code{WHERE} -clause that compares a date to a @code{TIMESTAMP}, @code{DATE}, or a -@code{DATETIME} column. (Relaxed form means that any punctuation character -may be used as the separator between parts. For example, @code{'1998-08-15'} -and @code{'1998#08#15'} are equivalent.) @strong{MySQL} can also convert a -string containing no separators (such as @code{'19980815'}), provided it -makes sense as a date. +You should run a compile command similar to the one that @code{make} displays, +except that you should remove the @code{-c} option near the end of the line +and add @code{-o udf_example.so} to the end of the line. (On some systems, +you may need to leave the @code{-c} on the command.) -The special date @code{'0000-00-00'} can be stored and retrieved as -@code{'0000-00-00'.} When using a @code{'0000-00-00'} date through -@strong{MyODBC}, it will automatically be converted to @code{NULL} in -@strong{MyODBC} Version 2.50.12 and above, because ODBC can't handle this kind of -date. +Once you compile a shared object containing UDFs, you must install it +and tell @strong{MySQL} about it. Compiling a shared object from +@file{udf_example.cc} produces a file named something like +@file{udf_example.so} (the exact name may vary from platform to platform). +Copy this file to some directory searched by @code{ld}, such as +@file{/usr/lib}. On many systems, you can set the @code{LD_LIBRARY} or +@code{LD_LIBRARY_PATH} environment variable to point at the directory where +you have your UDF function files. The @code{dlopen} manual page tells you +which variable you should use on your system. You should set this in +@code{mysql.server} or @code{safe_mysqld} and restart @code{mysqld}. -Because @strong{MySQL} performs the conversions described above, the following -statements work: +After the library is installed, notify @code{mysqld} about the new +functions with these commands: @example -mysql> INSERT INTO tbl_name (idate) VALUES (19970505); -mysql> INSERT INTO tbl_name (idate) VALUES ('19970505'); -mysql> INSERT INTO tbl_name (idate) VALUES ('97-05-05'); -mysql> INSERT INTO tbl_name (idate) VALUES ('1997.05.05'); -mysql> INSERT INTO tbl_name (idate) VALUES ('1997 05 05'); -mysql> INSERT INTO tbl_name (idate) VALUES ('0000-00-00'); - -mysql> SELECT idate FROM tbl_name WHERE idate >= '1997-05-05'; -mysql> SELECT idate FROM tbl_name WHERE idate >= 19970505; -mysql> SELECT mod(idate,100) FROM tbl_name WHERE idate >= 19970505; -mysql> SELECT idate FROM tbl_name WHERE idate >= '19970505'; +mysql> CREATE FUNCTION metaphon RETURNS STRING SONAME "udf_example.so"; +mysql> CREATE FUNCTION myfunc_double RETURNS REAL SONAME "udf_example.so"; +mysql> CREATE FUNCTION myfunc_int RETURNS INTEGER SONAME "udf_example.so"; +mysql> CREATE FUNCTION lookup RETURNS STRING SONAME "udf_example.so"; +mysql> CREATE FUNCTION reverse_lookup RETURNS STRING SONAME "udf_example.so"; @end example -@noindent -However, the following will not work: +Functions can be deleted using @code{DROP FUNCTION}: @example -mysql> SELECT idate FROM tbl_name WHERE STRCMP(idate,'19970505')=0; +mysql> DROP FUNCTION metaphon; +mysql> DROP FUNCTION myfunc_double; +mysql> DROP FUNCTION myfunc_int; +mysql> DROP FUNCTION lookup; +mysql> DROP FUNCTION reverse_lookup; @end example -@code{STRCMP()} is a string function, so it converts @code{idate} to -a string and performs a string comparison. It does not convert -@code{'19970505'} to a date and perform a date comparison. - -Note that @strong{MySQL} does no checking whether or not the date is -correct. If you store an incorrect date, such as @code{'1998-2-31'}, the -wrong date will be stored. If the date cannot be converted to any reasonable -value, a @code{0} is stored in the @code{DATE} field. This is mainly a speed -issue and we think it is up to the application to check the dates, and not -the server. - -@cindex timezone problems -@cindex problems, timezone -@tindex TZ environment variable -@tindex Environment variable, TZ -@node Timezone problems, Case sensitivity, Using DATE, Problems -@section Time Zone Problems - -If you have a problem with @code{SELECT NOW()} returning values in GMT and -not your local time, you have to set the @code{TZ} environment variable to -your current time zone. This should be done for the environment in which -the server runs, for example, in @code{safe_mysqld} or @code{mysql.server}. -@xref{Environment variables}. - -@cindex case sensitivity, in searches -@cindex searching, and case-sensitivity -@cindex Chinese -@cindex Big5 Chinese character encoding -@node Case sensitivity, Problems with NULL, Timezone problems, Problems -@section Case Sensitivity in Searches - -By default, @strong{MySQL} searches are case-insensitive (although there are -some character sets that are never case insensitive, such as @code{czech}). -That means that if you search with @code{col_name LIKE 'a%'}, you will get all -column values that start with @code{A} or @code{a}. If you want to make this -search case-sensitive, use something like @code{INDEX(col_name, "A")=0} to -check a prefix. Or use @code{STRCMP(col_name, "A") = 0} if the column value -must be exactly @code{"A"}. +The @code{CREATE FUNCTION} and @code{DROP FUNCTION} statements update the +system table @code{func} in the @code{mysql} database. The function's name, +type and shared library name are saved in the table. You must have the +@strong{insert} and @strong{delete} privileges for the @code{mysql} database +to create and drop functions. -Simple comparison operations (@code{>=, >, = , < , <=}, sorting and -grouping) are based on each character's ``sort value''. Characters with -the same sort value (like E, e and é) are treated as the same character! +You should not use @code{CREATE FUNCTION} to add a function that has already +been created. If you need to reinstall a function, you should remove it with +@code{DROP FUNCTION} and then reinstall it with @code{CREATE FUNCTION}. You +would need to do this, for example, if you recompile a new version of your +function, so that @code{mysqld} gets the new version. Otherwise the server +will continue to use the old version. -In older @strong{MySQL} versions @code{LIKE} comparisons where done on -the uppercase value of each character (E == e but E <> é). In newer -@strong{MySQL} versions @code{LIKE} works just like the other comparison -operators. +Active functions are reloaded each time the server starts, unless you start +@code{mysqld} with the @code{--skip-grant-tables} option. In this case, UDF +initialization is skipped and UDFs are unavailable. (An active function is +one that has been loaded with @code{CREATE FUNCTION} and not removed with +@code{DROP FUNCTION}.) -If you want a column always to be treated in case-sensitive fashion, -declare it as @code{BINARY}. @xref{CREATE TABLE, , @code{CREATE TABLE}}. +@cindex adding, native functions +@cindex native functions, adding +@cindex functions, native, adding +@node Adding native function, , Adding UDF, Adding functions +@section Adding a New Native Function -If you are using Chinese data in the so-called big5 encoding, you want to -make all character columns @code{BINARY}. This works because the sorting -order of big5 encoding characters is based on the order of ASCII codes. +The procedure for adding a new native function is described below. Note +that you cannot add native functions to a binary distribution because +the procedure involves modifying @strong{MySQL} source code. You must +compile @strong{MySQL} yourself from a source distribution. Also note +that if you migrate to another version of @strong{MySQL} (for example, +when a new version is released), you will need to repeat the procedure +with the new version. -@cindex @code{NULL} values, vs. empty values -@tindex NULL -@node Problems with NULL, Problems with alias, Case sensitivity, Problems -@section Problems with @code{NULL} Values +To add a new native @strong{MySQL} function, follow these steps: -The concept of the @code{NULL} value is a common source of confusion for -newcomers to SQL, who often think that @code{NULL} is the same thing as an -empty string @code{''}. This is not the case! For example, the following -statements are completely different: +@enumerate +@item +Add one line to @file{lex.h} that defines the function name in the +@code{sql_functions[]} array. +@item +If the function prototype is simple (just takes zero, one, two or three +arguments), you should in lex.h specify SYM(FUNC_ARG#) (where # is the +number of arguments) as the second argument in the +@code{sql_functions[]} array and add a function that creates a function +object in @file{item_create.cc}. Take a look at @code{"ABS"} and +@code{create_funcs_abs()} for an example of this. +If the function prototype is complicated (for example takes a variable number +of arguments), you should add two lines to @file{sql_yacc.yy}. One +indicates the preprocessor symbol that @code{yacc} should define (this +should be added at the beginning of the file). Then define the function +parameters and add an ``item'' with these parameters to the +@code{simple_expr} parsing rule. For an example, check all occurrences +of @code{ATAN} in @file{sql_yacc.yy} to see how this is done. +@item +In @file{item_func.h}, declare a class inheriting from @code{Item_num_func} or +@code{Item_str_func}, depending on whether your function returns a number or a +string. +@item +In @file{item_func.cc}, add one of the following declarations, depending +on whether you are defining a numeric or string function: @example -mysql> INSERT INTO my_table (phone) VALUES (NULL); -mysql> INSERT INTO my_table (phone) VALUES (""); +double Item_func_newname::val() +longlong Item_func_newname::val_int() +String *Item_func_newname::Str(String *str) @end example -Both statements insert a value into the @code{phone} column, but the first -inserts a @code{NULL} value and the second inserts an empty string. The -meaning of the first can be regarded as ``phone number is not known'' and the -meaning of the second can be regarded as ``she has no phone''. - -In SQL, the @code{NULL} value is always false in comparison to any -other value, even @code{NULL}. An expression that contains @code{NULL} -always produces a @code{NULL} value unless otherwise indicated in -the documentation for the operators and functions involved in the -expression. All columns in the following example return @code{NULL}: +If you inherit your object from any of the standard items (like +@code{Item_num_func} you probably only have to define one of the above +functions and let the parent object take care of the other functions. +For example, the @code{Item_str_func} class defines a @code{val()} function +that executes @code{atof()} on the value returned by @code{::str()}. +@item +You should probably also define the following object function: @example -mysql> SELECT NULL,1+NULL,CONCAT('Invisible',NULL); +void Item_func_newname::fix_length_and_dec() @end example +This function should at least calculate @code{max_length} based on the +given arguments. @code{max_length} is the maximum number of characters +the function may return. This function should also set @code{maybe_null += 0} if the main function can't return a @code{NULL} value. The +function can check if any of the function arguments can return +@code{NULL} by checking the arguments @code{maybe_null} variable. You +can take a look at @code{Item_func_mod::fix_length_and_dec} for a +typical example of how to do this. +@end enumerate -If you want to search for column values that are @code{NULL}, you -cannot use the @code{=NULL} test. The following statement returns no -rows, because @code{expr = NULL} is FALSE, for any expression: +All functions must be thread safe (In other words, don't use any global or +static variables in the functions without protecting them with mutexes). -@example -mysql> SELECT * FROM my_table WHERE phone = NULL; -@end example +If you want to return @code{NULL}, from @code{::val()}, @code{::val_int()} +or @code{::str()} you should set @code{null_value} to 1 and return 0. -To look for @code{NULL} values, you must use the @code{IS NULL} test. -The following shows how to find the @code{NULL} phone number and the -empty phone number: +For @code{::str()} object functions, there are some additional +considerations to be aware of: -@example -mysql> SELECT * FROM my_table WHERE phone IS NULL; -mysql> SELECT * FROM my_table WHERE phone = ""; -@end example +@itemize @bullet +@item +The @code{String *str} argument provides a string buffer that may be +used to hold the result. (For more information about the @code{String} type, +take a look at the @file{sql_string.h} file.) +@item +The @code{::str()} function should return the string that holds the result or +@code{(char*) 0} if the result is @code{NULL}. +@item +All current string functions try to avoid allocating any memory unless +absolutely necessary! +@end itemize -In @strong{MySQL}, as in many other SQL servers, you can't index -columns that can have @code{NULL} values. You must declare such columns -@code{NOT NULL}. Conversely, you cannot insert @code{NULL} into an indexed -column. +@cindex procedures, adding +@cindex adding, procedures +@cindex new procedures, adding +@node Adding procedures, ODBC, Adding functions, Top +@chapter Adding New Procedures to MySQL -@findex LOAD DATA INFILE -When reading data with @code{LOAD DATA INFILE}, empty columns are updated -with @code{''}. If you want a @code{NULL} value in a column, you should use -@code{\N} in the text file. The literal word @code{'NULL'} may also be used -under some circumstances. -@xref{LOAD DATA, , @code{LOAD DATA}}. +In @strong{MySQL}, you can define a procedure in C++ that can access and +modify the data in a query before it is sent to the client. The modification +can be done on row-by-row or @code{GROUP BY} level. -When using @code{ORDER BY}, @code{NULL} values are presented first. If you -sort in descending order using @code{DESC}, @code{NULL} values are presented -last. When using @code{GROUP BY}, all @code{NULL} values are regarded as -equal. +We have created an example procedure in @strong{MySQL} Version 3.23 to +show you what can be done. -To help with @code{NULL} handling, you can use the @code{IS NULL} and -@code{IS NOT NULL} operators and the @code{IFNULL()} function. +Additionally we recommend you to take a look at 'mylua', which you can find in the Contrib directory. @xref{Contrib}. Which this you can use the LUA +language to load a procedure at runtime into @code{mysqld}. -@cindex @code{TIMESTAMP}, and @code{NULL} values -@cindex @code{AUTO_INCREMENT}, and @code{NULL} values -@cindex @code{NULL} values, and @code{TIMESTAMP} columns -@cindex @code{NULL} values, and @code{AUTO_INCREMENT} columns -For some column types, @code{NULL} values are handled specially. If you -insert @code{NULL} into the first @code{TIMESTAMP} column of a table, the -current date and time is inserted. If you insert @code{NULL} into an -@code{AUTO_INCREMENT} column, the next number in the sequence is inserted. +@menu +* procedure analyse:: Procedure analyse +* Writing a procedure:: Writing a procedure. +@end menu -@tindex alias -@node Problems with alias, Deleting from related tables, Problems with NULL, Problems -@section Problems with @code{alias} +@node procedure analyse, Writing a procedure, Adding procedures, Adding procedures +@section Procedure Analyse -You can use an alias to refer to a column in the @code{GROUP BY}, -@code{ORDER BY}, or in the @code{HAVING} part. Aliases can also be used -to give columns better names: +@code{analyse([max elements,[max memory]])} -@example -SELECT SQRT(a*b) as rt FROM table_name GROUP BY rt HAVING rt > 0; -SELECT id,COUNT(*) AS cnt FROM table_name GROUP BY id HAVING cnt > 0; -SELECT id AS "Customer identity" FROM table_name; -@end example +This procedure is defined in the @file{sql/sql_analyse.cc}. This +examines the result from your query and returns an analysis of the +results: -Note that ANSI SQL doesn't allow you to refer to an alias in a -@code{WHERE} clause. This is because when the @code{WHERE} code is -executed the column value may not yet be determined. For example, the -following query is @strong{illegal}: +@itemize @bullet +@item +@code{max elements} (default 256) is the maximum number of distinct values +@code{analyse} will notice per column. This is used by @code{analyse} to check if +the optimal column type should be of type @code{ENUM}. +@item +@code{max memory} (default 8192) is the maximum memory @code{analyse} should +allocate per column while trying to find all distinct values. +@end itemize @example -SELECT id,COUNT(*) AS cnt FROM table_name WHERE cnt > 0 GROUP BY id; +SELECT ... FROM ... WHERE ... PROCEDURE ANALYSE([max elements,[max memory]]) @end example -The @code{WHERE} statement is executed to determine which rows should -be included in the @code{GROUP BY} part while @code{HAVING} is used to -decide which rows from the result set should be used. +@node Writing a procedure, , procedure analyse, Adding procedures +@section Writing a Procedure -@cindex deleting, rows -@cindex rows, deleting -@cindex tables, deleting rows -@node Deleting from related tables, No matching rows, Problems with alias, Problems -@section Deleting Rows from Related Tables +For the moment, the only documentation for this is the source. -As @strong{MySQL} doesn't support sub-selects or use of more than one table -in the @code{DELETE} statement, you should use the following approach to -delete rows from 2 related tables: +You can find all information about procedures by examining the following files: -@enumerate -@item -@code{SELECT} the rows based on some @code{WHERE} condition in the main table. -@item -@code{DELETE} the rows in the main table based on the same condition. -@item -@code{DELETE FROM related_table WHERE related_column IN (selected_rows)}. -@end enumerate +@itemize @bullet +@item @file{sql/sql_analyse.cc} +@item @file{sql/procedure.h} +@item @file{sql/procedure.cc} +@item @file{sql/sql_select.cc} +@end itemize -If the total number of characters in the query with -@code{related_column} is more than 1,048,576 (the default value of -@code{max_allowed_packet}, you should split it into smaller parts and -execute multiple @code{DELETE} statements. You will probably get the -fastest @code{DELETE} by only deleting 100-1000 @code{related_column} -id's per query if the @code{related_column} is an index. If the -@code{related_column} isn't an index, the speed is independent of the -number of arguments in the @code{IN} clause. +@cindex ODBC +@cindex Windows +@cindex MyODBC +@node ODBC, Common programs, Adding procedures, Top +@chapter MySQL ODBC Support -@cindex no matching rows -@cindex rows, matching problems -@node No matching rows, ALTER TABLE problems, Deleting from related tables, Problems -@section Solving Problems with No Matching Rows +@menu +* Installing MyODBC:: How to install MyODBC +* ODBC administrator:: How to fill in the various fields in the ODBC administrator program +* MyODBC connect parameters:: +* ODBC Problems:: How to report problems with @strong{MySQL} ODBC +* MyODBC clients:: Programs known to work with @strong{MyODBC} +* ODBC and last_insert_id:: How to get the value of an @code{AUTO_INCREMENT} column in ODBC +* MyODBC bug report:: Reporting problems with MyODBC +@end menu -If you have a complicated query that has many tables and that doesn't -return any rows, you should use the following procedure to find out what -is wrong with your query: +@strong{MySQL} provides support for ODBC by means of the @strong{MyODBC} +program. This chapter will teach you how to install @strong{MyODBC}, +and how to use it. Here, you will also find a list of common programs that +are known to work with @strong{MyODBC}. -@enumerate -@item -Test the query with @code{EXPLAIN} and check if you can find something that is -obviously wrong. @xref{EXPLAIN, , @code{EXPLAIN}}. +@node Installing MyODBC, ODBC administrator, ODBC, ODBC +@section How To Install MyODBC -@item -Select only those fields that are used in the @code{WHERE} clause. +@strong{MyODBC} is a 32-bit ODBC (2.50) level 0 (with level 1 and level +2 features) driver for connecting an ODBC-aware application to +@strong{MySQL}. @strong{MyODBC} works on Windows95, Windows98, NT, and +on most Unix platforms. -@item -Remove one table at a time from the query until it returns some rows. -If the tables are big, it's a good idea to use @code{LIMIT 10} with the query. +@strong{MyODBC} is in public domain, and you can find the newest version +at @uref{http://www.mysql.com/downloads/api-myodbc.html}. -@item -Do a @code{SELECT} for the column that should have matched a row against -the table that was last removed from the query. +If you have problem with @strong{MyODBC} and your program also works +with OLEDB, you should try the OLEDB driver that you can find in the +Contrib section. @xref{Contrib}. -@item -If you are comparing @code{FLOAT} or @code{DOUBLE} columns with numbers that -have decimals, you can't use @code{=}! This problem is common in most -computer languages because floating-point values are not exact values: +Normally you only need to install @strong{MyODBC} on Windows machines. +You only need @strong{MyODBC} for Unix if you have a program like +ColdFusion that is running on the Unix machine and uses ODBC to connect +to the databases. -@example -mysql> SELECT * FROM table_name WHERE float_column=3.5; - -> -mysql> SELECT * FROM table_name WHERE float_column between 3.45 and 3.55; -@end example +If you want to install @strong{MyODBC} on a Unix box, you will also need +an @strong{ODBC} manager. @strong{MyODBC} is known to work with +most of the Unix ODBC managers. You can find a list at these in the +@strong{ODBC}-related links section on the @strong{MySQL} useful links page. +@xref{Useful Links}. -In most cases, changing the @code{FLOAT} to a @code{DOUBLE} will fix this! - -@item -If you still can't figure out what's wrong, create a minimal test that can -be run with @code{mysql test < query.sql} that shows your problems. -You can create a test file with @code{mysqldump --quick database tables > query.sql}. Open the file in an editor, remove some insert lines (if there are -too many of these), and add your select statement at the end of the file. +To install @strong{MyODBC} on Windows, you should download the +appropriate @strong{MyODBC} .zip file (for Windows or NT/Win2000), +unpack it with @code{WINZIP}, or some similar program, and execute the +@code{SETUP.EXE} file. -Test that you still have your problem by doing: +On Windows/NT you may get the following error when trying to install +@strong{MyODBC}: @example -shell> mysqladmin create test2 -shell> mysql test2 < query.sql +An error occurred while copying C:\WINDOWS\SYSTEM\MFC30.DLL. Restart +Windows and try installing again (before running any applications which +use ODBC) @end example -Post the test file using @code{mysqlbug} to @email{mysql@@lists.mysql.com}. -@end enumerate +The problem in this case is that some other program is using ODBC and +because of how Windows is designed, you may not in this case be able to +install a new ODBC drivers with Microsoft's ODBC setup program. In most +cases you can continue by just pressing @code{Ignore} to copy the rest +of the MyODBC files and the final installation should still work. If +this doesn't work, the solution is to reboot your computer in ``safe +mode`` (Choose this by pressing F8 just before your machine starts +Windows during rebooting), install @strong{MyODBC}, and reboot to normal +mode. -@tindex ALTER TABLE -@node ALTER TABLE problems, Change column order, No matching rows, Problems -@section Problems with @code{ALTER TABLE}. +@itemize @bullet +@item +To make a connection to a Unix box from a Windows box, with an ODBC +application (one that doesn't support @strong{MySQL} natively), you must +first install @strong{MyODBC} on the Windows machine. +@item +The user and Windows machine must have the access privileges to the +@strong{MySQL} server on the Unix machine. This is set up with the +@code{GRANT} command. @xref{GRANT,,@code{GRANT}}. +@item +You must create an ODBC DSN entry as follows: -@code{ALTER TABLE} changes a table to the current character set. -If you during @code{ALTER TABLE} get a duplicate key error, then the cause -is either that the new character sets maps to keys to the same value -or that the table is corrupted, in which case you should run -@code{REPAIR TABLE} on the table. +@itemize @minus +@item +Open the Control Panel on the Windows machine. +@item +Double-click the ODBC Data Sources 32 bits icon. +@item +Click the tab User DSN. +@item +Click the button Add. +@item +Select @strong{MySQL} in the screen Create New Data Source and click +the Finish button. +@item +The @strong{MySQL} Driver default configuration screen is shown. +@xref{ODBC administrator}. +@end itemize -If @code{ALTER TABLE} dies with an error like this: +@item +Now start your application and select the ODBC driver with the DSN you +specified in the ODBC administrator. +@end itemize -@example -Error on rename of './database/name.frm' to './database/B-a.frm' (Errcode: 17) -@end example +Notice that there are other configuration options on the screen of +@strong{MySQL} (trace, don't prompt on connect, etc) that you can try if +you run into problems. -the problem may be that @strong{MySQL} has crashed in a previous @code{ALTER -TABLE} and there is an old table named @file{A-something} or -@file{B-something} lying around. In this case, go to the @strong{MySQL} data -directory and delete all files that have names starting with @code{A-} or -@code{B-}. (You may want to move them elsewhere instead of deleting them.) +@cindex ODBC, administrator +@node ODBC administrator, MyODBC connect parameters, Installing MyODBC, ODBC +@section How to Fill in the Various Fields in the ODBC Administrator Program -@code{ALTER TABLE} works the following way: +There are three possibilities for specifying the server name on +Windows95: @itemize @bullet -@item Create a new table named @file{A-xxx} with the requested changes. -@item All rows from the old table are copied to @file{A-xxx}. -@item The old table is renamed @file{B-xxx}. -@item @file{A-xxx} is renamed to your old table name. -@item @file{B-xxx} is deleted. -@end itemize - -If something goes wrong with the renaming operation, @strong{MySQL} tries to -undo the changes. If something goes seriously wrong (this shouldn't happen, -of course), @strong{MySQL} may leave the old table as @file{B-xxx}, but a -simple rename on the system level should get your data back. - -@cindex reordering, columns -@cindex columns, changing -@cindex changing, column order -@cindex tables, changing column order -@node Change column order, Temporary table problems, ALTER TABLE problems, Problems -@section How To Change the Order of Columns in a Table - -The whole point of SQL is to abstract the application from the data -storage format. You should always specify the order in which you wish to -retrieve your data. For example: +@item +Use the IP address of the server. +@item +Add a file @file{\windows\lmhosts} with the following information: @example -SELECT col_name1, col_name2, col_name3 FROM tbl_name; +ip hostname @end example -will return columns in the order @code{col_name1}, @code{col_name2}, @code{col_name3}, whereas: +For example: @example -SELECT col_name1, col_name3, col_name2 FROM tbl_name; +194.216.84.21 my_hostname @end example -will return columns in the order @code{col_name1}, @code{col_name3}, @code{col_name2}. - -You should @strong{NEVER}, in an application, use @code{SELECT *} and -retrieve the columns based on their position, because the order in which -columns are returned @strong{CANNOT} be guaranteed over time. A simple -change to your database may cause your application to fail rather -dramatically. - -If you want to change the order of columns anyway, you can do it as follows: - -@enumerate -@item -Create a new table with the columns in the right order. -@item -Execute -@code{INSERT INTO new_table SELECT fields-in-new_table-order FROM old_table}. -@item -Drop or rename @code{old_table}. -@item -@code{ALTER TABLE new_table RENAME old_table}. -@end enumerate - -@cindex temporary tables, problems -@node Temporary table problems, , Change column order, Problems -@section TEMPORARY TABLE problems - -The following are a list of the limitations with @code{TEMPORARY TABLES}. - -@itemize @bullet -@item -A temporary table can only be of type @code{HEAP}, @code{ISAM} or -@code{MyISAM}. @item -You can't use temporary tables more than once in the same query. -For example, the following doesn't work. +Configure the PC to use DNS. +@end itemize +Example of how to fill in the @code{ODBC setup}: @example -select * from temporary_table, temporary_table as t2; +Windows DSN name: test +Description: This is my test database +MySql Database: test +Server: 194.216.84.21 +User: monty +Password: my_password +Port: @end example -We plan to fix the above in 4.0. -@item -You can't use @code{RENAME} on a @code{TEMPORARY} table. -Note that @code{ALTER TABLE org_name RENAME new_name} works! - -We plan to fix the above in 4.0. -@end itemize - -@cindex problems, solving -@cindex solving, problems -@cindex databases, replicating -@node Common problems, Log files, Problems, Top -@chapter Solving Some Common Problems with MySQL +The value for the @code{Windows DSN name} field is any name that is unique +in your Windows ODBC setup. -@cindex replication -@menu -* Log Replication:: Database replication with update log -* Backup:: Database backups -* Multiple servers:: Running multiple @strong{MySQL} servers on the same machine -@end menu +You don't have to specify values for the @code{Server}, @code{User}, +@code{Password}, or @code{Port} fields in the ODBC setup screen. +However, if you do, the values will be used as the defaults later when +you attempt to make a connection. You have the option of changing the +values at that time. -In this chapter, you will find information to solve some of the more common -tasks with @strong{MySQL}. This includes making backups, running more than -one @strong{MySQL} server daemon on a single machine, and replicating a -database using the update or binary logs. +If the port number is not given, the default port (@value{default_port}) +is used. -@cindex database replication -@cindex replication, database -@node Log Replication, Backup, Common problems, Common problems -@section Database Replication with Update Log +If you specify the option @code{Read options from C:\my.cnf}, the groups +@code{client} and @code{odbc} will be read from the @file{C:\my.cnf} file. +You can use all options that are usable by @code{mysql_options()}. +@xref{mysql_options, , @code{mysql_options}}. -Now that master-slave internal replication is available starting in -Version 3.23.15, using the update log to implement replications is not -recommended. @xref{Replication}. +@node MyODBC connect parameters, ODBC Problems, ODBC administrator, ODBC +@section Connect parameters for MyODBC -However, it is still possible to replicate a database by using the -update log or the binary log. @xref{Update log}. This requires one -database that acts as a master (to which data changes are made) and one -or more other databases that act as slaves. To update a slave, just run -@code{mysql < update_log.*} or @code{mysqlbinlog binary_log.* | mysql}. -Supply host, user, and password options that are appropriate for the -slave database, and use the update log from the master database as -input. +One can specify the following parameters for @strong{MyODBC} on +the @code{[Servername]} section of an @code{ODBC.INI} file or +through the @code{InConnectionString} argument in the +@code{SQLDriverConnect()} call. -If you never delete anything from a table, you can use a @code{TIMESTAMP} -column to find out which rows have been inserted or changed in the table -since the last replication (by comparing the time when you did the -replication last time) and only copy these rows to the mirror. +@multitable @columnfractions .2 .2 .6 +@item @strong{Parameter} @tab @strong{Default value} @tab @strong{Comment} +@item user @tab ODBC (on Windows) @tab The username used to connect to @strong{MySQL}. +@item server @tab localhost @tab The hostname of the @strong{MySQL} server. +@item database @tab @tab The default database +@item option @tab 0 @tab A integer by which you can specify how @strong{MyODBC} should work. See below. +@item port @tab 3306 @tab The TCP/IP port to use if @code{server} is not @code{localhost}. +@item stmt @tab @tab A statement that will be executed when connection to @code{MySQL}. +@item password @tab @tab The password for the @code{server} @code{user} combination. +@item socket @tab @tab The socket or Windows pipe to connect to. +@end multitable -It is possible to make a two-way updating system using both the update -log (for deletes) and timestamps (on both sides). But in that case you -must be able to handle conflicts when the same data have been changed in -both ends. You probably want to keep the old version to help with -deciding what has been updated. +The option argument is used to tell @strong{MyODBC} that the client isn't 100% +ODBC compliant. On Windows, one normally sets the option flag by +toggling the different options on the connection screen but one can also +set this in the opton argument. The following options are listed in the +same order as they appear in the @strong{MyODBC} connect screen: -Because replication in this case is done with SQL statements, you should not -use the following functions in statements that update the database; they may -not return the same value as in the original database: +@multitable @columnfractions .1 .9 +@item @strong{Bit} @tab @strong{Meaning} +@item 1 @tab The client can't handle that @strong{MyODBC} returns the real width of a column. +@item 2 @tab The client can't handle that MySQL returns the true value of affected rows. If this flag is set then MySQL returns 'found rows' instead. One must have MySQL 3.21.14 or newer to get this to work. +@item 4 @tab Make a debug log in c:\myodbc.log. This is the same as putting @code{MYSQL_DEBUG=d:t:O,c::\myodbc.log} in @file{AUTOEXEC.BAT} +@item 8 @tab Don't set any packet limit for results and parameters. +@item 16 @tab Don't prompt for questions even if driver would like to prompt +@item 32 @tab Simulate a ODBC 1.0 driver in some context. +@item 64 @tab Ignore use of database name in 'database.table.column'. +@item 128 @tab Force use of ODBC manager cursors (experimental). +@item 256 @tab Disable the use of extended fetch (experimental) +@item 512 @tab Pad CHAR fields to full column length. +@item 1024 @tab SQLDescribeCol() will return fully qualifed column names +@item 2048 @tab Use the compressed server/client protocol +@item 4096 @tab Tell server to ignore space after function name and before @code{'('} (needed by PowerBuilder). This will make all function names keywords! +@item 8192 @tab Connect with named pipes to a @code{mysqld} server running on NT. +@item 16384 @tab Change LONGLONG columns to INT columns (Some applications can't handle LONGLONG). +@item 32768 @tab Return 'user' as Table_qualifier and Table_owner from SQLTables (experimental) +@item 65536 @tab Read parameters from the @code{client} and @code{odbc} groups from @file{my.cnf} +@item 131072 @tab Add some extra safety checks (should not bee needed but...) +@end multitable -@itemize @bullet -@item @code{DATABASE()} -@item @code{GET_LOCK()} and @code{RELEASE_LOCK()} -@item @code{RAND()} -@item @code{USER()}, @code{SYSTEM_USER()} or @code{SESSION_USER()} -@item @code{VERSION()}, @code{CONNECT_ID()} -@end itemize +If you want to have many options, you should add the above flags! For +example setting option to 12 (4+8) gives you debugging without package +limits! -All time functions are safe to use, as the timestamp is sent to the -mirror if needed. @code{LAST_INSERT_ID()} is also safe to use. +The default @file{MYODBC.DLL} is compiled for optimal performance. If +you want to to debug @strong{MyODBC} (for example to enable tracing), +you should instead use @code{MYODBCD.DLL}. To install this file, copy +@file{MYODBCD.DLL} over the installed @code{MYODBC.DLL} file. -@cindex databases, backups -@cindex backups -@node Backup, Multiple servers, Log Replication, Common problems -@section Database Backups +@node ODBC Problems, MyODBC clients, MyODBC connect parameters, ODBC +@section How to Report Problems with MyODBC -Because @strong{MySQL} tables are stored as files, it is easy to do a -backup. To get a consistent backup, do a @code{LOCK TABLES} on the -relevant tables followed by @code{FLUSH TABLES} for the tables. -@xref{LOCK TABLES, , @code{LOCK TABLES}}. -@xref{FLUSH, , @code{FLUSH}}. -You only need a read lock; this allows other threads to continue to -query the tables while you are making a copy of the files in the -database directory. The @code{FLUSH TABLE} is needed to ensure that -the all active index pages is written to disk before you start the backup. +@strong{MyODBC} has been tested with Access, Admndemo.exe, C++-Builder, +Borland Builder 4, Centura Team Developer (formerly Gupta SQL/Windows), +ColdFusion (on Solaris and NT with svc pack 5), Crystal Reports, +DataJunction, Delphi, ERwin, Excel, iHTML, FileMaker Pro, FoxPro, Notes +4.5/4.6, SBSS, Perl DBD-ODBC, Paradox, Powerbuilder, Powerdesigner 32 +bit, VC++, and Visual Basic. -If you want to make a SQL level backup of a table, you can use -@code{SELECT INTO OUTFILE} or @code{BACKUP TABLE}. @xref{SELECT}. -@xref{BACKUP TABLE}. +If you know of any other applications that work with @strong{MyODBC}, please +send mail to @email{myodbc@@lists.mysql.com} about this! -Another way to back up a database is to use the @code{mysqldump} program or -the @code{mysqlhotcopy script}. @xref{mysqldump, , @code{mysqldump}}. -@xref{mysqlhotcopy, , @code{mysqlhotcopy}}. +With some programs you may get an error like: +@code{Another user has modifies the record that you have modified}. In most +cases this can be solved by doing one of the following things: -@enumerate +@itemize @bullet @item -Do a full backup of your databases: +Add a primary key for the table if there isn't one already. +@item +Add a timestamp column if there isn't one already. +@item +Only use double float fields. Some programs may fail when they compare +single floats. +@end itemize -@example -shell> mysqldump --tab=/path/to/some/dir --opt --full +If the above doesn't help, you should do a @code{MyODBC} trace file and +try to figure out why things go wrong. -or +@node MyODBC clients, ODBC and last_insert_id, ODBC Problems, ODBC +@section Programs Known to Work with MyODBC -shell> mysqlhotcopy database /path/to/some/dir -@end example +Most programs should work with @strong{MyODBC}, but for each of those +listed below, we have tested it ourselves or received confirmation from +some user that it works: -You can also simply copy all table files (@file{*.frm}, @file{*.MYD}, and -@file{*.MYI} files) as long as the server isn't updating anything. -The script @code{mysqlhotcopy} does use this method. +@table @asis +@item @strong{Program} +@strong{Comment} +@cindex Access program +@item Access +To make Access work: +@itemize @bullet @item -@cindex log files, names -Stop @code{mysqld} if it's running, then start it with the -@code{--log-update[=file_name]} option. @xref{Update log}. The update -log file(s) provide you with the information you need to replicate -changes to the database that are made subsequent to the point at which -you executed @code{mysqldump}. -@end enumerate - -If you have to restore something, try to recover your tables using -@code{REPAIR TABLE} or @code{myisamchk -r} first. That should work in -99.9% of all cases. If @code{myisamchk} fails, try the following -procedure: (This will only work if you have started @strong{MySQL} with -@code{--log-update}. @xref{Update log}.): +If you are using Access 2000, you should get and install the newest +Microsoft MDAC (@code{Microsoft Data Access Components}) from +@uref{http://www.microsoft.com/data}. This will fix the following bug +in Access: when you export data to @strong{MySQL}, the table and column +names aren't specified. Another way to around this bug is to upgrade to +MyODBC Version 2.50.33 and @strong{MySQL} Version 3.23.x, which together +provide a workaround for this bug! -@enumerate +Note that if you are using @strong{MySQL} Version 3.22, you must to apply the +MDAC patch and use MyODBC 2.50.32 or 2.50.34 and above to go around +this problem. @item -Restore the original @code{mysqldump} backup. +Set the ``Return matching rows'' MyODBC option field when connecting to +@strong{MySQL}. @item -Execute the following command to re-run the updates in the binary log: - -@example -shell> mysqlbinlog hostname-bin.[0-9]* | mysql -@end example - -If you are using the update log you can use: - -@example -shell> ls -1 -t -r hostname.[0-9]* | xargs cat | mysql -@end example -@end enumerate - -@code{ls} is used to get all the update log files in the right order. - -You can also do selective backups with @code{SELECT * INTO OUTFILE 'file_name' -FROM tbl_name} and restore with @code{LOAD DATA INFILE 'file_name' REPLACE -...} To avoid duplicate records, you need a @code{PRIMARY KEY} or a -@code{UNIQUE} key in the table. The @code{REPLACE} keyword causes old records -to be replaced with new ones when a new record duplicates an old record on -a unique key value. - -If you get performance problems in making backups on your system, you can -solve this by setting up replication and do the backups on the slave -instead of on the master. @xref{Replication Intro}. - -If you are using a Veritas file system, you can do: +You should have a primary key in the table. If not, new or updated rows +may show up as @code{#Deleted#}. +@item +You should have a timestamp in all tables you want to be able to update. +For maximum portability @code{TIMESTAMP(14)} or simple @code{TIMESTAMP} +is recommended instead of other @code{TIMESTAMP(X)} variations. +@item +Only use double float fields. Access fails when comparing with single floats. +The symptom usually is that new or updated rows may show up as @code{#Deleted#} +or that you can't find or update rows. +@item +If you still get the error @code{Another user has changed your data} after +adding a @code{TIMESTAMP} column, the following trick may help you: -@enumerate -@item -Execute in a client (perl ?) @code{FLUSH TABLES WITH READ LOCK} +Don't use @code{table} data sheet view. Create instead a form with the +fields you want, and use that @code{form} data sheet view. You should +set the @code{DefaultValue} property for the @code{TIMESTAMP} column to +@code{NOW()}. It may be a good idea to hide the @code{TIMESTAMP} column +from view so your users are not confused. @item -Fork a shell or execute in another client @code{mount vxfs snapshot}. +Access on NT will report @code{BLOB} columns as @code{OLE OBJECTS}. If +you want to have @code{MEMO} columns instead, you should change the +column to @code{TEXT} with @code{ALTER TABLE}. @item -Execute in the first client @code{UNLOCK TABLES} +Access can't always handle @code{DATE} columns properly. If you have a problem +with these, change the columns to @code{DATETIME}. @item -Copy files from snapshot +In some cases, Access may generate illegal SQL queries that +@strong{MySQL} can't understand. You can fix this by selecting +@code{"Query|SQLSpecific|Pass-Through"} from the Access menu. @item -Unmount snapshot -@end enumerate - -@cindex multiple servers -@cindex servers, multiple -@cindex running, multiple servers -@node Multiple servers, , Backup, Common problems -@section Running Multiple MySQL Servers on the Same Machine - -There are circumstances when you might want to run multiple servers on the same -machine. For example, you might want to test a new @strong{MySQL} release -while leaving your existing production setup undisturbed. Or you might -be an Internet service provider that wants to provide independent -@strong{MySQL} installations for different customers. - -If you want to run multiple servers, the easiest way is to compile the servers -with different TCP/IP ports and socket files so they are not -both listening to the same TCP/IP port or socket file. @xref{mysqld_multi, , -@code{mysqld_multi}}. +If you have in Access a column defined as BYTE, Access will try to export this +as @code{TINYINT} instead of @code{TINYINT UNSIGNED}. This will give you +problems if you have values > 127 in the column! +@item +If you are using Access 7.0, You should use the option flag @code{Return +matching rows}. +@item +If you are using Access 2.0, You should use the option flags @code{Return +matching rows} and @code{Simulate ODBC 1.0}. +@end itemize -Assume an existing server is configured for the default port number and -socket file. Then configure the new server with a @code{configure} command -something like this: +@cindex ADO program +@item ADO +When you are coding with the ADO API and @strong{MyODBC} you need to put +attention in some default properties that aren't supported by the +@strong{MySQL} server. For example, using the @code{CursorLocation +Property} as @code{adUseServer} will return for the @code{RecordCount +Property} a result of -1. To have the right value, you need to set this +property to @code{adUseClient}, like is showing in the VB code below: @example -shell> ./configure --with-tcp-port=port_number \ - --with-unix-socket-path=file_name \ - --prefix=/usr/local/mysql-3.22.9 -@end example - -Here @code{port_number} and @code{file_name} should be different than the -default port number and socket file pathname, and the @code{--prefix} value -should specify an installation directory different than the one under which -the existing @strong{MySQL} installation is located. +Dim myconn As New ADODB.Connection +Dim myrs As New Recordset +Dim mySQL As String +Dim myrows As Long -You can check the socket used by any currently executing @strong{MySQL} server -with this command: +myconn.Open "DSN=MyODBCsample" +mySQL = "SELECT * from user" +myrs.Source = mySQL +Set myrs.ActiveConnection = myconn +myrs.CursorLocation = adUseClient +myrs.Open +myrows = myrs.RecordCount -@example -shell> mysqladmin -h hostname --port=port_number variables +myrs.Close +myconn.Close @end example -Note that if you specify ``@code{localhost}'' as a hostname, @code{mysqladmin} -will default to using Unix sockets instead of TCP/IP. +Another workaround is to use a @code{SELECT COUNT(*)} statement +for a similar query to get the correct row count. -If you have a @strong{MySQL} server running on the port you used, you will -get a list of some of the most important configurable variables in -@strong{MySQL}, including the socket name. +@item Active server pages (ASP) +You should use the option flag @code{Return matching rows}. -You don't have to recompile a new @strong{MySQL} server just to start with -a different port and socket. You can change the port and socket to be used -by specifying them at run time as options to @code{safe_mysqld}: - -@example -shell> /path/to/safe_mysqld --socket=file_name --port=port_number -@end example - -@code{mysqld_multi} can also take @code{safe_mysqld} (or @code{mysqld}) -as an argument and pass the options from a configuration file to -@code{safe_mysqld} and further to @code{mysqld}. - -If you run the new server on the same database directory as another -server with logging enabled, you should also specify the name of the log -files to @code{safe_mysqld} with @code{--log}, @code{--log-update}, or -@code{--log-slow-queries}. Otherwise, both servers may be trying to -write to the same log file. +@item BDE applications +To get these to work, you should set the option flags +@code{Don't optimize column widths} and @code{Return matching rows}. -@strong{WARNING}: Normally you should never have two servers that update -data in the same database! If your OS doesn't support fault-free system -locking, this may lead to unpleasant surprises! +@cindex Borland Buidler 4 program +@item Borland Builder 4 +When you start a query you can use the property @code{Active} or use the +method @code{Open}. Note that @code{Active} will start by automatically +issuing a @code{SELECT * FROM ...} query that may not be a good thing if +your tables are big! +@item ColdFusion (On Unix) +The following information is taken from the ColdFusion documentation: -If you want to use another database directory for the second server, you -can use the @code{--datadir=path} option to @code{safe_mysqld}. +Use the following information to configure ColdFusion Server for Linux +to use the unixODBC driver with @strong{MyODBC} for @strong{MySQL} data +sources. Allaire has verified that @strong{MyODBC} Version 2.50.26 +works with @strong{MySQL} Version 3.22.27 and ColdFusion for Linux. (Any +newer version should also work.) You can download @strong{MyODBC} at +@uref{http://www.mysql.com/downloads/api-myodbc.html} -@strong{NOTE} also that starting several @strong{MySQL} servers -(@code{mysqlds}) in different machines and letting them access one data -directory over @code{NFS} is generally a @strong{BAD IDEA}! The problem -is that the @code{NFS} will become the bottleneck with the speed. It is -not meant for such use. And last but not least, you would still have to -come up with a solution how to make sure that two or more @code{mysqlds} -are not interfering with each other. At the moment there is no platform -that would 100% reliable do the file locking (@code{lockd} daemon -usually) in every situation. Yet there would be one more possible risk -with @code{NFS}; it would make the work even more complicated for -@code{lockd} daemon to handle. So make it easy for your self and forget -about the idea. The working solution is to have one computer with an -operating system that efficiently handles threads and have several CPUs -in it. +@cindex ColdFusion program +ColdFusion Version 4.5.1 allows you to us the ColdFusion Administrator +to add the @strong{MySQL} data source. However, the driver is not +included with ColdFusion Version 4.5.1. Before the @strong{MySQL} driver +will appear in the ODBC datasources drop-down list, you must build and +copy the @strong{MyODBC} driver to +@file{/opt/coldfusion/lib/libmyodbc.so}. -When you want to connect to a @strong{MySQL} server that is running with -a different port than the port that is compiled into your client, you -can use one of the following methods: +The Contrib directory contains the program mydsn-xxx.zip which allows +you to build and remove the DSN registry file for the MyODBC driver +on Coldfusion applications. +@cindex DataJunction +@item DataJunction +You have to change it to output @code{VARCHAR} rather than @code{ENUM}, as +it exports the latter in a manner that causes @strong{MySQL} grief. +@cindex Excel +@item Excel +Works. Some tips: @itemize @bullet @item -Start the client with @code{--host 'hostname' --port=port_number} to connect -with TCP/IP, or @code{[--host localhost] --socket=file_name} to connect via -a Unix socket. - -@item -In your C or Perl programs, you can give the port or socket arguments -when connecting to the @strong{MySQL} server. - -@item -If your are using the Perl @code{DBD::mysql} module you can read the options -from the @strong{MySQL} option files. @xref{Option files}. - +If you have problems with dates, try to select them as strings using the +@code{CONCAT()} function. For example: @example -$dsn = "DBI:mysql:test;mysql_read_default_group=client;mysql_read_default_file=/usr/local/mysql/data/my.cnf" -$dbh = DBI->connect($dsn, $user, $password); +select CONCAT(rise_time), CONCAT(set_time) + from sunrise_sunset; @end example +Values retrieved as strings this way should be correctly recognized +as time values by Excel97. -@item -@tindex MYSQL_UNIX_PORT environment variable -@tindex MYSQL_TCP_PORT environment variable -@tindex environment variable, MYSQL_UNIX_PORT -@tindex environment variable, MYSQL_TCP_PORT -Set the @code{MYSQL_UNIX_PORT} and @code{MYSQL_TCP_PORT} environment variables -to point to the Unix socket and TCP/IP port before you start your clients. -If you normally use a specific socket or port, you should place commands -to set these environment variables in your @file{.login} file. -@xref{Environment variables}. @xref{Programs}. +The purpose of @code{CONCAT()} in this example is to fool ODBC into thinking +the column is of ``string type''. Without the @code{CONCAT()}, ODBC knows the +column is of time type, and Excel does not understand that. -@item -@tindex .my.cnf file -Specify the default socket and TCP/IP port in the @file{.my.cnf} file in your -home directory. @xref{Option files}. +Note that this is a bug in Excel, because it automatically converts a +string to a time. This would be great if the source was a text file, but +is plain stupid when the source is an ODBC connection that reports +exact types for each column. @end itemize +@cindex Word program +@item Word -@cindex Log files -@node Log files, Clients, Common problems, Top -@chapter The MySQL log files - -@strong{MySQL} has several different log files that can help you find -out what's going on inside @code{mysqld}: - -@multitable @columnfractions .3 .7 -@item The error log @tab Problems encountering starting, running or stopping @code{mysqld}. -@item The isam log @tab Logs all changes to the ISAM tables. Used only for debugging the isam code. -@item The query log @tab Established connections and executed queries. -@item The update log @tab Deprecated: Stores all statements that changes data -@item The binary log @tab Stores all statements that changes something. Used also for replication -@item The slow log @tab Stores all queries that took more than @code{long_query_time} to execute or didn't use indexes. -@end multitable +To retrieve data from @strong{MySQL} to Word/Excel documents, you need to +use the @code{MyODBC} driver and the Add-in Microsoft Query help. -All logs can be found in the @code{mysqld} data directory. You can -force @code{mysqld} to reopen the log files (or in some cases -switch to a new log) by executing @code{FLUSH LOGS}. @xref{FLUSH}. +For example, create a db with a table containing 2 columns of text: -@cindex error log -@cindex files, error log -@menu -* Error log:: -* Query log:: -* Update log:: -* Binary log:: -* Slow query log:: -@end menu +@itemize @bullet +@item +Insert rows using the @code{mysql} client command-line tool. +@item +Create a DSN file using the MyODBC driver, for example, my for the db above. +@item +Open the Word application. +@item +Create a blank new documentation. +@item +Using the tool bar called Database, press the button insert database. +@item +Press the button Get Data. +@item +At the right hand of the screen Get Data, press the button Ms Query. +@item +In the Ms Query create a New Data Source using the DSN file my. +@item +Select the new query. +@item +Select the columns that you want. +@item +Make a filter if you want. +@item +Make a Sort if you want. +@item +Select Return Data to Microsoft Word. +@item +Click Finish. +@item +Click Insert data and select the records. +@item +Click OK and you see the rows in your Word document. +@end itemize -@node Error log, Query log, Log files, Log files -@section The Error Log +@cindex odbcadmin program +@item odbcadmin +Test program for ODBC. +@cindex Delphi program +@item Delphi +You must use BDE Version 3.2 or newer. Set the `Don't optimize column width' +option field when connecting to @strong{MySQL}. -@code{mysqld} writes all errors to the stderr, which the -@code{safe_mysqld} script redirects to a file called -@code{'hostname'.err}. (On Windows, @code{mysqld} writes this directly -to @file{\mysql\data\mysql.err}). +Also, here is some potentially useful Delphi code that sets up both an +ODBC entry and a BDE entry for @strong{MyODBC} (the BDE entry requires a BDE +Alias Editor that is free at a Delphi Super Page near +you. (Thanks to Bryan Brunton @email{bryan@@flesherfab.com} for this): -This contains information indicating when @code{mysqld} was started and -stopped and also any critical errors found when running. If @code{mysqld} -dies unexpectedly and @code{safe_mysqld} needs to restart @code{mysqld}, -@code{safe_mysqld} will write a @code{restarted mysqld} row in this -file. This log also holds a warning if @code{mysqld} notices a table -that needs to be automatically checked or repaired. +@example +fReg:= TRegistry.Create; + fReg.OpenKey('\Software\ODBC\ODBC.INI\DocumentsFab', True); + fReg.WriteString('Database', 'Documents'); + fReg.WriteString('Description', ' '); + fReg.WriteString('Driver', 'C:\WINNT\System32\myodbc.dll'); + fReg.WriteString('Flag', '1'); + fReg.WriteString('Password', ''); + fReg.WriteString('Port', ' '); + fReg.WriteString('Server', 'xmark'); + fReg.WriteString('User', 'winuser'); + fReg.OpenKey('\Software\ODBC\ODBC.INI\ODBC Data Sources', True); + fReg.WriteString('DocumentsFab', 'MySQL'); + fReg.CloseKey; + fReg.Free; -On some operating systems, the error log will contain a stack trace -for where @code{mysqld} died. This can be used to find out where -@code{mysqld} died. @xref{Using stack trace}. + Memo1.Lines.Add('DATABASE NAME='); + Memo1.Lines.Add('USER NAME='); + Memo1.Lines.Add('ODBC DSN=DocumentsFab'); + Memo1.Lines.Add('OPEN MODE=READ/WRITE'); + Memo1.Lines.Add('BATCH COUNT=200'); + Memo1.Lines.Add('LANGDRIVER='); + Memo1.Lines.Add('MAX ROWS=-1'); + Memo1.Lines.Add('SCHEMA CACHE DIR='); + Memo1.Lines.Add('SCHEMA CACHE SIZE=8'); + Memo1.Lines.Add('SCHEMA CACHE TIME=-1'); + Memo1.Lines.Add('SQLPASSTHRU MODE=SHARED AUTOCOMMIT'); + Memo1.Lines.Add('SQLQRYMODE='); + Memo1.Lines.Add('ENABLE SCHEMA CACHE=FALSE'); + Memo1.Lines.Add('ENABLE BCD=FALSE'); + Memo1.Lines.Add('ROWSET SIZE=20'); + Memo1.Lines.Add('BLOBS TO CACHE=64'); + Memo1.Lines.Add('BLOB SIZE=32'); -@cindex query log -@cindex files, query log -@node Query log, Update log, Error log, Log files -@section The Query Log + AliasEditor.Add('DocumentsFab','MySQL',Memo1.Lines); +@end example -If you want to know what happens within @code{mysqld}, you should start -it with @code{--log[=file]}. This will log all connections and queries -to the log file (by default named @file{'hostname'.log}). This log can -be very useful when you suspect an error in a client and want to know -exactly what @code{mysqld} thought the client sent to it. +@cindex C++Builder +@item C++Builder +Tested with BDE Version 3.0. The only known problem is that when the table +schema changes, query fields are not updated. BDE, however, does not seem +to recognize primary keys, only the index PRIMARY, though this has not +been a problem. -By default, the @code{mysql.server} script starts the @strong{MySQL} -server with the @code{-l} option. If you need better performance when -you start using @strong{MySQL} in a production environment, you can -remove the @code{-l} option from @code{mysql.server} or change it to -@code{--log-binary}. +@item Vision +You should use the option flag @code{Return matching rows}. -The entries in this log are written as @code{mysqld} receives the questions. -This may be different than the order in which the statements are executed. -This is in contrast to the update log and the binary log which are written -after the query is executed, but before any locks are released. +@cindex Visual Basic +@item Visual Basic +To be able to update a table, you must define a primary key for the table. -@cindex update log -@cindex files, update log -@node Update log, Binary log, Query log, Log files -@section The Update Log +Visual Basic with ADO can't handle big integers. This means that some queries +like @code{SHOW PROCESSLIST} will not work properly. The fix is to set +add the option @code{OPTION=16834} in the ODBC connect string or set +the @code{Change BIGINT columns to INT} option in the MyODBC connect screen. +You may also want to set the @code{Return matching rows} option. -@strong{NOTE}: The update log is replaced by the binary -log. @xref{Binary log}. With this you can do anything that you can do -with the update log. +@item VisualInterDev +If you get the error @code{[Microsoft][ODBC Driver Manager] Driver does +not support this parameter} the reason may be that you have a +@code{BIGINT} in your result. Try setting the @code{Change BIGINT +columns to INT} option in the MyODBC connect screen. -When started with the @code{--log-update[=file_name]} option, -@code{mysqld} writes a log file containing all SQL commands that update -data. If no filename is given, it defaults to the name of the host -machine. If a filename is given, but it doesn't contain a path, the file -is written in the data directory. If @file{file_name} doesn't have an -extension, @code{mysqld} will create log file names like so: -@file{file_name.###}, where @code{###} is a number that is incremented each -time you execute @code{mysqladmin refresh}, execute @code{mysqladmin -flush-logs}, execute the @code{FLUSH LOGS} statement, or restart the server. +@item Visual Objects +You should use the option flag @code{Don't optimize column widths}. +@end table -@strong{NOTE:} For the above scheme to work, you should NOT create -your own files with the same filename as the update log + some extensions -that may be regarded as a number, in the directory used by the update log! +@cindex AUTO-INCREMENT, ODBC +@node ODBC and last_insert_id, MyODBC bug report, MyODBC clients, ODBC +@section How to Get the Value of an @code{AUTO_INCREMENT} Column in ODBC -If you use the @code{--log} or @code{-l} options, @code{mysqld} writes a -general log with a filename of @file{hostname.log}, and restarts and -refreshes do not cause a new log file to be generated (although it is closed -and reopened). In this case you can copy it (on Unix) by doing: +A common problem is how to get the value of an automatically generated ID +from an @code{INSERT}. With ODBC, you can do something like this (assuming +that @code{auto} is an @code{AUTO_INCREMENT} field): @example -mv hostname.log hostname-old.log -mysqladmin flush-logs -cp hostname-old.log to-backup-directory -rm hostname-old.log +INSERT INTO foo (auto,text) VALUES(NULL,'text'); +SELECT LAST_INSERT_ID(); @end example -Update logging is smart because it logs only statements that really update -data. So an @code{UPDATE} or a @code{DELETE} with a @code{WHERE} that finds no -rows is not written to the log. It even skips @code{UPDATE} statements that -set a column to the value it already has. +Or, if you are just going to insert the ID into another table, you can do this: -The update logging is done immediately after a query completes but before -any locks are released or any commit is done. This ensures that the log -will be logged in the execution order. +@example +INSERT INTO foo (auto,text) VALUES(NULL,'text'); +INSERT INTO foo2 (id,text) VALUES(LAST_INSERT_ID(),'text'); +@end example -If you want to update a database from update log files, you could do the -following (assuming your update logs have names of the form -@file{file_name.###}): +@xref{Getting unique ID}. +For the benefit of some ODBC applications (at least Delphi and Access), +the following query can be used to find a newly inserted row: @example -shell> ls -1 -t -r file_name.[0-9]* | xargs cat | mysql +SELECT * FROM tbl_name WHERE auto IS NULL; @end example -@code{ls} is used to get all the log files in the right order. +@cindex reporting, MyODBC problems +@cindex problems, ODBC +@cindex MyODBC, reporting problems +@node MyODBC bug report, , ODBC and last_insert_id, ODBC +@section Reporting Problems with MyODBC -This can be useful if you have to revert to backup files after a crash -and you want to redo the updates that occurred between the time of the backup -and the crash. +If you encounter difficulties with @strong{MyODBC}, you should start by +making a log file from the ODBC manager (the log you get when requesting +logs from ODBCADMIN) and a @strong{MyODBC} log. -@cindex binary log -@cindex files, binary log -@node Binary log, Slow query log, Update log, Log files -@section The Binary Log +To get a @strong{MyODBC} log, you need to do the following: -In the future the binary log will replace the update log, so we -recommend you to switch to this log format as soon as possible! +@enumerate +@item +Ensure that you are using @code{myodbcd.dll} and not @code{myodbc.dll}. +The easiest way to do this is to get @code{myodbcd.dll} from the MyODBC +distribution and copy it over the @code{myodbc.dll}, which is probably +in your @code{C:\windows\system32} or @code{C:\winnt\system32} directory. -The binary log contains all information that is available in the update -log in a more efficient format. It also contains information about how long -every query that updated the database took. +Note that you probably want to restore the old myodbc.dll file when you +have finished testing, as this is a lot faster than @code{myodbcd.dll}. +@item +Tag the `Trace MyODBC' option flag in the @strong{MyODBC} connect/configure +screen. The log will be written to file @file{C:\myodbc.log}. -The binary log is also used when you are replicating a slave from a master. -@xref{Replication}. +If the trace option is not remembered when you are going back to the +above screen, it means that you are not using the @code{myodbcd.dll} +driver (see above). +@item +Start your application and try to get it to fail. +@end enumerate -When started with the @code{--log-bin[=file_name]} option, @code{mysqld} -writes a log file containing all SQL commands that update data. If no -file name is given, it defaults to the name of the host machine followed -by @code{-bin}. If file name is given, but it doesn't contain a path, the -file is written in the data directory. +Check the @code{MyODBC trace file}, to find out what could be wrong. +You should be able to find out the issued queries by searching after +the string @code{>mysql_real_query} in the @file{myodbc.log} file. -You can use the following options to @code{mysqld} to affect what is logged -to the binary log: +You should also try duplicating the queries in the @code{mysql} monitor +or @code{admndemo} to find out if the error is MyODBC or @strong{MySQL}. -@multitable @columnfractions .4 .6 -@item @code{binlog-do-db=database_name} @tab -Tells the master it should log updates for the specified database, and -exclude all others not explicitly mentioned. -(Example: @code{binlog-do-db=some_database}) +If you find out something is wrong, please only send the relevant rows +(max 40 rows) to @email{myodbc@@lists.mysql.com}. Please never +send the whole MyODBC or ODBC log file! -@item @code{binlog-ignore-db=database_name} @tab -Tells the master that updates to the given database should not be logged -to the binary log (Example: @code{binlog-ignore-db=some_database}) -@end multitable +If you are unable to find out what's wrong, the last option is to +make an archive (tar or zip) that contains a MyODBC trace file, the ODBC +log file, and a README file that explains the problem. You can send this +to @uref{ftp://support.mysql.com/pub/mysql/secret}. Only we at +@strong{MySQL AB} will have access to the files you upload, and we will +be very discrete with the data! -To the binary log filename @code{mysqld} will append an extension that is a -number that is incremented each time you execute @code{mysqladmin -refresh}, execute @code{mysqladmin flush-logs}, execute the @code{FLUSH LOGS} -statement or restart the server. +If you can create a program that also shows this problem, please +upload this too! -To be able to know which different binary log files have been used, -@code{mysqld} will also create a binary log index file that -contains the name of all used binary log files. By default this has the -same name as the binary log file, with the extension @code{'.index'}. -You can change the name of the binary log index file with the -@code{--log-bin-index=[filename]} option. +If the program works with some other SQL server, you should make an ODBC log +file where you do exactly the same thing in the other SQL server. -If you are using replication, you should not delete old binary log -files until you are sure that no slave will ever need to use them. -One way to do this is to do @code{mysqladmin flush-logs} once a day and then -remove any logs that are more than 3 days old. +Remember that the more information you can supply to us, the more +likely it is that we can fix the problem! -You can examine the binary log file with the @code{mysqlbinlog} command. -For example, you can update a @strong{MySQL} server from the binary log -as follows: +@node Common programs, Problems, ODBC, Top +@chapter Using MySQL with Some Common Programs -@example -mysqlbinlog log-file | mysql -h server_name -@end example +@menu +* Apache:: Using @strong{MySQL} with Apache +* Borland C++:: +@end menu -You can also use the @code{mysqlbinlog} program to read the binary log -directly from a remote @strong{MySQL} server! +This chapter describes how to use @strong{MySQL} with some common programs. -@code{mysqlbinlog --help} will give you more information of how to use -this program! +In this chapter you will: -If you are using @code{BEGIN [WORK]} or @code{SET AUTOCOMMIT=0}, you must -use the @strong{MySQL} binary log for backups instead of the old update log. +@itemize @bullet +@item +Learn how to easily store your Apache log files in a @strong{MySQL} +database. +@item +Find some tips on how to compile @strong{MySQL} and @strong{MySQL}-based +programs using Borland C++. +@end itemize -The binary logging is done immediately after a query completes but before -any locks are released or any commit is done. This ensures that the log -will be logged in the execution order. +@cindex Apache +@node Apache, Borland C++, Common programs, Common programs +@section Using MySQL with Apache -All updates (@code{UPDATE}, @code{DELETE} or @code{INSERT}) that change -a transactional table (like BDB tables) are cached until a @code{COMMIT}. -Any updates to a non-transactional table are stored in the binary log at -once. Every thread will, on start, allocate a buffer of -@code{binlog_cache_size} to buffer queries. If a query is bigger than -this, the thread will open a temporary file to handle the bigger cache. -The temporary file will be deleted when the thread ends. +The Contrib section includes programs that let you authenticate your +users from a @strong{MySQL} database and also let you log your log files +into a @strong{MySQL} table. @xref{Contrib}. -The @code{max_binlog_cache_size} can be used to restrict the total size used -to cache a multi-transaction query. +You can change the Apache logging format to be easily readable by +@strong{MySQL} by putting the following into the Apache configuration file: -If you are using the update or binary log, concurrent inserts will -not work together with @code{CREATE ... INSERT} and @code{INSERT ... SELECT}. -This is to ensure that you can recreate an exact copy of your tables by -applying the log on a backup. +@example +LogFormat \ + "\"%h\",%@{%Y%m%d%H%M%S@}t,%>s,\"%b\",\"%@{Content-Type@}o\", \ + \"%U\",\"%@{Referer@}i\",\"%@{User-Agent@}i\"" +@end example -@cindex slow query log -@cindex files, slow query log -@node Slow query log, , Binary log, Log files -@section The Slow Query Log +In @strong{MySQL} you can do something like this: -When started with the @code{--log-slow-queries[=file_name]} option, -@code{mysqld} writes a log file containing all SQL commands that took -more than @code{long_query_time} to execute. The time to get the initial -table locks are not counted as execution time. +@example +LOAD DATA INFILE '/local/access_log' INTO TABLE table_name +FIELDS TERMINATED BY ',' OPTIONALLY ENCLOSED BY '"' ESCAPED BY '\\' +@end example -The slow query log is logged after the query is executed and after all -locks has been released. This may be different than the order in which -the statements are executed. +@cindex Borland C++ compiler +@node Borland C++, , Apache, Common programs +@section Borland C++ -If no file name is given, it defaults to the name of the host machine -suffixed with @code{-slow.log}. If a filename is given, but doesn't -contain a path, the file is written in the data directory. +You can compile the @strong{MySQL} Windows source with Borland C++ 5.02. +(The Windows source includes only projects for Microsoft VC++, for +Borland C++ you have to do the project files yourself). -The slow query log can be used to find queries that take a long time to -execute and are thus candidates for optimization. With a large log, that -can become a difficult task. You can pipe the slow query log through the -@code{mysqldumpslow} command to get a summary of the queries which -appear in the log. - -You are using @code{--log-long-format} then also queries that are not -using indexes are printed. @xref{Command-line options}. +One known problem with Borland C++ is that it uses a different structure +alignment than VC++. This means that you will run into problems if you +try to use the default @code{libmysql.dll} libraries (that was compiled +with VC++) with Borland C++. You can do one of the following to avoid +this problem. -@cindex database replication -@cindex replication, database -@cindex database mirroring -@cindex mirroring, database -You can also use the update logs when you have a mirrored database on -another host and you want to replicate the changes that have been made -to the master database. @xref{Log Replication}. +@itemize @bullet +@item +You can use the static @strong{MySQL} libraries for Borland C++ that you +can find on @uref{http://www.mysql.com/downloads/os-win32.html}. +@item +Only call @code{mysql_init()} with @code{NULL} as an argument, not a +pre-allocated MYSQL struct. +@end itemize -@cindex client tools -@cindex APIs -@cindex @code{mysqlclient} library -@cindex buffer sizes, client -@cindex library, @code{mysqlclient} -@node Clients, Comparisons, Log files, Top -@chapter MySQL APIs +@cindex problems, common errors +@cindex errors, common +@node Problems, Common problems, Common programs, Top +@chapter Problems and Common Errors @menu -* C:: @strong{MySQL} C API -* Perl:: @strong{MySQL} Perl API -* Eiffel:: @strong{MySQL} Eiffel wrapper -* Java:: @strong{MySQL} Java connectivity (JDBC) -* PHP:: @strong{MySQL} PHP API -* Cplusplus:: @strong{MySQL} C++ APIs -* Python:: @strong{MySQL} Python APIs -* Tcl:: @strong{MySQL} Tcl APIs +* What is crashing:: How to determine what is causing problems +* Crashing:: What to do if @strong{MySQL} keeps crashing +* Link errors:: Problems when linking with the @strong{MySQL} client library +* Common errors:: Some common errors when using @strong{MySQL} +* Full disk:: How @strong{MySQL} handles a full disk +* Multiple sql commands:: How to run SQL commands from a text file +* Temporary files:: Where @strong{MySQL} stores temporary files +* Problems with mysql.sock:: How to protect @file{/tmp/mysql.sock} +* Changing MySQL user:: How to run @strong{MySQL} as a normal user +* Resetting permissions:: How to reset a forgotten password. +* File permissions :: Problems with file permissions +* Not enough file handles:: File not found +* Using DATE:: Problems using @code{DATE} columns +* Timezone problems:: Timezone problems +* Case sensitivity:: Case sensitivity in searches +* Problems with NULL:: Problems with @code{NULL} values +* Problems with alias:: Problems with @code{alias} +* Deleting from related tables:: Deleting rows from related tables +* No matching rows:: Solving problems with no matching rows +* ALTER TABLE problems:: Problems with @code{ALTER TABLE}. +* Change column order:: How to change the order of columns in a table +* Temporary table problems:: @end menu -This chapter describes the APIs available for @strong{MySQL}, where to get -them, and how to use them. The C API is the most extensively covered, as it -was developed by the @strong{MySQL} team, and is the basis for most of the -other APIs. +This chapter lists some common problems and error messages that users have +run into. You will learn how to figure out what the problem is, and what +to do to solve it. You will also find proper solutions to some common +problems. -@cindex C API, datatypes -@cindex datatypes, C API -@node C, Perl, Clients, Clients -@section MySQL C API +@node What is crashing, Crashing, Problems, Problems +@section How to Determine What Is Causing Problems -@menu -* C API datatypes:: C API Datatypes -* C API function overview:: C API Function Overview -* C API functions:: C API Function Descriptions -* C API problems:: -* Thread-safe clients:: -@end menu +When you run into problems, the first thing you should do is to find out +which program / piece of equipment is causing problems: -The C API code is distributed with @strong{MySQL}. It is included in the -@code{mysqlclient} library and allows C programs to access a database. +@itemize @bullet +@item +If you have one of the following symptoms, then it is probably a hardware +(like memory, motherboard, CPU, or hard disk) or kernel problem: +@itemize @minus +@item +The keyboard doesn't work. This can normally be checked by pressing +Caps Lock. If the Caps Lock light doesn't change you have to replace +your keyboard. (Before doing this, you should try to reboot +your computer and check all cables to the keyboard.) +@item +The mouse pointer doesn't move. +@item +The machine doesn't answer to a remote machine's pings. +@item +Different, unrelated programs don't behave correctly. +@item +If your system rebooted unexpectedly (a faulty user level program should +NEVER be able to take down your system). +@end itemize -Many of the clients in the @strong{MySQL} source distribution are -written in C. If you are looking for examples that demonstrate how to -use the C API, take a look at these clients. You can find these in the -@code{clients} directory in the @strong{MySQL} source distribution. +In this case you should start by checking all your cables and run some +diagnostic tool to check your hardware! +You should also check if there are any patches, updates, or service +packs for your operating system that could likely solve your problems. +Check also that all your libraries (like glibc) are up to date. -Most of the other client APIs (all except Java) use the @code{mysqlclient} -library to communicate with the @strong{MySQL} server. This means that, for -example, you can take advantage of many of the same environment variables -that are used by other client programs, because they are referenced from the -library. See @ref{Programs}, for a list of these variables. +It's always good to use a machine with ECC memory to discover +memory problems early! +@item +If your keyboard is locked up, you may be able to fix this by +logging into your machine from another machine and execute +@code{kbd_mode -a} on it. -The client has a maximum communication buffer size. The size of the buffer -that is allocated initially (16K bytes) is automatically increased up to the -maximum size (the maximum is 16M). Because buffer sizes are increased -only as demand warrants, simply increasing the default maximum limit does not -in itself cause more resources to be used. This size check is mostly a check -for erroneous queries and communication packets. +@item +Please examine your system log file (/var/log/messages or similar) for +reasons for your problems. If you think the problem is in @strong{MySQL} +then you should also examine @strong{MySQL}'s log files. @xref{Update log}. -The communication buffer must be large enough to contain a single SQL -statement (for client-to-server traffic) and one row of returned data (for -server-to-client traffic). Each thread's communication buffer is dynamically -enlarged to handle any query or row up to the maximum limit. For example, if -you have @code{BLOB} values that contain up to 16M of data, you must have a -communication buffer limit of at least 16M (in both server and client). The -client's default maximum is 16M, but the default maximum in the server is -1M. You can increase this by changing the value of the -@code{max_allowed_packet} parameter when the server is started. @xref{Server -parameters}. +@item +If you don't think you have hardware problems, you should try to find +out which program is causing problems. -The @strong{MySQL} server shrinks each communication buffer to -@code{net_buffer_length} bytes after each query. For clients, the size of -the buffer associated with a connection is not decreased until the connection -is closed, at which time client memory is reclaimed. +Try using @code{top}, @code{ps}, @code{taskmanager}, or some similar program, +to check which program is taking all CPU or is locking the machine. -For programming with threads, consult the 'how to make a thread-safe -client' chapter. @xref{Thread-safe clients}. +@item +Check with @code{top}, @code{df}, or a similar program if you are out of +memory, disk space, open files, or some other critical resource. -@node C API datatypes, C API function overview, C, C -@subsection C API Datatypes -@table @code +@item +If the problem is some runaway process, you can always try to kill it. If it +doesn't want to die, there is probably a bug in the operating system. +@end itemize -@tindex MYSQL C type -@item MYSQL -This structure represents a handle to one database connection. It is -used for almost all @strong{MySQL} functions. +If after you have examined all other possibilities and you have +concluded that it's the @strong{MySQL} server or a @strong{MySQL} client +that is causing the problem, it's time to do a bug report for our +mailing list or our support team. In the bug report, try to give a +very detailed description of how the system is behaving and what you think is +happening. You should also state why you think it's @strong{MySQL} that +is causing the problems. Take into consideration all the situations in +this chapter. State any problems exactly how they appear when you +examine your system. Use the 'cut and paste' method for any output +and/or error messages from programs and/or log files! -@tindex MYSQL_RES C type -@item MYSQL_RES -This structure represents the result of a query that returns rows -(@code{SELECT}, @code{SHOW}, @code{DESCRIBE}, @code{EXPLAIN}). The -information returned from a query is called the @emph{result set} in the -remainder of this section. +Try to describe in detail which program is not working and all +symptoms you see! We have in the past received many bug reports that just +state "the system doesn't work". This doesn't provide us with any +information about what could be the problem. -@tindex MYSQL_ROW C type -@item MYSQL_ROW -This is a type-safe representation of one row of data. It is currently -implemented as an array of counted byte strings. (You cannot treat these as -null-terminated strings if field values may contain binary data, because such -values may contain null bytes internally.) Rows are obtained by calling -@code{mysql_fetch_row()}. +If a program fails, it's always useful to know: -@tindex MYSQL_FIELD C type -@item MYSQL_FIELD -This structure contains information about a field, such as the field's -name, type, and size. Its members are described in more detail below. -You may obtain the @code{MYSQL_FIELD} structures for each field by -calling @code{mysql_fetch_field()} repeatedly. Field values are not part of -this structure; they are contained in a @code{MYSQL_ROW} structure. +@itemize @bullet +@item +Has the program in question made a segmentation fault (core dumped)? +@item +Is the program taking the whole CPU? Check with @code{top}. Let the +program run for a while, it may be evaluating something heavy. +@item +If it's the @code{mysqld} server that is causing problems, can you +do @code{mysqladmin -u root ping} or @code{mysqladmin -u root processlist}? +@item +What does a client program say (try with @code{mysql}, for example) +when you try to connect to the @strong{MySQL} server? +Does the client jam? Do you get any output from the program? +@end itemize +When sending a bug report, you should of follow the outlines +described in this manual. @xref{Asking questions}. -@tindex MYSQL_FIELD_OFFSET C type -@item MYSQL_FIELD_OFFSET -This is a type-safe representation of an offset into a @strong{MySQL} field -list. (Used by @code{mysql_field_seek()}.) Offsets are field numbers -within a row, beginning at zero. +@cindex crash, repeated +@node Crashing, Link errors, What is crashing, Problems +@section What to Do if MySQL Keeps Crashing -@tindex my_ulonglong C type -@tindex my_ulonglong values, printing -@item my_ulonglong -The type used for the number of rows and for @code{mysql_affected_rows()}, -@code{mysql_num_rows()}, and @code{mysql_insert_id()}. This type provides a -range of @code{0} to @code{1.84e19}. +All @strong{MySQL} versions are tested on many platforms before they are +released. This doesn't mean that there aren't any bugs in +@strong{MySQL}, but it means if there are bugs, they are very few and can be +hard to find. If you have a problem, it will always help if you try to +find out exactly what crashes your system, as you will have a much better +chance of getting this fixed quickly. -On some systems, attempting to print a value of type @code{my_ulonglong} -will not work. To print such a value, convert it to @code{unsigned long} -and use a @code{%lu} print format. Example: -@example -printf (Number of rows: %lu\n", (unsigned long) mysql_num_rows(result)); -@end example -@end table +First, you should try to find out whether the problem is that the +@code{mysqld} daemon dies or whether your problem has to do with your +client. You can check how long your @code{mysqld} server has been up by +executing @code{mysqladmin version}. If @code{mysqld} has died, you may +find the reason for this in the file +@file{mysql-data-directory/`hostname`.err}. @xref{Error log}. -@noindent -The @code{MYSQL_FIELD} structure contains the members listed below: +Many crashes of @strong{MySQL} are caused by corrupted index / data +files. @strong{MySQL} will update the data on disk, with the +@code{write()} system call, after every SQL statement and before the +client is notified about the result. (This is not true if you are running +with @code{delayed_key_writes}, in which case only the data is written.) +This means that the data is safe even if @code{mysqld} crashes, as the OS will +ensure that the not flushed data is written to disk. You can force +@strong{MySQL} to sync everything to disk after every SQL command by +starting @code{mysqld} with @code{--flush}. -@table @code -@item char * name -The name of the field, as a null-terminated string. +The above means that normally you shouldn't get corrupted tables unless: -@item char * table -The name of the table containing this field, if it isn't a calculated field. -For calculated fields, the @code{table} value is an empty string. +@itemize @bullet +@item +Someone/something killed @code{mysqld} or the machine in the middle +of an update. +@item +You have found a bug in @code{mysqld} that caused it to die in the +middle of an update. +@item +Someone is manipulating the data/index files outside of @strong{mysqld} +without locking the table properly. +@item +If you are running many @code{mysqld} servers on the same data on a +system that doesn't support good file system locks (normally handled by +the @code{lockd} daemon ) or if you are running +multiple servers with @code{--skip-locking} +@item +You have a crashed index/data file that contains very wrong data that +got @code{mysqld} confused. +@item +You have found a bug in the data storage code. This isn't that likely, +but it's at least possible. In this case you can try to change the file +type to another database handler by using @code{ALTER TABLE} on a +repaired copy of the table! +@end itemize -@item char * def -The default value of this field, as a null-terminated string. This is set -only if you use @code{mysql_list_fields()}. +Because it is very difficult to know why something is crashing, first try to +check whether or not things that work for others crash for you. Please try +the following things: -@item enum enum_field_types type -The type of the field. -The @code{type} value may be one of the following: +@itemize @bullet +@item +Take down the @code{mysqld} daemon with @code{mysqladmin shutdown}, run +@code{myisamchk --silent --force */*.MYI} on all tables, and restart the +@code{mysqld} daemon. This will ensure that you are running from a clean +state. @xref{Maintenance}. -@multitable @columnfractions .3 .55 -@item @strong{Type value} @tab @strong{Type meaning} -@item @code{FIELD_TYPE_TINY} @tab @code{TINYINT} field -@item @code{FIELD_TYPE_SHORT} @tab @code{SMALLINT} field -@item @code{FIELD_TYPE_LONG} @tab @code{INTEGER} field -@item @code{FIELD_TYPE_INT24} @tab @code{MEDIUMINT} field -@item @code{FIELD_TYPE_LONGLONG} @tab @code{BIGINT} field -@item @code{FIELD_TYPE_DECIMAL} @tab @code{DECIMAL} or @code{NUMERIC} field -@item @code{FIELD_TYPE_FLOAT} @tab @code{FLOAT} field -@item @code{FIELD_TYPE_DOUBLE} @tab @code{DOUBLE} or @code{REAL} field -@item @code{FIELD_TYPE_TIMESTAMP} @tab @code{TIMESTAMP} field -@item @code{FIELD_TYPE_DATE} @tab @code{DATE} field -@item @code{FIELD_TYPE_TIME} @tab @code{TIME} field -@item @code{FIELD_TYPE_DATETIME} @tab @code{DATETIME} field -@item @code{FIELD_TYPE_YEAR} @tab @code{YEAR} field -@item @code{FIELD_TYPE_STRING} @tab String (@code{CHAR} or @code{VARCHAR}) field -@item @code{FIELD_TYPE_BLOB} @tab @code{BLOB} or @code{TEXT} field (use @code{max_length} to determine the maximum length) -@item @code{FIELD_TYPE_SET} @tab @code{SET} field -@item @code{FIELD_TYPE_ENUM} @tab @code{ENUM} field -@item @code{FIELD_TYPE_NULL} @tab @code{NULL}-type field -@item @code{FIELD_TYPE_CHAR} @tab Deprecated; use @code{FIELD_TYPE_TINY} instead -@end multitable +@item +Use @code{mysqld --log} and try to determine from the information in the log +whether or not some specific query kills the server. About 95% of all bugs are +related to a particular query! Normally this is one of the last queries in +the log file just before @strong{MySQL} restarted. @xref{Query log}. +If you can repeatadly kill @strong{MySQL} with one of the queries, even +when you have checked all tables just before doing the query, then you +have been able to locate the bug and should do a bug report for this! +@xref{Bug reports}. -You can use the @code{IS_NUM()} macro to test whether or not a field has a -numeric type. Pass the @code{type} value to @code{IS_NUM()} and it -will evaluate to TRUE if the field is numeric: +@item +Try to make a test case that we can use to reproduce the problem. +@xref{Reproduceable test case}. -@example -if (IS_NUM(field->type)) - printf("Field is numeric\n"); -@end example +@item +Try running the included mysql-test test and the @strong{MySQL} +benchmarks. @xref{MySQL test suite}. They should test @strong{MySQL} +rather well. You can also add code that to the benchmarks to simulates +your application! The benchmarks can be found in the @file{bench} +directory in the source distribution or, for a binary distribution, in +the @file{sql-bench} directory under your @strong{MySQL} installation +directory. -@item unsigned int length -The width of the field, as specified in the table definition. +@item +Try @code{fork_test.pl} and @code{fork2_test.pl}. -@item unsigned int max_length -The maximum width of the field for the result set (the length of the longest -field value for the rows actually in the result set). If you use -@code{mysql_store_result()} or @code{mysql_list_fields()}, this contains the -maximum length for the field. If you use @code{mysql_use_result()}, the -value of this variable is zero. +@item +If you configure @strong{MySQL} for debugging, it will be much easier to +gather information about possible errors if something goes wrong. +Reconfigure @strong{MySQL} with the @code{--with-debug} option or +@code{--with-debug=full} to @code{configure} and then recompile. +@xref{Debugging server}. -@item unsigned int flags -Different bit-flags for the field. The @code{flags} value may have zero -or more of the following bits set: +@item +Configuring @strong{MySQL} for debugging causes a safe memory allocator to be +included that can find some errors. It also provides a lot of output about +what is happening. -@multitable @columnfractions .3 .55 -@item @strong{Flag value} @tab @strong{Flag meaning} -@item @code{NOT_NULL_FLAG} @tab Field can't be @code{NULL} -@item @code{PRI_KEY_FLAG} @tab Field is part of a primary key -@item @code{UNIQUE_KEY_FLAG} @tab Field is part of a unique key -@item @code{MULTIPLE_KEY_FLAG} @tab Field is part of a non-unique key -@item @code{UNSIGNED_FLAG} @tab Field has the @code{UNSIGNED} attribute -@item @code{ZEROFILL_FLAG} @tab Field has the @code{ZEROFILL} attribute -@item @code{BINARY_FLAG} @tab Field has the @code{BINARY} attribute -@item @code{AUTO_INCREMENT_FLAG} @tab Field has the @code{AUTO_INCREMENT} -attribute -@item @code{ENUM_FLAG} @tab Field is an @code{ENUM} (deprecated) -@item @code{BLOB_FLAG} @tab Field is a @code{BLOB} or @code{TEXT} (deprecated) -@item @code{TIMESTAMP_FLAG} @tab Field is a @code{TIMESTAMP} (deprecated) -@end multitable +@item +Have you applied the latest patches for your operating system? -Use of the @code{BLOB_FLAG}, @code{ENUM_FLAG}, and @code{TIMESTAMP_FLAG} -flags is deprecated because they indicate the type of a field rather -than an attribute of its type. It is preferable to test -@code{field->type} against @code{FIELD_TYPE_BLOB}, -@code{FIELD_TYPE_ENUM}, or @code{FIELD_TYPE_TIMESTAMP} instead. +@item +Use the @code{--skip-locking} option to @code{mysqld}. On some systems, the +@code{lockd} lock manager does not work properly; the @code{--skip-locking} +option tells @code{mysqld} not to use external locking. (This means that you +cannot run 2 @code{mysqld} servers on the same data and that you must be +careful if you use @code{myisamchk}, but it may be instructive to try the +option as a test.) -@noindent -The example below illustrates a typical use of the @code{flags} value: +@item +Have you tried @code{mysqladmin -u root processlist} when @code{mysqld} +appears to be running but not responding? Sometimes @code{mysqld} is not +comatose even though you might think so. The problem may be that all +connections are in use, or there may be some internal lock problem. +@code{mysqladmin processlist} will usually be able to make a connection even +in these cases, and can provide useful information about the current number +of connections and their status. -@example -if (field->flags & NOT_NULL_FLAG) - printf("Field can't be null\n"); -@end example +@item +Run the command @code{mysqladmin -i 5 status} or @code{mysqladmin -i 5 +-r status} or in a separate window to produce statistics while you run +your other queries. -You may use the following convenience macros to determine the boolean -status of the @code{flags} value: +@item +Try the following: +@enumerate +@item +Start @code{mysqld} from @code{gdb} (or in another debugger). +@xref{Using gdb on mysqld}. -@multitable @columnfractions .3 .5 -@item @code{IS_NOT_NULL(flags)} @tab True if this field is defined as @code{NOT NULL} -@item @code{IS_PRI_KEY(flags)} @tab True if this field is a primary key -@item @code{IS_BLOB(flags)} @tab True if this field is a @code{BLOB} or @code{TEXT} (deprecated; test @code{field->type} instead) -@end multitable +@item +Run your test scripts. -@item unsigned int decimals -The number of decimals for numeric fields. -@end table +@item +Print the backtrace and the local variables at the 3 lowest levels. In gdb you +can do this with the following commands when @code{mysqld} has crashed inside +gdb: -@cindex C API, functions -@cindex functions, C API -@node C API function overview, C API functions, C API datatypes, C -@subsection C API Function Overview +@example +backtrace +info local +up +info local +up +info local +@end example -The functions available in the C API are listed below and are described in -greater detail in the next section. -@xref{C API functions}. +With gdb you can also examine which threads exist with @code{info +threads} and switch to a specific thread with @code{thread #}, where +@code{#} is the thread id. +@end enumerate -@multitable @columnfractions .3 .7 -@item @strong{mysql_affected_rows()} @tab -Returns the number of rows changed/deleted/inserted by the last @code{UPDATE}, -@code{DELETE}, or @code{INSERT} query. +@item +Try to simulate your application with a Perl script to force +@strong{MySQL} to crash or misbehave. -@item @strong{mysql_close()} @tab -Closes a server connection. +@item +Send a normal bug report. @xref{Bug reports}. Be even more detailed +than usual. Because @strong{MySQL} works for many people, it may be that the +crash results from something that exists only on your computer (for example, +an error that is related to your particular system libraries). +@item +If you have a problem with tables with dynamic-length rows and you are +not using @code{BLOB/TEXT} columns (but only @code{VARCHAR} columns), you +can try to change all @code{VARCHAR} to @code{CHAR} with @code{ALTER +TABLE}. This will force @strong{MySQL} to use fixed-size rows. +Fixed-size rows take a little extra space, but are much more tolerant to +corruption! -@item @strong{mysql_connect()} @tab -Connects to a @strong{MySQL} server. This function is deprecated; use -@code{mysql_real_connect()} instead. +The current dynamic row code has been in use at @strong{MySQL AB} for at +least 3 years without any problems, but by nature dynamic-length rows are +more prone to errors, so it may be a good idea to try the above to see if +it helps! +@end itemize -@item @strong{mysql_change_user()} @tab -Changes user and database on an open connection. +@cindex linking, errors +@cindex errors, linking +@cindex problems, linking +@node Link errors, Common errors, Crashing, Problems +@section Problems When Linking with the MySQL Client Library -@item @strong{mysql_character_set_name()} @tab -Returns the name of the default character set for the connection. +If you are linking your program and you get errors for unreferenced +symbols that start with @code{mysql_}, like the following: -@item @strong{mysql_create_db()} @tab -Creates a database. This function is deprecated; use the SQL command -@code{CREATE DATABASE} instead. +@example +/tmp/ccFKsdPa.o: In function `main': +/tmp/ccFKsdPa.o(.text+0xb): undefined reference to `mysql_init' +/tmp/ccFKsdPa.o(.text+0x31): undefined reference to `mysql_real_connect' +/tmp/ccFKsdPa.o(.text+0x57): undefined reference to `mysql_real_connect' +/tmp/ccFKsdPa.o(.text+0x69): undefined reference to `mysql_error' +/tmp/ccFKsdPa.o(.text+0x9a): undefined reference to `mysql_close' +@end example -@item @strong{mysql_data_seek()} @tab -Seeks to an arbitrary row in a query result set. +you should be able to solve this by adding @code{-Lpath-to-the-mysql-library +-lmysqlclient} @strong{LAST} on your link line. -@item @strong{mysql_debug()} @tab -Does a @code{DBUG_PUSH} with the given string. +If you get @code{undefined reference} errors for the @code{uncompress} +or @code{compress} function, add @code{-lgz} @strong{LAST} on your link +line and try again! -@item @strong{mysql_drop_db()} @tab -Drops a database. This function is deprecated; use the SQL command -@code{DROP DATABASE} instead. +If you get @code{undefined reference} errors for functions that should +exist on your system, like @code{connect}, check the man page for the +function in question, for which libraries you should add to the link +line! -@item @strong{mysql_dump_debug_info()} @tab -Makes the server write debug information to the log. +If you get @code{undefined reference} errors for functions that don't +exist on your system, like the following: -@item @strong{mysql_eof()} @tab -Determines whether or not the last row of a result set has been read. -This function is deprecated; @code{mysql_errno()} or @code{mysql_error()} -may be used instead. +@example +mf_format.o(.text+0x201): undefined reference to `__lxstat' +@end example -@item @strong{mysql_errno()} @tab -Returns the error number for the most recently invoked @strong{MySQL} function. +it usually means that your library is compiled on a system that is not +100 % compatible with yours. In this case you should download the +latest @strong{MySQL} source distribution and compile this yourself. +@xref{Installing source}. -@item @strong{mysql_error()} @tab -Returns the error message for the most recently invoked @strong{MySQL} function. +If you are trying to run a program and you then get errors for +unreferenced symbols that start with @code{mysql_} or that the +@code{mysqlclient} library can't be found, this means that your system +can't find the share @code{libmysqlclient.so} library. -@item @strong{mysql_real_escape_string()} @tab -Escapes special characters in a string for use in a SQL statement taking -into account the current charset of the connection. +The fix for this is to tell your system to search after shared +libraries where the library is located by one of the following methods: -@item @strong{mysql_escape_string()} @tab -Escapes special characters in a string for use in a SQL statement. +@itemize @bullet +@item +Add the path to the directory where you have @code{libmysqlclient.so} the +@code{LD_LIBRARY_PATH} environment variable. +@item +Add the path to the directory where you have @code{libmysqlclient.so} the +@code{LD_LIBRARY} environment variable. +@item +Copy @code{libmysqlclient.so} to some place that is searched by your system, +like @file{/lib}, and update the shared library information by executing +@code{ldconfig}. +@end itemize -@item @strong{mysql_fetch_field()} @tab -Returns the type of the next table field. +Another way to solve this problem is to link your program statically, with +@code{-static}, or by removing the dynamic @strong{MySQL} libraries +before linking your code. In the second case you should be +sure that no other programs are using the dynamic libraries! -@item @strong{mysql_fetch_field_direct()} @tab -Returns the type of a table field, given a field number. +@cindex errors, list of +@node Common errors, Full disk, Link errors, Problems +@section Some Common Errors When Using MySQL -@item @strong{mysql_fetch_fields()} @tab -Returns an array of all field structures. +@menu +* Error Access denied:: @code{Access denied} Error +* Gone away:: @code{MySQL server has gone away} error +* Can not connect to server:: @code{Can't connect to [local] MySQL server} error +* Blocked host:: @code{Host '...' is blocked} error +* Too many connections:: @code{Too many connections} error +* Non-transactional tables:: @code{Some non-transactional changed tables couldn't be rolled back} Error +* Out of memory:: @code{Out of memory} error +* Packet too large:: @code{Packet too large} error +* Communication errors:: Communication errors / Aborted connection +* Full table:: @code{The table is full} error +* Cannot create:: @code{Can't create/write to file} Error +* Commands out of sync:: @code{Commands out of sync} error in client +* Ignoring user:: @code{Ignoring user} error +* Cannot find table:: @code{Table 'xxx' doesn't exist} error +* Cannot initialize character set:: +@end menu -@item @strong{mysql_fetch_lengths()} @tab -Returns the lengths of all columns in the current row. +This section lists some errors that users frequently get. You will find +descriptions of the errors, and how to solve the problem here. -@item @strong{mysql_fetch_row()} @tab -Fetches the next row from the result set. +@cindex errors, access denied +@cindex problems, access denied errors +@cindex access denied errors +@node Error Access denied, Gone away, Common errors, Common errors +@subsection @code{Access denied} Error -@item @strong{mysql_field_seek()} @tab -Puts the column cursor on a specified column. +@xref{Privileges}, and especially. @xref{Access denied}. -@item @strong{mysql_field_count()} @tab -Returns the number of result columns for the most recent query. +@node Gone away, Can not connect to server, Error Access denied, Common errors +@subsection @code{MySQL server has gone away} Error -@item @strong{mysql_field_tell()} @tab -Returns the position of the field cursor used for the last -@code{mysql_fetch_field()}. +This section also covers the related @code{Lost connection to server +during query} error. -@item @strong{mysql_free_result()} @tab -Frees memory used by a result set. +The most common reason for the @code{MySQL server has gone away} error +is that the server timed out and closed the connection. By default, the +server closes the connection after 8 hours if nothing has happened. You +can change the time limit by setting the @code{wait_timeout} variable when +you start @code{mysqld}. -@item @strong{mysql_get_client_info()} @tab -Returns client version information. +Another common reason to receive the @code{MySQL server has gone away} error +is because you have issued a ``close'' on your @strong{MySQL} connection +and then tried to run a query on the closed connection. -@item @strong{mysql_get_host_info()} @tab -Returns a string describing the connection. +You can check that the @strong{MySQL} hasn't died by executing +@code{mysqladmin version} and examining the uptime. -@item @strong{mysql_get_proto_info()} @tab -Returns the protocol version used by the connection. +If you have a script, you just have to issue the query again for the client +to do an automatic reconnection. -@item @strong{mysql_get_server_info()} @tab -Returns the server version number. +You normally can get the following error codes in this case +(which one you get is OS-dependent): -@item @strong{mysql_info()} @tab -Returns information about the most recently executed query. +@multitable @columnfractions .3 .7 +@item @code{CR_SERVER_GONE_ERROR} @tab The client couldn't send a question to the +server. +@item @code{CR_SERVER_LOST} @tab The client didn't get an error when writing +to the server, but it didn't get a full answer (or any answer) to the question. +@end multitable -@item @strong{mysql_init()} @tab -Gets or initializes a @code{MYSQL} structure. +You can also get these errors if you send a query to the server that is +incorrect or too large. If @code{mysqld} gets a packet that is too large +or out of order, it assumes that something has gone wrong with the client and +closes the connection. If you need big queries (for example, if you are +working with big @code{BLOB} columns), you can increase the query limit by +starting @code{mysqld} with the @code{-O max_allowed_packet=#} option +(default 1M). The extra memory is allocated on demand, so @code{mysqld} will +use more memory only when you issue a big query or when @code{mysqld} must +return a big result row! -@item @strong{mysql_insert_id()} @tab -Returns the ID generated for an @code{AUTO_INCREMENT} column by the previous -query. +@node Can not connect to server, Blocked host, Gone away, Common errors +@subsection @code{Can't connect to [local] MySQL server} error -@item @strong{mysql_kill()} @tab -Kills a given thread. +A @strong{MySQL} client on Unix can connect to the @code{mysqld} server in two +different ways: Unix sockets, which connect through a file in the file +system (default @file{/tmp/mysqld.sock}) or TCP/IP, which connects +through a port number. Unix sockets are faster than TCP/IP but can only +be used when connecting to a server on the same computer. Unix sockets +are used if you don't specify a hostname or if you specify the special +hostname @code{localhost}. -@item @strong{mysql_list_dbs()} @tab -Returns database names matching a simple regular expression. +On Windows you can connect only with TCP/IP if the @code{mysqld} server +is running on Win95/Win98. If it's running on NT, you can also connect +with named pipes. The name of the named pipe is @strong{MySQL}. If you +don't give a hostname when connecting to @code{mysqld}, a @strong{MySQL} client +will first try to connect to the named pipe, and if this doesn't work it +will connect to the TCP/IP port. You can force the use of named pipes +on Windows by using @code{.} as the hostname. -@item @strong{mysql_list_fields()} @tab -Returns field names matching a simple regular expression. +The error (2002) @code{Can't connect to ...} normally means that there +isn't a @strong{MySQL} server running on the system or that you are +using a wrong socket file or TCP/IP port when trying to connect to the +@code{mysqld} server. -@item @strong{mysql_list_processes()} @tab -Returns a list of the current server threads. +Start by checking (using @code{ps} or the task manager on Windows) that +there is a process running named @code{mysqld} on your server! If there +isn't any @code{mysqld} process, you should start one. @xref{Starting +server}. -@item @strong{mysql_list_tables()} @tab -Returns table names matching a simple regular expression. +If a @code{mysqld} process is running, you can check the server by +trying these different connections (the port number and socket pathname +might be different in your setup, of course): -@item @strong{mysql_num_fields()} @tab -Returns the number of columns in a result set. +@example +shell> mysqladmin version +shell> mysqladmin variables +shell> mysqladmin -h `hostname` version variables +shell> mysqladmin -h `hostname` --port=3306 version +shell> mysqladmin -h 'ip for your host' version +shell> mysqladmin --socket=/tmp/mysql.sock version +@end example -@item @strong{mysql_num_rows()} @tab -Returns the number of rows in a result set. +Note the use of backquotes rather than forward quotes with the @code{hostname} +command; these cause the output of @code{hostname} (that is, the current +hostname) to be substituted into the @code{mysqladmin} command. -@item @strong{mysql_options()} @tab -Sets connect options for @code{mysql_connect()}. +Here are some reasons the @code{Can't connect to local MySQL server} +error might occur: -@item @strong{mysql_ping()} @tab -Checks whether or not the connection to the server is working, reconnecting -as necessary. +@itemize @bullet +@item +@code{mysqld} is not running. +@item +You are running on a system that uses MIT-pthreads. +If you are running on a system that doesn't have native threads, +@code{mysqld} uses the MIT-pthreads package. @xref{Which OS}. However, +all MIT-pthreads versions doesn't support Unix sockets. On a system +without sockets support you must always specify the hostname explicitly +when connecting to the server. Try using this command to check the +connection to the server: +@example +shell> mysqladmin -h `hostname` version +@end example +@item +Someone has removed the Unix socket that @code{mysqld} uses (default +@file{/tmp/mysqld.sock}). You might have a @code{cron} job that removes +the @strong{MySQL} socket (for example, a job that removes old files +from the @file{/tmp} directory). You can always run @code{mysqladmin +version} and check that the socket @code{mysqladmin} is trying to use +really exists. The fix in this case is to change the @code{cron} job to +not remove @file{mysqld.sock} or to place the socket somewhere else. You +can specify a different socket location at @strong{MySQL} configuration +time with this command: +@example +shell> ./configure --with-unix-socket-path=/path/to/socket +@end example +You can also start @code{safe_mysqld} with the +@code{--socket=/path/to/socket} option and set the environment variable +@code{MYSQL_UNIX_PORT} to the socket pathname before starting your +@strong{MySQL} clients. +@item +You have started the @code{mysqld} server with +the @code{--socket=/path/to/socket} option. If you change the socket +pathname for the server, you must also notify the @strong{MySQL} clients +about the new path. You can do this by setting the environment variable +@code{MYSQL_UNIX_PORT} to the socket pathname or by providing the socket path +as an argument to the clients. You can test the socket with this command: -@item @strong{mysql_query()} @tab -Executes a SQL query specified as a null-terminated string. +@example +shell> mysqladmin --socket=/path/to/socket version +@end example +@item +You are using Linux and one thread has died (core dumped). In this case +you must kill the other @code{mysqld} threads (for example, with the +@code{mysql_zap} script before you can start a new @strong{MySQL} +server. @xref{Crashing}. +@item +You may not have read and write privilege to either the directory that holds +the socket file or privilege to the socket file itself. In this case you +have to either change the privilege for the directory / file or restart +@code{mysqld} so that it uses a directory that you can access. +@end itemize -@item @strong{mysql_real_connect()} @tab -Connects to a @strong{MySQL} server. +If you get the error message @code{Can't connect to MySQL server on +some_hostname}, you can try the following things to find out what the +problem is : -@item @strong{mysql_real_query()} @tab -Executes a SQL query specified as a counted string. +@itemize @bullet +@item +Check if the server is up by doing @code{telnet your-host-name +tcp-ip-port-number} and press @code{RETURN} a couple of times. If there +is a @strong{MySQL} server running on this port you should get a +responses that includes the version number of the running @strong{MySQL} +server. If you get an error like @code{telnet: Unable to connect to +remote host: Connection refused}, then there is no server running on the +given port. +@item +Try connecting to the @code{mysqld} daemon on the local machine and check +the TCP/IP port that @code{mysqld} it's configured to use (variable @code{port}) with +@code{mysqladmin variables}. +@item +Check that your @code{mysqld} server is not started with the +@code{--skip-networking} option. +@end itemize -@item @strong{mysql_reload()} @tab -Tells the server to reload the grant tables. +@node Blocked host, Too many connections, Can not connect to server, Common errors +@subsection @code{Host '...' is blocked} Error -@item @strong{mysql_row_seek()} @tab -Seeks to a row in a result set, using value returned from -@code{mysql_row_tell()}. +If you get an error like this: -@item @strong{mysql_row_tell()} @tab -Returns the row cursor position. +@example +Host 'hostname' is blocked because of many connection errors. +Unblock with 'mysqladmin flush-hosts' +@end example -@item @strong{mysql_select_db()} @tab -Selects a database. +this means that @code{mysqld} has gotten a lot (@code{max_connect_errors}) +of connect requests from the host @code{'hostname'} that have been interrupted +in the middle. After @code{max_connect_errors} failed requests, @code{mysqld} +assumes that something is wrong (like an attack from a cracker), and +blocks the site from further connections until someone executes the command +@code{mysqladmin flush-hosts}. -@item @strong{mysql_shutdown()} @tab -Shuts down the database server. +By default, @code{mysqld} blocks a host after 10 connection errors. +You can easily adjust this by starting the server like this: -@item @strong{mysql_stat()} @tab -Returns the server status as a string. +@example +shell> safe_mysqld -O max_connect_errors=10000 & +@end example -@item @strong{mysql_store_result()} @tab -Retrieves a complete result set to the client. +Note that if you get this error message for a given host, you should first +check that there isn't anything wrong with TCP/IP connections from that +host. If your TCP/IP connections aren't working, it won't do you any good to +increase the value of the @code{max_connect_errors} variable! -@item @strong{mysql_thread_id()} @tab -Returns the current thread ID. +@node Too many connections, Non-transactional tables, Blocked host, Common errors +@subsection @code{Too many connections} Error -@item @strong{mysql_thread_save()} @tab -Returns 1 if the clients are compiled as thread-safe. +If you get the error @code{Too many connections} when you try to connect +to @strong{MySQL}, this means that there is already @code{max_connections} +clients connected to the @code{mysqld} server. -@item @strong{mysql_use_result()} @tab -Initiates a row-by-row result set retrieval. -@end multitable +If you need more connections than the default (100), then you should restart +@code{mysqld} with a bigger value for the @code{max_connections} variable. -To connect to the server, call @code{mysql_init()} to initialize a -connection handler, then call @code{mysql_real_connect()} with that -handler (along with other information such as the hostname, user name, -and password). Upon connection, @code{mysql_real_connect()} sets the -@code{reconnect} flag (part of the MYSQL structure) to a value of -@code{1}. This flag indicates, in the event that a query cannot be -performed because of a lost connection, to try reconnecting to the -server before giving up. When you are done with the connection, call -@code{mysql_close()} to terminate it. +Note that @code{mysqld} actually allows (@code{max_connections}+1) +clients to connect. The last connection is reserved for a user with the +@strong{process} privilege. By not giving this privilege to normal +users (they shouldn't need this), an administrator with this privilege +can log in and use @code{SHOW PROCESSLIST} to find out what could be +wrong. @xref{SHOW}. -While a connection is active, the client may send SQL queries to the server -using @code{mysql_query()} or @code{mysql_real_query()}. The difference -between the two is that @code{mysql_query()} expects the query to be -specified as a null-terminated string whereas @code{mysql_real_query()} -expects a counted string. If the string contains binary data (which may -include null bytes), you must use @code{mysql_real_query()}. +The maximum number of connects @strong{MySQL} is depending on how good +the thread library is on a given platform. Linux or Solaris should be +able to support 500-1000 simultaneous connections, depending on how much +RAM you have and what your clients are doing. -For each non-@code{SELECT} query (for example, @code{INSERT}, @code{UPDATE}, -@code{DELETE}), you can find out how many rows were changed (affected) -by calling @code{mysql_affected_rows()}. +@cindex Non-transactional tables +@node Non-transactional tables, Out of memory, Too many connections, Common errors +@subsection @code{Some non-transactional changed tables couldn't be rolled back} Error -For @code{SELECT} queries, you retrieve the selected rows as a result set. -(Note that some statements are @code{SELECT}-like in that they return rows. -These include @code{SHOW}, @code{DESCRIBE}, and @code{EXPLAIN}. They should -be treated the same way as @code{SELECT} statements.) +If you get the error/warning: @code{Warning: Some non-transactional +changed tables couldn't be rolled back} when trying to do a +@code{ROLLBACK}, this means that some of the tables you used in the +transaction didn't support transactions. These non-transactional tables +will not be affected by the @code{ROLLBACK} statement. -There are two ways for a client to process result sets. One way is to -retrieve the entire result set all at once by calling -@code{mysql_store_result()}. This function acquires from the server all the -rows returned by the query and stores them in the client. The second way is -for the client to initiate a row-by-row result set retrieval by calling -@code{mysql_use_result()}. This function initializes the retrieval, but does -not actually get any rows from the server. +The most typical case when this happens is when you have tried to create +a table of a type that is not supported by your @code{mysqld} binary. +If @code{mysqld} doesn't support a table type (or if the table type is +disabled by a startup option) , it will instead create the table type +with the table type that is most resembles to the one you requested, +probably @code{MyISAM}. -In both cases, you access rows by calling @code{mysql_fetch_row()}. With -@code{mysql_store_result()}, @code{mysql_fetch_row()} accesses rows that have -already been fetched from the server. With @code{mysql_use_result()}, -@code{mysql_fetch_row()} actually retrieves the row from the server. -Information about the size of the data in each row is available by -calling @code{mysql_fetch_lengths()}. +You can check the table type for a table by doing: -After you are done with a result set, call @code{mysql_free_result()} -to free the memory used for it. +@code{SHOW TABLE STATUS LIKE 'table_name'}. @xref{SHOW TABLE STATUS}. -The two retrieval mechanisms are complementary. Client programs should -choose the approach that is most appropriate for their requirements. -In practice, clients tend to use @code{mysql_store_result()} more -commonly. +You can check the extensions your @code{mysqld} binary supports by doing: -An advantage of @code{mysql_store_result()} is that because the rows have all -been fetched to the client, you not only can access rows sequentially, you -can move back and forth in the result set using @code{mysql_data_seek()} or -@code{mysql_row_seek()} to change the current row position within the result -set. You can also find out how many rows there are by calling -@code{mysql_num_rows()}. On the other hand, the memory requirements for -@code{mysql_store_result()} may be very high for large result sets and you -are more likely to encounter out-of-memory conditions. +@code{show variables like 'have_%'}. @xref{SHOW VARIABLES}. -An advantage of @code{mysql_use_result()} is that the client requires less -memory for the result set because it maintains only one row at a time (and -because there is less allocation overhead, @code{mysql_use_result()} can be -faster). Disadvantages are that you must process each row quickly to avoid -tying up the server, you don't have random access to rows within the result -set (you can only access rows sequentially), and you don't know how many rows -are in the result set until you have retrieved them all. Furthermore, you -@emph{must} retrieve all the rows even if you determine in mid-retrieval that -you've found the information you were looking for. +@node Out of memory, Packet too large, Non-transactional tables, Common errors +@subsection @code{Out of memory} Error -The API makes it possible for clients to respond appropriately to -queries (retrieving rows only as necessary) without knowing whether or -not the query is a @code{SELECT}. You can do this by calling -@code{mysql_store_result()} after each @code{mysql_query()} (or -@code{mysql_real_query()}). If the result set call succeeds, the query -was a @code{SELECT} and you can read the rows. If the result set call -fails, call @code{mysql_field_count()} to determine whether or not a -result was actually to be expected. If @code{mysql_field_count()} -returns zero, the query returned no data (indicating that it was an -@code{INSERT}, @code{UPDATE}, @code{DELETE}, etc.), and was not -expected to return rows. If @code{mysql_field_count()} is non-zero, the -query should have returned rows, but didn't. This indicates that the -query was a @code{SELECT} that failed. See the description for -@code{mysql_field_count()} for an example of how this can be done. +If you issue a query and get something like the following error: -Both @code{mysql_store_result()} and @code{mysql_use_result()} allow you to -obtain information about the fields that make up the result set (the number -of fields, their names and types, etc.). You can access field information -sequentially within the row by calling @code{mysql_fetch_field()} repeatedly, -or by field number within the row by calling -@code{mysql_fetch_field_direct()}. The current field cursor position may be -changed by calling @code{mysql_field_seek()}. Setting the field cursor -affects subsequent calls to @code{mysql_fetch_field()}. You can also get -information for fields all at once by calling @code{mysql_fetch_fields()}. +@example +mysql: Out of memory at line 42, 'malloc.c' +mysql: needed 8136 byte (8k), memory in use: 12481367 bytes (12189k) +ERROR 2008: MySQL client ran out of memory +@end example -For detecting and reporting errors, @strong{MySQL} provides access to error -information by means of the @code{mysql_errno()} and @code{mysql_error()} -functions. These return the error code or error message for the most -recently invoked function that can succeed or fail, allowing you to determine -when an error occurred and what it was. +note that the error refers to the @strong{MySQL} client @code{mysql}. The +reason for this error is simply that the client does not have enough memory to +store the whole result. -@node C API functions, C API problems, C API function overview, C -@subsection C API Function Descriptions +To remedy the problem, first check that your query is correct. Is it +reasonable that it should return so many rows? If so, +you can use @code{mysql --quick}, which uses @code{mysql_use_result()} +to retrieve the result set. This places less of a load on the client (but +more on the server). -@menu -* mysql_affected_rows:: @code{mysql_affected_rows()} -* mysql_close:: @code{mysql_close()} -* mysql_connect:: @code{mysql_connect()} -* mysql_change_user:: @code{mysql_change_user()} -* mysql_character_set_name:: @code{mysql_character_set_name()} -* mysql_create_db:: @code{mysql_create_db()} -* mysql_data_seek:: @code{mysql_data_seek()} -* mysql_debug:: @code{mysql_debug()} -* mysql_drop_db:: @code{mysql_drop_db()} -* mysql_dump_debug_info:: @code{mysql_dump_debug_info()} -* mysql_eof:: @code{mysql_eof()} -* mysql_errno:: @code{mysql_errno()} -* mysql_error:: @code{mysql_error()} -* mysql_escape_string:: @code{mysql_escape_string()} -* mysql_fetch_field:: @code{mysql_fetch_field()} -* mysql_fetch_fields:: @code{mysql_fetch_fields()} -* mysql_fetch_field_direct:: @code{mysql_fetch_field_direct()} -* mysql_fetch_lengths:: @code{mysql_fetch_lengths()} -* mysql_fetch_row:: @code{mysql_fetch_row()} -* mysql_field_count:: @code{mysql_field_count()} -* mysql_field_seek:: @code{mysql_field_seek()} -* mysql_field_tell:: @code{mysql_field_tell()} -* mysql_free_result:: @code{mysql_free_result()} -* mysql_get_client_info:: @code{mysql_get_client_info()} -* mysql_get_host_info:: @code{mysql_get_host_info()} -* mysql_get_proto_info:: @code{mysql_get_proto_info()} -* mysql_get_server_info:: @code{mysql_get_server_info()} -* mysql_info:: @code{mysql_info()} -* mysql_init:: @code{mysql_init()} -* mysql_insert_id:: @code{mysql_insert_id()} -* mysql_kill:: @code{mysql_kill()} -* mysql_list_dbs:: @code{mysql_list_dbs()} -* mysql_list_fields:: @code{mysql_list_fields()} -* mysql_list_processes:: @code{mysql_list_processes()} -* mysql_list_tables:: @code{mysql_list_tables()} -* mysql_num_fields:: @code{mysql_num_fields()} -* mysql_num_rows:: @code{mysql_num_rows()} -* mysql_options:: @code{mysql_options()} -* mysql_ping:: @code{mysql_ping()} -* mysql_query:: @code{mysql_query()} -* mysql_real_connect:: @code{mysql_real_connect()} -* mysql_real_escape_string:: @code{mysql_real_escape_string()} -* mysql_real_query:: @code{mysql_real_query()} -* mysql_reload:: @code{mysql_reload()} -* mysql_row_seek:: @code{mysql_row_seek()} -* mysql_row_tell:: @code{mysql_row_tell()} -* mysql_select_db:: @code{mysql_select_db()} -* mysql_shutdown:: @code{mysql_shutdown()} -* mysql_stat:: @code{mysql_stat()} -* mysql_store_result:: @code{mysql_store_result()} -* mysql_thread_id:: @code{mysql_thread_id()} -* mysql_use_result:: @code{mysql_use_result()} -@end menu - -In the descriptions below, a parameter or return value of @code{NULL} means -@code{NULL} in the sense of the C programming language, not a -@strong{MySQL} @code{NULL} value. +@node Packet too large, Communication errors, Out of memory, Common errors +@subsection @code{Packet too large} Error -Functions that return a value generally return a pointer or an integer. -Unless specified otherwise, functions returning a pointer return a -non-@code{NULL} value to indicate success or a @code{NULL} value to indicate -an error, and functions returning an integer return zero to indicate success -or non-zero to indicate an error. Note that ``non-zero'' means just that. -Unless the function description says otherwise, do not test against a value -other than zero: +When a @strong{MySQL} client or the @code{mysqld} server gets a packet bigger +than @code{max_allowed_packet} bytes, it issues a @code{Packet too large} +error and closes the connection. -@example -if (result) /* correct */ - ... error ... +If you are using the @code{mysql} client, you may specify a bigger buffer by +starting the client with @code{mysql --set-variable=max_allowed_packet=8M}. -if (result < 0) /* incorrect */ - ... error ... +If you are using other clients that do not allow you to specify the maximum +packet size (such as @code{DBI}), you need to set the packet size when you +start the server. You cau use a command-line option to @code{mysqld} to set +@code{max_allowed_packet} to a larger size. For example, if you are +expecting to store the full length of a @code{BLOB} into a table, you'll need +to start the server with the @code{--set-variable=max_allowed_packet=16M} +option. -if (result == -1) /* incorrect */ - ... error ... -@end example +@cindex aborted clients +@cindex aborted connection +@cindex connection, aborted +@node Communication errors, Full table, Packet too large, Common errors +@subsection Communication Errors / Aborted Connection -When a function returns an error, the @strong{Errors} subsection of the -function description lists the possible types of errors. You can -find out which of these occurred by calling @code{mysql_errno()}. -A string representation of the error may be obtained by calling -@code{mysql_error()}. +Starting with @code{MySQL 3.23.40} you only get the @code{Aborted +connection} error of you start @code{mysqld} with @code{--warnings}. -@findex @code{mysql_affected_rows()} -@node mysql_affected_rows, mysql_close, C API functions, C API functions -@subsubsection @code{mysql_affected_rows()} +If you find errors like the following in your error log. -@code{my_ulonglong mysql_affected_rows(MYSQL *mysql)} +@example +010301 14:38:23 Aborted connection 854 to db: 'users' user: 'josh' +@end example -@subsubheading Description +@xref{Error log}. -Returns the number of rows changed by the last @code{UPDATE}, deleted by -the last @code{DELETE} or inserted by the last @code{INSERT} -statement. May be called immediately after @code{mysql_query()} for -@code{UPDATE}, @code{DELETE}, or @code{INSERT} statements. For -@code{SELECT} statements, @code{mysql_affected_rows()} works like -@code{mysql_num_rows()}. +This means that something of the following has happened: -@subsubheading Return Values +@itemize @bullet +@item +The client program did not call @code{mysql_close()} before exit. +@item +The client had been sleeping more than @code{wait_timeout} or +@code{interactive_timeout} without doing any requests. @xref{SHOW +VARIABLES}. +@item +The client program ended abruptly in the middle of the transfer. +@end itemize -An integer greater than zero indicates the number of rows affected or -retrieved. Zero indicates that no records where updated for an -@code{UPDATE} statement, no rows matched the @code{WHERE} clause in the -query or that no query has yet been executed. -1 indicates that the -query returned an error or that, for a @code{SELECT} query, -@code{mysql_affected_rows()} was called prior to calling -@code{mysql_store_result()}. +When the above happens, the server variable @code{Aborted_clients} is +incremented. -@subsubheading Errors +The server variable @code{Aborted_connects} is incremented when: -None. +@itemize @bullet +@item +When a connection packet doesn't contain the right information. +@item +When the user didn't have privileges to connect to a database. +@item +When a user uses a wrong password. +@item +When it takes more than @code{connect_timeout} seconds to get +a connect package. +@end itemize -@subsubheading Example +Note that the above could indicate that someone is trying to break into +your database! -@example -mysql_query(&mysql,"UPDATE products SET cost=cost*1.25 WHERE group=10"); -printf("%ld products updated",(long) mysql_affected_rows(&mysql)); -@end example +@xref{SHOW VARIABLES}. -If one specifies the flag @code{CLIENT_FOUND_ROWS} when connecting to -@code{mysqld}, @code{mysql_affected_rows()} will return the number of -rows matched by the @code{WHERE} statement for @code{UPDATE} statements. +Other reasons for problems with Aborted clients / Aborted connections. +@itemize @bullet +@item +Usage of duplex Ethernet protocol, both half and full with +Linux. Many Linux Ethernet drivers have this bug. You should test +for this bug by transferring a huge file via ftp between these two +machines. If a transfer goes in burst-pause-burst-pause ... mode then +you are experiencing a Linux duplex syndrome. The only solution to +this problem is switching of both half and full duplexing on hubs +and switches. +@item +Some problem with the thread library that causes interrupts on reads. +@item +Badly configured TCP/IP. +@item +Faulty Ethernets or hubs or switches, cables ... This can be diagnosed +properly only by replacing hardware. +@end itemize -Note that when one uses a @code{REPLACE} command, -@code{mysql_affected_rows()} will return 2 if the new row replaced and -old row. This is because in this case one row was inserted and then the -duplicate was deleted. -@findex @code{mysql_close()} -@node mysql_close, mysql_connect, mysql_affected_rows, C API functions -@subsubsection @code{mysql_close()} +@cindex table is full +@node Full table, Cannot create, Communication errors, Common errors +@subsection @code{The table is full} Error -@code{void mysql_close(MYSQL *mysql)} +This error occurs in older @strong{MySQL} versions when an in-memory temporary +table becomes larger than @code{tmp_table_size} bytes. To avoid this +problem, you can use the @code{-O tmp_table_size=#} option to +@code{mysqld} to increase the temporary table size or use the SQL +option @code{SQL_BIG_TABLES} before you issue the problematic +query. @xref{SET OPTION, , @code{SET OPTION}}. -@subsubheading Description -Closes a previously opened connection. @code{mysql_close()} also deallocates -the connection handle pointed to by @code{mysql} if the handle was allocated -automatically by @code{mysql_init()} or @code{mysql_connect()}. +You can also start @code{mysqld} with the @code{--big-tables} option. +This is exactly the same as using @code{SQL_BIG_TABLES} for all queries. -@subsubheading Return Values +In @strong{MySQL} Version 3.23, in-memory temporary tables will automatically be +converted to a disk-based @code{MyISAM} table after the table size gets +bigger than @code{tmp_table_size}. -None. +@cindex can't create/write to file +@node Cannot create, Commands out of sync, Full table, Common errors +@subsection @code{Can't create/write to file} Error -@subsubheading Errors +If you get an error for some queries of type: -None. +@example +Can't create/write to file '\\sqla3fe_0.ism'. +@end example -@findex @code{mysql_connect()} -@node mysql_connect, mysql_change_user, mysql_close, C API functions -@subsubsection @code{mysql_connect()} +this means that @strong{MySQL} can't create a temporary file for the +result set in the given temporary directory. (The above error is a +typical error message on Windows, and the Unix error message is similar.) +The fix is to start @code{mysqld} with @code{--tmpdir=path} or to add to your option +file: -@code{MYSQL *mysql_connect(MYSQL *mysql, const char *host, const char *user, const char *passwd)} +@example +[mysqld] +tmpdir=C:/temp +@end example -@subsubheading Description +assuming that the @file{c:\\temp} directory exists. @xref{Option files}. -This function is deprecated. It is preferable to use -@code{mysql_real_connect()} instead. +Check also the error code that you get with @code{perror}. One reason +may also be a disk full error; -@code{mysql_connect()} attempts to establish a connection to a @strong{MySQL} -database engine running on @code{host}. @code{mysql_connect()} must complete -successfully before you can execute any of the other API functions, with the -exception of @code{mysql_get_client_info()}. +@example +shell> perror 28 +Error code 28: No space left on device +@end example -The meanings of the parameters are the same as for the corresponding -parameters for @code{mysql_real_connect()} with the difference that the -connection parameter may be @code{NULL}. In this case the C API -allocates memory for the connection structure automatically and frees it -when you call @code{mysql_close()}. The disadvantage of this approach is -that you can't retrieve an error message if the connection fails. (To -get error information from @code{mysql_errno()} or @code{mysql_error()}, -you must provide a valid @code{MYSQL} pointer.) +@cindex commands out of sync +@node Commands out of sync, Ignoring user, Cannot create, Common errors +@subsection @code{Commands out of sync} Error in Client -@subsubheading Return Values +If you get @code{Commands out of sync; You can't run this command now} +in your client code, you are calling client functions in the wrong order! -Same as for @code{mysql_real_connect()}. +This can happen, for example, if you are using @code{mysql_use_result()} and +try to execute a new query before you have called @code{mysql_free_result()}. +It can also happen if you try to execute two queries that return data without +a @code{mysql_use_result()} or @code{mysql_store_result()} in between. -@subsubheading Errors +@node Ignoring user, Cannot find table, Commands out of sync, Common errors +@subsection @code{Ignoring user} Error -Same as for @code{mysql_real_connect()}. +If you get the following error: -@findex @code{mysql_change_user()} -@node mysql_change_user, mysql_character_set_name, mysql_connect, C API functions -@subsubsection @code{mysql_change_user()} +@code{Found wrong password for user: 'some_user@@some_host'; Ignoring user} -@code{my_bool mysql_change_user(MYSQL *mysql, const char *user, const -char *password, const char *db)} +this means that when @code{mysqld} was started or when it reloaded the +permissions tables, it found an entry in the @code{user} table with +an invalid password. As a result, the entry is simply ignored by the +permission system. -@subsubheading Description +Possible causes of and fixes for this problem: -Changes the user and causes the database specified by @code{db} to -become the default (current) database on the connection specified by -@code{mysql}. In subsequent queries, this database is the default for -table references that do not include an explicit database specifier. +@itemize @bullet +@item +You may be running a new version of @code{mysqld} with an old +@code{user} table. +You can check this by executing @code{mysqlshow mysql user} to see if +the password field is shorter than 16 characters. If so, you can correct this +condition by running the @code{scripts/add_long_password} script. -This function was introduced in @strong{MySQL} Version 3.23.3. +@item +The user has an old password (8 characters long) and you didn't start +@code{mysqld} with the @code{--old-protocol} option. +Update the user in the @code{user} table with a new password or +restart @code{mysqld} with @code{--old-protocol}. -@code{mysql_change_user()} fails unless the connected user can be -authenticated or if he doesn't have permission to use the database. In -this case the user and database are not changed +@item +@findex PASSWORD() +You have specified a password in the @code{user} table without using the +@code{PASSWORD()} function. Use @code{mysql} to update the user in the +@code{user} table with a new password. Make sure to use the @code{PASSWORD()} +function: -The @code{db} parameter may be set to @code{NULL} if you don't want to have a -default database. +@example +mysql> update user set password=PASSWORD('your password') + where user='XXX'; +@end example +@end itemize -@subsubheading Return values +@node Cannot find table, Cannot initialize character set, Ignoring user, Common errors +@subsection @code{Table 'xxx' doesn't exist} Error -Zero for success. Non-zero if an error occurred. +If you get the error @code{Table 'xxx' doesn't exist} or @code{Can't +find file: 'xxx' (errno: 2)}, this means that no table exists +in the current database with the name @code{xxx}. -@subsubheading Errors +Note that as @strong{MySQL} uses directories and files to store databases and +tables, the database and table names are @strong{case sensitive}! +(On Windows the databases and tables names are not case sensitive, but all +references to a given table within a query must use the same case!) -The same that you can get from @code{mysql_real_connect()}. +You can check which tables you have in the current database with +@code{SHOW TABLES}. @xref{SHOW, , @code{SHOW}}. -@table @code -@item CR_COMMANDS_OUT_OF_SYNC -Commands were executed in an improper order. -@item CR_SERVER_GONE_ERROR -The @strong{MySQL} server has gone away. -@item CR_SERVER_LOST -The connection to the server was lost during the query. -@item CR_UNKNOWN_ERROR -An unknown error occurred. -@item ER_UNKNOWN_COM_ERROR -The @strong{MySQL} server doesn't implement this command (probably an old server) -@item ER_ACCESS_DENIED_ERROR -The user or password was wrong. -@item ER_BAD_DB_ERROR -The database didn't exist. -@item ER_DBACCESS_DENIED_ERROR -The user did not have access rights to the database. -@item ER_WRONG_DB_NAME -The database name was too long. -@end table +@cindex multibyte character sets +@node Cannot initialize character set, , Cannot find table, Common errors +@subsection @code{Can@'t initialize character set xxx} error. -@subsubheading Example +If you get an error like: @example -if (mysql_change_user(&mysql, "user", "password", "new_database")) -@{ - fprintf(stderr, "Failed to change user. Error: %s\n", - mysql_error(&mysql)); -@} +MySQL Connection Failed: Can't initialize character set xxx @end example -@findex @code{mysql_character_set_name()} -@node mysql_character_set_name, mysql_create_db, mysql_change_user, C API functions -@subsubsection @code{mysql_character_set_name()} +This means one of the following things: -@code{const char *mysql_character_set_name(MYSQL *mysql)} +@itemize @bullet +@item +The character set is a multi-byte character set and you have not support +for the character set in the client. -@subsubheading Description +In this case you need to recompile the client with +@code{--with-charset=xxx} or with @code{--with-extra-charsets=xxx}. +@xref{configure options}. -Returns the default character set for the current connection. +All standard @strong{MySQL} binaries are compiled with +@code{--with-extra-character-sets=complex} which will enable support for +all multi-byte character sets. @xref{Character sets}. -@subsubheading Return Values +@item +The character set is a simple character set which is not compiled into +@code{mysqld} and the character set definition files is not in the place +where the client expect to find them. -The default character set +In this case you need to: -@subsubheading Errors -None. +@itemize @bullet +@item +Recompile the client with support for the character set. +@xref{configure options}. +@item +Specify to the client where the character set definition files are. For many +client you can do this with the +@code{--character-sets-dir=path-to-charset-dir} option. +@item +Copy the character definition files to the path where the client expect them +to be. +@end itemize +@end itemize +@cindex full disk +@cindex disk full +@node Full disk, Multiple sql commands, Common errors, Problems +@section How MySQL Handles a Full Disk -@findex @code{mysql_create_db()} -@node mysql_create_db, mysql_data_seek, mysql_character_set_name, C API functions -@subsubsection @code{mysql_create_db()} +@noindent +When a disk-full condition occurs, @strong{MySQL} does the following: -@code{int mysql_create_db(MYSQL *mysql, const char *db)} +@itemize @bullet +@item +It checks once every minute to see whether or not there is enough space to +write the current row. If there is enough space, it continues as if nothing had +happened. +@item +Every 6 minutes it writes an entry to the log file warning about the disk +full condition. +@end itemize -@subsubheading Description -Creates the database named by the @code{db} parameter. +@noindent +To alleviate the problem, you can take the following actions: -This function is deprecated. It is preferable to use @code{mysql_query()} -to issue a SQL @code{CREATE DATABASE} statement instead. +@itemize @bullet +@item +To continue, you only have to free enough disk space to insert all records. +@item +To abort the thread, you must send a @code{mysqladmin kill} to the thread. +The thread will be aborted the next time it checks the disk (in 1 minute). +@item +Note that other threads may be waiting for the table that caused the disk +full condition. If you have several ``locked'' threads, killing the one +thread that is waiting on the disk-full condition will allow the other +threads to continue. +@end itemize -@subsubheading Return Values +Exceptions to the above behaveour is when you use @code{REPAIR} or +@code{OPTIMIZE} or when the indexes are created in a batch after an +@code{LOAD DATA INFILE} or after an @code{ALTER TABLE} statement. -Zero if the database was created successfully. Non-zero if an error -occurred. +All of the above commands may use big temporary files that left to +themself would cause big problems for the rest of the system. If +@strong{MySQL} gets disk full while doing any of the above operations, +it will remove the big temporary files and mark the table as crashed +(except for @code{ALTER TABLE}, in which the old table will be left +unchanged). -@subsubheading Errors -@table @code +@node Multiple sql commands, Temporary files, Full disk, Problems +@section How to Run SQL Commands from a Text File -@item CR_COMMANDS_OUT_OF_SYNC -Commands were executed in an improper order. +The @code{mysql} client typically is used interactively, like this: -@item CR_SERVER_GONE_ERROR -The @strong{MySQL} server has gone away. +@example +shell> mysql database +@end example -@item CR_SERVER_LOST -The connection to the server was lost during the query. +However, it's also possible to put your SQL commands in a file and tell +@code{mysql} to read its input from that file. To do so, create a text +file @file{text_file} that contains the commands you wish to execute. +Then invoke @code{mysql} as shown below: -@item CR_UNKNOWN_ERROR -An unknown error occurred. -@end table +@example +shell> mysql database < text_file +@end example -@subsubheading Example +You can also start your text file with a @code{USE db_name} statement. In +this case, it is unnecessary to specify the database name on the command +line: @example -if(mysql_create_db(&mysql, "my_database")) -@{ - fprintf(stderr, "Failed to create new database. Error: %s\n", - mysql_error(&mysql)); -@} +shell> mysql < text_file @end example -@findex @code{mysql_data_seek()} -@node mysql_data_seek, mysql_debug, mysql_create_db, C API functions -@subsubsection @code{mysql_data_seek()} +@xref{Programs}. -@code{void mysql_data_seek(MYSQL_RES *result, unsigned long long offset)} +@node Temporary files, Problems with mysql.sock, Multiple sql commands, Problems +@section Where MySQL Stores Temporary Files -@subsubheading Description -Seeks to an arbitrary row in a query result set. This requires that the -result set structure contains the entire result of the query, so -@code{mysql_data_seek()} may be used in conjunction only with -@code{mysql_store_result()}, not with @code{mysql_use_result()}. +@strong{MySQL} uses the value of the @code{TMPDIR} environment variable as +the pathname of the directory in which to store temporary files. If you don't +have @code{TMPDIR} set, @strong{MySQL} uses the system default, which is +normally @file{/tmp} or @file{/usr/tmp}. If the file system containing your +temporary file directory is too small, you should edit @code{safe_mysqld} to +set @code{TMPDIR} to point to a directory in a file system where you have +enough space! You can also set the temporary directory using the +@code{--tmpdir} option to @code{mysqld}. -The offset should be a value in the range from 0 to -@code{mysql_num_rows(result)-1}. +@strong{MySQL} creates all temporary files as hidden files. This ensures +that the temporary files will be removed if @code{mysqld} is terminated. The +disadvantage of using hidden files is that you will not see a big temporary +file that fills up the file system in which the temporary file directory is +located. -@subsubheading Return Values +When sorting (@code{ORDER BY} or @code{GROUP BY}), @strong{MySQL} normally +uses one or two temporary files. The maximum disk-space needed is: -None. +@example +(length of what is sorted + sizeof(database pointer)) +* number of matched rows +* 2 +@end example -@subsubheading Errors -None. +@code{sizeof(database pointer)} is usually 4, but may grow in the future for +really big tables. -@findex @code{mysql_debug()} -@node mysql_debug, mysql_drop_db, mysql_data_seek, C API functions -@subsubsection @code{mysql_debug()} +For some @code{SELECT} queries, @strong{MySQL} also creates temporary SQL +tables. These are not hidden and have names of the form @file{SQL_*}. -@code{void mysql_debug(char *debug)} +@code{ALTER TABLE} creates a temporary table in the same directory as +the original table. -@subsubheading Description -Does a @code{DBUG_PUSH} with the given string. @code{mysql_debug()} uses the -Fred Fish debug library. To use this function, you must compile the client -library to support debugging. -@xref{Debugging server}. @xref{Debugging client}. - -@subsubheading Return Values - -None. - -@subsubheading Errors -None. - -@subsubheading Example +@cindex @code{mysql.sock}, protection +@cindex deletion, @code{mysql.sock} +@node Problems with mysql.sock, Changing MySQL user, Temporary files, Problems +@section How to Protect @file{/tmp/mysql.sock} from Being Deleted -The call shown below causes the client library to generate a trace file in -@file{/tmp/client.trace} on the client machine: +If you have problems with the fact that anyone can delete the +@strong{MySQL} communication socket @file{/tmp/mysql.sock}, you can, +on most versions of Unix, protect your @file{/tmp} file system by setting +the @code{sticky} bit on it. Log in as @code{root} and do the following: @example -mysql_debug("d:t:O,/tmp/client.trace"); +shell> chmod +t /tmp @end example -@findex @code{mysql_drop_db()} -@node mysql_drop_db, mysql_dump_debug_info, mysql_debug, C API functions -@subsubsection @code{mysql_drop_db()} - -@code{int mysql_drop_db(MYSQL *mysql, const char *db)} - -@subsubheading Description -Drops the database named by the @code{db} parameter. - -This function is deprecated. It is preferable to use @code{mysql_query()} -to issue a SQL @code{DROP DATABASE} statement instead. +This will protect your @file{/tmp} file system so that files can be deleted +only by their owners or the superuser (@code{root}). -@subsubheading Return Values +You can check if the @code{sticky} bit is set by executing @code{ls -ld /tmp}. +If the last permission bit is @code{t}, the bit is set. -Zero if the database was dropped successfully. Non-zero if an error -occurred. +@cindex starting, @code{mysqld} +@cindex @code{mysqld}, starting +@node Changing MySQL user, Resetting permissions, Problems with mysql.sock, Problems +@section How to Run MySQL As a Normal User -@subsubheading Errors +The @strong{MySQL} server @code{mysqld} can be started and run by any user. +In order to change @code{mysqld} to run as a Unix user @code{user_name}, you must +do the following: -@table @code -@item CR_COMMANDS_OUT_OF_SYNC -Commands were executed in an improper order. -@item CR_SERVER_GONE_ERROR -The @strong{MySQL} server has gone away. -@item CR_SERVER_LOST -The connection to the server was lost during the query. -@item CR_UNKNOWN_ERROR -An unknown error occurred. -@end table +@enumerate +@item +Stop the server if it's running (use @code{mysqladmin shutdown}). -@subsubheading Example +@item +Change the database directories and files so that @code{user_name} has +privileges to read and write files in them (you may need to do this as +the Unix @code{root} user): @example -if(mysql_drop_db(&mysql, "my_database")) - fprintf(stderr, "Failed to drop the database: Error: %s\n", - mysql_error(&mysql)); +shell> chown -R user_name /path/to/mysql/datadir @end example -@findex @code{mysql_dump_debug_info()} -@node mysql_dump_debug_info, mysql_eof, mysql_drop_db, C API functions -@subsubsection @code{mysql_dump_debug_info()} +If directories or files within the @strong{MySQL} data directory are +symlinks, you'll also need to follow those links and change the directories +and files they point to. @code{chown -R} may not follow symlinks for +you. -@code{int mysql_dump_debug_info(MYSQL *mysql)} +@item +Start the server as user @code{user_name}, or, if you are using +@strong{MySQL} Version 3.22 or later, start @code{mysqld} as the Unix @code{root} +user and use the @code{--user=user_name} option. @code{mysqld} will switch +to run as the Unix user @code{user_name} before accepting any connections. -@subsubheading Description +@item +To start the server as the given user name automatically at system +startup time, add a @code{user} line that specifies the user name to +the @code{[mysqld]} group of the @file{/etc/my.cnf} option file or the +@file{my.cnf} option file in the server's data directory. For example: -Instructs the server to write some debug information to the log. The -connected user must have the @strong{process} privilege for this to work. +@example +[mysqld] +user=user_name +@end example +@end enumerate -@subsubheading Return values +At this point, your @code{mysqld} process should be running fine and dandy as +the Unix user @code{user_name}. One thing hasn't changed, though: the +contents of the permissions tables. By default (right after running the +permissions table install script @code{mysql_install_db}), the @strong{MySQL} +user @code{root} is the only user with permission to access the @code{mysql} +database or to create or drop databases. Unless you have changed those +permissions, they still hold. This shouldn't stop you from accessing +@strong{MySQL} as the @strong{MySQL} @code{root} user when you're logged in +as a Unix user other than @code{root}; just specify the @code{-u root} option +to the client program. -Zero if the command was successful. Non-zero if an error occurred. +Note that accessing @strong{MySQL} as @code{root}, by supplying @code{-u +root} on the command line, has @emph{nothing} to do with @strong{MySQL} running +as the Unix @code{root} user, or, indeed, as another Unix user. The access +permissions and user names of @strong{MySQL} are completely separate from +Unix user names. The only connection with Unix user names is that if you +don't provide a @code{-u} option when you invoke a client program, the client +will try to connect using your Unix login name as your @strong{MySQL} user +name. -@subsubheading Errors -@table @code -@item CR_COMMANDS_OUT_OF_SYNC -Commands were executed in an improper order. -@item CR_SERVER_GONE_ERROR -The @strong{MySQL} server has gone away. -@item CR_SERVER_LOST -The connection to the server was lost during the query. -@item CR_UNKNOWN_ERROR -An unknown error occurred. -@end table +If your Unix box itself isn't secured, you should probably at least put a +password on the @strong{MySQL} @code{root} users in the access tables. +Otherwise, any user with an account on that machine can run @code{mysql -u +root db_name} and do whatever he likes. -@findex @code{mysql_eof()} -@node mysql_eof, mysql_errno, mysql_dump_debug_info, C API functions -@subsubsection @code{mysql_eof()} +@cindex passwords, forgotten +@cindex passwords, resetting +@cindex root user, password resetting +@node Resetting permissions, File permissions , Changing MySQL user, Problems +@section How to Reset a Forgotten Password -@code{my_bool mysql_eof(MYSQL_RES *result)} +If you have forgotten the @code{root} user password for @strong{MySQL}, you +can restore it with the following procedure: -@subsubheading Description +@enumerate +@item +Take down the @code{mysqld} server by sending a @code{kill} (not @code{kill +-9}) to the @code{mysqld} server. The pid is stored in a @code{.pid} +file, which is normally in the @strong{MySQL} database directory: -This function is deprecated. @code{mysql_errno()} or @code{mysql_error()} -may be used instead. +@example +kill `cat /mysql-data-directory/hostname.pid` +@end example -@code{mysql_eof()} determines whether or not the last row of a result -set has been read. +You must be either the Unix @code{root} user or the same user the server +runs as to do this. -If you acquire a result set from a successful call to -@code{mysql_store_result()}, the client receives the entire set in one -operation. In this case, a @code{NULL} return from @code{mysql_fetch_row()} -always means the end of the result set has been reached and it is -unnecessary to call @code{mysql_eof()}. +@item +Restart @code{mysqld} with the @code{--skip-grant-tables} option. +@item +Connect to the @code{mysqld} server with @code{mysql -h hostname mysql} and change +the password with a @code{GRANT} command. @xref{GRANT,,@code{GRANT}}. +You can also do this with +@code{mysqladmin -h hostname -u user password 'new password'} +@item +Load the privilege tables with: @code{mysqladmin -h hostname +flush-privileges} or with the SQL command @code{FLUSH PRIVILEGES}. +@end enumerate -On the other hand, if you use @code{mysql_use_result()} to initiate a result -set retrieval, the rows of the set are obtained from the server one by one as -you call @code{mysql_fetch_row()} repeatedly. Because an error may occur on -the connection during this process, a @code{NULL} return value from -@code{mysql_fetch_row()} does not necessarily mean the end of the result set -was reached normally. In this case, you can use @code{mysql_eof()} to -determine what happened. @code{mysql_eof()} returns a non-zero value if the -end of the result set was reached and zero if an error occurred. +Note that after you started @code{mysqld} with @code{--skip-grant-tables}, +any usage of @code{GRANT} commands will give you an @code{Unknown command} +error until you have executed @code{FLUSH PRIVILEGES}. -Historically, @code{mysql_eof()} predates the standard @strong{MySQL} error -functions @code{mysql_errno()} and @code{mysql_error()}. Because those error -functions provide the same information, their use is preferred over -@code{mysql_eof()}, which is now deprecated. (In fact, they provide more -information, because @code{mysql_eof()} returns only a boolean value whereas -the error functions indicate a reason for the error when one occurs.) +@cindex files, permissions +@cindex error mesaages, can't find file +@cindex files, not found message +@node File permissions , Not enough file handles, Resetting permissions, Problems +@section Problems with File Permissions -@subsubheading Return Values +If you have problems with file permissions, for example, if @code{mysql} +issues the following error message when you create a table: -Zero if no error occurred. Non-zero if the end of the result set has been -reached. +@example +ERROR: Can't find file: 'path/with/filename.frm' (Errcode: 13) +@end example -@subsubheading Errors -None. +@tindex UMASK environment variable +@tindex Environment variable, UMASK +then the environment variable @code{UMASK} might be set incorrectly when +@code{mysqld} starts up. The default umask value is @code{0660}. You can +change this behavior by starting @code{safe_mysqld} as follows: -@subsubheading Example +@example +shell> UMASK=384 # = 600 in octal +shell> export UMASK +shell> /path/to/safe_mysqld & +@end example -The following example shows how you might use @code{mysql_eof()}: +@tindex UMASK_DIR environment variable +@tindex Environment variable, UMASK_DIR +By default @strong{MySQL} will create database and @code{RAID} +directories with permission type 0700. You can modify this behavior by +setting the @code{UMASK_DIR} variable. If you set this, new +directories are created with the combined @code{UMASK} and +@code{UMASK_DIR}. For example, if you want to give group access to +all new directories, you can do: @example -mysql_query(&mysql,"SELECT * FROM some_table"); -result = mysql_use_result(&mysql); -while((row = mysql_fetch_row(result))) -@{ - // do something with data -@} -if(!mysql_eof(result)) // mysql_fetch_row() failed due to an error -@{ - fprintf(stderr, "Error: %s\n", mysql_error(&mysql)); -@} +shell> UMASK_DIR=504 # = 770 in octal +shell> export UMASK_DIR +shell> /path/to/safe_mysqld & @end example -However, you can achieve the same effect with the standard @strong{MySQL} -error functions: +In @strong{MySQL} Version 3.23.25 and above, @strong{MySQL} assumes that the +value for @code{UMASK} and @code{UMASK_DIR} is in octal if it starts +with a zero. + +@xref{Environment variables}. + +@node Not enough file handles, Using DATE, File permissions , Problems +@section File Not Found + +If you get @code{ERROR '...' not found (errno: 23)}, @code{Can't open +file: ... (errno: 24)}, or any other error with @code{errno 23} or +@code{errno 24} from @strong{MySQL}, it means that you haven't allocated +enough file descriptors for @strong{MySQL}. You can use the +@code{perror} utility to get a description of what the error number +means: @example -mysql_query(&mysql,"SELECT * FROM some_table"); -result = mysql_use_result(&mysql); -while((row = mysql_fetch_row(result))) -@{ - // do something with data -@} -if(mysql_errno(&mysql)) // mysql_fetch_row() failed due to an error -@{ - fprintf(stderr, "Error: %s\n", mysql_error(&mysql)); -@} +shell> perror 23 +File table overflow +shell> perror 24 +Too many open files +shell> perror 11 +Resource temporarily unavailable @end example -@findex @code{mysql_errno()} -@node mysql_errno, mysql_error, mysql_eof, C API functions -@subsubsection @code{mysql_errno()} +The problem here is that @code{mysqld} is trying to keep open too many +files simultaneously. You can either tell @code{mysqld} not to open so +many files at once or increase the number of file descriptors +available to @code{mysqld}. -@code{unsigned int mysql_errno(MYSQL *mysql)} +To tell @code{mysqld} to keep open fewer files at a time, you can make +the table cache smaller by using the @code{-O table_cache=32} option to +@code{safe_mysqld} (the default value is 64). Reducing the value of +@code{max_connections} will also reduce the number of open files (the +default value is 90). -@subsubheading Description +@tindex ulimit +To change the number of file descriptors available to @code{mysqld}, you +can use the option @code{--open-files-limit=#} to @code{safe_mysqld} or +@code{-O open-files-limit=#} to @code{mysqld}. @xref{SHOW VARIABLES}. +The easiest way to do that is to add the option to your option file. +@xref{Option files}. If you have an old @code{mysqld} version that +doesn't support this, you can edit the @code{safe_mysqld} script. There +is a commented-out line @code{ulimit -n 256} in the script. You can +remove the @code{'#'} character to uncomment this line, and change the +number 256 to affect the number of file descriptors available to +@code{mysqld}. -For the connection specified by @code{mysql}, @code{mysql_errno()} returns -the error code for the most recently invoked API function that can succeed -or fail. A return value of zero means that no error occurred. Client error -message numbers are listed in the @strong{MySQL} @file{errmsg.h} header file. -Server error message numbers are listed in @file{mysqld_error.h}. In the -@strong{MySQL} source distribution you can find a complete list of -error messages and error numbers in the file @file{Docs/mysqld_error.txt}. +@code{ulimit} (and @code{open-files-limit}) can increase the number of +file descriptors, but only up to the limit imposed by the operating +system. There is also a 'hard' limit that can only be overrided if you +start @code{safe_mysqld} or @code{mysqld} as root (Just remember that +you need to also use the @code{--user=..} option in this case). If you +need to increase the OS limit on the number of file descriptors +available to each process, consult the documentation for your operating +system. -@subsubheading Return Values +Note that if you run the @code{tcsh} shell, @code{ulimit} will not work! +@code{tcsh} will also report incorrect values when you ask for the current +limits! In this case you should start @code{safe_mysqld} with @code{sh}! -An error code value. Zero if no error occurred. +@findex DATE +@cindex DATE columns, problems +@cindex problems, @code{DATE} columns +@node Using DATE, Timezone problems, Not enough file handles, Problems +@section Problems Using @code{DATE} Columns -@subsubheading Errors -None. +The format of a @code{DATE} value is @code{'YYYY-MM-DD'}. According to ANSI +SQL, no other format is allowed. You should use this format in @code{UPDATE} +expressions and in the WHERE clause of @code{SELECT} statements. For +example: -@findex @code{mysql_error()} -@node mysql_error, mysql_escape_string, mysql_errno, C API functions -@subsubsection @code{mysql_error()} +@example +mysql> SELECT * FROM tbl_name WHERE date >= '1997-05-05'; +@end example -@code{char *mysql_error(MYSQL *mysql)} +As a convenience, @strong{MySQL} automatically converts a date to a number if +the date is used in a numeric context (and vice versa). It is also smart +enough to allow a ``relaxed'' string form when updating and in a @code{WHERE} +clause that compares a date to a @code{TIMESTAMP}, @code{DATE}, or a +@code{DATETIME} column. (Relaxed form means that any punctuation character +may be used as the separator between parts. For example, @code{'1998-08-15'} +and @code{'1998#08#15'} are equivalent.) @strong{MySQL} can also convert a +string containing no separators (such as @code{'19980815'}), provided it +makes sense as a date. -@subsubheading Description +The special date @code{'0000-00-00'} can be stored and retrieved as +@code{'0000-00-00'.} When using a @code{'0000-00-00'} date through +@strong{MyODBC}, it will automatically be converted to @code{NULL} in +@strong{MyODBC} Version 2.50.12 and above, because ODBC can't handle this kind of +date. -For the connection specified by @code{mysql}, @code{mysql_error()} returns -the error message for the most recently invoked API function that can succeed -or fail. An empty string (@code{""}) is returned if no error occurred. -This means the following two tests are equivalent: +Because @strong{MySQL} performs the conversions described above, the following +statements work: @example -if(mysql_errno(&mysql)) -@{ - // an error occurred -@} +mysql> INSERT INTO tbl_name (idate) VALUES (19970505); +mysql> INSERT INTO tbl_name (idate) VALUES ('19970505'); +mysql> INSERT INTO tbl_name (idate) VALUES ('97-05-05'); +mysql> INSERT INTO tbl_name (idate) VALUES ('1997.05.05'); +mysql> INSERT INTO tbl_name (idate) VALUES ('1997 05 05'); +mysql> INSERT INTO tbl_name (idate) VALUES ('0000-00-00'); -if(mysql_error(&mysql)[0] != '\0') -@{ - // an error occurred -@} +mysql> SELECT idate FROM tbl_name WHERE idate >= '1997-05-05'; +mysql> SELECT idate FROM tbl_name WHERE idate >= 19970505; +mysql> SELECT mod(idate,100) FROM tbl_name WHERE idate >= 19970505; +mysql> SELECT idate FROM tbl_name WHERE idate >= '19970505'; @end example -The language of the client error messages may be changed by -recompiling the @strong{MySQL} client library. Currently you can choose -error messages in several different languages. -@xref{Languages}. +@noindent +However, the following will not work: -@subsubheading Return Values +@example +mysql> SELECT idate FROM tbl_name WHERE STRCMP(idate,'19970505')=0; +@end example -A character string that describes the error. An empty string if no error -occurred. +@code{STRCMP()} is a string function, so it converts @code{idate} to +a string and performs a string comparison. It does not convert +@code{'19970505'} to a date and perform a date comparison. -@subsubheading Errors -None. +Note that @strong{MySQL} does no checking whether or not the date is +correct. If you store an incorrect date, such as @code{'1998-2-31'}, the +wrong date will be stored. If the date cannot be converted to any reasonable +value, a @code{0} is stored in the @code{DATE} field. This is mainly a speed +issue and we think it is up to the application to check the dates, and not +the server. -@findex @code{mysql_escape_string()} -@node mysql_escape_string, mysql_fetch_field, mysql_error, C API functions -@subsubsection @code{mysql_escape_string()} +@cindex timezone problems +@cindex problems, timezone +@tindex TZ environment variable +@tindex Environment variable, TZ +@node Timezone problems, Case sensitivity, Using DATE, Problems +@section Time Zone Problems -You should use @code{mysql_real_escape_string()} instead! +If you have a problem with @code{SELECT NOW()} returning values in GMT and +not your local time, you have to set the @code{TZ} environment variable to +your current time zone. This should be done for the environment in which +the server runs, for example, in @code{safe_mysqld} or @code{mysql.server}. +@xref{Environment variables}. -This is identical to @code{mysql_real_escape_string()} except that it -takes the connection as the first -argument. @code{mysql_real_escape_string()} will escape the string -according to the current character set while -@code{mysql_escape_string()} does not respect the current charset -setting. +@cindex case sensitivity, in searches +@cindex searching, and case-sensitivity +@cindex Chinese +@cindex Big5 Chinese character encoding +@node Case sensitivity, Problems with NULL, Timezone problems, Problems +@section Case Sensitivity in Searches -@findex @code{mysql_fetch_field()} -@node mysql_fetch_field, mysql_fetch_fields, mysql_escape_string, C API functions -@subsubsection @code{mysql_fetch_field()} +By default, @strong{MySQL} searches are case-insensitive (although there are +some character sets that are never case insensitive, such as @code{czech}). +That means that if you search with @code{col_name LIKE 'a%'}, you will get all +column values that start with @code{A} or @code{a}. If you want to make this +search case-sensitive, use something like @code{INDEX(col_name, "A")=0} to +check a prefix. Or use @code{STRCMP(col_name, "A") = 0} if the column value +must be exactly @code{"A"}. -@code{MYSQL_FIELD *mysql_fetch_field(MYSQL_RES *result)} +Simple comparison operations (@code{>=, >, = , < , <=}, sorting and +grouping) are based on each character's ``sort value''. Characters with +the same sort value (like E, e and é) are treated as the same character! -@subsubheading Description +In older @strong{MySQL} versions @code{LIKE} comparisons where done on +the uppercase value of each character (E == e but E <> é). In newer +@strong{MySQL} versions @code{LIKE} works just like the other comparison +operators. -Returns the definition of one column of a result set as a @code{MYSQL_FIELD} -structure. Call this function repeatedly to retrieve information about all -columns in the result set. @code{mysql_fetch_field()} returns @code{NULL} -when no more fields are left. +If you want a column always to be treated in case-sensitive fashion, +declare it as @code{BINARY}. @xref{CREATE TABLE, , @code{CREATE TABLE}}. -@code{mysql_fetch_field()} is reset to return information about the first -field each time you execute a new @code{SELECT} query. The field returned by -@code{mysql_fetch_field()} is also affected by calls to -@code{mysql_field_seek()}. +If you are using Chinese data in the so-called big5 encoding, you want to +make all character columns @code{BINARY}. This works because the sorting +order of big5 encoding characters is based on the order of ASCII codes. -If you've called @code{mysql_query()} to perform a @code{SELECT} on a table -but have not called @code{mysql_store_result()}, @strong{MySQL} returns the -default blob length (8K bytes) if you call @code{mysql_fetch_field()} to ask -for the length of a @code{BLOB} field. (The 8K size is chosen because -@strong{MySQL} doesn't know the maximum length for the @code{BLOB}. This -should be made configurable sometime.) Once you've retrieved the result set, -@code{field->max_length} contains the length of the largest value for this -column in the specific query. +@cindex @code{NULL} values, vs. empty values +@tindex NULL +@node Problems with NULL, Problems with alias, Case sensitivity, Problems +@section Problems with @code{NULL} Values -@subsubheading Return Values +The concept of the @code{NULL} value is a common source of confusion for +newcomers to SQL, who often think that @code{NULL} is the same thing as an +empty string @code{''}. This is not the case! For example, the following +statements are completely different: -The @code{MYSQL_FIELD} structure for the current column. @code{NULL} -if no columns are left. +@example +mysql> INSERT INTO my_table (phone) VALUES (NULL); +mysql> INSERT INTO my_table (phone) VALUES (""); +@end example -@subsubheading Errors -None. +Both statements insert a value into the @code{phone} column, but the first +inserts a @code{NULL} value and the second inserts an empty string. The +meaning of the first can be regarded as ``phone number is not known'' and the +meaning of the second can be regarded as ``she has no phone''. -@subsubheading Example +In SQL, the @code{NULL} value is always false in comparison to any +other value, even @code{NULL}. An expression that contains @code{NULL} +always produces a @code{NULL} value unless otherwise indicated in +the documentation for the operators and functions involved in the +expression. All columns in the following example return @code{NULL}: @example -MYSQL_FIELD *field; - -while((field = mysql_fetch_field(result))) -@{ - printf("field name %s\n", field->name); -@} +mysql> SELECT NULL,1+NULL,CONCAT('Invisible',NULL); @end example -@findex @code{mysql_fetch_fields()} -@node mysql_fetch_fields, mysql_fetch_field_direct, mysql_fetch_field, C API functions -@subsubsection @code{mysql_fetch_fields()} +If you want to search for column values that are @code{NULL}, you +cannot use the @code{=NULL} test. The following statement returns no +rows, because @code{expr = NULL} is FALSE, for any expression: -@code{MYSQL_FIELD *mysql_fetch_fields(MYSQL_RES *result)} +@example +mysql> SELECT * FROM my_table WHERE phone = NULL; +@end example -@subsubheading Description - -Returns an array of all @code{MYSQL_FIELD} structures for a result set. -Each structure provides the field definition for one column of the result -set. - -@subsubheading Return Values - -An array of @code{MYSQL_FIELD} structures for all columns of a result set. - -@subsubheading Errors -None. - -@subsubheading Example +To look for @code{NULL} values, you must use the @code{IS NULL} test. +The following shows how to find the @code{NULL} phone number and the +empty phone number: @example -unsigned int num_fields; -unsigned int i; -MYSQL_FIELD *fields; - -num_fields = mysql_num_fields(result); -fields = mysql_fetch_fields(result); -for(i = 0; i < num_fields; i++) -@{ - printf("Field %u is %s\n", i, fields[i].name); -@} +mysql> SELECT * FROM my_table WHERE phone IS NULL; +mysql> SELECT * FROM my_table WHERE phone = ""; @end example -@findex @code{mysql_fetch_field_direct()} -@node mysql_fetch_field_direct, mysql_fetch_lengths, mysql_fetch_fields, C API functions -@subsubsection @code{mysql_fetch_field_direct()} - -@code{MYSQL_FIELD *mysql_fetch_field_direct(MYSQL_RES *result, unsigned int fieldnr)} +In @strong{MySQL}, as in many other SQL servers, you can't index +columns that can have @code{NULL} values. You must declare such columns +@code{NOT NULL}. Conversely, you cannot insert @code{NULL} into an indexed +column. -@subsubheading Description +@findex LOAD DATA INFILE +When reading data with @code{LOAD DATA INFILE}, empty columns are updated +with @code{''}. If you want a @code{NULL} value in a column, you should use +@code{\N} in the text file. The literal word @code{'NULL'} may also be used +under some circumstances. +@xref{LOAD DATA, , @code{LOAD DATA}}. -Given a field number @code{fieldnr} for a column within a result set, returns -that column's field definition as a @code{MYSQL_FIELD} structure. You may use -this function to retrieve the definition for an arbitrary column. The value -of @code{fieldnr} should be in the range from 0 to -@code{mysql_num_fields(result)-1}. +When using @code{ORDER BY}, @code{NULL} values are presented first. If you +sort in descending order using @code{DESC}, @code{NULL} values are presented +last. When using @code{GROUP BY}, all @code{NULL} values are regarded as +equal. -@subsubheading Return Values +To help with @code{NULL} handling, you can use the @code{IS NULL} and +@code{IS NOT NULL} operators and the @code{IFNULL()} function. -The @code{MYSQL_FIELD} structure for the specified column. +@cindex @code{TIMESTAMP}, and @code{NULL} values +@cindex @code{AUTO_INCREMENT}, and @code{NULL} values +@cindex @code{NULL} values, and @code{TIMESTAMP} columns +@cindex @code{NULL} values, and @code{AUTO_INCREMENT} columns +For some column types, @code{NULL} values are handled specially. If you +insert @code{NULL} into the first @code{TIMESTAMP} column of a table, the +current date and time is inserted. If you insert @code{NULL} into an +@code{AUTO_INCREMENT} column, the next number in the sequence is inserted. -@subsubheading Errors -None. +@tindex alias +@node Problems with alias, Deleting from related tables, Problems with NULL, Problems +@section Problems with @code{alias} -@subsubheading Example +You can use an alias to refer to a column in the @code{GROUP BY}, +@code{ORDER BY}, or in the @code{HAVING} part. Aliases can also be used +to give columns better names: @example -unsigned int num_fields; -unsigned int i; -MYSQL_FIELD *field; - -num_fields = mysql_num_fields(result); -for(i = 0; i < num_fields; i++) -@{ - field = mysql_fetch_field_direct(result, i); - printf("Field %u is %s\n", i, field->name); -@} +SELECT SQRT(a*b) as rt FROM table_name GROUP BY rt HAVING rt > 0; +SELECT id,COUNT(*) AS cnt FROM table_name GROUP BY id HAVING cnt > 0; +SELECT id AS "Customer identity" FROM table_name; @end example -@findex @code{mysql_fetch_lengths()} -@node mysql_fetch_lengths, mysql_fetch_row, mysql_fetch_field_direct, C API functions -@subsubsection @code{mysql_fetch_lengths()} - -@code{unsigned long *mysql_fetch_lengths(MYSQL_RES *result)} - -@subsubheading Description - -Returns the lengths of the columns of the current row within a result set. -If you plan to copy field values, this length information is also useful for -optimization, because you can avoid calling @code{strlen()}. In addition, if -the result set contains binary data, you @emph{must} use this function to -determine the size of the data, because @code{strlen()} returns incorrect -results for any field containing null characters. - -The length for empty columns and for columns containing @code{NULL} values is -zero. To see how to distinguish these two cases, see the description for -@code{mysql_fetch_row()}. +Note that ANSI SQL doesn't allow you to refer to an alias in a +@code{WHERE} clause. This is because when the @code{WHERE} code is +executed the column value may not yet be determined. For example, the +following query is @strong{illegal}: -@subsubheading Return Values +@example +SELECT id,COUNT(*) AS cnt FROM table_name WHERE cnt > 0 GROUP BY id; +@end example -An array of unsigned long integers representing the size of each column (not -including any terminating null characters). -@code{NULL} if an error occurred. +The @code{WHERE} statement is executed to determine which rows should +be included in the @code{GROUP BY} part while @code{HAVING} is used to +decide which rows from the result set should be used. -@subsubheading Errors -@code{mysql_fetch_lengths()} is valid only for the current row of the result -set. It returns @code{NULL} if you call it before calling -@code{mysql_fetch_row()} or after retrieving all rows in the result. +@cindex deleting, rows +@cindex rows, deleting +@cindex tables, deleting rows +@node Deleting from related tables, No matching rows, Problems with alias, Problems +@section Deleting Rows from Related Tables -@subsubheading Example +As @strong{MySQL} doesn't support sub-selects or use of more than one table +in the @code{DELETE} statement, you should use the following approach to +delete rows from 2 related tables: -@example -MYSQL_ROW row; -unsigned long *lengths; -unsigned int num_fields; -unsigned int i; +@enumerate +@item +@code{SELECT} the rows based on some @code{WHERE} condition in the main table. +@item +@code{DELETE} the rows in the main table based on the same condition. +@item +@code{DELETE FROM related_table WHERE related_column IN (selected_rows)}. +@end enumerate -row = mysql_fetch_row(result); -if (row) -@{ - num_fields = mysql_num_fields(result); - lengths = mysql_fetch_lengths(result); - for(i = 0; i < num_fields; i++) - @{ - printf("Column %u is %lu bytes in length.\n", i, lengths[i]); - @} -@} -@end example +If the total number of characters in the query with +@code{related_column} is more than 1,048,576 (the default value of +@code{max_allowed_packet}, you should split it into smaller parts and +execute multiple @code{DELETE} statements. You will probably get the +fastest @code{DELETE} by only deleting 100-1000 @code{related_column} +id's per query if the @code{related_column} is an index. If the +@code{related_column} isn't an index, the speed is independent of the +number of arguments in the @code{IN} clause. -@findex @code{mysql_fetch_row()} -@node mysql_fetch_row, mysql_field_count, mysql_fetch_lengths, C API functions -@subsubsection @code{mysql_fetch_row()} +@cindex no matching rows +@cindex rows, matching problems +@node No matching rows, ALTER TABLE problems, Deleting from related tables, Problems +@section Solving Problems with No Matching Rows -@code{MYSQL_ROW mysql_fetch_row(MYSQL_RES *result)} +If you have a complicated query that has many tables and that doesn't +return any rows, you should use the following procedure to find out what +is wrong with your query: -@subsubheading Description +@enumerate +@item +Test the query with @code{EXPLAIN} and check if you can find something that is +obviously wrong. @xref{EXPLAIN, , @code{EXPLAIN}}. -Retrieves the next row of a result set. When used after -@code{mysql_store_result()}, @code{mysql_fetch_row()} returns @code{NULL} -when there are no more rows to retrieve. When used after -@code{mysql_use_result()}, @code{mysql_fetch_row()} returns @code{NULL} when -there are no more rows to retrieve or if an error occurred. +@item +Select only those fields that are used in the @code{WHERE} clause. -The number of values in the row is given by @code{mysql_num_fields(result)}. -If @code{row} holds the return value from a call to @code{mysql_fetch_row()}, -pointers to the values are accessed as @code{row[0]} to -@code{row[mysql_num_fields(result)-1]}. @code{NULL} values in the row are -indicated by @code{NULL} pointers. +@item +Remove one table at a time from the query until it returns some rows. +If the tables are big, it's a good idea to use @code{LIMIT 10} with the query. -The lengths of the field values in the row may be obtained by calling -@code{mysql_fetch_lengths()}. Empty fields and fields containing -@code{NULL} both have length 0; you can distinguish these by checking -the pointer for the field value. If the pointer is @code{NULL}, the field -is @code{NULL}; otherwise the field is empty. +@item +Do a @code{SELECT} for the column that should have matched a row against +the table that was last removed from the query. -@subsubheading Return Values +@item +If you are comparing @code{FLOAT} or @code{DOUBLE} columns with numbers that +have decimals, you can't use @code{=}! This problem is common in most +computer languages because floating-point values are not exact values: -A @code{MYSQL_ROW} structure for the next row. @code{NULL} if -there are no more rows to retrieve or if an error occurred. +@example +mysql> SELECT * FROM table_name WHERE float_column=3.5; + -> +mysql> SELECT * FROM table_name WHERE float_column between 3.45 and 3.55; +@end example -@subsubheading Errors +In most cases, changing the @code{FLOAT} to a @code{DOUBLE} will fix this! -@table @code -@item CR_SERVER_LOST -The connection to the server was lost during the query. -@item CR_UNKNOWN_ERROR -An unknown error occurred. -@end table +@item +If you still can't figure out what's wrong, create a minimal test that can +be run with @code{mysql test < query.sql} that shows your problems. +You can create a test file with @code{mysqldump --quick database tables > query.sql}. Open the file in an editor, remove some insert lines (if there are +too many of these), and add your select statement at the end of the file. -@subsubheading Example +Test that you still have your problem by doing: @example -MYSQL_ROW row; -unsigned int num_fields; -unsigned int i; - -num_fields = mysql_num_fields(result); -while ((row = mysql_fetch_row(result))) -@{ - unsigned long *lengths; - lengths = mysql_fetch_lengths(result); - for(i = 0; i < num_fields; i++) - @{ - printf("[%.*s] ", (int) lengths[i], row[i] ? row[i] : "NULL"); - @} - printf("\n"); -@} +shell> mysqladmin create test2 +shell> mysql test2 < query.sql @end example -@findex @code{mysql_field_count()} -@node mysql_field_count, mysql_field_seek, mysql_fetch_row, C API functions -@subsubsection @code{mysql_field_count()} +Post the test file using @code{mysqlbug} to @email{mysql@@lists.mysql.com}. +@end enumerate -@code{unsigned int mysql_field_count(MYSQL *mysql)} +@tindex ALTER TABLE +@node ALTER TABLE problems, Change column order, No matching rows, Problems +@section Problems with @code{ALTER TABLE}. -If you are using a version of @strong{MySQL} earlier than Version 3.22.24, you -should use @code{unsigned int mysql_num_fields(MYSQL *mysql)} instead. +@code{ALTER TABLE} changes a table to the current character set. +If you during @code{ALTER TABLE} get a duplicate key error, then the cause +is either that the new character sets maps to keys to the same value +or that the table is corrupted, in which case you should run +@code{REPAIR TABLE} on the table. -@subsubheading Description +If @code{ALTER TABLE} dies with an error like this: -Returns the number of columns for the most recent query on the connection. +@example +Error on rename of './database/name.frm' to './database/B-a.frm' (Errcode: 17) +@end example -The normal use of this function is when @code{mysql_store_result()} -returned @code{NULL} (and thus you have no result set pointer). -In this case, you can call @code{mysql_field_count()} to -determine whether or not @code{mysql_store_result()} should have produced a -non-empty result. This allows the client program to take proper action -without knowing whether or not the query was a @code{SELECT} (or -@code{SELECT}-like) statement. The example shown below illustrates how this -may be done. +the problem may be that @strong{MySQL} has crashed in a previous @code{ALTER +TABLE} and there is an old table named @file{A-something} or +@file{B-something} lying around. In this case, go to the @strong{MySQL} data +directory and delete all files that have names starting with @code{A-} or +@code{B-}. (You may want to move them elsewhere instead of deleting them.) -@xref{NULL mysql_store_result, , @code{NULL mysql_store_result()}}. +@code{ALTER TABLE} works the following way: -@subsubheading Return Values +@itemize @bullet +@item Create a new table named @file{A-xxx} with the requested changes. +@item All rows from the old table are copied to @file{A-xxx}. +@item The old table is renamed @file{B-xxx}. +@item @file{A-xxx} is renamed to your old table name. +@item @file{B-xxx} is deleted. +@end itemize -An unsigned integer representing the number of fields in a result set. +If something goes wrong with the renaming operation, @strong{MySQL} tries to +undo the changes. If something goes seriously wrong (this shouldn't happen, +of course), @strong{MySQL} may leave the old table as @file{B-xxx}, but a +simple rename on the system level should get your data back. -@subsubheading Errors -None. +@cindex reordering, columns +@cindex columns, changing +@cindex changing, column order +@cindex tables, changing column order +@node Change column order, Temporary table problems, ALTER TABLE problems, Problems +@section How To Change the Order of Columns in a Table -@subsubheading Example +The whole point of SQL is to abstract the application from the data +storage format. You should always specify the order in which you wish to +retrieve your data. For example: @example -MYSQL_RES *result; -unsigned int num_fields; -unsigned int num_rows; - -if (mysql_query(&mysql,query_string)) -@{ - // error -@} -else // query succeeded, process any data returned by it -@{ - result = mysql_store_result(&mysql); - if (result) // there are rows - @{ - num_fields = mysql_num_fields(result); - // retrieve rows, then call mysql_free_result(result) - @} - else // mysql_store_result() returned nothing; should it have? - @{ - if(mysql_field_count(&mysql) == 0) - @{ - // query does not return data - // (it was not a SELECT) - num_rows = mysql_affected_rows(&mysql); - @} - else // mysql_store_result() should have returned data - @{ - fprintf(stderr, "Error: %s\n", mysql_error(&mysql)); - @} - @} -@} +SELECT col_name1, col_name2, col_name3 FROM tbl_name; @end example -An alternative is to replace the @code{mysql_field_count(&mysql)} call with -@code{mysql_errno(&mysql)}. In this case, you are checking directly for an -error from @code{mysql_store_result()} rather than inferring from the value -of @code{mysql_field_count()} whether or not the statement was a -@code{SELECT}. +will return columns in the order @code{col_name1}, @code{col_name2}, @code{col_name3}, whereas: -@findex @code{mysql_field_seek()} -@node mysql_field_seek, mysql_field_tell, mysql_field_count, C API functions -@subsubsection @code{mysql_field_seek()} +@example +SELECT col_name1, col_name3, col_name2 FROM tbl_name; +@end example -@code{MYSQL_FIELD_OFFSET mysql_field_seek(MYSQL_RES *result, MYSQL_FIELD_OFFSET offset)} +will return columns in the order @code{col_name1}, @code{col_name3}, @code{col_name2}. -@subsubheading Description +You should @strong{NEVER}, in an application, use @code{SELECT *} and +retrieve the columns based on their position, because the order in which +columns are returned @strong{CANNOT} be guaranteed over time. A simple +change to your database may cause your application to fail rather +dramatically. -Sets the field cursor to the given offset. The next call to -@code{mysql_fetch_field()} will retrieve the field definition of the column -associated with that offset. +If you want to change the order of columns anyway, you can do it as follows: -To seek to the beginning of a row, pass an @code{offset} value of zero. +@enumerate +@item +Create a new table with the columns in the right order. +@item +Execute +@code{INSERT INTO new_table SELECT fields-in-new_table-order FROM old_table}. +@item +Drop or rename @code{old_table}. +@item +@code{ALTER TABLE new_table RENAME old_table}. +@end enumerate -@subsubheading Return Values +@cindex temporary tables, problems +@node Temporary table problems, , Change column order, Problems +@section TEMPORARY TABLE problems -The previous value of the field cursor. +The following are a list of the limitations with @code{TEMPORARY TABLES}. -@subsubheading Errors -None. +@itemize @bullet +@item +A temporary table can only be of type @code{HEAP}, @code{ISAM} or +@code{MyISAM}. +@item +You can't use temporary tables more than once in the same query. +For example, the following doesn't work. -@findex @code{mysql_field_tell()} -@node mysql_field_tell, mysql_free_result, mysql_field_seek, C API functions -@subsubsection @code{mysql_field_tell()} +@example +select * from temporary_table, temporary_table as t2; +@end example -@code{MYSQL_FIELD_OFFSET mysql_field_tell(MYSQL_RES *result)} +We plan to fix the above in 4.0. +@item +You can't use @code{RENAME} on a @code{TEMPORARY} table. +Note that @code{ALTER TABLE org_name RENAME new_name} works! -@subsubheading Description +We plan to fix the above in 4.0. +@end itemize -Returns the position of the field cursor used for the last -@code{mysql_fetch_field()}. This value can be used as an argument to -@code{mysql_field_seek()}. +@cindex problems, solving +@cindex solving, problems +@cindex databases, replicating +@node Common problems, Log files, Problems, Top +@chapter Solving Some Common Problems with MySQL -@subsubheading Return Values +@cindex replication +@menu +* Log Replication:: Database replication with update log +* Backup:: Database backups +* Multiple servers:: Running multiple @strong{MySQL} servers on the same machine +@end menu -The current offset of the field cursor. +In this chapter, you will find information to solve some of the more common +tasks with @strong{MySQL}. This includes making backups, running more than +one @strong{MySQL} server daemon on a single machine, and replicating a +database using the update or binary logs. -@subsubheading Errors -None. +@cindex database replication +@cindex replication, database +@node Log Replication, Backup, Common problems, Common problems +@section Database Replication with Update Log -@findex @code{mysql_free_result()} -@node mysql_free_result, mysql_get_client_info, mysql_field_tell, C API functions -@subsubsection @code{mysql_free_result()} +Now that master-slave internal replication is available starting in +Version 3.23.15, using the update log to implement replications is not +recommended. @xref{Replication}. -@code{void mysql_free_result(MYSQL_RES *result)} +However, it is still possible to replicate a database by using the +update log or the binary log. @xref{Update log}. This requires one +database that acts as a master (to which data changes are made) and one +or more other databases that act as slaves. To update a slave, just run +@code{mysql < update_log.*} or @code{mysqlbinlog binary_log.* | mysql}. +Supply host, user, and password options that are appropriate for the +slave database, and use the update log from the master database as +input. -@subsubheading Description +If you never delete anything from a table, you can use a @code{TIMESTAMP} +column to find out which rows have been inserted or changed in the table +since the last replication (by comparing the time when you did the +replication last time) and only copy these rows to the mirror. -Frees the memory allocated for a result set by @code{mysql_store_result()}, -@code{mysql_use_result()}, @code{mysql_list_dbs()}, etc. When you are done -with a result set, you must free the memory it uses by calling -@code{mysql_free_result()}. +It is possible to make a two-way updating system using both the update +log (for deletes) and timestamps (on both sides). But in that case you +must be able to handle conflicts when the same data have been changed in +both ends. You probably want to keep the old version to help with +deciding what has been updated. -@subsubheading Return Values +Because replication in this case is done with SQL statements, you should not +use the following functions in statements that update the database; they may +not return the same value as in the original database: -None. +@itemize @bullet +@item @code{DATABASE()} +@item @code{GET_LOCK()} and @code{RELEASE_LOCK()} +@item @code{RAND()} +@item @code{USER()}, @code{SYSTEM_USER()} or @code{SESSION_USER()} +@item @code{VERSION()}, @code{CONNECT_ID()} +@end itemize -@subsubheading Errors -None. +All time functions are safe to use, as the timestamp is sent to the +mirror if needed. @code{LAST_INSERT_ID()} is also safe to use. -@findex @code{mysql_get_client_info()} -@node mysql_get_client_info, mysql_get_host_info, mysql_free_result, C API functions -@subsubsection @code{mysql_get_client_info()} +@cindex databases, backups +@cindex backups +@node Backup, Multiple servers, Log Replication, Common problems +@section Database Backups -@code{char *mysql_get_client_info(void)} +Because @strong{MySQL} tables are stored as files, it is easy to do a +backup. To get a consistent backup, do a @code{LOCK TABLES} on the +relevant tables followed by @code{FLUSH TABLES} for the tables. +@xref{LOCK TABLES, , @code{LOCK TABLES}}. +@xref{FLUSH, , @code{FLUSH}}. +You only need a read lock; this allows other threads to continue to +query the tables while you are making a copy of the files in the +database directory. The @code{FLUSH TABLE} is needed to ensure that +the all active index pages is written to disk before you start the backup. -@subsubheading Description +If you want to make a SQL level backup of a table, you can use +@code{SELECT INTO OUTFILE} or @code{BACKUP TABLE}. @xref{SELECT}. +@xref{BACKUP TABLE}. -Returns a string that represents the client library version. +Another way to back up a database is to use the @code{mysqldump} program or +the @code{mysqlhotcopy script}. @xref{mysqldump, , @code{mysqldump}}. +@xref{mysqlhotcopy, , @code{mysqlhotcopy}}. -@subsubheading Return Values +@enumerate +@item +Do a full backup of your databases: -A character string that represents the @strong{MySQL} client library version. +@example +shell> mysqldump --tab=/path/to/some/dir --opt --full -@subsubheading Errors -None. +or -@findex @code{mysql_get_host_info()} -@node mysql_get_host_info, mysql_get_proto_info, mysql_get_client_info, C API functions -@subsubsection @code{mysql_get_host_info()} +shell> mysqlhotcopy database /path/to/some/dir +@end example -@code{char *mysql_get_host_info(MYSQL *mysql)} +You can also simply copy all table files (@file{*.frm}, @file{*.MYD}, and +@file{*.MYI} files) as long as the server isn't updating anything. +The script @code{mysqlhotcopy} does use this method. -@subsubheading Description +@item +@cindex log files, names +Stop @code{mysqld} if it's running, then start it with the +@code{--log-update[=file_name]} option. @xref{Update log}. The update +log file(s) provide you with the information you need to replicate +changes to the database that are made subsequent to the point at which +you executed @code{mysqldump}. +@end enumerate -Returns a string describing the type of connection in use, including the -server host name. +If you have to restore something, try to recover your tables using +@code{REPAIR TABLE} or @code{myisamchk -r} first. That should work in +99.9% of all cases. If @code{myisamchk} fails, try the following +procedure: (This will only work if you have started @strong{MySQL} with +@code{--log-update}. @xref{Update log}.): -@subsubheading Return Values +@enumerate +@item +Restore the original @code{mysqldump} backup. +@item +Execute the following command to re-run the updates in the binary log: -A character string representing the server host name and the connection type. +@example +shell> mysqlbinlog hostname-bin.[0-9]* | mysql +@end example -@subsubheading Errors -None. +If you are using the update log you can use: -@findex @code{mysql_get_proto_info()} -@node mysql_get_proto_info, mysql_get_server_info, mysql_get_host_info, C API functions -@subsubsection @code{mysql_get_proto_info()} +@example +shell> ls -1 -t -r hostname.[0-9]* | xargs cat | mysql +@end example +@end enumerate -@code{unsigned int mysql_get_proto_info(MYSQL *mysql)} +@code{ls} is used to get all the update log files in the right order. -@subsubheading Description +You can also do selective backups with @code{SELECT * INTO OUTFILE 'file_name' +FROM tbl_name} and restore with @code{LOAD DATA INFILE 'file_name' REPLACE +...} To avoid duplicate records, you need a @code{PRIMARY KEY} or a +@code{UNIQUE} key in the table. The @code{REPLACE} keyword causes old records +to be replaced with new ones when a new record duplicates an old record on +a unique key value. -Returns the protocol version used by current connection. +If you get performance problems in making backups on your system, you can +solve this by setting up replication and do the backups on the slave +instead of on the master. @xref{Replication Intro}. -@subsubheading Return Values +If you are using a Veritas file system, you can do: -An unsigned integer representing the protocol version used by the current -connection. +@enumerate +@item +Execute in a client (perl ?) @code{FLUSH TABLES WITH READ LOCK} +@item +Fork a shell or execute in another client @code{mount vxfs snapshot}. +@item +Execute in the first client @code{UNLOCK TABLES} +@item +Copy files from snapshot +@item +Unmount snapshot +@end enumerate -@subsubheading Errors -None. +@cindex multiple servers +@cindex servers, multiple +@cindex running, multiple servers +@node Multiple servers, , Backup, Common problems +@section Running Multiple MySQL Servers on the Same Machine -@findex @code{mysql_get_server_info()} -@node mysql_get_server_info, mysql_info, mysql_get_proto_info, C API functions -@subsubsection @code{mysql_get_server_info()} +There are circumstances when you might want to run multiple servers on the same +machine. For example, you might want to test a new @strong{MySQL} release +while leaving your existing production setup undisturbed. Or you might +be an Internet service provider that wants to provide independent +@strong{MySQL} installations for different customers. -@code{char *mysql_get_server_info(MYSQL *mysql)} +If you want to run multiple servers, the easiest way is to compile the servers +with different TCP/IP ports and socket files so they are not +both listening to the same TCP/IP port or socket file. @xref{mysqld_multi, , +@code{mysqld_multi}}. -@subsubheading Description +Assume an existing server is configured for the default port number and +socket file. Then configure the new server with a @code{configure} command +something like this: -Returns a string that represents the server version number. +@example +shell> ./configure --with-tcp-port=port_number \ + --with-unix-socket-path=file_name \ + --prefix=/usr/local/mysql-3.22.9 +@end example -@subsubheading Return Values +Here @code{port_number} and @code{file_name} should be different than the +default port number and socket file pathname, and the @code{--prefix} value +should specify an installation directory different than the one under which +the existing @strong{MySQL} installation is located. -A character string that represents the server version number. +You can check the socket used by any currently executing @strong{MySQL} server +with this command: -@subsubheading Errors -None. +@example +shell> mysqladmin -h hostname --port=port_number variables +@end example -@findex @code{mysql_info()} -@node mysql_info, mysql_init, mysql_get_server_info, C API functions -@subsubsection @code{mysql_info()} +Note that if you specify ``@code{localhost}'' as a hostname, @code{mysqladmin} +will default to using Unix sockets instead of TCP/IP. -@code{char *mysql_info(MYSQL *mysql)} +If you have a @strong{MySQL} server running on the port you used, you will +get a list of some of the most important configurable variables in +@strong{MySQL}, including the socket name. -@subsubheading Description +You don't have to recompile a new @strong{MySQL} server just to start with +a different port and socket. You can change the port and socket to be used +by specifying them at run time as options to @code{safe_mysqld}: -Retrieves a string providing information about the most recently executed -query, but only for the statements listed below. For other statements, -@code{mysql_info()} returns @code{NULL}. The format of the string varies -depending on the type of query, as described below. The numbers are -illustrative only; the string will contain values appropriate for the query. +@example +shell> /path/to/safe_mysqld --socket=file_name --port=port_number +@end example -@table @code -@item INSERT INTO ... SELECT ... -String format: @code{Records: 100 Duplicates: 0 Warnings: 0} -@item INSERT INTO ... VALUES (...),(...),(...)... -String format: @code{Records: 3 Duplicates: 0 Warnings: 0} -@item LOAD DATA INFILE ... -String format: @code{Records: 1 Deleted: 0 Skipped: 0 Warnings: 0} -@item ALTER TABLE -String format: @code{Records: 3 Duplicates: 0 Warnings: 0} -@item UPDATE -String format: @code{Rows matched: 40 Changed: 40 Warnings: 0} -@end table +@code{mysqld_multi} can also take @code{safe_mysqld} (or @code{mysqld}) +as an argument and pass the options from a configuration file to +@code{safe_mysqld} and further to @code{mysqld}. -Note that @code{mysql_info()} returns a non-@code{NULL} value for the -@code{INSERT ... VALUES} statement only if multiple value lists are -specified in the statement. +If you run the new server on the same database directory as another +server with logging enabled, you should also specify the name of the log +files to @code{safe_mysqld} with @code{--log}, @code{--log-update}, or +@code{--log-slow-queries}. Otherwise, both servers may be trying to +write to the same log file. -@subsubheading Return Values +@strong{WARNING}: Normally you should never have two servers that update +data in the same database! If your OS doesn't support fault-free system +locking, this may lead to unpleasant surprises! -A character string representing additional information about the most -recently executed query. @code{NULL} if no information is available for the -query. +If you want to use another database directory for the second server, you +can use the @code{--datadir=path} option to @code{safe_mysqld}. -@subsubheading Errors -None. +@strong{NOTE} also that starting several @strong{MySQL} servers +(@code{mysqlds}) in different machines and letting them access one data +directory over @code{NFS} is generally a @strong{BAD IDEA}! The problem +is that the @code{NFS} will become the bottleneck with the speed. It is +not meant for such use. And last but not least, you would still have to +come up with a solution how to make sure that two or more @code{mysqlds} +are not interfering with each other. At the moment there is no platform +that would 100% reliable do the file locking (@code{lockd} daemon +usually) in every situation. Yet there would be one more possible risk +with @code{NFS}; it would make the work even more complicated for +@code{lockd} daemon to handle. So make it easy for your self and forget +about the idea. The working solution is to have one computer with an +operating system that efficiently handles threads and have several CPUs +in it. -@findex @code{mysql_init()} -@node mysql_init, mysql_insert_id, mysql_info, C API functions -@subsubsection @code{mysql_init()} +When you want to connect to a @strong{MySQL} server that is running with +a different port than the port that is compiled into your client, you +can use one of the following methods: -@code{MYSQL *mysql_init(MYSQL *mysql)} +@itemize @bullet +@item +Start the client with @code{--host 'hostname' --port=port_number} to connect +with TCP/IP, or @code{[--host localhost] --socket=file_name} to connect via +a Unix socket. -@subsubheading Description +@item +In your C or Perl programs, you can give the port or socket arguments +when connecting to the @strong{MySQL} server. -Allocates or initializes a @code{MYSQL} object suitable for -@code{mysql_real_connect()}. If @code{mysql} is a @code{NULL} pointer, the -function allocates, initializes, and returns a new object. Otherwise the -object is initialized and the address of the object is returned. If -@code{mysql_init()} allocates a new object, it will be freed when -@code{mysql_close()} is called to close the connection. +@item +If your are using the Perl @code{DBD::mysql} module you can read the options +from the @strong{MySQL} option files. @xref{Option files}. -@subsubheading Return Values +@example +$dsn = "DBI:mysql:test;mysql_read_default_group=client;mysql_read_default_file=/usr/local/mysql/data/my.cnf" +$dbh = DBI->connect($dsn, $user, $password); +@end example -An initialized @code{MYSQL*} handle. @code{NULL} if there was -insufficient memory to allocate a new object. +@item +@tindex MYSQL_UNIX_PORT environment variable +@tindex MYSQL_TCP_PORT environment variable +@tindex environment variable, MYSQL_UNIX_PORT +@tindex environment variable, MYSQL_TCP_PORT +Set the @code{MYSQL_UNIX_PORT} and @code{MYSQL_TCP_PORT} environment variables +to point to the Unix socket and TCP/IP port before you start your clients. +If you normally use a specific socket or port, you should place commands +to set these environment variables in your @file{.login} file. +@xref{Environment variables}. @xref{Programs}. -@subsubheading Errors -In case of insufficient memory, @code{NULL} is returned. +@item +@tindex .my.cnf file +Specify the default socket and TCP/IP port in the @file{.my.cnf} file in your +home directory. @xref{Option files}. +@end itemize -@findex @code{mysql_insert_id()} -@node mysql_insert_id, mysql_kill, mysql_init, C API functions -@subsubsection @code{mysql_insert_id()} +@cindex Log files +@node Log files, Clients, Common problems, Top +@chapter The MySQL log files -@code{my_ulonglong mysql_insert_id(MYSQL *mysql)} +@strong{MySQL} has several different log files that can help you find +out what's going on inside @code{mysqld}: -@subsubheading Description +@multitable @columnfractions .3 .7 +@item The error log @tab Problems encountering starting, running or stopping @code{mysqld}. +@item The isam log @tab Logs all changes to the ISAM tables. Used only for debugging the isam code. +@item The query log @tab Established connections and executed queries. +@item The update log @tab Deprecated: Stores all statements that changes data +@item The binary log @tab Stores all statements that changes something. Used also for replication +@item The slow log @tab Stores all queries that took more than @code{long_query_time} to execute or didn't use indexes. +@end multitable -Returns the ID generated for an @code{AUTO_INCREMENT} column by the previous -query. Use this function after you have performed an @code{INSERT} query -into a table that contains an @code{AUTO_INCREMENT} field. +All logs can be found in the @code{mysqld} data directory. You can +force @code{mysqld} to reopen the log files (or in some cases +switch to a new log) by executing @code{FLUSH LOGS}. @xref{FLUSH}. -Note that @code{mysql_insert_id()} returns @code{0} if the previous query -does not generate an @code{AUTO_INCREMENT} value. If you need to save -the value for later, be sure to call @code{mysql_insert_id()} immediately -after the query that generates the value. +@cindex error log +@cindex files, error log +@menu +* Error log:: +* Query log:: +* Update log:: +* Binary log:: +* Slow query log:: +@end menu -Also note that the value of the SQL @code{LAST_INSERT_ID()} function always -contains the most recently generated @code{AUTO_INCREMENT} value, and is -not reset between queries because the value of that function is maintained -in the server. +@node Error log, Query log, Log files, Log files +@section The Error Log -@subsubheading Return Values +@code{mysqld} writes all errors to the stderr, which the +@code{safe_mysqld} script redirects to a file called +@code{'hostname'.err}. (On Windows, @code{mysqld} writes this directly +to @file{\mysql\data\mysql.err}). -The value of the @code{AUTO_INCREMENT} field that was updated by the previous -query. Returns zero if there was no previous query on the connection or if -the query did not update an @code{AUTO_INCREMENT} value. +This contains information indicating when @code{mysqld} was started and +stopped and also any critical errors found when running. If @code{mysqld} +dies unexpectedly and @code{safe_mysqld} needs to restart @code{mysqld}, +@code{safe_mysqld} will write a @code{restarted mysqld} row in this +file. This log also holds a warning if @code{mysqld} notices a table +that needs to be automatically checked or repaired. -@subsubheading Errors -None. +On some operating systems, the error log will contain a stack trace +for where @code{mysqld} died. This can be used to find out where +@code{mysqld} died. @xref{Using stack trace}. -@findex @code{mysql_kill()} -@node mysql_kill, mysql_list_dbs, mysql_insert_id, C API functions -@subsubsection @code{mysql_kill()} +@cindex query log +@cindex files, query log +@node Query log, Update log, Error log, Log files +@section The Query Log -@code{int mysql_kill(MYSQL *mysql, unsigned long pid)} +If you want to know what happens within @code{mysqld}, you should start +it with @code{--log[=file]}. This will log all connections and queries +to the log file (by default named @file{'hostname'.log}). This log can +be very useful when you suspect an error in a client and want to know +exactly what @code{mysqld} thought the client sent to it. -@subsubheading Description +By default, the @code{mysql.server} script starts the @strong{MySQL} +server with the @code{-l} option. If you need better performance when +you start using @strong{MySQL} in a production environment, you can +remove the @code{-l} option from @code{mysql.server} or change it to +@code{--log-binary}. -Asks the server to kill the thread specified by @code{pid}. +The entries in this log are written as @code{mysqld} receives the questions. +This may be different than the order in which the statements are executed. +This is in contrast to the update log and the binary log which are written +after the query is executed, but before any locks are released. -@subsubheading Return Values +@cindex update log +@cindex files, update log +@node Update log, Binary log, Query log, Log files +@section The Update Log -Zero for success. Non-zero if an error occurred. +@strong{NOTE}: The update log is replaced by the binary +log. @xref{Binary log}. With this you can do anything that you can do +with the update log. -@subsubheading Errors +When started with the @code{--log-update[=file_name]} option, +@code{mysqld} writes a log file containing all SQL commands that update +data. If no filename is given, it defaults to the name of the host +machine. If a filename is given, but it doesn't contain a path, the file +is written in the data directory. If @file{file_name} doesn't have an +extension, @code{mysqld} will create log file names like so: +@file{file_name.###}, where @code{###} is a number that is incremented each +time you execute @code{mysqladmin refresh}, execute @code{mysqladmin +flush-logs}, execute the @code{FLUSH LOGS} statement, or restart the server. -@table @code -@item CR_COMMANDS_OUT_OF_SYNC -Commands were executed in an improper order. -@item CR_SERVER_GONE_ERROR -The @strong{MySQL} server has gone away. -@item CR_SERVER_LOST -The connection to the server was lost during the query. -@item CR_UNKNOWN_ERROR -An unknown error occurred. -@end table +@strong{NOTE:} For the above scheme to work, you should NOT create +your own files with the same filename as the update log + some extensions +that may be regarded as a number, in the directory used by the update log! -@findex @code{mysql_list_dbs()} -@node mysql_list_dbs, mysql_list_fields, mysql_kill, C API functions -@subsubsection @code{mysql_list_dbs()} +If you use the @code{--log} or @code{-l} options, @code{mysqld} writes a +general log with a filename of @file{hostname.log}, and restarts and +refreshes do not cause a new log file to be generated (although it is closed +and reopened). In this case you can copy it (on Unix) by doing: -@code{MYSQL_RES *mysql_list_dbs(MYSQL *mysql, const char *wild)} +@example +mv hostname.log hostname-old.log +mysqladmin flush-logs +cp hostname-old.log to-backup-directory +rm hostname-old.log +@end example -@subsubheading Description +Update logging is smart because it logs only statements that really update +data. So an @code{UPDATE} or a @code{DELETE} with a @code{WHERE} that finds no +rows is not written to the log. It even skips @code{UPDATE} statements that +set a column to the value it already has. -Returns a result set consisting of database names on the server that match -the simple regular expression specified by the @code{wild} parameter. -@code{wild} may contain the wild-card characters @samp{%} or @samp{_}, or may -be a @code{NULL} pointer to match all databases. Calling -@code{mysql_list_dbs()} is similar to executing the query @code{SHOW -databases [LIKE wild]}. +The update logging is done immediately after a query completes but before +any locks are released or any commit is done. This ensures that the log +will be logged in the execution order. -You must free the result set with @code{mysql_free_result()}. +If you want to update a database from update log files, you could do the +following (assuming your update logs have names of the form +@file{file_name.###}): -@subsubheading Return Values +@example +shell> ls -1 -t -r file_name.[0-9]* | xargs cat | mysql +@end example -A @code{MYSQL_RES} result set for success. @code{NULL} if an error occurred. +@code{ls} is used to get all the log files in the right order. -@subsubheading Errors +This can be useful if you have to revert to backup files after a crash +and you want to redo the updates that occurred between the time of the backup +and the crash. -@table @code -@item CR_COMMANDS_OUT_OF_SYNC -Commands were executed in an improper order. -@item CR_OUT_OF_MEMORY -Out of memory. -@item CR_SERVER_GONE_ERROR -The @strong{MySQL} server has gone away. -@item CR_SERVER_LOST -The connection to the server was lost during the query. -@item CR_UNKNOWN_ERROR -An unknown error occurred. -@end table +@cindex binary log +@cindex files, binary log +@node Binary log, Slow query log, Update log, Log files +@section The Binary Log -@findex @code{mysql_list_fields()} -@node mysql_list_fields, mysql_list_processes, mysql_list_dbs, C API functions -@subsubsection @code{mysql_list_fields()} +In the future the binary log will replace the update log, so we +recommend you to switch to this log format as soon as possible! -@code{MYSQL_RES *mysql_list_fields(MYSQL *mysql, const char *table, const char *wild)} +The binary log contains all information that is available in the update +log in a more efficient format. It also contains information about how long +every query that updated the database took. -@subsubheading Description +The binary log is also used when you are replicating a slave from a master. +@xref{Replication}. -Returns a result set consisting of field names in the given table that match -the simple regular expression specified by the @code{wild} parameter. -@code{wild} may contain the wild-card characters @samp{%} or @samp{_}, or may -be a @code{NULL} pointer to match all fields. Calling -@code{mysql_list_fields()} is similar to executing the query @code{SHOW -COLUMNS FROM tbl_name [LIKE wild]}. +When started with the @code{--log-bin[=file_name]} option, @code{mysqld} +writes a log file containing all SQL commands that update data. If no +file name is given, it defaults to the name of the host machine followed +by @code{-bin}. If file name is given, but it doesn't contain a path, the +file is written in the data directory. -Note that it's recommended that you use @code{SHOW COLUMNS FROM tbl_name} -instead of @code{mysql_list_fields()}. +You can use the following options to @code{mysqld} to affect what is logged +to the binary log: -You must free the result set with @code{mysql_free_result()}. +@multitable @columnfractions .4 .6 +@item @code{binlog-do-db=database_name} @tab +Tells the master it should log updates for the specified database, and +exclude all others not explicitly mentioned. +(Example: @code{binlog-do-db=some_database}) -@subsubheading Return Values +@item @code{binlog-ignore-db=database_name} @tab +Tells the master that updates to the given database should not be logged +to the binary log (Example: @code{binlog-ignore-db=some_database}) +@end multitable -A @code{MYSQL_RES} result set for success. @code{NULL} if an error occurred. +To the binary log filename @code{mysqld} will append an extension that is a +number that is incremented each time you execute @code{mysqladmin +refresh}, execute @code{mysqladmin flush-logs}, execute the @code{FLUSH LOGS} +statement or restart the server. -@subsubheading Errors +To be able to know which different binary log files have been used, +@code{mysqld} will also create a binary log index file that +contains the name of all used binary log files. By default this has the +same name as the binary log file, with the extension @code{'.index'}. +You can change the name of the binary log index file with the +@code{--log-bin-index=[filename]} option. -@table @code -@item CR_COMMANDS_OUT_OF_SYNC -Commands were executed in an improper order. -@item CR_SERVER_GONE_ERROR -The @strong{MySQL} server has gone away. -@item CR_SERVER_LOST -The connection to the server was lost during the query. -@item CR_UNKNOWN_ERROR -An unknown error occurred. -@end table +If you are using replication, you should not delete old binary log +files until you are sure that no slave will ever need to use them. +One way to do this is to do @code{mysqladmin flush-logs} once a day and then +remove any logs that are more than 3 days old. -@findex @code{mysql_list_processes()} -@node mysql_list_processes, mysql_list_tables, mysql_list_fields, C API functions -@subsubsection @code{mysql_list_processes()} +You can examine the binary log file with the @code{mysqlbinlog} command. +For example, you can update a @strong{MySQL} server from the binary log +as follows: -@code{MYSQL_RES *mysql_list_processes(MYSQL *mysql)} +@example +mysqlbinlog log-file | mysql -h server_name +@end example -@subsubheading Description +You can also use the @code{mysqlbinlog} program to read the binary log +directly from a remote @strong{MySQL} server! -Returns a result set describing the current server threads. This is the same -kind of information as that reported by @code{mysqladmin processlist} or -a @code{SHOW PROCESSLIST} query. +@code{mysqlbinlog --help} will give you more information of how to use +this program! -You must free the result set with @code{mysql_free_result()}. +If you are using @code{BEGIN [WORK]} or @code{SET AUTOCOMMIT=0}, you must +use the @strong{MySQL} binary log for backups instead of the old update log. -@subsubheading Return Values +The binary logging is done immediately after a query completes but before +any locks are released or any commit is done. This ensures that the log +will be logged in the execution order. -A @code{MYSQL_RES} result set for success. @code{NULL} if an error occurred. +All updates (@code{UPDATE}, @code{DELETE} or @code{INSERT}) that change +a transactional table (like BDB tables) are cached until a @code{COMMIT}. +Any updates to a non-transactional table are stored in the binary log at +once. Every thread will, on start, allocate a buffer of +@code{binlog_cache_size} to buffer queries. If a query is bigger than +this, the thread will open a temporary file to handle the bigger cache. +The temporary file will be deleted when the thread ends. -@subsubheading Errors +The @code{max_binlog_cache_size} can be used to restrict the total size used +to cache a multi-transaction query. -@table @code -@item CR_COMMANDS_OUT_OF_SYNC -Commands were executed in an improper order. -@item CR_SERVER_GONE_ERROR -The @strong{MySQL} server has gone away. -@item CR_SERVER_LOST -The connection to the server was lost during the query. -@item CR_UNKNOWN_ERROR -An unknown error occurred. -@end table +If you are using the update or binary log, concurrent inserts will +not work together with @code{CREATE ... INSERT} and @code{INSERT ... SELECT}. +This is to ensure that you can recreate an exact copy of your tables by +applying the log on a backup. -@findex @code{mysql_list_tables()} -@node mysql_list_tables, mysql_num_fields, mysql_list_processes, C API functions -@subsubsection @code{mysql_list_tables()} +@cindex slow query log +@cindex files, slow query log +@node Slow query log, , Binary log, Log files +@section The Slow Query Log -@code{MYSQL_RES *mysql_list_tables(MYSQL *mysql, const char *wild)} +When started with the @code{--log-slow-queries[=file_name]} option, +@code{mysqld} writes a log file containing all SQL commands that took +more than @code{long_query_time} to execute. The time to get the initial +table locks are not counted as execution time. -@subsubheading Description +The slow query log is logged after the query is executed and after all +locks has been released. This may be different than the order in which +the statements are executed. -Returns a result set consisting of table names in the current database that -match the simple regular expression specified by the @code{wild} parameter. -@code{wild} may contain the wild-card characters @samp{%} or @samp{_}, or may -be a @code{NULL} pointer to match all tables. Calling -@code{mysql_list_tables()} is similar to executing the query @code{SHOW -tables [LIKE wild]}. +If no file name is given, it defaults to the name of the host machine +suffixed with @code{-slow.log}. If a filename is given, but doesn't +contain a path, the file is written in the data directory. -You must free the result set with @code{mysql_free_result()}. +The slow query log can be used to find queries that take a long time to +execute and are thus candidates for optimization. With a large log, that +can become a difficult task. You can pipe the slow query log through the +@code{mysqldumpslow} command to get a summary of the queries which +appear in the log. -@subsubheading Return Values +You are using @code{--log-long-format} then also queries that are not +using indexes are printed. @xref{Command-line options}. -A @code{MYSQL_RES} result set for success. @code{NULL} if an error occurred. +@cindex database replication +@cindex replication, database +@cindex database mirroring +@cindex mirroring, database +You can also use the update logs when you have a mirrored database on +another host and you want to replicate the changes that have been made +to the master database. @xref{Log Replication}. -@subsubheading Errors +@cindex client tools +@cindex APIs +@cindex @code{mysqlclient} library +@cindex buffer sizes, client +@cindex library, @code{mysqlclient} +@node Clients, MySQL internals, Log files, Top +@chapter MySQL APIs -@table @code -@item CR_COMMANDS_OUT_OF_SYNC -Commands were executed in an improper order. -@item CR_SERVER_GONE_ERROR -The @strong{MySQL} server has gone away. -@item CR_SERVER_LOST -The connection to the server was lost during the query. -@item CR_UNKNOWN_ERROR -An unknown error occurred. -@end table +@menu +* C:: @strong{MySQL} C API +* Perl:: @strong{MySQL} Perl API +* Eiffel:: @strong{MySQL} Eiffel wrapper +* Java:: @strong{MySQL} Java connectivity (JDBC) +* PHP:: @strong{MySQL} PHP API +* Cplusplus:: @strong{MySQL} C++ APIs +* Python:: @strong{MySQL} Python APIs +* Tcl:: @strong{MySQL} Tcl APIs +@end menu -@findex @code{mysql_num_fields()} -@findex @code{mysql_field_count()} -@node mysql_num_fields, mysql_num_rows, mysql_list_tables, C API functions -@subsubsection @code{mysql_num_fields()} +This chapter describes the APIs available for @strong{MySQL}, where to get +them, and how to use them. The C API is the most extensively covered, as it +was developed by the @strong{MySQL} team, and is the basis for most of the +other APIs. -@code{unsigned int mysql_num_fields(MYSQL_RES *result)} +@cindex C API, datatypes +@cindex datatypes, C API +@node C, Perl, Clients, Clients +@section MySQL C API -or +@menu +* C API datatypes:: C API Datatypes +* C API function overview:: C API Function Overview +* C API functions:: C API Function Descriptions +* C API problems:: +* Thread-safe clients:: +@end menu -@code{unsigned int mysql_num_fields(MYSQL *mysql)} +The C API code is distributed with @strong{MySQL}. It is included in the +@code{mysqlclient} library and allows C programs to access a database. -The second form doesn't work on @strong{MySQL} Version 3.22.24 or newer. To pass a -@code{MYSQL*} argument, you must use -@code{unsigned int mysql_field_count(MYSQL *mysql)} instead. +Many of the clients in the @strong{MySQL} source distribution are +written in C. If you are looking for examples that demonstrate how to +use the C API, take a look at these clients. You can find these in the +@code{clients} directory in the @strong{MySQL} source distribution. -@subsubheading Description +Most of the other client APIs (all except Java) use the @code{mysqlclient} +library to communicate with the @strong{MySQL} server. This means that, for +example, you can take advantage of many of the same environment variables +that are used by other client programs, because they are referenced from the +library. See @ref{Programs}, for a list of these variables. -Returns the number of columns in a result set. +The client has a maximum communication buffer size. The size of the buffer +that is allocated initially (16K bytes) is automatically increased up to the +maximum size (the maximum is 16M). Because buffer sizes are increased +only as demand warrants, simply increasing the default maximum limit does not +in itself cause more resources to be used. This size check is mostly a check +for erroneous queries and communication packets. -Note that you can get the number of columns either from a pointer to a result -set or to a connection handle. You would use the connection handle if -@code{mysql_store_result()} or @code{mysql_use_result()} returned -@code{NULL} (and thus you have no result set pointer). In this case, you can -call @code{mysql_field_count()} to determine whether or not -@code{mysql_store_result()} should have produced a non-empty result. This -allows the client program to take proper action without knowing whether or -not the query was a @code{SELECT} (or @code{SELECT}-like) statement. The -example shown below illustrates how this may be done. +The communication buffer must be large enough to contain a single SQL +statement (for client-to-server traffic) and one row of returned data (for +server-to-client traffic). Each thread's communication buffer is dynamically +enlarged to handle any query or row up to the maximum limit. For example, if +you have @code{BLOB} values that contain up to 16M of data, you must have a +communication buffer limit of at least 16M (in both server and client). The +client's default maximum is 16M, but the default maximum in the server is +1M. You can increase this by changing the value of the +@code{max_allowed_packet} parameter when the server is started. @xref{Server +parameters}. -@xref{NULL mysql_store_result, , @code{NULL mysql_store_result()}}. +The @strong{MySQL} server shrinks each communication buffer to +@code{net_buffer_length} bytes after each query. For clients, the size of +the buffer associated with a connection is not decreased until the connection +is closed, at which time client memory is reclaimed. -@subsubheading Return Values +For programming with threads, consult the 'how to make a thread-safe +client' chapter. @xref{Thread-safe clients}. -An unsigned integer representing the number of fields in a result set. +@node C API datatypes, C API function overview, C, C +@subsection C API Datatypes +@table @code -@subsubheading Errors -None. +@tindex MYSQL C type +@item MYSQL +This structure represents a handle to one database connection. It is +used for almost all @strong{MySQL} functions. -@subsubheading Example +@tindex MYSQL_RES C type +@item MYSQL_RES +This structure represents the result of a query that returns rows +(@code{SELECT}, @code{SHOW}, @code{DESCRIBE}, @code{EXPLAIN}). The +information returned from a query is called the @emph{result set} in the +remainder of this section. -@example -MYSQL_RES *result; -unsigned int num_fields; -unsigned int num_rows; +@tindex MYSQL_ROW C type +@item MYSQL_ROW +This is a type-safe representation of one row of data. It is currently +implemented as an array of counted byte strings. (You cannot treat these as +null-terminated strings if field values may contain binary data, because such +values may contain null bytes internally.) Rows are obtained by calling +@code{mysql_fetch_row()}. -if (mysql_query(&mysql,query_string)) -@{ - // error -@} -else // query succeeded, process any data returned by it -@{ - result = mysql_store_result(&mysql); - if (result) // there are rows - @{ - num_fields = mysql_num_fields(result); - // retrieve rows, then call mysql_free_result(result) - @} - else // mysql_store_result() returned nothing; should it have? - @{ - if (mysql_errno(&mysql)) - @{ - fprintf(stderr, "Error: %s\n", mysql_error(&mysql)); - @} - else if (mysql_field_count(&mysql) == 0) - @{ - // query does not return data - // (it was not a SELECT) - num_rows = mysql_affected_rows(&mysql); - @} - @} -@} +@tindex MYSQL_FIELD C type +@item MYSQL_FIELD +This structure contains information about a field, such as the field's +name, type, and size. Its members are described in more detail below. +You may obtain the @code{MYSQL_FIELD} structures for each field by +calling @code{mysql_fetch_field()} repeatedly. Field values are not part of +this structure; they are contained in a @code{MYSQL_ROW} structure. + + +@tindex MYSQL_FIELD_OFFSET C type +@item MYSQL_FIELD_OFFSET +This is a type-safe representation of an offset into a @strong{MySQL} field +list. (Used by @code{mysql_field_seek()}.) Offsets are field numbers +within a row, beginning at zero. + +@tindex my_ulonglong C type +@tindex my_ulonglong values, printing +@item my_ulonglong +The type used for the number of rows and for @code{mysql_affected_rows()}, +@code{mysql_num_rows()}, and @code{mysql_insert_id()}. This type provides a +range of @code{0} to @code{1.84e19}. + +On some systems, attempting to print a value of type @code{my_ulonglong} +will not work. To print such a value, convert it to @code{unsigned long} +and use a @code{%lu} print format. Example: +@example +printf (Number of rows: %lu\n", (unsigned long) mysql_num_rows(result)); @end example +@end table -An alternative (if you KNOW that your query should have returned a result set) -is to replace the @code{mysql_errno(&mysql)} call with a check if -@code{mysql_field_count(&mysql)} is = 0. This will only happen if something -went wrong. +@noindent +The @code{MYSQL_FIELD} structure contains the members listed below: -@findex @code{mysql_num_rows()} -@node mysql_num_rows, mysql_options, mysql_num_fields, C API functions -@subsubsection @code{mysql_num_rows()} +@table @code +@item char * name +The name of the field, as a null-terminated string. -@code{my_ulonglong mysql_num_rows(MYSQL_RES *result)} +@item char * table +The name of the table containing this field, if it isn't a calculated field. +For calculated fields, the @code{table} value is an empty string. -@subsubheading Description +@item char * def +The default value of this field, as a null-terminated string. This is set +only if you use @code{mysql_list_fields()}. -Returns the number of rows in the result set. +@item enum enum_field_types type +The type of the field. +The @code{type} value may be one of the following: -The use of @code{mysql_num_rows()} depends on whether you use -@code{mysql_store_result()} or @code{mysql_use_result()} to return the result -set. If you use @code{mysql_store_result()}, @code{mysql_num_rows()} may be -called immediately. If you use @code{mysql_use_result()}, -@code{mysql_num_rows()} will not return the correct value until all the rows -in the result set have been retrieved. +@multitable @columnfractions .3 .55 +@item @strong{Type value} @tab @strong{Type meaning} +@item @code{FIELD_TYPE_TINY} @tab @code{TINYINT} field +@item @code{FIELD_TYPE_SHORT} @tab @code{SMALLINT} field +@item @code{FIELD_TYPE_LONG} @tab @code{INTEGER} field +@item @code{FIELD_TYPE_INT24} @tab @code{MEDIUMINT} field +@item @code{FIELD_TYPE_LONGLONG} @tab @code{BIGINT} field +@item @code{FIELD_TYPE_DECIMAL} @tab @code{DECIMAL} or @code{NUMERIC} field +@item @code{FIELD_TYPE_FLOAT} @tab @code{FLOAT} field +@item @code{FIELD_TYPE_DOUBLE} @tab @code{DOUBLE} or @code{REAL} field +@item @code{FIELD_TYPE_TIMESTAMP} @tab @code{TIMESTAMP} field +@item @code{FIELD_TYPE_DATE} @tab @code{DATE} field +@item @code{FIELD_TYPE_TIME} @tab @code{TIME} field +@item @code{FIELD_TYPE_DATETIME} @tab @code{DATETIME} field +@item @code{FIELD_TYPE_YEAR} @tab @code{YEAR} field +@item @code{FIELD_TYPE_STRING} @tab String (@code{CHAR} or @code{VARCHAR}) field +@item @code{FIELD_TYPE_BLOB} @tab @code{BLOB} or @code{TEXT} field (use @code{max_length} to determine the maximum length) +@item @code{FIELD_TYPE_SET} @tab @code{SET} field +@item @code{FIELD_TYPE_ENUM} @tab @code{ENUM} field +@item @code{FIELD_TYPE_NULL} @tab @code{NULL}-type field +@item @code{FIELD_TYPE_CHAR} @tab Deprecated; use @code{FIELD_TYPE_TINY} instead +@end multitable -@subsubheading Return Values +You can use the @code{IS_NUM()} macro to test whether or not a field has a +numeric type. Pass the @code{type} value to @code{IS_NUM()} and it +will evaluate to TRUE if the field is numeric: -The number of rows in the result set. +@example +if (IS_NUM(field->type)) + printf("Field is numeric\n"); +@end example -@subsubheading Errors -None. +@item unsigned int length +The width of the field, as specified in the table definition. -@findex @code{mysql_options()} -@node mysql_options, mysql_ping, mysql_num_rows, C API functions -@subsubsection @code{mysql_options()} +@item unsigned int max_length +The maximum width of the field for the result set (the length of the longest +field value for the rows actually in the result set). If you use +@code{mysql_store_result()} or @code{mysql_list_fields()}, this contains the +maximum length for the field. If you use @code{mysql_use_result()}, the +value of this variable is zero. -@code{int mysql_options(MYSQL *mysql, enum mysql_option option, const char *arg)} +@item unsigned int flags +Different bit-flags for the field. The @code{flags} value may have zero +or more of the following bits set: -@subsubheading Description +@multitable @columnfractions .3 .55 +@item @strong{Flag value} @tab @strong{Flag meaning} +@item @code{NOT_NULL_FLAG} @tab Field can't be @code{NULL} +@item @code{PRI_KEY_FLAG} @tab Field is part of a primary key +@item @code{UNIQUE_KEY_FLAG} @tab Field is part of a unique key +@item @code{MULTIPLE_KEY_FLAG} @tab Field is part of a non-unique key +@item @code{UNSIGNED_FLAG} @tab Field has the @code{UNSIGNED} attribute +@item @code{ZEROFILL_FLAG} @tab Field has the @code{ZEROFILL} attribute +@item @code{BINARY_FLAG} @tab Field has the @code{BINARY} attribute +@item @code{AUTO_INCREMENT_FLAG} @tab Field has the @code{AUTO_INCREMENT} +attribute +@item @code{ENUM_FLAG} @tab Field is an @code{ENUM} (deprecated) +@item @code{BLOB_FLAG} @tab Field is a @code{BLOB} or @code{TEXT} (deprecated) +@item @code{TIMESTAMP_FLAG} @tab Field is a @code{TIMESTAMP} (deprecated) +@end multitable -Can be used to set extra connect options and affect behavior for a connection. -This function may be called multiple times to set several options. +Use of the @code{BLOB_FLAG}, @code{ENUM_FLAG}, and @code{TIMESTAMP_FLAG} +flags is deprecated because they indicate the type of a field rather +than an attribute of its type. It is preferable to test +@code{field->type} against @code{FIELD_TYPE_BLOB}, +@code{FIELD_TYPE_ENUM}, or @code{FIELD_TYPE_TIMESTAMP} instead. -@code{mysql_options()} should be called after @code{mysql_init()} and before -@code{mysql_connect()} or @code{mysql_real_connect()}. +@noindent +The example below illustrates a typical use of the @code{flags} value: -The @code{option} argument is the option that you want to set; the @code{arg} -argument is the value for the option. If the option is an integer, then -@code{arg} should point to the value of the integer. +@example +if (field->flags & NOT_NULL_FLAG) + printf("Field can't be null\n"); +@end example -Possible options values: +You may use the following convenience macros to determine the boolean +status of the @code{flags} value: -@multitable @columnfractions .25 .25 .5 -@item @strong{Option} @tab @strong{Argument type} @tab @strong{Function} -@item @code{MYSQL_OPT_CONNECT_TIMEOUT} @tab @code{unsigned int *} @tab Connect timeout in seconds. -@item @code{MYSQL_OPT_COMPRESS} @tab Not used @tab Use the compressed client/server protocol. -@item @code{MYSQL_OPT_NAMED_PIPE} @tab Not used @tab Use named pipes to connect to a @strong{MySQL} server on NT. -@item @code{MYSQL_INIT_COMMAND} @tab @code{char *} @tab Command to execute when connecting to the @strong{MySQL} server. Will automatically be re-executed when reconnecting. -@item @code{MYSQL_READ_DEFAULT_FILE} @tab @code{char *} @tab Read options from the named option file instead of from @file{my.cnf}. -@item @code{MYSQL_READ_DEFAULT_GROUP} @tab @code{char *} @tab Read options from the named group from @file{my.cnf} or the file specified with @code{MYSQL_READ_DEFAULT_FILE}. +@multitable @columnfractions .3 .5 +@item @code{IS_NOT_NULL(flags)} @tab True if this field is defined as @code{NOT NULL} +@item @code{IS_PRI_KEY(flags)} @tab True if this field is a primary key +@item @code{IS_BLOB(flags)} @tab True if this field is a @code{BLOB} or @code{TEXT} (deprecated; test @code{field->type} instead) @end multitable -Note that the group @code{client} is always read if you use -@code{MYSQL_READ_DEFAULT_FILE} or @code{MYSQL_READ_DEFAULT_GROUP}. +@item unsigned int decimals +The number of decimals for numeric fields. +@end table -The specified group in the option file may contain the following options: +@cindex C API, functions +@cindex functions, C API +@node C API function overview, C API functions, C API datatypes, C +@subsection C API Function Overview + +The functions available in the C API are listed below and are described in +greater detail in the next section. +@xref{C API functions}. @multitable @columnfractions .3 .7 -@item @code{connect_timeout} @tab Connect timeout in seconds. On Linux this timeout is also used for waiting for the first answer from the server. -@item @code{compress} @tab Use the compressed client/server protocol. -@item @code{database} @tab Connect to this database if no database was specified in the connect command. -@item @code{debug} @tab Debug options. -@item @code{host} @tab Default host name. -@item @code{init-command} @tab Command to execute when connecting to @strong{MySQL} server. Will automatically be re-executed when reconnecting. -@item @code{interactive-timeout} @tab Same as specifying @code{CLIENT_INTERACTIVE} to @code{mysql_real_connect()}. @xref{mysql_real_connect}. -@item @code{password} @tab Default password. -@item @code{pipe} @tab Use named pipes to connect to a @strong{MySQL} server on NT. -@item @code{port} @tab Default port number. -@item @code{return-found-rows} @tab Tell @code{mysql_info()} to return found rows instead of updated rows when using @code{UPDATE}. -@item @code{socket} @tab Default socket number. -@item -@item @code{user} @tab Default user. -@end multitable +@item @strong{mysql_affected_rows()} @tab +Returns the number of rows changed/deleted/inserted by the last @code{UPDATE}, +@code{DELETE}, or @code{INSERT} query. -Note that @code{timeout} has been replaced by @code{connect_timeout}, but -@code{timeout} will still work for a while. +@item @strong{mysql_close()} @tab +Closes a server connection. -For more information about option files, see @ref{Option files}. +@item @strong{mysql_connect()} @tab +Connects to a @strong{MySQL} server. This function is deprecated; use +@code{mysql_real_connect()} instead. -@subsubheading Return Values +@item @strong{mysql_change_user()} @tab +Changes user and database on an open connection. -Zero for success. Non-zero if you used an unknown option. +@item @strong{mysql_character_set_name()} @tab +Returns the name of the default character set for the connection. -@subsubheading Example +@item @strong{mysql_create_db()} @tab +Creates a database. This function is deprecated; use the SQL command +@code{CREATE DATABASE} instead. -@example -MYSQL mysql; +@item @strong{mysql_data_seek()} @tab +Seeks to an arbitrary row in a query result set. -mysql_init(&mysql); -mysql_options(&mysql,MYSQL_OPT_COMPRESS,0); -mysql_options(&mysql,MYSQL_READ_DEFAULT_GROUP,"odbc"); -if (!mysql_real_connect(&mysql,"host","user","passwd","database",0,NULL,0)) -@{ - fprintf(stderr, "Failed to connect to database: Error: %s\n", - mysql_error(&mysql)); -@} -@end example +@item @strong{mysql_debug()} @tab +Does a @code{DBUG_PUSH} with the given string. -The above requests the client to use the compressed client/server protocol and -read the additional options from the @code{odbc} section in the @code{my.cnf} -file. +@item @strong{mysql_drop_db()} @tab +Drops a database. This function is deprecated; use the SQL command +@code{DROP DATABASE} instead. -@findex @code{mysql_ping()} -@node mysql_ping, mysql_query, mysql_options, C API functions -@subsubsection @code{mysql_ping()} +@item @strong{mysql_dump_debug_info()} @tab +Makes the server write debug information to the log. -@code{int mysql_ping(MYSQL *mysql)} +@item @strong{mysql_eof()} @tab +Determines whether or not the last row of a result set has been read. +This function is deprecated; @code{mysql_errno()} or @code{mysql_error()} +may be used instead. -@subsubheading Description +@item @strong{mysql_errno()} @tab +Returns the error number for the most recently invoked @strong{MySQL} function. -Checks whether or not the connection to the server is working. If it has gone -down, an automatic reconnection is attempted. +@item @strong{mysql_error()} @tab +Returns the error message for the most recently invoked @strong{MySQL} function. -This function can be used by clients that remain idle for a long while, -to check whether or not the server has closed the connection and reconnect -if necessary. +@item @strong{mysql_real_escape_string()} @tab +Escapes special characters in a string for use in a SQL statement taking +into account the current charset of the connection. -@subsubheading Return Values +@item @strong{mysql_escape_string()} @tab +Escapes special characters in a string for use in a SQL statement. -Zero if the server is alive. Non-zero if an error occurred. +@item @strong{mysql_fetch_field()} @tab +Returns the type of the next table field. -@subsubheading Errors +@item @strong{mysql_fetch_field_direct()} @tab +Returns the type of a table field, given a field number. -@table @code -@item CR_COMMANDS_OUT_OF_SYNC -Commands were executed in an improper order. -@item CR_SERVER_GONE_ERROR -The @strong{MySQL} server has gone away. -@item CR_UNKNOWN_ERROR -An unknown error occurred. -@end table +@item @strong{mysql_fetch_fields()} @tab +Returns an array of all field structures. -@findex @code{mysql_query()} -@node mysql_query, mysql_real_connect, mysql_ping, C API functions -@subsubsection @code{mysql_query()} +@item @strong{mysql_fetch_lengths()} @tab +Returns the lengths of all columns in the current row. -@code{int mysql_query(MYSQL *mysql, const char *query)} +@item @strong{mysql_fetch_row()} @tab +Fetches the next row from the result set. -@subsubheading Description -Executes the SQL query pointed to by the null-terminated string @code{query}. -The query must consist of a single SQL statement. You should not add -a terminating semicolon (@samp{;}) or @code{\g} to the statement. +@item @strong{mysql_field_seek()} @tab +Puts the column cursor on a specified column. -@code{mysql_query()} cannot be used for queries that contain binary data; you -should use @code{mysql_real_query()} instead. (Binary data may contain the -@samp{\0} character, which @code{mysql_query()} interprets as the end of the -query string.) +@item @strong{mysql_field_count()} @tab +Returns the number of result columns for the most recent query. -If you want to know if the query should return a result set or not, you can -use @code{mysql_field_count()} to check for this. -@xref{mysql_field_count, , @code{mysql_field_count}}. +@item @strong{mysql_field_tell()} @tab +Returns the position of the field cursor used for the last +@code{mysql_fetch_field()}. -@subsubheading Return Values +@item @strong{mysql_free_result()} @tab +Frees memory used by a result set. -Zero if the query was successful. Non-zero if an error occurred. +@item @strong{mysql_get_client_info()} @tab +Returns client version information. -@subsubheading Errors +@item @strong{mysql_get_host_info()} @tab +Returns a string describing the connection. -@table @code -@item CR_COMMANDS_OUT_OF_SYNC -Commands were executed in an improper order. -@item CR_SERVER_GONE_ERROR -The @strong{MySQL} server has gone away. -@item CR_SERVER_LOST -The connection to the server was lost during the query. -@item CR_UNKNOWN_ERROR -An unknown error occurred. -@end table +@item @strong{mysql_get_proto_info()} @tab +Returns the protocol version used by the connection. -@findex @code{mysql_real_connect()} -@node mysql_real_connect, mysql_real_escape_string, mysql_query, C API functions -@subsubsection @code{mysql_real_connect()} +@item @strong{mysql_get_server_info()} @tab +Returns the server version number. -@code{MYSQL *mysql_real_connect(MYSQL *mysql, const char *host, - const char *user, const char *passwd, const char *db, - unsigned int port, const char *unix_socket, - unsigned int client_flag)} +@item @strong{mysql_info()} @tab +Returns information about the most recently executed query. -@subsubheading Description +@item @strong{mysql_init()} @tab +Gets or initializes a @code{MYSQL} structure. -@code{mysql_real_connect()} attempts to establish a connection to a -@strong{MySQL} database engine running on @code{host}. -@code{mysql_real_connect()} must complete successfully before you can execute -any of the other API functions, with the exception of -@code{mysql_get_client_info()}. +@item @strong{mysql_insert_id()} @tab +Returns the ID generated for an @code{AUTO_INCREMENT} column by the previous +query. -The parameters are specified as follows: +@item @strong{mysql_kill()} @tab +Kills a given thread. -@itemize @bullet -@item -The first parameter should be the address of an existing @code{MYSQL} -structure. Before calling @code{mysql_real_connect()} you must call -@code{mysql_init()} to initialize the @code{MYSQL} structure. You can -change a lot of connect options with the @code{mysql_options()} -call. @xref{mysql_options}. +@item @strong{mysql_list_dbs()} @tab +Returns database names matching a simple regular expression. -@item -The value of @code{host} may be either a hostname or an IP address. If -@code{host} is @code{NULL} or the string @code{"localhost"}, a connection to -the local host is assumed. If the OS supports sockets (Unix) or named pipes -(Windows), they are used instead of TCP/IP to connect to the server. +@item @strong{mysql_list_fields()} @tab +Returns field names matching a simple regular expression. -@item -The @code{user} parameter contains the user's @strong{MySQL} login ID. If -@code{user} is @code{NULL}, the current user is assumed. Under Unix, this is -the current login name. Under Windows ODBC, the current user name must be -specified explicitly. -@xref{ODBC administrator}. +@item @strong{mysql_list_processes()} @tab +Returns a list of the current server threads. -@item -The @code{passwd} parameter contains the password for @code{user}. If -@code{passwd} is @code{NULL}, only entries in the @code{user} table for the -user that have a blank (empty) password field will be checked for a match. This -allows the database administrator to set up the @strong{MySQL} privilege -system in such a way that users get different privileges depending on whether -or not they have specified a password. +@item @strong{mysql_list_tables()} @tab +Returns table names matching a simple regular expression. -NOTE: Do not attempt to encrypt the password before calling -@code{mysql_real_connect()}; password encryption is handled automatically by -the client API. +@item @strong{mysql_num_fields()} @tab +Returns the number of columns in a result set. -@item -@code{db} is the database name. -If @code{db} is not @code{NULL}, the connection will set the default -database to this value. +@item @strong{mysql_num_rows()} @tab +Returns the number of rows in a result set. -@item -If @code{port} is not 0, the value will be used as the port number -for the TCP/IP connection. Note that the @code{host} parameter -determines the type of the connection. +@item @strong{mysql_options()} @tab +Sets connect options for @code{mysql_connect()}. -@item -If @code{unix_socket} is not @code{NULL}, the string specifies the -socket or named pipe that should be used. Note that the @code{host} -parameter determines the type of the connection. +@item @strong{mysql_ping()} @tab +Checks whether or not the connection to the server is working, reconnecting +as necessary. -@item -The value of @code{client_flag} is usually 0, but can be set to a combination -of the following flags in very special circumstances: +@item @strong{mysql_query()} @tab +Executes a SQL query specified as a null-terminated string. -@multitable @columnfractions .25 .7 -@item @strong{Flag name} @tab @strong{Flag meaning} -@code{mysqld} to be more ODBC-friendly. -@item @code{CLIENT_COMPRESS} @tab Use compression protocol. -@item @code{CLIENT_FOUND_ROWS} @tab Return the number of found (matched) rows, not the number of affected rows. -@item @code{CLIENT_IGNORE_SPACE} @tab Allow spaces after function names. Makes all functions names reserved words. -@item @code{CLIENT_INTERACTIVE} @tab Allow @code{interactive_timeout} seconds (instead of @code{wait_timeout} seconds) of inactivity before closing the connection. -@item @code{CLIENT_NO_SCHEMA} @tab Don't allow the @code{db_name.tbl_name.col_name} syntax. This is for ODBC. It causes the parser to generate an error if you use that syntax, which is useful for trapping bugs in some ODBC programs. -@item @code{CLIENT_ODBC} @tab The client is an ODBC client. This changes -@item @code{CLIENT_SSL} @tab Use SSL (encrypted protocol). -@end multitable -@end itemize +@item @strong{mysql_real_connect()} @tab +Connects to a @strong{MySQL} server. -@subsubheading Return Values +@item @strong{mysql_real_query()} @tab +Executes a SQL query specified as a counted string. -A @code{MYSQL*} connection handle if the connection was successful, -@code{NULL} if the connection was unsuccessful. For a successful connection, -the return value is the same as the value of the first parameter, unless you -pass @code{NULL} for that parameter. +@item @strong{mysql_reload()} @tab +Tells the server to reload the grant tables. -@subsubheading Errors +@item @strong{mysql_row_seek()} @tab +Seeks to a row in a result set, using value returned from +@code{mysql_row_tell()}. -@table @code -@item CR_CONN_HOST_ERROR -Failed to connect to the @strong{MySQL} server. +@item @strong{mysql_row_tell()} @tab +Returns the row cursor position. -@item CR_CONNECTION_ERROR -Failed to connect to the local @strong{MySQL} server. +@item @strong{mysql_select_db()} @tab +Selects a database. -@item CR_IPSOCK_ERROR -Failed to create an IP socket. +@item @strong{mysql_shutdown()} @tab +Shuts down the database server. -@item CR_OUT_OF_MEMORY -Out of memory. +@item @strong{mysql_stat()} @tab +Returns the server status as a string. -@item CR_SOCKET_CREATE_ERROR -Failed to create a Unix socket. +@item @strong{mysql_store_result()} @tab +Retrieves a complete result set to the client. -@item CR_UNKNOWN_HOST -Failed to find the IP address for the hostname. +@item @strong{mysql_thread_id()} @tab +Returns the current thread ID. -@item CR_VERSION_ERROR -A protocol mismatch resulted from attempting to connect to a server with a -client library that uses a different protocol version. This can happen if you -use a very old client library to connect to a new server that wasn't started -with the @code{--old-protocol} option. +@item @strong{mysql_thread_save()} @tab +Returns 1 if the clients are compiled as thread-safe. -@item CR_NAMEDPIPEOPEN_ERROR -Failed to create a named pipe on Windows. +@item @strong{mysql_use_result()} @tab +Initiates a row-by-row result set retrieval. +@end multitable -@item CR_NAMEDPIPEWAIT_ERROR -Failed to wait for a named pipe on Windows. +To connect to the server, call @code{mysql_init()} to initialize a +connection handler, then call @code{mysql_real_connect()} with that +handler (along with other information such as the hostname, user name, +and password). Upon connection, @code{mysql_real_connect()} sets the +@code{reconnect} flag (part of the MYSQL structure) to a value of +@code{1}. This flag indicates, in the event that a query cannot be +performed because of a lost connection, to try reconnecting to the +server before giving up. When you are done with the connection, call +@code{mysql_close()} to terminate it. -@item CR_NAMEDPIPESETSTATE_ERROR -Failed to get a pipe handler on Windows. +While a connection is active, the client may send SQL queries to the server +using @code{mysql_query()} or @code{mysql_real_query()}. The difference +between the two is that @code{mysql_query()} expects the query to be +specified as a null-terminated string whereas @code{mysql_real_query()} +expects a counted string. If the string contains binary data (which may +include null bytes), you must use @code{mysql_real_query()}. -@item CR_SERVER_LOST -If @code{connect_timeout} > 0 and it took longer then @code{connect_timeout} -seconds to connect to the server or if the server died while executing the -@code{init-command}. +For each non-@code{SELECT} query (for example, @code{INSERT}, @code{UPDATE}, +@code{DELETE}), you can find out how many rows were changed (affected) +by calling @code{mysql_affected_rows()}. -@end table +For @code{SELECT} queries, you retrieve the selected rows as a result set. +(Note that some statements are @code{SELECT}-like in that they return rows. +These include @code{SHOW}, @code{DESCRIBE}, and @code{EXPLAIN}. They should +be treated the same way as @code{SELECT} statements.) -@subsubheading Example +There are two ways for a client to process result sets. One way is to +retrieve the entire result set all at once by calling +@code{mysql_store_result()}. This function acquires from the server all the +rows returned by the query and stores them in the client. The second way is +for the client to initiate a row-by-row result set retrieval by calling +@code{mysql_use_result()}. This function initializes the retrieval, but does +not actually get any rows from the server. -@example -MYSQL mysql; +In both cases, you access rows by calling @code{mysql_fetch_row()}. With +@code{mysql_store_result()}, @code{mysql_fetch_row()} accesses rows that have +already been fetched from the server. With @code{mysql_use_result()}, +@code{mysql_fetch_row()} actually retrieves the row from the server. +Information about the size of the data in each row is available by +calling @code{mysql_fetch_lengths()}. -mysql_init(&mysql); -mysql_options(&mysql,MYSQL_READ_DEFAULT_GROUP,"your_prog_name"); -if (!mysql_real_connect(&mysql,"host","user","passwd","database",0,NULL,0)) -@{ - fprintf(stderr, "Failed to connect to database: Error: %s\n", - mysql_error(&mysql)); -@} -@end example +After you are done with a result set, call @code{mysql_free_result()} +to free the memory used for it. -By using @code{mysql_options()} the @strong{MySQL} library will read the -@code{[client]} and @code{your_prog_name} sections in the @code{my.cnf} -file which will ensure that your program will work, even if someone has -set up @strong{MySQL} in some non-standard way. +The two retrieval mechanisms are complementary. Client programs should +choose the approach that is most appropriate for their requirements. +In practice, clients tend to use @code{mysql_store_result()} more +commonly. -Note that upon connection, @code{mysql_real_connect()} sets the @code{reconnect} -flag (part of the MYSQL structure) to a value of @code{1}. This flag indicates, -in the event that a query cannot be performed because of a lost connection, to -try reconnecting to the server before giving up. +An advantage of @code{mysql_store_result()} is that because the rows have all +been fetched to the client, you not only can access rows sequentially, you +can move back and forth in the result set using @code{mysql_data_seek()} or +@code{mysql_row_seek()} to change the current row position within the result +set. You can also find out how many rows there are by calling +@code{mysql_num_rows()}. On the other hand, the memory requirements for +@code{mysql_store_result()} may be very high for large result sets and you +are more likely to encounter out-of-memory conditions. -@findex @code{mysql_real_escape_string()} -@node mysql_real_escape_string, mysql_real_query, mysql_real_connect, C API functions -@subsubsection @code{mysql_real_escape_string()} +An advantage of @code{mysql_use_result()} is that the client requires less +memory for the result set because it maintains only one row at a time (and +because there is less allocation overhead, @code{mysql_use_result()} can be +faster). Disadvantages are that you must process each row quickly to avoid +tying up the server, you don't have random access to rows within the result +set (you can only access rows sequentially), and you don't know how many rows +are in the result set until you have retrieved them all. Furthermore, you +@emph{must} retrieve all the rows even if you determine in mid-retrieval that +you've found the information you were looking for. -@code{unsigned int mysql_real_escape_string(MYSQL *mysql, char *to, const char *from, unsigned int length)} +The API makes it possible for clients to respond appropriately to +queries (retrieving rows only as necessary) without knowing whether or +not the query is a @code{SELECT}. You can do this by calling +@code{mysql_store_result()} after each @code{mysql_query()} (or +@code{mysql_real_query()}). If the result set call succeeds, the query +was a @code{SELECT} and you can read the rows. If the result set call +fails, call @code{mysql_field_count()} to determine whether or not a +result was actually to be expected. If @code{mysql_field_count()} +returns zero, the query returned no data (indicating that it was an +@code{INSERT}, @code{UPDATE}, @code{DELETE}, etc.), and was not +expected to return rows. If @code{mysql_field_count()} is non-zero, the +query should have returned rows, but didn't. This indicates that the +query was a @code{SELECT} that failed. See the description for +@code{mysql_field_count()} for an example of how this can be done. -@subsubheading Description +Both @code{mysql_store_result()} and @code{mysql_use_result()} allow you to +obtain information about the fields that make up the result set (the number +of fields, their names and types, etc.). You can access field information +sequentially within the row by calling @code{mysql_fetch_field()} repeatedly, +or by field number within the row by calling +@code{mysql_fetch_field_direct()}. The current field cursor position may be +changed by calling @code{mysql_field_seek()}. Setting the field cursor +affects subsequent calls to @code{mysql_fetch_field()}. You can also get +information for fields all at once by calling @code{mysql_fetch_fields()}. -This function is used to create a legal SQL string that you can use in a -SQL statement. @xref{String syntax}. +For detecting and reporting errors, @strong{MySQL} provides access to error +information by means of the @code{mysql_errno()} and @code{mysql_error()} +functions. These return the error code or error message for the most +recently invoked function that can succeed or fail, allowing you to determine +when an error occurred and what it was. -The string in @code{from} is encoded to an escaped SQL string, taking -into account the current character set of the connection. The result is placed -in @code{to} and a terminating null byte is appended. Characters -encoded are @code{NUL} (ASCII 0), @samp{\n}, @samp{\r}, @samp{\}, -@samp{'}, @samp{"}, and Control-Z (@pxref{Literals}). +@node C API functions, C API problems, C API function overview, C +@subsection C API Function Descriptions -The string pointed to by @code{from} must be @code{length} bytes long. You -must allocate the @code{to} buffer to be at least @code{length*2+1} bytes -long. (In the worse case, each character may need to be encoded as using two -bytes, and you need room for the terminating null byte.) When -@code{mysql_escape_string()} returns, the contents of @code{to} will be a -null-terminated string. The return value is the length of the encoded -string, not including the terminating null character. +@menu +* mysql_affected_rows:: @code{mysql_affected_rows()} +* mysql_close:: @code{mysql_close()} +* mysql_connect:: @code{mysql_connect()} +* mysql_change_user:: @code{mysql_change_user()} +* mysql_character_set_name:: @code{mysql_character_set_name()} +* mysql_create_db:: @code{mysql_create_db()} +* mysql_data_seek:: @code{mysql_data_seek()} +* mysql_debug:: @code{mysql_debug()} +* mysql_drop_db:: @code{mysql_drop_db()} +* mysql_dump_debug_info:: @code{mysql_dump_debug_info()} +* mysql_eof:: @code{mysql_eof()} +* mysql_errno:: @code{mysql_errno()} +* mysql_error:: @code{mysql_error()} +* mysql_escape_string:: @code{mysql_escape_string()} +* mysql_fetch_field:: @code{mysql_fetch_field()} +* mysql_fetch_fields:: @code{mysql_fetch_fields()} +* mysql_fetch_field_direct:: @code{mysql_fetch_field_direct()} +* mysql_fetch_lengths:: @code{mysql_fetch_lengths()} +* mysql_fetch_row:: @code{mysql_fetch_row()} +* mysql_field_count:: @code{mysql_field_count()} +* mysql_field_seek:: @code{mysql_field_seek()} +* mysql_field_tell:: @code{mysql_field_tell()} +* mysql_free_result:: @code{mysql_free_result()} +* mysql_get_client_info:: @code{mysql_get_client_info()} +* mysql_get_host_info:: @code{mysql_get_host_info()} +* mysql_get_proto_info:: @code{mysql_get_proto_info()} +* mysql_get_server_info:: @code{mysql_get_server_info()} +* mysql_info:: @code{mysql_info()} +* mysql_init:: @code{mysql_init()} +* mysql_insert_id:: @code{mysql_insert_id()} +* mysql_kill:: @code{mysql_kill()} +* mysql_list_dbs:: @code{mysql_list_dbs()} +* mysql_list_fields:: @code{mysql_list_fields()} +* mysql_list_processes:: @code{mysql_list_processes()} +* mysql_list_tables:: @code{mysql_list_tables()} +* mysql_num_fields:: @code{mysql_num_fields()} +* mysql_num_rows:: @code{mysql_num_rows()} +* mysql_options:: @code{mysql_options()} +* mysql_ping:: @code{mysql_ping()} +* mysql_query:: @code{mysql_query()} +* mysql_real_connect:: @code{mysql_real_connect()} +* mysql_real_escape_string:: @code{mysql_real_escape_string()} +* mysql_real_query:: @code{mysql_real_query()} +* mysql_reload:: @code{mysql_reload()} +* mysql_row_seek:: @code{mysql_row_seek()} +* mysql_row_tell:: @code{mysql_row_tell()} +* mysql_select_db:: @code{mysql_select_db()} +* mysql_shutdown:: @code{mysql_shutdown()} +* mysql_stat:: @code{mysql_stat()} +* mysql_store_result:: @code{mysql_store_result()} +* mysql_thread_id:: @code{mysql_thread_id()} +* mysql_use_result:: @code{mysql_use_result()} +@end menu -@subsubheading Example +In the descriptions below, a parameter or return value of @code{NULL} means +@code{NULL} in the sense of the C programming language, not a +@strong{MySQL} @code{NULL} value. + +Functions that return a value generally return a pointer or an integer. +Unless specified otherwise, functions returning a pointer return a +non-@code{NULL} value to indicate success or a @code{NULL} value to indicate +an error, and functions returning an integer return zero to indicate success +or non-zero to indicate an error. Note that ``non-zero'' means just that. +Unless the function description says otherwise, do not test against a value +other than zero: @example -char query[1000],*end; +if (result) /* correct */ + ... error ... -end = strmov(query,"INSERT INTO test_table values("); -*end++ = '\''; -end += mysql_real_escape_string(&mysql, end,"What's this",11); -*end++ = '\''; -*end++ = ','; -*end++ = '\''; -end += mysql_real_escape_string(&mysql, end,"binary data: \0\r\n",16); -*end++ = '\''; -*end++ = ')'; +if (result < 0) /* incorrect */ + ... error ... -if (mysql_real_query(&mysql,query,(unsigned int) (end - query))) -@{ - fprintf(stderr, "Failed to insert row, Error: %s\n", - mysql_error(&mysql)); -@} +if (result == -1) /* incorrect */ + ... error ... @end example -The @code{strmov()} function used in the example is included in the -@code{mysqlclient} library and works like @code{strcpy()} but returns a -pointer to the terminating null of the first parameter. +When a function returns an error, the @strong{Errors} subsection of the +function description lists the possible types of errors. You can +find out which of these occurred by calling @code{mysql_errno()}. +A string representation of the error may be obtained by calling +@code{mysql_error()}. + +@findex @code{mysql_affected_rows()} +@node mysql_affected_rows, mysql_close, C API functions, C API functions +@subsubsection @code{mysql_affected_rows()} + +@code{my_ulonglong mysql_affected_rows(MYSQL *mysql)} + +@subsubheading Description + +Returns the number of rows changed by the last @code{UPDATE}, deleted by +the last @code{DELETE} or inserted by the last @code{INSERT} +statement. May be called immediately after @code{mysql_query()} for +@code{UPDATE}, @code{DELETE}, or @code{INSERT} statements. For +@code{SELECT} statements, @code{mysql_affected_rows()} works like +@code{mysql_num_rows()}. @subsubheading Return Values -The length of the value placed into @code{to}, not including the -terminating null character. +An integer greater than zero indicates the number of rows affected or +retrieved. Zero indicates that no records where updated for an +@code{UPDATE} statement, no rows matched the @code{WHERE} clause in the +query or that no query has yet been executed. -1 indicates that the +query returned an error or that, for a @code{SELECT} query, +@code{mysql_affected_rows()} was called prior to calling +@code{mysql_store_result()}. @subsubheading Errors + None. -@findex @code{mysql_real_query()} -@node mysql_real_query, mysql_reload, mysql_real_escape_string, C API functions -@subsubsection @code{mysql_real_query()} +@subsubheading Example -@code{int mysql_real_query(MYSQL *mysql, const char *query, unsigned int length)} +@example +mysql_query(&mysql,"UPDATE products SET cost=cost*1.25 WHERE group=10"); +printf("%ld products updated",(long) mysql_affected_rows(&mysql)); +@end example -@subsubheading Description +If one specifies the flag @code{CLIENT_FOUND_ROWS} when connecting to +@code{mysqld}, @code{mysql_affected_rows()} will return the number of +rows matched by the @code{WHERE} statement for @code{UPDATE} statements. -Executes the SQL query pointed to by @code{query}, which should be a string -@code{length} bytes long. The query must consist of a single SQL statement. -You should not add a terminating semicolon (@samp{;}) or @code{\g} to the -statement. +Note that when one uses a @code{REPLACE} command, +@code{mysql_affected_rows()} will return 2 if the new row replaced and +old row. This is because in this case one row was inserted and then the +duplicate was deleted. -You @emph{must} use @code{mysql_real_query()} rather than -@code{mysql_query()} for queries that contain binary data, because binary data -may contain the @samp{\0} character. In addition, @code{mysql_real_query()} -is faster than @code{mysql_query()} because it does not call @code{strlen()} on -the query string. +@findex @code{mysql_close()} +@node mysql_close, mysql_connect, mysql_affected_rows, C API functions +@subsubsection @code{mysql_close()} -If you want to know if the query should return a result set or not, you can -use @code{mysql_field_count()} to check for this. -@xref{mysql_field_count, @code{mysql_field_count}}. +@code{void mysql_close(MYSQL *mysql)} + +@subsubheading Description +Closes a previously opened connection. @code{mysql_close()} also deallocates +the connection handle pointed to by @code{mysql} if the handle was allocated +automatically by @code{mysql_init()} or @code{mysql_connect()}. @subsubheading Return Values -Zero if the query was successful. Non-zero if an error occurred. +None. @subsubheading Errors -@table @code -@item CR_COMMANDS_OUT_OF_SYNC -Commands were executed in an improper order. -@item CR_SERVER_GONE_ERROR -The @strong{MySQL} server has gone away. -@item CR_SERVER_LOST -The connection to the server was lost during the query. -@item CR_UNKNOWN_ERROR -An unknown error occurred. -@end table +None. -@findex @code{mysql_reload()} -@node mysql_reload, mysql_row_seek, mysql_real_query, C API functions -@subsubsection @code{mysql_reload()} +@findex @code{mysql_connect()} +@node mysql_connect, mysql_change_user, mysql_close, C API functions +@subsubsection @code{mysql_connect()} -@code{int mysql_reload(MYSQL *mysql)} +@code{MYSQL *mysql_connect(MYSQL *mysql, const char *host, const char *user, const char *passwd)} @subsubheading Description -Asks the @strong{MySQL} server to reload the grant tables. The -connected user must have the @strong{reload} privilege. +This function is deprecated. It is preferable to use +@code{mysql_real_connect()} instead. -This function is deprecated. It is preferable to use @code{mysql_query()} -to issue a SQL @code{FLUSH PRIVILEGES} statement instead. +@code{mysql_connect()} attempts to establish a connection to a @strong{MySQL} +database engine running on @code{host}. @code{mysql_connect()} must complete +successfully before you can execute any of the other API functions, with the +exception of @code{mysql_get_client_info()}. + +The meanings of the parameters are the same as for the corresponding +parameters for @code{mysql_real_connect()} with the difference that the +connection parameter may be @code{NULL}. In this case the C API +allocates memory for the connection structure automatically and frees it +when you call @code{mysql_close()}. The disadvantage of this approach is +that you can't retrieve an error message if the connection fails. (To +get error information from @code{mysql_errno()} or @code{mysql_error()}, +you must provide a valid @code{MYSQL} pointer.) @subsubheading Return Values -Zero for success. Non-zero if an error occurred. +Same as for @code{mysql_real_connect()}. @subsubheading Errors -@table @code -@item CR_COMMANDS_OUT_OF_SYNC -Commands were executed in an improper order. -@item CR_SERVER_GONE_ERROR -The @strong{MySQL} server has gone away. -@item CR_SERVER_LOST -The connection to the server was lost during the query. -@item CR_UNKNOWN_ERROR -An unknown error occurred. -@end table +Same as for @code{mysql_real_connect()}. -@findex @code{mysql_row_seek()} -@node mysql_row_seek, mysql_row_tell, mysql_reload, C API functions -@subsubsection @code{mysql_row_seek()} +@findex @code{mysql_change_user()} +@node mysql_change_user, mysql_character_set_name, mysql_connect, C API functions +@subsubsection @code{mysql_change_user()} -@code{MYSQL_ROW_OFFSET mysql_row_seek(MYSQL_RES *result, MYSQL_ROW_OFFSET offset)} +@code{my_bool mysql_change_user(MYSQL *mysql, const char *user, const +char *password, const char *db)} @subsubheading Description -Sets the row cursor to an arbitrary row in a query result set. This requires -that the result set structure contains the entire result of the query, so -@code{mysql_row_seek()} may be used in conjunction only with -@code{mysql_store_result()}, not with @code{mysql_use_result()}. - -The offset should be a value returned from a call to @code{mysql_row_tell()} -or to @code{mysql_row_seek()}. This value is not simply a row number; if you -want to seek to a row within a result set using a row number, use -@code{mysql_data_seek()} instead. - -@subsubheading Return Values - -The previous value of the row cursor. This value may be passed to a -subsequent call to @code{mysql_row_seek()}. - -@subsubheading Errors -None. - -@findex @code{mysql_row_tell()} -@node mysql_row_tell, mysql_select_db, mysql_row_seek, C API functions -@subsubsection @code{mysql_row_tell()} - -@code{MYSQL_ROW_OFFSET mysql_row_tell(MYSQL_RES *result)} - -@subsubheading Description - -Returns the current position of the row cursor for the last -@code{mysql_fetch_row()}. This value can be used as an argument to -@code{mysql_row_seek()}. - -You should use @code{mysql_row_tell()} only after @code{mysql_store_result()}, -not after @code{mysql_use_result()}. - -@subsubheading Return Values -The current offset of the row cursor. - -@subsubheading Errors -None. - -@findex @code{mysql_select_db()} -@node mysql_select_db, mysql_shutdown, mysql_row_tell, C API functions -@subsubsection @code{mysql_select_db()} - -@code{int mysql_select_db(MYSQL *mysql, const char *db)} +Changes the user and causes the database specified by @code{db} to +become the default (current) database on the connection specified by +@code{mysql}. In subsequent queries, this database is the default for +table references that do not include an explicit database specifier. -@subsubheading Description +This function was introduced in @strong{MySQL} Version 3.23.3. -Causes the database specified by @code{db} to become the default (current) -database on the connection specified by @code{mysql}. In subsequent queries, -this database is the default for table references that do not include an -explicit database specifier. +@code{mysql_change_user()} fails unless the connected user can be +authenticated or if he doesn't have permission to use the database. In +this case the user and database are not changed -@code{mysql_select_db()} fails unless the connected user can be authenticated -as having permission to use the database. +The @code{db} parameter may be set to @code{NULL} if you don't want to have a +default database. -@subsubheading Return Values +@subsubheading Return values Zero for success. Non-zero if an error occurred. @subsubheading Errors +The same that you can get from @code{mysql_real_connect()}. + @table @code @item CR_COMMANDS_OUT_OF_SYNC Commands were executed in an improper order. @@ -41416,129 +41564,161 @@ The @strong{MySQL} server has gone away. The connection to the server was lost during the query. @item CR_UNKNOWN_ERROR An unknown error occurred. +@item ER_UNKNOWN_COM_ERROR +The @strong{MySQL} server doesn't implement this command (probably an old server) +@item ER_ACCESS_DENIED_ERROR +The user or password was wrong. +@item ER_BAD_DB_ERROR +The database didn't exist. +@item ER_DBACCESS_DENIED_ERROR +The user did not have access rights to the database. +@item ER_WRONG_DB_NAME +The database name was too long. @end table -@findex @code{mysql_shutdown()} -@node mysql_shutdown, mysql_stat, mysql_select_db, C API functions -@subsubsection @code{mysql_shutdown()} +@subsubheading Example -@code{int mysql_shutdown(MYSQL *mysql)} +@example +if (mysql_change_user(&mysql, "user", "password", "new_database")) +@{ + fprintf(stderr, "Failed to change user. Error: %s\n", + mysql_error(&mysql)); +@} +@end example + +@findex @code{mysql_character_set_name()} +@node mysql_character_set_name, mysql_create_db, mysql_change_user, C API functions +@subsubsection @code{mysql_character_set_name()} + +@code{const char *mysql_character_set_name(MYSQL *mysql)} @subsubheading Description -Asks the database server to shut down. The connected user must have -@strong{shutdown} privileges. +Returns the default character set for the current connection. @subsubheading Return Values -Zero for success. Non-zero if an error occurred. +The default character set @subsubheading Errors +None. -@table @code -@item CR_COMMANDS_OUT_OF_SYNC -Commands were executed in an improper order. -@item CR_SERVER_GONE_ERROR -The @strong{MySQL} server has gone away. -@item CR_SERVER_LOST -The connection to the server was lost during the query. -@item CR_UNKNOWN_ERROR -An unknown error occurred. -@end table -@findex @code{mysql_stat()} -@node mysql_stat, mysql_store_result, mysql_shutdown, C API functions -@subsubsection @code{mysql_stat()} +@findex @code{mysql_create_db()} +@node mysql_create_db, mysql_data_seek, mysql_character_set_name, C API functions +@subsubsection @code{mysql_create_db()} -@code{char *mysql_stat(MYSQL *mysql)} +@code{int mysql_create_db(MYSQL *mysql, const char *db)} @subsubheading Description +Creates the database named by the @code{db} parameter. -Returns a character string containing information similar to that provided by -the @code{mysqladmin status} command. This includes uptime in seconds and -the number of running threads, questions, reloads, and open tables. +This function is deprecated. It is preferable to use @code{mysql_query()} +to issue a SQL @code{CREATE DATABASE} statement instead. @subsubheading Return Values -A character string describing the server status. @code{NULL} if an -error occurred. +Zero if the database was created successfully. Non-zero if an error +occurred. @subsubheading Errors - @table @code + @item CR_COMMANDS_OUT_OF_SYNC Commands were executed in an improper order. + @item CR_SERVER_GONE_ERROR The @strong{MySQL} server has gone away. + @item CR_SERVER_LOST The connection to the server was lost during the query. + @item CR_UNKNOWN_ERROR An unknown error occurred. @end table -@findex @code{mysql_store_result()} -@node mysql_store_result, mysql_thread_id, mysql_stat, C API functions -@subsubsection @code{mysql_store_result()} +@subsubheading Example -@code{MYSQL_RES *mysql_store_result(MYSQL *mysql)} +@example +if(mysql_create_db(&mysql, "my_database")) +@{ + fprintf(stderr, "Failed to create new database. Error: %s\n", + mysql_error(&mysql)); +@} +@end example + +@findex @code{mysql_data_seek()} +@node mysql_data_seek, mysql_debug, mysql_create_db, C API functions +@subsubsection @code{mysql_data_seek()} + +@code{void mysql_data_seek(MYSQL_RES *result, unsigned long long offset)} @subsubheading Description +Seeks to an arbitrary row in a query result set. This requires that the +result set structure contains the entire result of the query, so +@code{mysql_data_seek()} may be used in conjunction only with +@code{mysql_store_result()}, not with @code{mysql_use_result()}. -You must call @code{mysql_store_result()} or @code{mysql_use_result()} -for every query that successfully retrieves data (@code{SELECT}, -@code{SHOW}, @code{DESCRIBE}, @code{EXPLAIN}). +The offset should be a value in the range from 0 to +@code{mysql_num_rows(result)-1}. -You don't have to call @code{mysql_store_result()} or -@code{mysql_use_result()} for other queries, but it will not do any -harm or cause any notable performance if you call @code{mysql_store_result()} -in all cases. You can detect if the query didn't have a result set by -checking if @code{mysql_store_result()} returns 0 (more about this later one). +@subsubheading Return Values -If you want to know if the query should return a result set or not, you can -use @code{mysql_field_count()} to check for this. -@xref{mysql_field_count, @code{mysql_field_count}}. +None. -@code{mysql_store_result()} reads the entire result of a query to the client, -allocates a @code{MYSQL_RES} structure, and places the result into this -structure. +@subsubheading Errors +None. -@code{mysql_store_results()} returns a null pointer if the query didn't return -a result set (if the query was, for example, an @code{INSERT} statement). +@findex @code{mysql_debug()} +@node mysql_debug, mysql_drop_db, mysql_data_seek, C API functions +@subsubsection @code{mysql_debug()} -@code{mysql_store_results()} also returns a null pointer if reading of the -result set failed. You can check if you got an error by checking if -@code{mysql_error()} doesn't return a null pointer, if -@code{mysql_errno()} returns <> 0, or if @code{mysql_field_count()} -returns <> 0. +@code{void mysql_debug(char *debug)} -An empty result set is returned if there are no rows returned. (An empty -result set differs from a null pointer as a return value.) +@subsubheading Description +Does a @code{DBUG_PUSH} with the given string. @code{mysql_debug()} uses the +Fred Fish debug library. To use this function, you must compile the client +library to support debugging. +@xref{Debugging server}. @xref{Debugging client}. -Once you have called @code{mysql_store_result()} and got a result back -that isn't a null pointer, you may call @code{mysql_num_rows()} to find -out how many rows are in the result set. +@subsubheading Return Values -You can call @code{mysql_fetch_row()} to fetch rows from the result set, -or @code{mysql_row_seek()} and @code{mysql_row_tell()} to obtain or -set the current row position within the result set. +None. -You must call @code{mysql_free_result()} once you are done with the result -set. +@subsubheading Errors +None. -@xref{NULL mysql_store_result, , @code{NULL mysql_store_result()}}. +@subsubheading Example + +The call shown below causes the client library to generate a trace file in +@file{/tmp/client.trace} on the client machine: + +@example +mysql_debug("d:t:O,/tmp/client.trace"); +@end example + +@findex @code{mysql_drop_db()} +@node mysql_drop_db, mysql_dump_debug_info, mysql_debug, C API functions +@subsubsection @code{mysql_drop_db()} + +@code{int mysql_drop_db(MYSQL *mysql, const char *db)} + +@subsubheading Description +Drops the database named by the @code{db} parameter. + +This function is deprecated. It is preferable to use @code{mysql_query()} +to issue a SQL @code{DROP DATABASE} statement instead. @subsubheading Return Values -A @code{MYSQL_RES} result structure with the results. @code{NULL} if -an error occurred. +Zero if the database was dropped successfully. Non-zero if an error +occurred. @subsubheading Errors @table @code @item CR_COMMANDS_OUT_OF_SYNC Commands were executed in an improper order. -@item CR_OUT_OF_MEMORY -Out of memory. @item CR_SERVER_GONE_ERROR The @strong{MySQL} server has gone away. @item CR_SERVER_LOST @@ -41547,83 +41727,33 @@ The connection to the server was lost during the query. An unknown error occurred. @end table -@findex @code{mysql_thread_id()} -@node mysql_thread_id, mysql_use_result, mysql_store_result, C API functions -@subsubsection @code{mysql_thread_id()} - -@code{unsigned long mysql_thread_id(MYSQL *mysql)} - -@subsubheading Description - -Returns the thread ID of the current connection. This value can be used as -an argument to @code{mysql_kill()} to kill the thread. - -If the connection is lost and you reconnect with @code{mysql_ping()}, the -thread ID will change. This means you should not get the thread ID and store -it for later. You should get it when you need it. - -@subsubheading Return Values - -The thread ID of the current connection. +@subsubheading Example -@subsubheading Errors -None. +@example +if(mysql_drop_db(&mysql, "my_database")) + fprintf(stderr, "Failed to drop the database: Error: %s\n", + mysql_error(&mysql)); +@end example -@findex @code{mysql_use_result()} -@node mysql_use_result, , mysql_thread_id, C API functions -@subsubsection @code{mysql_use_result()} +@findex @code{mysql_dump_debug_info()} +@node mysql_dump_debug_info, mysql_eof, mysql_drop_db, C API functions +@subsubsection @code{mysql_dump_debug_info()} -@code{MYSQL_RES *mysql_use_result(MYSQL *mysql)} +@code{int mysql_dump_debug_info(MYSQL *mysql)} @subsubheading Description -You must call @code{mysql_store_result()} or @code{mysql_use_result()} for -every query that successfully retrieves data (@code{SELECT}, @code{SHOW}, -@code{DESCRIBE}, @code{EXPLAIN}). - -@code{mysql_use_result()} initiates a result set retrieval but does not -actually read the result set into the client like @code{mysql_store_result()} -does. Instead, each row must be retrieved individually by making calls to -@code{mysql_fetch_row()}. This reads the result of a query directly from the -server without storing it in a temporary table or local buffer, which is -somewhat faster and uses much less memory than @code{mysql_store_result()}. -The client will only allocate memory for the current row and a communication -buffer that may grow up to @code{max_allowed_packet} bytes. - -On the other hand, you shouldn't use @code{mysql_use_result()} if you are -doing a lot of processing for each row on the client side, or if the output -is sent to a screen on which the user may type a @code{^S} (stop scroll). -This will tie up the server and prevent other threads from updating any -tables from which the data is being fetched. - -When using @code{mysql_use_result()}, you must execute -@code{mysql_fetch_row()} until a @code{NULL} value is returned, otherwise the -unfetched rows will be returned as part of the result set for your next -query. The C API will give the error @code{Commands out of sync; You can't -run this command now} if you forget to do this! - -You may not use @code{mysql_data_seek()}, @code{mysql_row_seek()}, -@code{mysql_row_tell()}, @code{mysql_num_rows()}, or -@code{mysql_affected_rows()} with a result returned from -@code{mysql_use_result()}, nor may you issue other queries until the -@code{mysql_use_result()} has finished. (However, after you have fetched all -the rows, @code{mysql_num_rows()} will accurately return the number of rows -fetched.) - -You must call @code{mysql_free_result()} once you are done with the result -set. +Instructs the server to write some debug information to the log. The +connected user must have the @strong{process} privilege for this to work. -@subsubheading Return Values +@subsubheading Return values -A @code{MYSQL_RES} result structure. @code{NULL} if an error occurred. +Zero if the command was successful. Non-zero if an error occurred. @subsubheading Errors - @table @code @item CR_COMMANDS_OUT_OF_SYNC Commands were executed in an improper order. -@item CR_OUT_OF_MEMORY -Out of memory. @item CR_SERVER_GONE_ERROR The @strong{MySQL} server has gone away. @item CR_SERVER_LOST @@ -41632,9910 +41762,9861 @@ The connection to the server was lost during the query. An unknown error occurred. @end table -@node C API problems, Thread-safe clients, C API functions, C -@subsection Common questions and problems when using the C API - -@tindex @code{mysql_query()} -@tindex @code{mysql_store_result()} -@menu -* NULL mysql_store_result:: -* Query results:: -* Getting unique ID:: -* C API linking problems:: -@end menu - -@node NULL mysql_store_result, Query results, C API problems, C API problems -@subsubsection Why Is It that After @code{mysql_query()} Returns Success, @code{mysql_store_result()} Sometimes Returns @code{NULL?} - -It is possible for @code{mysql_store_result()} to return @code{NULL} -following a successful call to @code{mysql_query()}. When this happens, it -means one of the following conditions occurred: - -@itemize @bullet -@item -There was a @code{malloc()} failure (for example, if the result set was too -large). - -@item -The data couldn't be read (an error occurred on the connection). - -@item -The query returned no data (for example, it was an @code{INSERT}, -@code{UPDATE}, or @code{DELETE}). -@end itemize - -You can always check whether or not the statement should have produced a -non-empty result by calling @code{mysql_field_count()}. If -@code{mysql_field_count()} returns zero, the result is empty and the last -query was a statement that does not return values (for example, an -@code{INSERT} or a @code{DELETE}). If @code{mysql_field_count()} returns a -non-zero value, the statement should have produced a non-empty result. -See the description of the @code{mysql_field_count()} function for an -example. +@findex @code{mysql_eof()} +@node mysql_eof, mysql_errno, mysql_dump_debug_info, C API functions +@subsubsection @code{mysql_eof()} -You can test for an error by calling @code{mysql_error()} or -@code{mysql_errno()}. +@code{my_bool mysql_eof(MYSQL_RES *result)} -@cindex queries, C API results -@menu -* Query results:: -* Getting unique ID:: -* C API linking problems:: -@end menu +@subsubheading Description -@node Query results, Getting unique ID, NULL mysql_store_result, C API problems -@subsubsection What Results Can I Get From a Query? +This function is deprecated. @code{mysql_errno()} or @code{mysql_error()} +may be used instead. -In addition to the result set returned by a query, you can also get the -following information: +@code{mysql_eof()} determines whether or not the last row of a result +set has been read. -@itemize @bullet -@item -@code{mysql_affected_rows()} returns the number of rows affected by the last -query when doing an @code{INSERT}, @code{UPDATE}, or @code{DELETE}. An -exception is that if @code{DELETE} is used without a @code{WHERE} clause, the -table is re-created empty, which is much faster! In this case, -@code{mysql_affected_rows()} returns zero for the number of records -affected. +If you acquire a result set from a successful call to +@code{mysql_store_result()}, the client receives the entire set in one +operation. In this case, a @code{NULL} return from @code{mysql_fetch_row()} +always means the end of the result set has been reached and it is +unnecessary to call @code{mysql_eof()}. -@item -@code{mysql_num_rows()} returns the number of rows in a result set. With -@code{mysql_store_result()}, @code{mysql_num_rows()} may be called as soon as -@code{mysql_store_result()} returns. With @code{mysql_use_result()}, -@code{mysql_num_rows()} may be called only after you have fetched all the -rows with @code{mysql_fetch_row()}. +On the other hand, if you use @code{mysql_use_result()} to initiate a result +set retrieval, the rows of the set are obtained from the server one by one as +you call @code{mysql_fetch_row()} repeatedly. Because an error may occur on +the connection during this process, a @code{NULL} return value from +@code{mysql_fetch_row()} does not necessarily mean the end of the result set +was reached normally. In this case, you can use @code{mysql_eof()} to +determine what happened. @code{mysql_eof()} returns a non-zero value if the +end of the result set was reached and zero if an error occurred. -@item -@code{mysql_insert_id()} returns the ID generated by the last -query that inserted a row into a table with an @code{AUTO_INCREMENT} index. -@xref{mysql_insert_id, , @code{mysql_insert_id()}}. +Historically, @code{mysql_eof()} predates the standard @strong{MySQL} error +functions @code{mysql_errno()} and @code{mysql_error()}. Because those error +functions provide the same information, their use is preferred over +@code{mysql_eof()}, which is now deprecated. (In fact, they provide more +information, because @code{mysql_eof()} returns only a boolean value whereas +the error functions indicate a reason for the error when one occurs.) -@item -Some queries (@code{LOAD DATA INFILE ...}, @code{INSERT INTO -... SELECT ...}, @code{UPDATE}) return additional information. The result is -returned by @code{mysql_info()}. See the description for @code{mysql_info()} -for the format of the string that it returns. @code{mysql_info()} returns a -@code{NULL} pointer if there is no additional information. -@end itemize +@subsubheading Return Values -@cindex unique ID -@cindex last row, unique ID -@cindex ID, unique -@cindex tables, unique ID for last row -@node Getting unique ID, C API linking problems, Query results, C API problems -@subsubsection How Can I Get the Unique ID for the Last Inserted Row? +Zero if no error occurred. Non-zero if the end of the result set has been +reached. -If you insert a record in a table containing a column that has the -@code{AUTO_INCREMENT} attribute, you can get the most recently generated -ID by calling the @code{mysql_insert_id()} function. +@subsubheading Errors +None. -You can also retrieve the ID by using the @code{LAST_INSERT_ID()} function in -a query string that you pass to @code{mysql_query()}. +@subsubheading Example -You can check if an @code{AUTO_INCREMENT} index is used by executing -the following code. This also checks if the query was an @code{INSERT} with -an @code{AUTO_INCREMENT} index: +The following example shows how you might use @code{mysql_eof()}: @example -if (mysql_error(&mysql)[0] == 0 && - mysql_num_fields(result) == 0 && - mysql_insert_id(&mysql) != 0) +mysql_query(&mysql,"SELECT * FROM some_table"); +result = mysql_use_result(&mysql); +while((row = mysql_fetch_row(result))) @{ - used_id = mysql_insert_id(&mysql); + // do something with data +@} +if(!mysql_eof(result)) // mysql_fetch_row() failed due to an error +@{ + fprintf(stderr, "Error: %s\n", mysql_error(&mysql)); @} @end example -The most recently generated ID is maintained in the server on a -per-connection basis. It will not be changed by another client. It will not -even be changed if you update another @code{AUTO_INCREMENT} column with a -non-magic value (that is, a value that is not @code{NULL} and not @code{0}). - -If you want to use the ID that was generated for one table and insert -it into a second table, you can use SQL statements like this: +However, you can achieve the same effect with the standard @strong{MySQL} +error functions: @example -INSERT INTO foo (auto,text) - VALUES(NULL,'text'); # generate ID by inserting NULL -INSERT INTO foo2 (id,text) - VALUES(LAST_INSERT_ID(),'text'); # use ID in second table +mysql_query(&mysql,"SELECT * FROM some_table"); +result = mysql_use_result(&mysql); +while((row = mysql_fetch_row(result))) +@{ + // do something with data +@} +if(mysql_errno(&mysql)) // mysql_fetch_row() failed due to an error +@{ + fprintf(stderr, "Error: %s\n", mysql_error(&mysql)); +@} @end example -@cindex linking, problems -@cindex C API, linking problems -@node C API linking problems, , Getting unique ID, C API problems -@subsubsection Problems Linking with the C API +@findex @code{mysql_errno()} +@node mysql_errno, mysql_error, mysql_eof, C API functions +@subsubsection @code{mysql_errno()} -When linking with the C API, the following errors may occur on some systems: +@code{unsigned int mysql_errno(MYSQL *mysql)} -@example -gcc -g -o client test.o -L/usr/local/lib/mysql -lmysqlclient -lsocket -lnsl +@subsubheading Description -Undefined first referenced - symbol in file -floor /usr/local/lib/mysql/libmysqlclient.a(password.o) -ld: fatal: Symbol referencing errors. No output written to client -@end example +For the connection specified by @code{mysql}, @code{mysql_errno()} returns +the error code for the most recently invoked API function that can succeed +or fail. A return value of zero means that no error occurred. Client error +message numbers are listed in the @strong{MySQL} @file{errmsg.h} header file. +Server error message numbers are listed in @file{mysqld_error.h}. In the +@strong{MySQL} source distribution you can find a complete list of +error messages and error numbers in the file @file{Docs/mysqld_error.txt}. -If this happens on your system, you must include the math library by -adding @code{-lm} to the end of the compile/link line. +@subsubheading Return Values -@cindex clients, thread-safe -@cindex thread-safe clients -@node Thread-safe clients, , C API problems, C -@subsection How to Make a Thread-safe Client +An error code value. Zero if no error occurred. -The client library is almost thread safe. The biggest problem is -that the subroutines in @file{net.c} that read from sockets are not -interrupt safe. This was done with the thought that you might want to -have your own alarm that can break a long read to a server. If you -install interrupt handlers for the @code{SIGPIPE} interrupt, -the socket handling should be thread safe. +@subsubheading Errors +None. -In the older binaries we distribute on our Web site, the client -libraries are not normally compiled with the thread-safe option (the -Windows binaries are by default compiled to be thread safe). -Newer binary distributions should have both a normal and a -thread-safe client library. - -To get a really thread-safe client where you can interrupt the client -from other threads and set timeouts when talking with the @strong{MySQL} -server, you should use the @code{-lmysys}, @code{-lstring}, and @code{-ldbug} -libraries and the @code{net_serv.o} code that the server uses. - -If you don't need interrupts or timeouts, you can just compile a thread -safe client library @code{(mysqlclient_r)} and use this. @xref{C,, -MySQL C API}. In this case you don't have to worry about the -@code{net_serv.o} object file or the other @strong{MySQL} libraries. +@findex @code{mysql_error()} +@node mysql_error, mysql_escape_string, mysql_errno, C API functions +@subsubsection @code{mysql_error()} -When using a threaded client and you want to use timeouts and interrupts, -you can make great use of the routines in the @file{thr_alarm.c} file. -If you are using routines from the @code{mysys} library, the only thing -you must remember is to call @code{my_init()} first! +@code{char *mysql_error(MYSQL *mysql)} -All functions except @code{mysql_real_connect()} are by default -thread safe. The following notes describe how to compile a thread safe -client library and use it in a thread-safe manner. (The notes below for -@code{mysql_real_connect()} actually apply to @code{mysql_connect()} as -well, but because @code{mysql_connect()} is deprecated, you should be -using @code{mysql_real_connect()} anyway.) +@subsubheading Description -To make @code{mysql_real_connect()} thread safe, you must recompile the -client library with this command: +For the connection specified by @code{mysql}, @code{mysql_error()} returns +the error message for the most recently invoked API function that can succeed +or fail. An empty string (@code{""}) is returned if no error occurred. +This means the following two tests are equivalent: @example -shell> ./configure --with-thread-safe-client +if(mysql_errno(&mysql)) +@{ + // an error occurred +@} + +if(mysql_error(&mysql)[0] != '\0') +@{ + // an error occurred +@} @end example -This will create a thread-safe client library @code{libmysqlclient_r}. -@code{--with-thread-safe-client}. This library is thread safe per -connection. You can let two threads share the same connection as long -as you do the following: +The language of the client error messages may be changed by +recompiling the @strong{MySQL} client library. Currently you can choose +error messages in several different languages. +@xref{Languages}. -@itemize @bullet -@item -Two threads can't send a query to the @strong{MySQL} at the same time on -the same connection. In particular, you have to ensure that between a -@code{mysql_query()} and @code{mysql_store_result()} no other thread is using -the same connection. -@item -Many threads can access different result sets that are retrieved with -@code{mysql_store_result()}. -@item -If you use @code{mysql_use_result}, you have to ensure that no other thread -is asking anything on the same connection until the result set is closed. -However, it really is best for threaded clients that share the same -connection to use @code{mysql_use_result()}. -@item -If you want to use multiple threads on the same connection, you must -have a mutex lock around your @code{mysql_query()} and -@code{mysql_store_result()} call combination. Once -@code{mysql_store_result()} is ready, the lock can be released and other -threads may query the same connection. -@item -If you program with POSIX threads, you can use -@code{pthread_mutex_lock()} and @code{pthread_mutex_unlock()} to -establish and release a mutex lock. -@end itemize +@subsubheading Return Values -You may get some errors because of undefined symbols when linking your -client with @code{mysqlclient_r}. In most cases this is because you haven't -included the thread libraries on the link/compile line. +A character string that describes the error. An empty string if no error +occurred. -@cindex APIs, Perl -@cindex Perl API -@node Perl, Eiffel, C, Clients -@section MySQL Perl API +@subsubheading Errors +None. -This section documents the Perl @code{DBI} interface. The former interface -was called @code{mysqlperl}. @code{DBI}/@code{DBD} now is the -recommended Perl interface, so @code{mysqlperl} is obsolete and is not -documented here. +@findex @code{mysql_escape_string()} +@node mysql_escape_string, mysql_fetch_field, mysql_error, C API functions +@subsubsection @code{mysql_escape_string()} -@menu -* DBI with DBD:: @code{DBI} with @code{DBD::mysql} -* Perl DBI Class:: The @code{DBI} interface -* DBI-info:: More @code{DBI}/@code{DBD} information -@end menu +You should use @code{mysql_real_escape_string()} instead! -@cindex @code{DBI} interface -@node DBI with DBD, Perl DBI Class, Perl, Perl -@subsection @code{DBI} with @code{DBD::mysql} +This is identical to @code{mysql_real_escape_string()} except that it +takes the connection as the first +argument. @code{mysql_real_escape_string()} will escape the string +according to the current character set while +@code{mysql_escape_string()} does not respect the current charset +setting. -@code{DBI} is a generic interface for many databases. That means that -you can write a script that works with many different database engines -without change. You need a DataBase Driver (DBD) defined for each -database type. For @strong{MySQL}, this driver is called -@code{DBD::mysql}. +@findex @code{mysql_fetch_field()} +@node mysql_fetch_field, mysql_fetch_fields, mysql_escape_string, C API functions +@subsubsection @code{mysql_fetch_field()} -For more information on the Perl5 DBI, please visit the @code{DBI} Web -page and read the documentation: -@example -@uref{http://www.symbolstone.org/technology/perl/DBI/index.html} -@end example -For more information on Object Oriented Programming -(OOP) as defined in Perl5, see the Perl OOP page: -@example -@uref{http://language.perl.com/info/documentation.html} -@end example +@code{MYSQL_FIELD *mysql_fetch_field(MYSQL_RES *result)} -Note that if you want to use transactions with Perl, you need to have -@code{Msql-Mysql-modules} version 1.2216 or newer. +@subsubheading Description -Installation instructions for @strong{MySQL} Perl support are given in -@ref{Perl support}. +Returns the definition of one column of a result set as a @code{MYSQL_FIELD} +structure. Call this function repeatedly to retrieve information about all +columns in the result set. @code{mysql_fetch_field()} returns @code{NULL} +when no more fields are left. -@cindex @code{DBI} Perl module -@node Perl DBI Class, DBI-info, DBI with DBD, Perl -@subsection The @code{DBI} Interface +@code{mysql_fetch_field()} is reset to return information about the first +field each time you execute a new @code{SELECT} query. The field returned by +@code{mysql_fetch_field()} is also affected by calls to +@code{mysql_field_seek()}. -@noindent -@strong{Portable DBI Methods} +If you've called @code{mysql_query()} to perform a @code{SELECT} on a table +but have not called @code{mysql_store_result()}, @strong{MySQL} returns the +default blob length (8K bytes) if you call @code{mysql_fetch_field()} to ask +for the length of a @code{BLOB} field. (The 8K size is chosen because +@strong{MySQL} doesn't know the maximum length for the @code{BLOB}. This +should be made configurable sometime.) Once you've retrieved the result set, +@code{field->max_length} contains the length of the largest value for this +column in the specific query. -@multitable @columnfractions .3 .7 -@item @code{connect} @tab Establishes a connection to a database server. -@item @code{disconnect} @tab Disconnects from the database server. -@item @code{prepare} @tab Prepares a SQL statement for execution. -@item @code{execute} @tab Executes prepared statements. -@item @code{do} @tab Prepares and executes a SQL statement. -@item @code{quote} @tab Quotes string or @code{BLOB} values to be inserted. -@item @code{fetchrow_array} @tab Fetches the next row as an array of fields. -@item @code{fetchrow_arrayref} @tab Fetches next row as a reference array of fields. -@item @code{fetchrow_hashref} @tab Fetches next row as a reference to a hashtable. -@item @code{fetchall_arrayref} @tab Fetches all data as an array of arrays. -@item @code{finish} @tab Finishes a statement and lets the system free resources. -@item @code{rows} @tab Returns the number of rows affected. -@item @code{data_sources} @tab Returns an array of databases available on localhost. -@item @code{ChopBlanks} @tab Controls whether @code{fetchrow_*} methods trim spaces. -@item @code{NUM_OF_PARAMS} @tab The number of placeholders in the prepared statement. -@item @code{NULLABLE} @tab Which columns can be @code{NULL}. -@item @code{trace} @tab Perform tracing for debugging. -@end multitable +@subsubheading Return Values -@noindent -@strong{MySQL-specific Methods} +The @code{MYSQL_FIELD} structure for the current column. @code{NULL} +if no columns are left. -@multitable @columnfractions .3 .7 -@item @code{insertid} @tab The latest @code{AUTO_INCREMENT} value. -@item @code{is_blob} @tab Which columns are @code{BLOB} values. -@item @code{is_key} @tab Which columns are keys. -@item @code{is_num} @tab Which columns are numeric. -@item @code{is_pri_key} @tab Which columns are primary keys. -@item @code{is_not_null} @tab Which columns CANNOT be @code{NULL}. See @code{NULLABLE}. -@item @code{length} @tab Maximum possible column sizes. -@item @code{max_length} @tab Maximum column sizes actually present in result. -@item @code{NAME} @tab Column names. -@item @code{NUM_OF_FIELDS} @tab Number of fields returned. -@item @code{table} @tab Table names in returned set. -@item @code{type} @tab All column types. -@end multitable +@subsubheading Errors +None. -The Perl methods are described in more detail in the following sections. -Variables used for method return values have these meanings: +@subsubheading Example -@table @code -@item $dbh -Database handle +@example +MYSQL_FIELD *field; -@item $sth -Statement handle +while((field = mysql_fetch_field(result))) +@{ + printf("field name %s\n", field->name); +@} +@end example -@item $rc -Return code (often a status) +@findex @code{mysql_fetch_fields()} +@node mysql_fetch_fields, mysql_fetch_field_direct, mysql_fetch_field, C API functions +@subsubsection @code{mysql_fetch_fields()} -@item $rv -Return value (often a row count) -@end table +@code{MYSQL_FIELD *mysql_fetch_fields(MYSQL_RES *result)} -@noindent -@strong{Portable DBI Methods} +@subsubheading Description -@table @code +Returns an array of all @code{MYSQL_FIELD} structures for a result set. +Each structure provides the field definition for one column of the result +set. + +@subsubheading Return Values + +An array of @code{MYSQL_FIELD} structures for all columns of a result set. + +@subsubheading Errors +None. + +@subsubheading Example -@findex DBI->connect() -@findex connect() DBI method -@item connect($data_source, $username, $password) -Use the @code{connect} method to make a database connection to the data -source. The @code{$data_source} value should begin with -@code{DBI:driver_name:}. -Example uses of @code{connect} with the @code{DBD::mysql} driver: @example -$dbh = DBI->connect("DBI:mysql:$database", $user, $password); -$dbh = DBI->connect("DBI:mysql:$database:$hostname", - $user, $password); -$dbh = DBI->connect("DBI:mysql:$database:$hostname:$port", - $user, $password); +unsigned int num_fields; +unsigned int i; +MYSQL_FIELD *fields; + +num_fields = mysql_num_fields(result); +fields = mysql_fetch_fields(result); +for(i = 0; i < num_fields; i++) +@{ + printf("Field %u is %s\n", i, fields[i].name); +@} @end example -If the user name and/or password are undefined, @code{DBI} uses the -values of the @code{DBI_USER} and @code{DBI_PASS} environment variables, -respectively. If you don't specify a hostname, it defaults to -@code{'localhost'}. If you don't specify a port number, it defaults to the -default @strong{MySQL} port (@value{default_port}). -As of @code{Msql-Mysql-modules} Version 1.2009, -the @code{$data_source} value allows certain modifiers: +@findex @code{mysql_fetch_field_direct()} +@node mysql_fetch_field_direct, mysql_fetch_lengths, mysql_fetch_fields, C API functions +@subsubsection @code{mysql_fetch_field_direct()} -@table @code -@item mysql_read_default_file=file_name -Read @file{filename} as an option file. For information on option files, -see @ref{Option files}. +@code{MYSQL_FIELD *mysql_fetch_field_direct(MYSQL_RES *result, unsigned int fieldnr)} -@item mysql_read_default_group=group_name -The default group when reading an option file is normally the -@code{[client]} group. By specifying the @code{mysql_read_default_group} -option, the default group becomes the @code{[group_name]} group. +@subsubheading Description -@item mysql_compression=1 -Use compressed communication between the client and server (@strong{MySQL} -Version 3.22.3 or later). +Given a field number @code{fieldnr} for a column within a result set, returns +that column's field definition as a @code{MYSQL_FIELD} structure. You may use +this function to retrieve the definition for an arbitrary column. The value +of @code{fieldnr} should be in the range from 0 to +@code{mysql_num_fields(result)-1}. -@item mysql_socket=/path/to/socket -Specify the pathname of the Unix socket that is used to connect -to the server (@strong{MySQL} Version 3.21.15 or later). -@end table +@subsubheading Return Values -Multiple modifiers may be given; each must be preceded by a semicolon. +The @code{MYSQL_FIELD} structure for the specified column. -For example, if you want to avoid hardcoding the user name and password into -a @code{DBI} script, you can take them from the user's @file{~/.my.cnf} -option file instead by writing your @code{connect} call like this: +@subsubheading Errors +None. + +@subsubheading Example @example -$dbh = DBI->connect("DBI:mysql:$database" - . ";mysql_read_default_file=$ENV@{HOME@}/.my.cnf", - $user, $password); +unsigned int num_fields; +unsigned int i; +MYSQL_FIELD *field; + +num_fields = mysql_num_fields(result); +for(i = 0; i < num_fields; i++) +@{ + field = mysql_fetch_field_direct(result, i); + printf("Field %u is %s\n", i, field->name); +@} @end example -This call will read options defined for the @code{[client]} group in the -option file. If you wanted to do the same thing but use options specified -for the @code{[perl]} group as well, you could use this: +@findex @code{mysql_fetch_lengths()} +@node mysql_fetch_lengths, mysql_fetch_row, mysql_fetch_field_direct, C API functions +@subsubsection @code{mysql_fetch_lengths()} -@example -$dbh = DBI->connect("DBI:mysql:$database" - . ";mysql_read_default_file=$ENV@{HOME@}/.my.cnf" - . ";mysql_read_default_group=perl", - $user, $password); -@end example +@code{unsigned long *mysql_fetch_lengths(MYSQL_RES *result)} -@findex DBI->disconnect -@findex disconnect DBI method -@item disconnect -The @code{disconnect} method disconnects the database handle from the database. -This is typically called right before you exit from the program. -Example: -@example -$rc = $dbh->disconnect; -@end example +@subsubheading Description -@findex DBI->prepare() -@findex prepare() DBI method -@item prepare($statement) -Prepares a SQL statement for execution by the database engine -and returns a statement handle @code{($sth)}, which you can use to invoke -the @code{execute} method. -Typically you handle @code{SELECT} statements (and @code{SELECT}-like statements -such as @code{SHOW}, @code{DESCRIBE}, and @code{EXPLAIN}) by means of -@code{prepare} and @code{execute}. -Example: -@example -$sth = $dbh->prepare($statement) - or die "Can't prepare $statement: $dbh->errstr\n"; -@end example +Returns the lengths of the columns of the current row within a result set. +If you plan to copy field values, this length information is also useful for +optimization, because you can avoid calling @code{strlen()}. In addition, if +the result set contains binary data, you @emph{must} use this function to +determine the size of the data, because @code{strlen()} returns incorrect +results for any field containing null characters. -@findex DBI->execute -@findex execute DBI method -@item execute -The @code{execute} method executes a prepared statement. For -non-@code{SELECT} statements, @code{execute} returns the number of rows -affected. If no rows are affected, @code{execute} returns @code{"0E0"}, -which Perl treats as zero but regards as true. If an error occurs, -@code{execute} returns @code{undef}. For @code{SELECT} statements, -@code{execute} only starts the SQL query in the database; you need to use one -of the @code{fetch_*} methods described below to retrieve the data. -Example: -@example -$rv = $sth->execute - or die "can't execute the query: $sth->errstr; -@end example +The length for empty columns and for columns containing @code{NULL} values is +zero. To see how to distinguish these two cases, see the description for +@code{mysql_fetch_row()}. -@findex DBI->do() -@findex do() DBI method -@item do($statement) -The @code{do} method prepares and executes a SQL statement and returns the -number of rows affected. If no rows are affected, @code{do} returns -@code{"0E0"}, which Perl treats as zero but regards as true. This method is -generally used for non-@code{SELECT} statements that cannot be prepared in -advance (due to driver limitations) or that do not need to be executed more -than once (inserts, deletes, etc.). Example: -@example -$rv = $dbh->do($statement) - or die "Can't execute $statement: $dbh- >errstr\n"; -@end example +@subsubheading Return Values -Generally the 'do' statement is MUCH faster (and is preferable) -than prepare/execute for statements that don't contain parameters. +An array of unsigned long integers representing the size of each column (not +including any terminating null characters). +@code{NULL} if an error occurred. -@findex DBI->quote() -@findex quote() DBI method -@cindex quoting strings -@cindex strings, quoting -@item quote($string) -The @code{quote} method is used to "escape" any special characters contained in -the string and to add the required outer quotation marks. -Example: -@example -$sql = $dbh->quote($string) -@end example +@subsubheading Errors +@code{mysql_fetch_lengths()} is valid only for the current row of the result +set. It returns @code{NULL} if you call it before calling +@code{mysql_fetch_row()} or after retrieving all rows in the result. -@findex DBI->fetchrow_array -@findex fetchrow_array DBI method -@item fetchrow_array -This method fetches the next row of data and returns it as an array of -field values. Example: -@example -while(@@row = $sth->fetchrow_array) @{ - print qw($row[0]\t$row[1]\t$row[2]\n); -@} -@end example +@subsubheading Example -@findex DBI->fetchrow_arrayref -@findex fetchrow_arrayref DBI method -@item fetchrow_arrayref -This method fetches the next row of data and returns it as a reference -to an array of field values. Example: @example -while($row_ref = $sth->fetchrow_arrayref) @{ - print qw($row_ref->[0]\t$row_ref->[1]\t$row_ref->[2]\n); -@} -@end example +MYSQL_ROW row; +unsigned long *lengths; +unsigned int num_fields; +unsigned int i; -@findex DBI->fetchrow_hashref -@findex fetchrow_hashref DBI method -@item fetchrow_hashref -This method fetches a row of data and returns a reference to a hash -table containing field name/value pairs. This method is not nearly as -efficient as using array references as demonstrated above. Example: -@example -while($hash_ref = $sth->fetchrow_hashref) @{ - print qw($hash_ref->@{firstname@}\t$hash_ref->@{lastname@}\t\ - $hash_ref- > title@}\n); +row = mysql_fetch_row(result); +if (row) +@{ + num_fields = mysql_num_fields(result); + lengths = mysql_fetch_lengths(result); + for(i = 0; i < num_fields; i++) + @{ + printf("Column %u is %lu bytes in length.\n", i, lengths[i]); + @} @} @end example -@findex DBI->fetchall_arrayref -@findex fetchall_arrayref DBI method -@item fetchall_arrayref -This method is used to get all the data (rows) to be returned from the -SQL statement. It returns a reference to an array of references to arrays -for each row. You access or print the data by using a nested -loop. Example: -@example -my $table = $sth->fetchall_arrayref - or die "$sth->errstr\n"; -my($i, $j); -for $i ( 0 .. $#@{$table@} ) @{ - for $j ( 0 .. $#@{$table->[$i]@} ) @{ - print "$table->[$i][$j]\t"; - @} - print "\n"; -@} -@end example +@findex @code{mysql_fetch_row()} +@node mysql_fetch_row, mysql_field_count, mysql_fetch_lengths, C API functions +@subsubsection @code{mysql_fetch_row()} -@findex DBI->finish -@findex finish DBI method -@item finish -Indicates that no more data will be fetched from this statement -handle. You call this method to free up the statement handle and any -system resources associated with it. Example: -@example -$rc = $sth->finish; -@end example +@code{MYSQL_ROW mysql_fetch_row(MYSQL_RES *result)} -@findex DBI->rows -@findex rows DBI method -@item rows -Returns the number of rows changed (updated, deleted, etc.) by the last -command. This is usually used after a non-@code{SELECT} @code{execute} -statement. Example: -@example -$rv = $sth->rows; -@end example +@subsubheading Description -@findex DBI->@{NULLABLE@} -@findex NULLABLE DBI method -@item NULLABLE -Returns a reference to an array of boolean values; for each element of -the array, a value of TRUE indicates that this -column may contain @code{NULL} values. -Example: -@example -$null_possible = $sth->@{NULLABLE@}; -@end example +Retrieves the next row of a result set. When used after +@code{mysql_store_result()}, @code{mysql_fetch_row()} returns @code{NULL} +when there are no more rows to retrieve. When used after +@code{mysql_use_result()}, @code{mysql_fetch_row()} returns @code{NULL} when +there are no more rows to retrieve or if an error occurred. -@findex DBI->@{NUM_OF_FIELDS@} -@findex NUM_OF_FIELDS DBI method -@item NUM_OF_FIELDS -This attribute indicates -the number of fields returned by a @code{SELECT} or @code{SHOW FIELDS} -statement. You may use this for checking whether a statement returned a -result: A zero value indicates a non-@code{SELECT} statement like -@code{INSERT}, @code{DELETE}, or @code{UPDATE}. -Example: -@example -$nr_of_fields = $sth->@{NUM_OF_FIELDS@}; -@end example +The number of values in the row is given by @code{mysql_num_fields(result)}. +If @code{row} holds the return value from a call to @code{mysql_fetch_row()}, +pointers to the values are accessed as @code{row[0]} to +@code{row[mysql_num_fields(result)-1]}. @code{NULL} values in the row are +indicated by @code{NULL} pointers. -@findex DBI->data_sources() -@findex data_sources() DBI method -@item data_sources($driver_name) -This method returns an array containing names of databases available to the -@strong{MySQL} server on the host @code{'localhost'}. -Example: -@example -@@dbs = DBI->data_sources("mysql"); -@end example +The lengths of the field values in the row may be obtained by calling +@code{mysql_fetch_lengths()}. Empty fields and fields containing +@code{NULL} both have length 0; you can distinguish these by checking +the pointer for the field value. If the pointer is @code{NULL}, the field +is @code{NULL}; otherwise the field is empty. -@findex DBI->@{ChopBlanks@} -@findex ChopBlanks DBI method -@item ChopBlanks -This attribute determines whether the @code{fetchrow_*} methods will chop -leading and trailing blanks from the returned values. -Example: -@example -$sth->@{'ChopBlanks'@} =1; -@end example +@subsubheading Return Values -@findex DBI->trace -@findex trace DBI method -@item trace($trace_level) -@itemx trace($trace_level, $trace_filename) -The @code{trace} method enables or disables tracing. When invoked as a -@code{DBI} class method, it affects tracing for all handles. When invoked as -a database or statement handle method, it affects tracing for the given -handle (and any future children of the handle). Setting @code{$trace_level} -to 2 provides detailed trace information. Setting @code{$trace_level} to 0 -disables tracing. Trace output goes to the standard error output by -default. If @code{$trace_filename} is specified, the file is opened in -append mode and output for @emph{all} traced handles is written to that -file. Example: -@example -DBI->trace(2); # trace everything -DBI->trace(2,"/tmp/dbi.out"); # trace everything to - # /tmp/dbi.out -$dth->trace(2); # trace this database handle -$sth->trace(2); # trace this statement handle -@end example +A @code{MYSQL_ROW} structure for the next row. @code{NULL} if +there are no more rows to retrieve or if an error occurred. -@tindex DBI_TRACE environment variable -@tindex Environment variable, DBI_TRACE -You can also enable @code{DBI} tracing by setting the @code{DBI_TRACE} -environment variable. Setting it to a numeric value is equivalent to calling -@code{DBI->(value)}. Setting it to a pathname is equivalent to calling -@code{DBI->(2,value)}. +@subsubheading Errors +@table @code +@item CR_SERVER_LOST +The connection to the server was lost during the query. +@item CR_UNKNOWN_ERROR +An unknown error occurred. @end table -@noindent -@strong{MySQL-specific Methods} - -The methods shown below are @strong{MySQL}-specific and not part of the -@code{DBI} standard. Several of them are now deprecated: -@code{is_blob}, @code{is_key}, @code{is_num}, @code{is_pri_key}, -@code{is_not_null}, @code{length}, @code{max_length}, and @code{table}. -Where @code{DBI}-standard alternatives exist, they are noted below: +@subsubheading Example -@table @code -@findex DBI->@{insertid@} -@findex insertid DBI method -@tindex AUTO_INCREMENT, using with DBI -@item insertid -If you use the @code{AUTO_INCREMENT} feature of @strong{MySQL}, the new -auto-incremented values will be stored here. -Example: @example -$new_id = $sth->@{insertid@}; +MYSQL_ROW row; +unsigned int num_fields; +unsigned int i; + +num_fields = mysql_num_fields(result); +while ((row = mysql_fetch_row(result))) +@{ + unsigned long *lengths; + lengths = mysql_fetch_lengths(result); + for(i = 0; i < num_fields; i++) + @{ + printf("[%.*s] ", (int) lengths[i], row[i] ? row[i] : "NULL"); + @} + printf("\n"); +@} @end example -As an alternative, you can use @code{$dbh->@{'mysql_insertid'@}}. +@findex @code{mysql_field_count()} +@node mysql_field_count, mysql_field_seek, mysql_fetch_row, C API functions +@subsubsection @code{mysql_field_count()} -@findex DBI->@{is_blob@} -@findex is_blob DBI method -@item is_blob -Returns a reference to an array of boolean values; for each element of the -array, a value of TRUE indicates that the -respective column is a @code{BLOB}. -Example: -@example -$keys = $sth->@{is_blob@}; -@end example +@code{unsigned int mysql_field_count(MYSQL *mysql)} -@findex DBI->@{is_key@} -@findex is_key DBI method -@item is_key -Returns a reference to an array of boolean values; for each element of the -array, a value of TRUE indicates that the -respective column is a key. -Example: -@example -$keys = $sth->@{is_key@}; -@end example +If you are using a version of @strong{MySQL} earlier than Version 3.22.24, you +should use @code{unsigned int mysql_num_fields(MYSQL *mysql)} instead. -@findex DBI->@{is_num@} -@findex is_num DBI method -@item is_num -Returns a reference to an array of boolean values; for each element of the -array, a value of TRUE indicates that the -respective column contains numeric values. -Example: -@example -$nums = $sth->@{is_num@}; -@end example +@subsubheading Description -@findex DBI->@{is_pri_key@} -@findex is_pri_key DBI method -@item is_pri_key -Returns a reference to an array of boolean values; for each element of the -array, a value of TRUE indicates that the respective column is a primary key. -Example: -@example -$pri_keys = $sth->@{is_pri_key@}; -@end example +Returns the number of columns for the most recent query on the connection. -@findex DBI->@{is_not_null@} -@findex is_not_null DBI method -@item is_not_null -Returns a reference to an array of boolean values; for each element of the -array, a value of FALSE indicates that this column may contain @code{NULL} -values. -Example: -@example -$not_nulls = $sth->@{is_not_null@}; -@end example +The normal use of this function is when @code{mysql_store_result()} +returned @code{NULL} (and thus you have no result set pointer). +In this case, you can call @code{mysql_field_count()} to +determine whether or not @code{mysql_store_result()} should have produced a +non-empty result. This allows the client program to take proper action +without knowing whether or not the query was a @code{SELECT} (or +@code{SELECT}-like) statement. The example shown below illustrates how this +may be done. -@code{is_not_null} is deprecated; it is preferable to use the -@code{NULLABLE} attribute (described above), because that is a DBI standard. +@xref{NULL mysql_store_result, , @code{NULL mysql_store_result()}}. -@findex DBI->@{length@} -@findex length DBI method -@findex DBI->@{max_length@} -@findex max_length DBI method -@item length -@itemx max_length -Each of these methods returns a reference to an array of column sizes. The -@code{length} array indicates the maximum possible sizes that each column may -be (as declared in the table description). The @code{max_length} array -indicates the maximum sizes actually present in the result table. Example: +@subsubheading Return Values -@example -$lengths = $sth->@{length@}; -$max_lengths = $sth->@{max_length@}; -@end example +An unsigned integer representing the number of fields in a result set. -@findex DBI->@{NAME@} -@findex NAME DBI method -@item NAME -Returns a reference to an array of column names. -Example: -@example -$names = $sth->@{NAME@}; -@end example +@subsubheading Errors +None. -@findex DBI->@{table@} -@findex table DBI method -@item table -Returns a reference to an array of table names. -Example: -@example -$tables = $sth->@{table@}; -@end example +@subsubheading Example -@findex DBI->@{type@} -@findex type DBI method -@item type -Returns a reference to an array of column types. -Example: @example -$types = $sth->@{type@}; +MYSQL_RES *result; +unsigned int num_fields; +unsigned int num_rows; + +if (mysql_query(&mysql,query_string)) +@{ + // error +@} +else // query succeeded, process any data returned by it +@{ + result = mysql_store_result(&mysql); + if (result) // there are rows + @{ + num_fields = mysql_num_fields(result); + // retrieve rows, then call mysql_free_result(result) + @} + else // mysql_store_result() returned nothing; should it have? + @{ + if(mysql_field_count(&mysql) == 0) + @{ + // query does not return data + // (it was not a SELECT) + num_rows = mysql_affected_rows(&mysql); + @} + else // mysql_store_result() should have returned data + @{ + fprintf(stderr, "Error: %s\n", mysql_error(&mysql)); + @} + @} +@} @end example -@end table +An alternative is to replace the @code{mysql_field_count(&mysql)} call with +@code{mysql_errno(&mysql)}. In this case, you are checking directly for an +error from @code{mysql_store_result()} rather than inferring from the value +of @code{mysql_field_count()} whether or not the statement was a +@code{SELECT}. -@cindex @code{DBI/DBD} -@node DBI-info, , Perl DBI Class, Perl -@subsection More @code{DBI}/@code{DBD} Information +@findex @code{mysql_field_seek()} +@node mysql_field_seek, mysql_field_tell, mysql_field_count, C API functions +@subsubsection @code{mysql_field_seek()} -You can use the @code{perldoc} command to get more information about -@code{DBI}. +@code{MYSQL_FIELD_OFFSET mysql_field_seek(MYSQL_RES *result, MYSQL_FIELD_OFFSET offset)} -@example -perldoc DBI -perldoc DBI::FAQ -perldoc DBD::mysql -@end example +@subsubheading Description -You can also use the @code{pod2man}, @code{pod2html}, etc., tools to -translate to other formats. +Sets the field cursor to the given offset. The next call to +@code{mysql_fetch_field()} will retrieve the field definition of the column +associated with that offset. -You can find the latest @code{DBI} information at -the @code{DBI} Web page: -@example -@uref{http://www.symbolstone.org/technology/perl/DBI/index.html} -@end example +To seek to the beginning of a row, pass an @code{offset} value of zero. -@cindex Eiffel Wrapper -@cindex wrappers, Eiffel -@node Eiffel, Java, Perl, Clients -@section MySQL Eiffel wrapper +@subsubheading Return Values -The @strong{MySQL} @uref{http://www.mysql.com/Downloads/Contrib/,Contrib directory} -contains an Eiffel wrapper written by Michael Ravits. +The previous value of the field cursor. -You can also find this at: -@url{http://www.netpedia.net/hosting/newplayer/} +@subsubheading Errors +None. -@cindex Java connectivity -@cindex JDBC -@node Java, PHP, Eiffel, Clients -@section MySQL Java Connectivity (JDBC) +@findex @code{mysql_field_tell()} +@node mysql_field_tell, mysql_free_result, mysql_field_seek, C API functions +@subsubsection @code{mysql_field_tell()} -There are 2 supported JDBC drivers for @strong{MySQL} (the mm driver and -the Reisin JDBC driver). You can find a copy of the mm driver at -@uref{http://mmmysql.sourceforge.net/} or -@uref{http://www.mysql.com/Downloads/Contrib/} and the Reisin driver at -@uref{http://www.caucho.com/projects/jdbc-mysql/index.xtp} For -documentation consult any JDBC documentation and the driver's own -documentation for @strong{MySQL}-specific features. +@code{MYSQL_FIELD_OFFSET mysql_field_tell(MYSQL_RES *result)} -@cindex PHP API -@node PHP, Cplusplus, Java, Clients -@section MySQL PHP API +@subsubheading Description -PHP is a server-side, HTML-embedded scripting language that may be used to -create dynamic Web pages. It contains support for accessing several -databases, including @strong{MySQL}. PHP may be run as a separate program -or compiled as a module for use with the Apache Web server. +Returns the position of the field cursor used for the last +@code{mysql_fetch_field()}. This value can be used as an argument to +@code{mysql_field_seek()}. -The distribution and documentation are available at the -@uref{http://www.php.net/, PHP web site}. +@subsubheading Return Values -@menu -* PHP problems:: Common problems with MySQL and PHP -@end menu +The current offset of the field cursor. -@node PHP problems, , PHP, PHP -@subsection Common Problems with MySQL and PHP +@subsubheading Errors +None. -@itemize @bullet -@item Error: "Maximum Execution Time Exceeded" -This is a PHP limit; Go into the @file{php3.ini} file and set the maximum -execution time up from 30 seconds to something higher, as needed. -It is also not a bad idea to double the ram allowed per script to 16MB instead of -8 MB. -@item Error: "Fatal error: Call to unsupported or undefined function mysql_connect() in .." -This means that your PHP version isn't compiled with @strong{MySQL} support. -You can either compile a dynamic @strong{MySQL} module and load it into PHP or -recompile PHP with built-in @strong{MySQL} support. This is described in -detail in the PHP manual. -@item Error: "undefined reference to `uncompress'" -This means that the client library is compiled with support for a compressed -client/server protocol. The fix is to add @code{-lz} last when linking -with @code{-lmysqlclient}. -@end itemize +@findex @code{mysql_free_result()} +@node mysql_free_result, mysql_get_client_info, mysql_field_tell, C API functions +@subsubsection @code{mysql_free_result()} -@cindex C++ APIs -@node Cplusplus, Python, PHP, Clients -@section MySQL C++ APIs +@code{void mysql_free_result(MYSQL_RES *result)} -Two APIs are available in the @strong{MySQL} -@uref{http://www.mysql.com/Downloads/Contrib/,Contrib directory}. +@subsubheading Description -@cindex Python APIs -@node Python, Tcl, Cplusplus, Clients -@section MySQL Python APIs +Frees the memory allocated for a result set by @code{mysql_store_result()}, +@code{mysql_use_result()}, @code{mysql_list_dbs()}, etc. When you are done +with a result set, you must free the memory it uses by calling +@code{mysql_free_result()}. -The @strong{MySQL} @uref{http://www.mysql.com/Downloads/Contrib/,Contrib directory} -contains a Python interface written by Joseph Skinner. +@subsubheading Return Values -You can also use the Python interface to iODBC to access a -@strong{MySQL} server. -@uref{http://starship.skyport.net/~lemburg/,mxODBC} +None. -@cindex Tcl APIs -@node Tcl, , Python, Clients -@section MySQL Tcl APIs +@subsubheading Errors +None. -@uref{http://www.binevolve.com/~tdarugar/tcl-sql/, Tcl at binevolve}. -The -@uref{http://www.mysql.com/Downloads/Contrib,Contrib directory} contains a Tcl -interface that is based on msqltcl 1.50. +@findex @code{mysql_get_client_info()} +@node mysql_get_client_info, mysql_get_host_info, mysql_free_result, C API functions +@subsubsection @code{mysql_get_client_info()} -@cindex databases, MySQL vs. others -@cindex comparisons, MySQL vs. others -@node Comparisons, MySQL internals, Clients, Top -@chapter How MySQL Compares to Other Databases +@code{char *mysql_get_client_info(void)} -@menu -* Compare mSQL:: How @strong{MySQL} compares to @code{mSQL} -* Compare PostgreSQL:: How @strong{MySQL} compares with PostgreSQL -@end menu +@subsubheading Description -This chapter compares @strong{MySQL} to other popular databases. +Returns a string that represents the client library version. -This chapter has been written by the @strong{MySQL} developers, so it -should be read with that in mind. There are no factual errors contained -in this chapter that we know of. If you find something which you believe -to be an error, please contact us about it at @email{docs@@mysql.com}. +@subsubheading Return Values -For a list of all supported limits, functions, and types, see the -@code{crash-me} Web page at -@uref{http://www.mysql.com/information/crash-me.php}. +A character string that represents the @strong{MySQL} client library version. -@node Compare mSQL, Compare PostgreSQL, Comparisons, Comparisons -@section How MySQL Compares to @code{mSQL} +@subsubheading Errors +None. -@table @strong -@item Performance +@findex @code{mysql_get_host_info()} +@node mysql_get_host_info, mysql_get_proto_info, mysql_get_client_info, C API functions +@subsubsection @code{mysql_get_host_info()} -For a true comparison of speed, consult the growing @strong{MySQL} benchmark -suite. @xref{Benchmarks}. +@code{char *mysql_get_host_info(MYSQL *mysql)} -Because there is no thread creation overhead, a small parser, few features, and -simple security, @code{mSQL} should be quicker at: +@subsubheading Description -@itemize @bullet -@item -Tests that perform repeated connects and disconnects, running a very simple -query during each connection. -@item -@code{INSERT} operations into very simple tables with few columns and keys. -@item -@code{CREATE TABLE} and @code{DROP TABLE}. -@item -@code{SELECT} on something that isn't an index. (A table scan is very -easy.) -@end itemize +Returns a string describing the type of connection in use, including the +server host name. -Because these operations are so simple, it is hard to be better at them when -you have a higher startup overhead. After the connection is established, -@strong{MySQL} should perform much better. +@subsubheading Return Values -On the other hand, @strong{MySQL} is much faster than @code{mSQL} (and -most other SQL implementations) on the following: +A character string representing the server host name and the connection type. -@itemize @bullet -@item -Complex @code{SELECT} operations. -@item -Retrieving large results (@strong{MySQL} has a better, faster, and safer -protocol). -@item -Tables with variable-length strings, because @strong{MySQL} has more efficient -handling and can have indexes on @code{VARCHAR} columns. -@item -Handling tables with many columns. -@item -Handling tables with large record lengths. -@item -@code{SELECT} with many expressions. -@item -@code{SELECT} on large tables. -@item -Handling many connections at the same time. @strong{MySQL} is fully -multi-threaded. Each connection has its own thread, which means that -no thread has to wait for another (unless a thread is modifying -a table another thread wants to access). In @code{mSQL}, once one connection -is established, all others must wait until the first has finished, regardless -of whether the connection is running a query that is short or long. When the -first connection terminates, the next can be served, while all the others wait -again, etc. -@item -Joins. -@code{mSQL} can become pathologically slow if you change the order of tables -in a @code{SELECT}. In the benchmark suite, a time more than 15000 times -slower than @strong{MySQL} was seen. This is due to @code{mSQL}'s lack of a -join optimizer to order tables in the optimal order. However, if you put the -tables in exactly the right order in @code{mSQL}2 and the @code{WHERE} is -simple and uses index columns, the join will be relatively fast! -@xref{Benchmarks}. -@item -@code{ORDER BY} and @code{GROUP BY}. -@item -@code{DISTINCT}. -@item -Using @code{TEXT} or @code{BLOB} columns. -@end itemize +@subsubheading Errors +None. -@item SQL Features +@findex @code{mysql_get_proto_info()} +@node mysql_get_proto_info, mysql_get_server_info, mysql_get_host_info, C API functions +@subsubsection @code{mysql_get_proto_info()} -@itemize @bullet -@item @code{GROUP BY} and @code{HAVING}. -@code{mSQL} does not support @code{GROUP BY} at all. -@strong{MySQL} supports a full @code{GROUP BY} with both @code{HAVING} and -the following functions: @code{COUNT()}, @code{AVG()}, @code{MIN()}, -@code{MAX()}, @code{SUM()}, and @code{STD()}. @code{COUNT(*)} is optimized to -return very quickly if the @code{SELECT} retrieves from one table, no other -columns are retrieved, and there is no @code{WHERE} clause. @code{MIN()} and -@code{MAX()} may take string arguments. +@code{unsigned int mysql_get_proto_info(MYSQL *mysql)} -@item @code{INSERT} and @code{UPDATE} with calculations. -@strong{MySQL} can do calculations in an @code{INSERT} or @code{UPDATE}. -For example: -@example -mysql> UPDATE SET x=x*10+y WHERE x<20; -@end example +@subsubheading Description -@item Aliasing. -@strong{MySQL} has column aliasing. +Returns the protocol version used by current connection. -@item Qualifying column names. -In @strong{MySQL}, if a column name is unique among the tables used in a -query, you do not have to use the full qualifier. +@subsubheading Return Values -@item @code{SELECT} with functions. -@strong{MySQL} has many functions (too many to list here; see @ref{Functions}). +An unsigned integer representing the protocol version used by the current +connection. -@end itemize +@subsubheading Errors +None. -@item Disk Space Efficiency -That is, how small can you make your tables? +@findex @code{mysql_get_server_info()} +@node mysql_get_server_info, mysql_info, mysql_get_proto_info, C API functions +@subsubsection @code{mysql_get_server_info()} -@strong{MySQL} has very precise types, so you can create tables that take -very little space. An example of a useful @strong{MySQL} datatype is the -@code{MEDIUMINT} that is 3 bytes long. If you have 100,000,000 records, -saving even one byte per record is very important. +@code{char *mysql_get_server_info(MYSQL *mysql)} -@code{mSQL2} has a more limited set of column types, so it is -more difficult to get small tables. -@item Stability -This is harder to judge objectively. For a discussion of @strong{MySQL} -stability, see @ref{Stability}. +@subsubheading Description -We have no experience with @code{mSQL} stability, so we cannot say -anything about that. +Returns a string that represents the server version number. -@item Price -Another important issue is the license. @strong{MySQL} has a -more flexible license than @code{mSQL}, and is also less expensive than -@code{mSQL}. Whichever product you choose to use, remember to at least -consider paying for a license or e-mail support. (You are required to get -a license if you include @strong{MySQL} with a product that you sell, -of course.) +@subsubheading Return Values -@item Perl Interfaces -@strong{MySQL} has basically the same interfaces to Perl as @code{mSQL} with -some added features. +A character string that represents the server version number. -@item JDBC (Java) -@strong{MySQL} currently has a lot of different JDBC drivers: -@itemize @bullet -@item -The mm driver: A type 4 JDBC driver by Mark Matthews -@email{mmatthew@@ecn.purdue.edu}. This is released under the LGPL. -@item -The Resin driver. This is a commercial JDBC driver released under open -source. @uref{http://www.caucho.com/projects/jdbc-mysql/index.xtp} -@item -The gwe driver: A Java interface by GWE technologies (not supported anymore). -@item -The jms driver: An improved gwe driver by Xiaokun Kelvin ZHU -@email{X.Zhu@@brad.ac.uk} (not supported anymore). -@item -The twz driver: A type 4 JDBC driver by Terrence W. Zellers -@email{zellert@@voicenet.com}. This is commercial but is free for private -and educational use (not supported anymore). -@end itemize +@subsubheading Errors +None. -The recommended driver is the mm driver. The Resin driver may also be -good (at least the benchmarks looks good), but we haven't received that much -information about this yet. +@findex @code{mysql_info()} +@node mysql_info, mysql_init, mysql_get_server_info, C API functions +@subsubsection @code{mysql_info()} -We know that @code{mSQL} has a JDBC driver, but we have too little experience -with it to compare. +@code{char *mysql_info(MYSQL *mysql)} -@item Rate of Development -@strong{MySQL} has a very small team of developers, but we are quite -used to coding C and C++ very rapidly. Because threads, functions, -@code{GROUP BY}, and so on are still not implemented in @code{mSQL}, it -has a lot of catching up to do. To get some perspective on this, you -can view the @code{mSQL} @file{HISTORY} file for the last year and -compare it with the News section of the @strong{MySQL} Reference Manual -(@pxref{News}). It should be pretty obvious which one has developed -most rapidly. +@subsubheading Description -@item Utility Programs -Both @code{mSQL} and @strong{MySQL} have many interesting third-party -tools. Because it is very easy to port upward (from @code{mSQL} to -@strong{MySQL}), almost all the interesting applications that are available for -@code{mSQL} are also available for @strong{MySQL}. +Retrieves a string providing information about the most recently executed +query, but only for the statements listed below. For other statements, +@code{mysql_info()} returns @code{NULL}. The format of the string varies +depending on the type of query, as described below. The numbers are +illustrative only; the string will contain values appropriate for the query. -@strong{MySQL} comes with a simple @code{msql2mysql} program that fixes -differences in spelling between @code{mSQL} and @strong{MySQL} for the -most-used C API functions. -For example, it changes instances of @code{msqlConnect()} to -@code{mysql_connect()}. Converting a client program from @code{mSQL} to -@strong{MySQL} usually takes a couple of minutes. +@table @code +@item INSERT INTO ... SELECT ... +String format: @code{Records: 100 Duplicates: 0 Warnings: 0} +@item INSERT INTO ... VALUES (...),(...),(...)... +String format: @code{Records: 3 Duplicates: 0 Warnings: 0} +@item LOAD DATA INFILE ... +String format: @code{Records: 1 Deleted: 0 Skipped: 0 Warnings: 0} +@item ALTER TABLE +String format: @code{Records: 3 Duplicates: 0 Warnings: 0} +@item UPDATE +String format: @code{Rows matched: 40 Changed: 40 Warnings: 0} @end table -@menu -* Using mSQL tools:: How to convert @code{mSQL} tools for @strong{MySQL} -* Protocol differences:: How @code{mSQL} and @strong{MySQL} client/server communications protocols differ -* Syntax differences:: How @code{mSQL} 2.0 SQL syntax differs from @strong{MySQL} -@end menu +Note that @code{mysql_info()} returns a non-@code{NULL} value for the +@code{INSERT ... VALUES} statement only if multiple value lists are +specified in the statement. -@cindex MySQL tools, conversion -@cindex converting, tools -@cindex tools, converting -@node Using mSQL tools, Protocol differences, Compare mSQL, Compare mSQL -@subsection How to Convert @code{mSQL} Tools for MySQL +@subsubheading Return Values -According to our experience, it would just take a few hours to convert tools -such as @code{msql-tcl} and @code{msqljava} that use the -@code{mSQL} C API so that they work with the @strong{MySQL} C API. +A character string representing additional information about the most +recently executed query. @code{NULL} if no information is available for the +query. -The conversion procedure is: +@subsubheading Errors +None. -@enumerate -@item -Run the shell script @code{msql2mysql} on the source. This requires the -@code{replace} program, which is distributed with @strong{MySQL}. -@item -Compile. -@item -Fix all compiler errors. -@end enumerate +@findex @code{mysql_init()} +@node mysql_init, mysql_insert_id, mysql_info, C API functions +@subsubsection @code{mysql_init()} -Differences between the @code{mSQL} C API and the @strong{MySQL} C API are: -@itemize @bullet -@item -@strong{MySQL} uses a @code{MYSQL} structure as a connection type (@code{mSQL} -uses an @code{int}). -@item -@code{mysql_connect()} takes a pointer to a @code{MYSQL} structure as a -parameter. It is easy to define one globally or to use @code{malloc()} to get -one. -@code{mysql_connect()} also takes two parameters for specifying the user and -password. You may set these to @code{NULL, NULL} for default use. -@item -@code{mysql_error()} takes the @code{MYSQL} structure as a parameter. Just add -the parameter to your old @code{msql_error()} code if you are porting old code. -@item -@strong{MySQL} returns an error number and a text error message for all -errors. @code{mSQL} returns only a text error message. -@item -Some incompatibilities exist as a result of @strong{MySQL} supporting -multiple connections to the server from the same process. -@end itemize +@code{MYSQL *mysql_init(MYSQL *mysql)} -@cindex communications protocols -@cindex mSQL vs. MySQL -@node Protocol differences, Syntax differences, Using mSQL tools, Compare mSQL -@subsection How @code{mSQL} and MySQL Client/Server Communications Protocols Differ +@subsubheading Description -There are enough differences that it is impossible (or at least not easy) -to support both. +Allocates or initializes a @code{MYSQL} object suitable for +@code{mysql_real_connect()}. If @code{mysql} is a @code{NULL} pointer, the +function allocates, initializes, and returns a new object. Otherwise the +object is initialized and the address of the object is returned. If +@code{mysql_init()} allocates a new object, it will be freed when +@code{mysql_close()} is called to close the connection. -The most significant ways in which the @strong{MySQL} protocol differs -from the @code{mSQL} protocol are listed below: +@subsubheading Return Values -@itemize @bullet -@item -A message buffer may contain many result rows. -@item -The message buffers are dynamically enlarged if the query or the -result is bigger than the current buffer, up to a configurable server and -client limit. -@item -All packets are numbered to catch duplicated or missing packets. -@item -All column values are sent in ASCII. The lengths of columns and rows are sent -in packed binary coding (1, 2, or 3 bytes). -@item -@strong{MySQL} can read in the result unbuffered (without having to store the -full set in the client). -@item -If a single read/write takes more than 30 seconds, the server closes -the connection. -@item -If a connection is idle for 8 hours, the server closes the connection. -@end itemize +An initialized @code{MYSQL*} handle. @code{NULL} if there was +insufficient memory to allocate a new object. -@node Syntax differences, , Protocol differences, Compare mSQL -@subsection How @code{mSQL} 2.0 SQL Syntax Differs from MySQL +@subsubheading Errors +In case of insufficient memory, @code{NULL} is returned. -@noindent -@strong{Column types} +@findex @code{mysql_insert_id()} +@node mysql_insert_id, mysql_kill, mysql_init, C API functions +@subsubsection @code{mysql_insert_id()} -@table @code -@item @strong{MySQL} -Has the following additional types (among others; -@pxref{CREATE TABLE, , @code{CREATE TABLE}}): -@itemize @bullet -@item -@code{ENUM} type for one of a set of strings. -@item -@code{SET} type for many of a set of strings. -@item -@code{BIGINT} type for 64-bit integers. -@end itemize -@item -@strong{MySQL} also supports -the following additional type attributes: -@itemize @bullet -@item -@code{UNSIGNED} option for integer columns. -@item -@code{ZEROFILL} option for integer columns. -@item -@code{AUTO_INCREMENT} option for integer columns that are a -@code{PRIMARY KEY}. -@xref{mysql_insert_id, , @code{mysql_insert_id()}}. -@item -@code{DEFAULT} value for all columns. -@end itemize -@item mSQL2 -@code{mSQL} column types correspond to the @strong{MySQL} types shown below: -@multitable @columnfractions .15 .85 -@item @code{mSQL} @strong{type} @tab @strong{Corresponding @strong{MySQL} type} -@item @code{CHAR(len)} @tab @code{CHAR(len)} -@item @code{TEXT(len)} @tab @code{TEXT(len)}. @code{len} is the maximal length. -And @code{LIKE} works. -@item @code{INT} @tab @code{INT}. With many more options! -@item @code{REAL} @tab @code{REAL}. Or @code{FLOAT}. Both 4- and 8-byte versions are available. -@item @code{UINT} @tab @code{INT UNSIGNED} -@item @code{DATE} @tab @code{DATE}. Uses ANSI SQL format rather than @code{mSQL}'s own format. -@item @code{TIME} @tab @code{TIME} -@item @code{MONEY} @tab @code{DECIMAL(12,2)}. A fixed-point value with two decimals. -@end multitable -@end table +@code{my_ulonglong mysql_insert_id(MYSQL *mysql)} -@noindent -@strong{Index Creation} +@subsubheading Description -@table @code -@item @strong{MySQL} -Indexes may be specified at table creation time with the @code{CREATE TABLE} -statement. -@item mSQL -Indexes must be created after the table has been created, with separate -@code{CREATE INDEX} statements. -@end table +Returns the ID generated for an @code{AUTO_INCREMENT} column by the previous +query. Use this function after you have performed an @code{INSERT} query +into a table that contains an @code{AUTO_INCREMENT} field. -@noindent -@strong{To Insert a Unique Identifier into a Table} +Note that @code{mysql_insert_id()} returns @code{0} if the previous query +does not generate an @code{AUTO_INCREMENT} value. If you need to save +the value for later, be sure to call @code{mysql_insert_id()} immediately +after the query that generates the value. -@table @code -@item @strong{MySQL} -Use @code{AUTO_INCREMENT} as a column type -specifier. -@xref{mysql_insert_id, , @code{mysql_insert_id()}}. -@item mSQL -Create a @code{SEQUENCE} on a table and select the @code{_seq} column. -@end table +Also note that the value of the SQL @code{LAST_INSERT_ID()} function always +contains the most recently generated @code{AUTO_INCREMENT} value, and is +not reset between queries because the value of that function is maintained +in the server. -@noindent -@strong{To Obtain a Unique Identifier for a Row} +@subsubheading Return Values -@table @code -@item @strong{MySQL} -Add a @code{PRIMARY KEY} or @code{UNIQUE} key to the table and use this. -New in Version 3.23.11: If the @code{PRIMARY} or @code{UNIQUE} key consists of only one -column and this is of type integer, one can also refer to it as -@code{_rowid}. -@item mSQL -Use the @code{_rowid} column. Observe that @code{_rowid} may change over time -depending on many factors. -@end table +The value of the @code{AUTO_INCREMENT} field that was updated by the previous +query. Returns zero if there was no previous query on the connection or if +the query did not update an @code{AUTO_INCREMENT} value. -@noindent -@strong{To Get the Time a Column Was Last Modified} +@subsubheading Errors +None. -@table @code -@item @strong{MySQL} -Add a @code{TIMESTAMP} column to the table. This column is automatically set -to the current date and time for @code{INSERT} or @code{UPDATE} statements if -you don't give the column a value or if you give it a @code{NULL} value. +@findex @code{mysql_kill()} +@node mysql_kill, mysql_list_dbs, mysql_insert_id, C API functions +@subsubsection @code{mysql_kill()} -@item mSQL -Use the @code{_timestamp} column. -@end table +@code{int mysql_kill(MYSQL *mysql, unsigned long pid)} -@noindent -@strong{@code{NULL} Value Comparisons} +@subsubheading Description -@table @code -@item @strong{MySQL} -@strong{MySQL} follows -ANSI SQL, and a comparison with @code{NULL} is always @code{NULL}. -@item mSQL -In @code{mSQL}, @code{NULL = NULL} is TRUE. You -must change @code{=NULL} to @code{IS NULL} and @code{<>NULL} to -@code{IS NOT NULL} when porting old code from @code{mSQL} to @strong{MySQL}. -@end table +Asks the server to kill the thread specified by @code{pid}. -@noindent -@strong{String Comparisons} +@subsubheading Return Values -@table @code -@item @strong{MySQL} -Normally, string comparisons are performed in case-independent fashion with -the sort order determined by the current character set (ISO-8859-1 Latin1 by -default). If you don't like this, declare your columns with the -@code{BINARY} attribute, which causes comparisons to be done according to the -ASCII order used on the @strong{MySQL} server host. -@item mSQL -All string comparisons are performed in case-sensitive fashion with -sorting in ASCII order. -@end table +Zero for success. Non-zero if an error occurred. -@noindent -@strong{Case-insensitive Searching} +@subsubheading Errors @table @code -@item @strong{MySQL} -@code{LIKE} is a case-insensitive or case-sensitive operator, depending on -the columns involved. If possible, @strong{MySQL} uses indexes if the -@code{LIKE} argument doesn't start with a wild-card character. -@item mSQL -Use @code{CLIKE}. +@item CR_COMMANDS_OUT_OF_SYNC +Commands were executed in an improper order. +@item CR_SERVER_GONE_ERROR +The @strong{MySQL} server has gone away. +@item CR_SERVER_LOST +The connection to the server was lost during the query. +@item CR_UNKNOWN_ERROR +An unknown error occurred. @end table -@noindent -@strong{Handling of Trailing Spaces} +@findex @code{mysql_list_dbs()} +@node mysql_list_dbs, mysql_list_fields, mysql_kill, C API functions +@subsubsection @code{mysql_list_dbs()} -@table @code -@item @strong{MySQL} -Strips all spaces at the end of @code{CHAR} and @code{VARCHAR} -columns. Use a @code{TEXT} column if this behavior is not desired. -@item mSQL -Retains trailing space. -@end table +@code{MYSQL_RES *mysql_list_dbs(MYSQL *mysql, const char *wild)} -@noindent -@strong{@code{WHERE} Clauses} +@subsubheading Description -@table @code -@item @strong{MySQL} -@strong{MySQL} correctly prioritizes everything (@code{AND} is evaluated -before @code{OR}). To get @code{mSQL} behavior in @strong{MySQL}, use -parentheses (as shown in an example below). -@item mSQL -Evaluates everything from left to right. This means that some logical -calculations with more than three arguments cannot be expressed in any -way. It also means you must change some queries when you upgrade to -@strong{MySQL}. You do this easily by adding parentheses. Suppose you -have the following @code{mSQL} query: -@example -mysql> SELECT * FROM table WHERE a=1 AND b=2 OR a=3 AND b=4; -@end example -To make @strong{MySQL} evaluate this the way that @code{mSQL} would, -you must add parentheses: -@example -mysql> SELECT * FROM table WHERE (a=1 AND (b=2 OR (a=3 AND (b=4)))); -@end example -@end table +Returns a result set consisting of database names on the server that match +the simple regular expression specified by the @code{wild} parameter. +@code{wild} may contain the wild-card characters @samp{%} or @samp{_}, or may +be a @code{NULL} pointer to match all databases. Calling +@code{mysql_list_dbs()} is similar to executing the query @code{SHOW +databases [LIKE wild]}. -@noindent -@strong{Access Control} +You must free the result set with @code{mysql_free_result()}. + +@subsubheading Return Values + +A @code{MYSQL_RES} result set for success. @code{NULL} if an error occurred. + +@subsubheading Errors @table @code -@item @strong{MySQL} -Has tables to store grant (permission) options per user, host, and -database. @xref{Privileges}. -@item mSQL -Has a file @file{mSQL.acl} in which you can grant read/write privileges for -users. -@item +@item CR_COMMANDS_OUT_OF_SYNC +Commands were executed in an improper order. +@item CR_OUT_OF_MEMORY +Out of memory. +@item CR_SERVER_GONE_ERROR +The @strong{MySQL} server has gone away. +@item CR_SERVER_LOST +The connection to the server was lost during the query. +@item CR_UNKNOWN_ERROR +An unknown error occurred. @end table -@cindex PostgreSQL/MySQL, overview -@node Compare PostgreSQL, , Compare mSQL, Comparisons -@section How MySQL Compares to PostgreSQL +@findex @code{mysql_list_fields()} +@node mysql_list_fields, mysql_list_processes, mysql_list_dbs, C API functions +@subsubsection @code{mysql_list_fields()} -When reading the following, please note that both products are -continually evolving. We at @strong{MySQL AB} and the PostgreSQL -developers are both working on making our respective database as good as -possible, so we are both a serious choice to any commercial database. +@code{MYSQL_RES *mysql_list_fields(MYSQL *mysql, const char *table, const char *wild)} -The following comparison is made by us at MySQL AB. We have tried to be -as accurate and fair as possible, but because we don't have a full -knowledge of all PostgreSQL features while we know MySQL througly, we -may have got some things wrong. We will however correct these when they -come to our attention. +@subsubheading Description -We would first like to note that @code{PostgreSQL} and @strong{MySQL} -are both widely used products, but with different design goals, even if -we are both striving to be ANSI SQL compatible. This means that for -some applications @strong{MySQL} is more suitable and for others -@code{PostgreSQL} is more suitable. When choosing which database to -use, you should first check if the database's feature set satisfies your -application. If you need speed, @strong{MySQL} is probably your best -choice. If you need some of the extra features that only @code{PostgreSQL} -can offer, you should use @code{PostgreSQL}. +Returns a result set consisting of field names in the given table that match +the simple regular expression specified by the @code{wild} parameter. +@code{wild} may contain the wild-card characters @samp{%} or @samp{_}, or may +be a @code{NULL} pointer to match all fields. Calling +@code{mysql_list_fields()} is similar to executing the query @code{SHOW +COLUMNS FROM tbl_name [LIKE wild]}. -@cindex PostgreSQL/MySQL, strategies -@menu -* MySQL-PostgreSQL goals:: MySQL and PostgreSQL development strategies -* MySQL-PostgreSQL features:: Featurevise Comparison of MySQL and PostgreSQL -* MySQL-PostgreSQL benchmarks:: Benchmarking MySQL and PostgreSQL -@end menu +Note that it's recommended that you use @code{SHOW COLUMNS FROM tbl_name} +instead of @code{mysql_list_fields()}. -@node MySQL-PostgreSQL goals, MySQL-PostgreSQL features, Compare PostgreSQL, Compare PostgreSQL -@subsection MySQL and PostgreSQL development strategies +You must free the result set with @code{mysql_free_result()}. -When adding things to MySQL we take pride to do an optimal, definite -solution. The code should be so good that we shouldn't have any need to -change it in the foreseeable future. We also do not like to sacrifice -speed for features but instead will do our utmost to find a solution -that will give maximal throughput. This means that development will take -a little longer, but the end result will be well worth this. This kind -of development is only possible because all server code are checked by -one of a few (currently two) persons before it's included in the -@strong{MySQL} server. +@subsubheading Return Values -We at MySQL AB believe in frequent releases to be able to push out new -features quickly to our users. Because of this we do a new small release -about every 3 weeks, which a major branch every year. All releases are -throughly tested with our testing tools on a lot of different platforms. +A @code{MYSQL_RES} result set for success. @code{NULL} if an error occurred. -PostgreSQL is based on a kernel with lots of contributors. In this setup -it makes sense to prioritize adding a lot of new features, instead of -implementing them optimally, because one can always optimize things -later if there arises a need for this. +@subsubheading Errors -Another big difference between @strong{MySQL} and PostgreSQL is that -nearly all of the code in the MySQL server are coded by developers that -are employed by MySQL AB and are still working on the server code. The -exceptions are the transaction engines and the regexp library. +@table @code +@item CR_COMMANDS_OUT_OF_SYNC +Commands were executed in an improper order. +@item CR_SERVER_GONE_ERROR +The @strong{MySQL} server has gone away. +@item CR_SERVER_LOST +The connection to the server was lost during the query. +@item CR_UNKNOWN_ERROR +An unknown error occurred. +@end table -This is in sharp contrast to the PostgreSQL code where the majority of -the code is coded by a big group of people with different backgrounds. -It was only recently that the PostgreSQL developers announced that they -current developer group had finally had time to take a look at all -the code in the current PostgreSQL release. +@findex @code{mysql_list_processes()} +@node mysql_list_processes, mysql_list_tables, mysql_list_fields, C API functions +@subsubsection @code{mysql_list_processes()} -Both of the above development methods has it's own merits and drawbacks. -We here at @strong{MySQL AB} think of course that our model is better -because our model gives better code consistence, more optimal and -reusable code and, in our opinion, fewer bugs. Because we are the -authors of the @strong{MySQL} server code we are better able to -coordinate new features and releases. +@code{MYSQL_RES *mysql_list_processes(MYSQL *mysql)} -@cindex PostgreSQL/MySQL, features -@node MySQL-PostgreSQL features, MySQL-PostgreSQL benchmarks, MySQL-PostgreSQL goals, Compare PostgreSQL -@subsection Featurevise Comparison of MySQL and PostgreSQL +@subsubheading Description -On the @uref{http://www.mysql.com/information/crash-me.php, crash-me} -page you can find a list of those database constructs and limits that -one can detect automatically with a program. Note however that a lot of -the numerical limits may be changed with startup options for respective -database. The above web page is however extremely useful when you want to -ensure that your applications works with many different databases or -when you want to convert your application from one datbase to another. +Returns a result set describing the current server threads. This is the same +kind of information as that reported by @code{mysqladmin processlist} or +a @code{SHOW PROCESSLIST} query. -@strong{MySQL} offers the following advantages over PostgreSQL: +You must free the result set with @code{mysql_free_result()}. -@itemize @bullet -@item -@code{MySQL} is generally much faster than PostgreSQL. -@xref{MySQL-PostgreSQL benchmarks}. -@item -Because @strong{MySQL} has a much larger user base than PostgreSQL the -code is more tested and has historically been more stable than -PostgreSQL. @strong{MySQL} is the much more used in production -environments than PostgreSQL, mostly thanks to that @strong{MySQL AB}, -former TCX DataKonsult AB, has provided top quality commercial support -for @strong{MySQL} from the day it was released, whereas until recently -PostgreSQL was unsupported. -@item -@strong{MySQL} works on more platforms than PostgreSQL. @xref{Which OS}. -@item -@strong{MySQL} works better on Windows; @strong{MySQL} is running as a -native windows application (a service on NT/Win2000/WinXP), while -PostgreSQL is run under the cygwin emulation. We have heard that -PostgreSQL is not yet that stable on windows but we haven't been able to -verify this ourselves. -@item -@strong{MySQL} has more API to other languages and is supported by more -programs than PostgreSQL. @xref{Contrib}. -@item -@strong{MySQL} works on 24/7 heavy duty systems. In most circumstances -you never have to run any cleanups on @code{MySQL}. PostgreSQL doesn't -yet support 24/7 systems because you have have to run @code{vacuum()} -once in a while to reclaim space from @code{UPDATE} and @code{DELETE} -commands and to perform statistics analyzes that are critical to get -good performance with PostgreSQL. Vacuum is also needed after adding -a lot of new rows to a table. On a busy system with lots of changes -vacuum must be run very frequently, in the worst cases even many times a -day. During the @code{vacuum()} run, which may take hours if the -database is big, the database is from a production standpoint -practically dead. The PostgreSQL team has fixing this on their TODO, -but we assume that this is not an easy thing to fix permanently. -@item -A working, tested replication feature used by sites like -@uref{http://finance.yahoo.com, Yahoo finance}, -@uref{http://www.mobile.de/,mobile.de} and -@uref{http://www.slashdot.org,Slashdot}. -@item -Included in the @strong{MySQL} distribution is included two different -testing suits (@file{mysql-test-run} and -@uref{http://www.mysql.com/information/crash-me.php,crash-me}) and a -benchmark suite. The test system is actively updated with code to test -each new feature and almost all repeatable bugs that comes to our -attention. We test @strong{MySQL} with these on a lot of platforms -before every release. These tests are more sofisticated than anything -have seen from PostgreSQL and ensures that the @strong{MySQL} code keeps -at a high standard. -@item -There are far moore books in print on @strong{MySQL} than on PostgreSQL. -O'Reilly, Sams, Que, and New Riders are all major publishers with books -about MySQL. All @strong{MySQL} features is also documented in the -@strong{MySQL} on-line manual because when a feature is implemented, the -@strong{MySQL} developers are required to document it before it's -included in the source. -@item -@strong{MySQL} has supports more of the standard ODBC functions than -@code{PostgreSQL}. -@item -@strong{MySQL} has a much more sophisticated @code{ALTER TABLE}. -@item -@strong{MySQL} has support for tables without transactions for -applications that need all speed they can get. The tables may be memory -based,@code{HEAP} tables or disk based @code{MyISAM}. @xref{Table types}. -@item -@strong{MySQL} has support for 3 different table handles that support -transactions (@code{BDB} and @code{InnoDB}). Because -every transaction engine performs differently under different -conditions, this gives the application writer more options to find an -optimal solution for his/her setup. @xref{Table types}. -@item -@code{MERGE} tables gives you a unique way to instantly make a view over -a set of identical tables and use these as one. This is perfectly for -systems where you have log files that you order for example by month. -@xref{MERGE}. -@item -The option to compress read-only tables, but still have direct access to -the rows in the table, gives you better performance by minimizing disk -reads. This is very useful when you are archiving -things.@xref{myisampack}. -@item -@strong{MySQL} has internal support for text search. @xref{Fulltext Search}. -@item -You can access many databases from the same connection (depending of course -on your privileges). -@item -@strong{MySQL} is coded from the start with multi-threading while -PostgreSQL uses processes. Because context switching and access to -common storage areas is much faster between threads, than are separate -processes, this gives @strong{MySQL} a big speed advantage in multi-user -applications and also makes it easier for @strong{MySQL} to take full -advantage of symmetric multiprocessor systems (SMP). -@item -@strong{MySQL} has a much more sophisticated privilege system than -PostgreSQL. While PostgreSQL only supports @code{INSERT}, -@code{SELECT}, @code{update/delete} grants per user on a database or a -table @strong{MySQL} allows you to define a full set of different -privileges on database, table and columns level. @strong{MySQL} also allows -you to specify the privilege on host+user combinations. @xref{GRANT}. -@item -@strong{MySQL} supports a compressed server/client protocol which -improves performance over slow links. -@item -@strong{MySQL} employs the table handler concept and is the only -relational database we know of built around this concept. This allows -different low level table types to be swapped into the SQL engine, each -table type optimized for a different performance characteristics. -@item -All @code{MySQL} table types (except @strong{InnoDB}) are implemented as -files (ie: one table per file), which makes it really easy to backup, -move, delete and even symlink databases and tables when the server is -down. -@item -Tools to repair and optimize @strong{MyISAM} tables (the most common -@strong{MySQL} table type). A repair tool is only needed when a -physical corruption of a data file happens, usually from a hardware -failure. It allows a majority of the data to be recovered. -@item -Upgrading @strong{MySQL} is painless. When you are upgrading @strong{MySQL}, -you don't need to dump/restore your data, as you have to do with most -PostgreSQL upgrades. -@end itemize +@subsubheading Return Values -Drawbacks with @strong{MySQL} compared to PostgreSQL: +A @code{MYSQL_RES} result set for success. @code{NULL} if an error occurred. -@itemize @bullet -@item -The transaction support in @strong{MySQL} is not yet as well tested as -PostgreSQL's system. -@item -Because @strong{MySQL} uses threads, which are still a moving target on -many OS, one must either use binaries from -@uref{http://www.mysql.com/downloads} or carefully follow our -instructions on -@uref{http://www.mysql.com/doc/I/n/Installing_source.html} to get an -optimal binary that works in all cases. -@item -Table locking, as used by the non-transactional @code{MyISAM} tables, is -in many cases faster than page locks, row locks or versioning. The -drawback however is that if one doesn't take into account how table -locks works, a single long-running query can block a table for updates -for a long time. This can usable be avoided when designing the -application. If not, one can always switch the trouble table to use one -of the transactional table types. @xref{Table locking}. -@item -With UDF (user defined functions) one can extend @strong{MySQL} with -both normal SQL functions and aggregates, but this is not as easy or as -flexible as in PostgreSQL. @xref{Adding functions}. -@item -Updates and deletes that goes over multiple tables is harder to do in -@strong{MySQL}. (Will be fixed in @strong{MySQL} 4.0 with multi-table -@code{DELETE} and multi-table @code{UPDATE} and in @strong{MySQL} 4.1 -with @code{SUB-SELECT}) -@end itemize +@subsubheading Errors -PostgreSQL offers currently the following advantages over @strong{MySQL}: +@table @code +@item CR_COMMANDS_OUT_OF_SYNC +Commands were executed in an improper order. +@item CR_SERVER_GONE_ERROR +The @strong{MySQL} server has gone away. +@item CR_SERVER_LOST +The connection to the server was lost during the query. +@item CR_UNKNOWN_ERROR +An unknown error occurred. +@end table -Note that because we know the @strong{MySQL} road map, we have included -in the following table the version when @strong{MySQL} should support -this feature. Unfortunately we couldn't do this for previous comparison, -because we don't know the PostgreSQL roadmap. +@findex @code{mysql_list_tables()} +@node mysql_list_tables, mysql_num_fields, mysql_list_processes, C API functions +@subsubsection @code{mysql_list_tables()} -@multitable @columnfractions .70 .30 -@item @strong{Feature} @tab @strong{MySQL version} -@item Subselects @tab 4.1 -@item Foreign keys @tab 4.0 and 4.1 -@item Views. @tab 4.2 -@item Stored procedures in multiple languages @tab 4.1 -@item Extensible type system. @tab Not planed -@item Unions @tab 4.0. -@item Full join. @tab 4.0 or 4.1. -@item Triggers. @tab 4.1 -@item Constrainst @tab 4.1 -@item Cursors @tab 4.1 or 4.2 -@item Extensible index types like R-trees @tab R-trees are planned to 4.2 -@item Inherited tables @tab Not planned -@end multitable +@code{MYSQL_RES *mysql_list_tables(MYSQL *mysql, const char *wild)} -Other reasons to use PostgreSQL: +@subsubheading Description -@itemize @bullet -@item -Standard usage is in PostgreSQL closer to ANSI SQL in some cases. -@item -One can get speed up PostgreSQL by coding things as stored procedures. -@item -Bigger team of developers that contributes to the server. -@end itemize +Returns a result set consisting of table names in the current database that +match the simple regular expression specified by the @code{wild} parameter. +@code{wild} may contain the wild-card characters @samp{%} or @samp{_}, or may +be a @code{NULL} pointer to match all tables. Calling +@code{mysql_list_tables()} is similar to executing the query @code{SHOW +tables [LIKE wild]}. -Drawbacks with PostgreSQL compared to @strong{MySQL}: +You must free the result set with @code{mysql_free_result()}. -@itemize @bullet -@item -@code{Vaccum()} makes PostgreSQL hard to use in a 24/7 environment. -@item -Only transactional tables. -@item -Much slower insert/delete/update. -@end itemize +@subsubheading Return Values -For a complete list of drawbacks, you should also examine the first table -in this section. +A @code{MYSQL_RES} result set for success. @code{NULL} if an error occurred. -@cindex PostgreSQL/MySQL, benchmarks -@node MySQL-PostgreSQL benchmarks, , MySQL-PostgreSQL features, Compare PostgreSQL -@subsection Benchmarking MySQL and PostgreSQL +@subsubheading Errors -The only open source benchmark, that we know of, that can be used to -benchmark @strong{MySQL} and PostgreSQL (and other databases) is our -own. It can be found at: -@uref{http://www.mysql.com/information/benchmarks.html}. +@table @code +@item CR_COMMANDS_OUT_OF_SYNC +Commands were executed in an improper order. +@item CR_SERVER_GONE_ERROR +The @strong{MySQL} server has gone away. +@item CR_SERVER_LOST +The connection to the server was lost during the query. +@item CR_UNKNOWN_ERROR +An unknown error occurred. +@end table -We have many times asked the PostgreSQL developers and some PostgreSQL -users to help us extend this benchmark to make the definitive benchmark -for databases, but unfortunately we haven't got any feedback for this. +@findex @code{mysql_num_fields()} +@findex @code{mysql_field_count()} +@node mysql_num_fields, mysql_num_rows, mysql_list_tables, C API functions +@subsubsection @code{mysql_num_fields()} -We, the @strong{MySQL} developers, have because of this spent a lot of -hours to get maximum performance from PostgreSQL for the benchmarks, but -because we don't know PostgreSQL intimately we are sure that there are -things that we have missed. We have on the benchmark page documented -exactly how we did run the benchmark so that it should be easy for -anyone to repeat and verify our results. +@code{unsigned int mysql_num_fields(MYSQL_RES *result)} -The benchmarks are usually run with and without the @code{--fast} -option. When run with @code{--fast} we are trying to use every trick -the server can do to get the code to execute as fast as possible. -The idea is that the normal run should show how the server would work in -a default setup and the @code{--fast} run shows how the server would do -if the application developer would use extensions in the server to make -his application run faster. +or -When running with PostgreSQL and @code{--fast} we do a @code{vacuum()} -between after every major table update and drop table to make the database -in perfect shape for the following selects. The time for vacuum() is -measured separately. +@code{unsigned int mysql_num_fields(MYSQL *mysql)} -When running with PostgreSQL 7.1.1 we could however not run with -@code{--fast} because during the insert test, the postmaster (the -PostgreSQL deamon) died and the database was so corrupted that it was -impossible to restart postmaster. (The details about the machine we run -the benchmark can be found on the benchmark page). After this happened -twice, we decided to postpone the @code{--fast} test until next -PostgreSQL release. +The second form doesn't work on @strong{MySQL} Version 3.22.24 or newer. To pass a +@code{MYSQL*} argument, you must use +@code{unsigned int mysql_field_count(MYSQL *mysql)} instead. -Before going to the other benchmarks we know of, We would like to give -some background to benchmarks: +@subsubheading Description -It's very easy to write a test that shows ANY database to be best -database in the world, by just restricting the test to something the -database is very good at and not test anything that the database is not -good at; If one after this publish the result with a single figure -things is even easier. +Returns the number of columns in a result set. -This would be like we would measure the speed of @strong{MySQL} compared -to PostgreSQL by looking at the summary time of the MySQL benchmarks on -our web page. Based on this @strong{MySQL} would be more than 40 times -faster than PostgreSQL, something that is of course not true. We could -make things even worse by just taking the test where PostgreSQL performs -worst and claim that @strong{MySQL} is more than 2000 times faster than -PostgreSQL. +Note that you can get the number of columns either from a pointer to a result +set or to a connection handle. You would use the connection handle if +@code{mysql_store_result()} or @code{mysql_use_result()} returned +@code{NULL} (and thus you have no result set pointer). In this case, you can +call @code{mysql_field_count()} to determine whether or not +@code{mysql_store_result()} should have produced a non-empty result. This +allows the client program to take proper action without knowing whether or +not the query was a @code{SELECT} (or @code{SELECT}-like) statement. The +example shown below illustrates how this may be done. -The case is that @strong{MySQL} does a lot of optimizations that -PostgreSQL doesn't do and the other way around. An SQL optimizer is a -very complex thing and a company could spend years on just making the -optimizer faster and faster. +@xref{NULL mysql_store_result, , @code{NULL mysql_store_result()}}. -When looking at the benchmark results you should look for things that -you do in your application and just use these results to decide which -database would be best suited for your application. The benchmark -results also shows things a particular database is not good at and should -give you a notion about things to avoid and what you may have to do in -other ways. +@subsubheading Return Values -We know of two benchmark tests that claims that PostgreSQL performers -better than @strong{MySQL}. These both where multi-user tests, a test -that we here at @strong{MySQL AB} haven't had time to write and include in -the benchmark suite, mainly because it's a big task to do this in a -manner that is fair against all databases. +An unsigned integer representing the number of fields in a result set. -One is the benchmark paid for by -@uref{http://www.greatbridge.com/about/press.php?content_id=4,Great -Bridge}. +@subsubheading Errors +None. -This is the worst benchmark we have ever seen anyone ever conduct. This -was not only tuned to only test what PostgreSQL is absolutely best at, -it was also totally unfair against every other database involved in the -test. +@subsubheading Example -@strong{NOTE}: We know that not even some of the main PostgreSQL -developers did like the way Great Bridge conducted the benchmark, so we -don't blame them for the way the benchmark was made. +@example +MYSQL_RES *result; +unsigned int num_fields; +unsigned int num_rows; -This benchmark has been condemned in a lot of postings and newsgroups so -we will here just shortly repeat some things that where wrong with it. +if (mysql_query(&mysql,query_string)) +@{ + // error +@} +else // query succeeded, process any data returned by it +@{ + result = mysql_store_result(&mysql); + if (result) // there are rows + @{ + num_fields = mysql_num_fields(result); + // retrieve rows, then call mysql_free_result(result) + @} + else // mysql_store_result() returned nothing; should it have? + @{ + if (mysql_errno(&mysql)) + @{ + fprintf(stderr, "Error: %s\n", mysql_error(&mysql)); + @} + else if (mysql_field_count(&mysql) == 0) + @{ + // query does not return data + // (it was not a SELECT) + num_rows = mysql_affected_rows(&mysql); + @} + @} +@} +@end example -@itemize @bullet -@item -The tests where run with an expensive commercial tool, that makes it -impossible for an open source company like us to verify the benchmarks, -or even check how the benchmark where really done. The tool is not even -a true benchmark tool, but a application/setup testing tool. To refer -this as STANDARD benchmark tool is to stretch the truth a long way. -@item -Great Bridge admitted that they had optimized the PostgreSQL database -(with vacuum() before the test) and tuned the startup for the tests, -something they hadn't done for any of the other databases involved. To -say "This process optimizes indexes and frees up disk space a bit. The -optimized indexes boost performance by some margin". Our benchmarks -clearly indicates that the difference in running a lot of selects on a -database with and without vacuum() can easily differ by a factor of 10. -@item -The test results where also strange; The AS3AP test documentation -mentions that the test does: +An alternative (if you KNOW that your query should have returned a result set) +is to replace the @code{mysql_errno(&mysql)} call with a check if +@code{mysql_field_count(&mysql)} is = 0. This will only happen if something +went wrong. -"selections, simple joins, projections, aggregates, one-tuple updates, -and bulk updates" +@findex @code{mysql_num_rows()} +@node mysql_num_rows, mysql_options, mysql_num_fields, C API functions +@subsubsection @code{mysql_num_rows()} -PostgreSQL is good at doing selects and joins (especially after a -vacuum()), but doesn't perform as well on inserts/updates; The -benchmarks seem to indicate that only SELECTs where done (or very few -updates) . This could easily explain they good results for PostgreSQL in -this test. The bad results for MySQL will be obvious a bit down in this -document. -@item -They did run the so called benchmark from a Windows machine against a -Linux machine over ODBC, a setup that no normal database user would ever -do when running a heavy multi-user application. This tested more the -ODBC driver and the Windows protocol used between the clients than the -database itself. -@item -When running the database against Oracle and MS-SQL (Great Bridge has -indirectly indicated that the databases they used in the test), -they didn't use the native protocol but instead ODBC. Anyone that has -ever used Oracle, knows that all real application uses the native -interface instead of ODBC. Doing a test through ODBC and claiming that -the results had anything to do with using the database for real can't -be regarded as fair play. They should have done two tests with and -without ODBC to provide the right facts (after having got experts to tune -all involved databases of course). -@item -They refer to the TPC-C tests, but doesn't anywhere mention that the -tests they did where not a true TPC-C test and they where not even -allowed to call it a TPC-C test. A TPC-C test can only be conducted by -the rules approved by the @uref{http://www.tpc.org,TPC-council}. Great -Bridge didn't do that. By doing this they have both violated the TPC -trademark and miscredited their own benchmarks. The rules set by the -TPC-council are very strict to ensure that no one can produce false -results or make unprovable statements. Apparently Great Bridge wasn't -interested in doing this. -@item -After the first test, we contacted Great Bridge and mentioned to them -some of the obvious mistakes they had done with @strong{MySQL}; Running -with a debug version of our ODBC driver, running on a Linux system that -wasn't optimized for threads, using an old MySQL version when there was -a recommended newer one available, not starting @strong{MySQL} with the -right options for heavy multi-user use (the default installation of -MySQL is tuned for minimal resource use). Great Bridge did run a new -test, with our optimized ODBC driver and with better startup options for -MySQL, but refused to either use our updated glibc library or our -standard binary (used by 80% of our users), which was statically linked -with a fixed glibc library. +@code{my_ulonglong mysql_num_rows(MYSQL_RES *result)} -According to what we know, Great Bridge did nothing to ensure that the -other databases where setup correctly to run good in their test -environment. We are sure however that they didn't contact Oracle or -Microsoft to ask for their advice in this matter ;) -@item -The benchmark was paid for by Great Bridge, and they decided to publish -only partial chosen results (instead of publishing it all). -@end itemize +@subsubheading Description -Tim Perdue, a long time PostgreSQL fan and a reluctant MySQL user -published a comparison on -@uref{http://www.phpbuilder.com/columns/tim20001112.php3,phpbuider}. +Returns the number of rows in the result set. -When we got aware of the comparison, we phoned Tim Perdue about this -because there was a lot of strange things in his results. For example, -he claimed that MySQL had a problem with five users in his tests, when we -know that there are users with similar machines as his that are using -MySQL with 2000 simultaneous connections doing 400 queries per second (In -this case the limit was the web bandwidth, not the database). +The use of @code{mysql_num_rows()} depends on whether you use +@code{mysql_store_result()} or @code{mysql_use_result()} to return the result +set. If you use @code{mysql_store_result()}, @code{mysql_num_rows()} may be +called immediately. If you use @code{mysql_use_result()}, +@code{mysql_num_rows()} will not return the correct value until all the rows +in the result set have been retrieved. -It sounded like he was using a Linux kernel that either had some -problems with many threads (Linux kernels before 2.4 had a problem with -this but we have documented how to fix this and Tim should be aware of -this problem). The other possible problem could have been an old glibc -library and that Tim didn't use a MySQL binary from our site, which is -linked with a corrected glibc library, but had compiled a version of his -own with. In any of the above cases, the symptom would have been exactly -what Tim had measured. +@subsubheading Return Values -We asked Tim if we could get access to his data so that we could repeat -the benchmark and if he could check the MySQL version on the machine to -find out what was wrong and he promised to come back to us about this. -He has not done that yet. +The number of rows in the result set. -Because of this we can't put any trust in this benchmark either :( +@subsubheading Errors +None. -Conclusion: +@findex @code{mysql_options()} +@node mysql_options, mysql_ping, mysql_num_rows, C API functions +@subsubsection @code{mysql_options()} -The only benchmarks that exist today that anyone can download and run -against @strong{MySQL}and PostgreSQL is the MySQL benchmarks. We here -at @strong{MySQL} believe that open source databases should be tested -with open source tools! This is the only way to ensure that no one -does tests that nobody can reproduce and use this to claim that a -database is better than another. Without knowing all the facts it's -impossible to answer the claims of the tester. +@code{int mysql_options(MYSQL *mysql, enum mysql_option option, const char *arg)} -The thing we find strange is that every test we have seen about -PostgreSQL, that is impossible to reproduce, claims that PostgreSQL is -better in most cases while our tests, which anyone can reproduce, -clearly shows otherwise. With this we don't want to say that PostgreSQL -isn't good at many things (It is!) We would just like to see a fair test -where they are very good so that we could get some friendly competition -going! +@subsubheading Description -For more information about our benchmarks suite see @xref{MySQL -Benchmarks}. +Can be used to set extra connect options and affect behavior for a connection. +This function may be called multiple times to set several options. -We are working on an even better benchmark suite, including much better -documentation of what the individual tests really do and how to add more -tests to the suite. +@code{mysql_options()} should be called after @code{mysql_init()} and before +@code{mysql_connect()} or @code{mysql_real_connect()}. -@cindex internals -@cindex threads -@node MySQL internals, Environment variables, Comparisons, Top -@chapter MySQL Internals +The @code{option} argument is the option that you want to set; the @code{arg} +argument is the value for the option. If the option is an integer, then +@code{arg} should point to the value of the integer. -This chapter describes a lot of things that you need to know when -working on the @strong{MySQL} code. If you plan to contribute to MySQL -development, want to have access to the bleeding-edge in-between -versions code, or just want to keep track of development, follow the -instructions in @xref{Installing source tree}. If you are interested in MySQL -internals, you should also subscribe to @email{internals@@lists.mysql.com}. -This is a relatively low traffic list, in comparison with -@email{mysql@@lists.mysql.com}. - -@menu -* MySQL threads:: MySQL threads -* MySQL test suite:: MySQL test suite -@end menu - -@node MySQL threads, MySQL test suite, MySQL internals, MySQL internals -@section MySQL Threads +Possible options values: -The @strong{MySQL} server creates the following threads: +@multitable @columnfractions .25 .25 .5 +@item @strong{Option} @tab @strong{Argument type} @tab @strong{Function} +@item @code{MYSQL_OPT_CONNECT_TIMEOUT} @tab @code{unsigned int *} @tab Connect timeout in seconds. +@item @code{MYSQL_OPT_COMPRESS} @tab Not used @tab Use the compressed client/server protocol. +@item @code{MYSQL_OPT_NAMED_PIPE} @tab Not used @tab Use named pipes to connect to a @strong{MySQL} server on NT. +@item @code{MYSQL_INIT_COMMAND} @tab @code{char *} @tab Command to execute when connecting to the @strong{MySQL} server. Will automatically be re-executed when reconnecting. +@item @code{MYSQL_READ_DEFAULT_FILE} @tab @code{char *} @tab Read options from the named option file instead of from @file{my.cnf}. +@item @code{MYSQL_READ_DEFAULT_GROUP} @tab @code{char *} @tab Read options from the named group from @file{my.cnf} or the file specified with @code{MYSQL_READ_DEFAULT_FILE}. +@end multitable -@itemize @bullet +Note that the group @code{client} is always read if you use +@code{MYSQL_READ_DEFAULT_FILE} or @code{MYSQL_READ_DEFAULT_GROUP}. -@item -The TCP/IP connection thread handles all connection requests and -creates a new dedicated thread to handle the authentication and -and SQL query processing for each connection. +The specified group in the option file may contain the following options: +@multitable @columnfractions .3 .7 +@item @code{connect_timeout} @tab Connect timeout in seconds. On Linux this timeout is also used for waiting for the first answer from the server. +@item @code{compress} @tab Use the compressed client/server protocol. +@item @code{database} @tab Connect to this database if no database was specified in the connect command. +@item @code{debug} @tab Debug options. +@item @code{host} @tab Default host name. +@item @code{init-command} @tab Command to execute when connecting to @strong{MySQL} server. Will automatically be re-executed when reconnecting. +@item @code{interactive-timeout} @tab Same as specifying @code{CLIENT_INTERACTIVE} to @code{mysql_real_connect()}. @xref{mysql_real_connect}. +@item @code{password} @tab Default password. +@item @code{pipe} @tab Use named pipes to connect to a @strong{MySQL} server on NT. +@item @code{port} @tab Default port number. +@item @code{return-found-rows} @tab Tell @code{mysql_info()} to return found rows instead of updated rows when using @code{UPDATE}. +@item @code{socket} @tab Default socket number. @item -On Windows NT there is a named pipe handler thread that does the same work as -the TCP/IP connection thread on named pipe connect requests. +@item @code{user} @tab Default user. +@end multitable -@item -The signal thread handles all signals. This thread also normally handles -alarms and calls @code{process_alarm()} to force timeouts on connections -that have been idle too long. +Note that @code{timeout} has been replaced by @code{connect_timeout}, but +@code{timeout} will still work for a while. -@item -If @code{mysqld} is compiled with @code{-DUSE_ALARM_THREAD}, a dedicated -thread that handles alarms is created. This is only used on some systems where -there are problems with @code{sigwait()} or if one wants to use the -@code{thr_alarm()} code in ones application without a dedicated signal -handling thread. +For more information about option files, see @ref{Option files}. -@item -If one uses the @code{--flush_time=#} option, a dedicated thread is created -to flush all tables at the given interval. +@subsubheading Return Values -@item -Every connection has its own thread. +Zero for success. Non-zero if you used an unknown option. -@item -Every different table on which one uses @code{INSERT DELAYED} gets its -own thread. +@subsubheading Example -@item -If you use @code{--master-host}, a slave replication thread will be -started to read and apply updates from the master. -@end itemize +@example +MYSQL mysql; -@code{mysqladmin processlist} only shows the connection, @code{INSERT DELAYED}, -and replication threads. +mysql_init(&mysql); +mysql_options(&mysql,MYSQL_OPT_COMPRESS,0); +mysql_options(&mysql,MYSQL_READ_DEFAULT_GROUP,"odbc"); +if (!mysql_real_connect(&mysql,"host","user","passwd","database",0,NULL,0)) +@{ + fprintf(stderr, "Failed to connect to database: Error: %s\n", + mysql_error(&mysql)); +@} +@end example -@cindex mysqltest, MySQL Test Suite -@cindex testing mysqld, mysqltest -@node MySQL test suite, , MySQL threads, MySQL internals -@section MySQL Test Suite +The above requests the client to use the compressed client/server protocol and +read the additional options from the @code{odbc} section in the @code{my.cnf} +file. -Until recently, our main full-coverage test suite was based on proprietary -customer data and for that reason has not been publicly available. The only -publicly available part of our testing process consisted of the @code{crash-me} -test, a Perl DBI/DBD benchmark found in the @code{sql-bench} directory, and -miscellaneous tests located in @code{tests} directory. The lack of a -standardized publicly available test suite has made it difficult for our users, -as well developers, to do regression tests on the @strong{MySQL} code. To -address this problem, we have created a new test system that is included in -the source and binary distributions starting in Version 3.23.29. +@findex @code{mysql_ping()} +@node mysql_ping, mysql_query, mysql_options, C API functions +@subsubsection @code{mysql_ping()} -The current set of test cases doesn't test everything in @strong{MySQL}, but it -should catch most obvious bugs in the SQL processing code, OS/library -issues, and is quite thorough in testing replication. Our eventual goal -is to have the tests cover 100% of the code. We welcome contributions -to our test suite. You may especially want to contribute tests that -examine the functionality critical to your system, as this will ensure -that all future @strong{MySQL} releases will work well with your -applications. +@code{int mysql_ping(MYSQL *mysql)} -@menu -* running mysqltest:: -* extending mysqltest:: -* Reporting mysqltest bugs:: -@end menu +@subsubheading Description -@node running mysqltest, extending mysqltest, MySQL test suite, MySQL test suite -@subsection Running the MySQL Test Suite +Checks whether or not the connection to the server is working. If it has gone +down, an automatic reconnection is attempted. -The test system consist of a test language interpreter -(@code{mysqltest}), a shell script to run all -tests(@code{mysql-test-run}), the actual test cases written in a special -test language, and their expected results. To run the test suite on -your system after a build, type @code{make test} or -@code{mysql-test/mysql-test-run} from the source root. If you have -installed a binary distribution, @code{cd} to the install root -(eg. @code{/usr/local/mysql}), and do @code{scripts/mysql-test-run}. -All tests should succeed. If not, you should try to find out why and -report the problem if this is a bug in @strong{MySQL}. -@xref{Reporting mysqltest bugs}. +This function can be used by clients that remain idle for a long while, +to check whether or not the server has closed the connection and reconnect +if necessary. -If you have a copy of @code{mysqld} running on the machine where you want to -run the test suite you do not have to stop it, as long as it is not using -ports @code{9306} and @code{9307}. If one of those ports is taken, you should -edit @code{mysql-test-run} and change the values of the master and/or slave -port to one that is available. +@subsubheading Return Values -You can run one individual test case with -@code{mysql-test/mysql-test-run test_name}. +Zero if the server is alive. Non-zero if an error occurred. -If one test fails, you should test running @code{mysql-test-run} with -the @code{--force} option to check if any other tests fails. +@subsubheading Errors -@node extending mysqltest, Reporting mysqltest bugs, running mysqltest, MySQL test suite -@subsection Extending the MySQL Test Suite +@table @code +@item CR_COMMANDS_OUT_OF_SYNC +Commands were executed in an improper order. +@item CR_SERVER_GONE_ERROR +The @strong{MySQL} server has gone away. +@item CR_UNKNOWN_ERROR +An unknown error occurred. +@end table -You can use the @code{mysqltest} language to write your own test cases. -Unfortunately, we have not yet written full documentation for it - we plan to -do this shortly. You can, however, look at our current test cases and use -them as an example. The following points should help you get started: +@findex @code{mysql_query()} +@node mysql_query, mysql_real_connect, mysql_ping, C API functions +@subsubsection @code{mysql_query()} -@itemize -@item -The tests are located in @code{mysql-test/t/*.test} +@code{int mysql_query(MYSQL *mysql, const char *query)} -@item -A test case consists of @code{;} terminated statements and is similar to the -input of @code{mysql} command line client. A statement by default is a query -to be sent to @strong{MySQL} server, unless it is recognized as internal -command ( eg. @code{sleep} ). +@subsubheading Description +Executes the SQL query pointed to by the null-terminated string @code{query}. +The query must consist of a single SQL statement. You should not add +a terminating semicolon (@samp{;}) or @code{\g} to the statement. -@item -All queries that produce results, e.g. @code{SELECT}, @code{SHOW}, -@code{EXPLAIN}, etc., must be preceded with @code{@@/path/to/result/file}. The -file must contain the expected results. An easy way to generate the result -file is to run @code{mysqltest -r < t/test-case-name.test} from -@code{mysql-test} directory, and then edit the generated result files, if -needed, to adjust them to the expected output. In that case, be very careful -about not adding or deleting any invisible characters - make sure to only -change the text and/or delete lines. If you have to insert a line, make sure -the fields are separated with a hard tab, and there is a hard tab at the end. -You may want to use @code{od -c} to make sure your text editor has not messed -anything up during edit. We, of course, hope that you will never have to edit -the output of @code{mysqltest -r} as you only have to do it when you find a -bug. +@code{mysql_query()} cannot be used for queries that contain binary data; you +should use @code{mysql_real_query()} instead. (Binary data may contain the +@samp{\0} character, which @code{mysql_query()} interprets as the end of the +query string.) -@item -To be consistent with our setup, you should put your result files in -@code{mysql-test/r} directory and name them @code{test_name.result}. If the -test produces more than one result, you should use @code{test_name.a.result}, -@code{test_name.b.result}, etc. +If you want to know if the query should return a result set or not, you can +use @code{mysql_field_count()} to check for this. +@xref{mysql_field_count, , @code{mysql_field_count}}. -@item -If a statement returns an error, you should on the line before the statement -specify with the @code{--error error-number}. The error number can be -a list of possible error numbers separated with @code{','}. +@subsubheading Return Values -@item -If you are writing a replication test case, you should on the first line of -the test file, put @code{source include/master-slave.inc;}. To switch between -master and slave, use @code{connection master;} and @code{connection slave;}. -If you need to do something on an alternate connection, you can do -@code{connection master1;} for the master, and @code{connection slave1;} for -the slave. +Zero if the query was successful. Non-zero if an error occurred. -@item -If you need to do something in a loop, you can use something like this: -@example -let $1=1000; -while ($1) -@{ - # do your queries here - dec $1; -@} -@end example +@subsubheading Errors -@item -To sleep between queries, use the @code{sleep} command. It supports fractions -of a second, so you can do @code{sleep 1.3;}, for example, to sleep 1.3 -seconds. +@table @code +@item CR_COMMANDS_OUT_OF_SYNC +Commands were executed in an improper order. +@item CR_SERVER_GONE_ERROR +The @strong{MySQL} server has gone away. +@item CR_SERVER_LOST +The connection to the server was lost during the query. +@item CR_UNKNOWN_ERROR +An unknown error occurred. +@end table -@item -To run the slave with additional options for your test case, put them -in the command-line format in @code{mysql-test/t/test_name-slave.opt}. For -the master, put them in @code{mysql-test/t/test_name-master.opt}. +@findex @code{mysql_real_connect()} +@node mysql_real_connect, mysql_real_escape_string, mysql_query, C API functions +@subsubsection @code{mysql_real_connect()} -@item -If you have a question about the test suite, or have a test case to contribute, -e-mail to @email{internals@@lists.mysql.com}. As the list does not accept -attachments, you should ftp all the relevant files to: -@url{ftp://support.mysql.com/pub/mysql/Incoming} +@code{MYSQL *mysql_real_connect(MYSQL *mysql, const char *host, + const char *user, const char *passwd, const char *db, + unsigned int port, const char *unix_socket, + unsigned int client_flag)} -@end itemize +@subsubheading Description -@node Reporting mysqltest bugs, , extending mysqltest, MySQL test suite -@subsection Reporting bugs in the MySQL Test Suite +@code{mysql_real_connect()} attempts to establish a connection to a +@strong{MySQL} database engine running on @code{host}. +@code{mysql_real_connect()} must complete successfully before you can execute +any of the other API functions, with the exception of +@code{mysql_get_client_info()}. -If your @strong{MySQL} version doesn't pass the test suite you should -do the following: +The parameters are specified as follows: @itemize @bullet @item -Don't send a bug report before you have found out as much as possible of -what when wrong! When you do it, please use the @code{mysqlbug} script -so that we can get information about your system and @code{MySQL} -version. @xref{Bug reports}. -@item -Make sure to include the output of @code{mysql-test-run}, as well as -contents of all @code{.reject} files in @code{mysql-test/r} directory. -@item -If a test in the test suite fails, check if the test fails also when run -by its own: - -@example -cd mysql-test -mysql-test-run --local test-name -@end example - -If this fails, then you should configure @strong{MySQL} with -@code{--with-debug} and run @code{mysql-test-run} with the -@code{--debug} option. If this also fails send the trace file -@file{var/tmp/master.trace} to ftp://support.mysql.com/pub/mysql/secret -so that we can examine it. Please remember to also include a full -description of your system, the version of the mysqld binary and how you -compiled it. +The first parameter should be the address of an existing @code{MYSQL} +structure. Before calling @code{mysql_real_connect()} you must call +@code{mysql_init()} to initialize the @code{MYSQL} structure. You can +change a lot of connect options with the @code{mysql_options()} +call. @xref{mysql_options}. @item -Try also to run @code{mysql-test-run} with the @code{--force} option to -see if there is any other test that fails. +The value of @code{host} may be either a hostname or an IP address. If +@code{host} is @code{NULL} or the string @code{"localhost"}, a connection to +the local host is assumed. If the OS supports sockets (Unix) or named pipes +(Windows), they are used instead of TCP/IP to connect to the server. @item -If you have compiled @strong{MySQL} yourself, check our manual for how -to compile @strong{MySQL} on your platform or, preferable, use one of -the binaries we have compiled for you at -@uref{http://www.mysql.com/downloads/}. All our standard binaries should -pass the test suite ! +The @code{user} parameter contains the user's @strong{MySQL} login ID. If +@code{user} is @code{NULL}, the current user is assumed. Under Unix, this is +the current login name. Under Windows ODBC, the current user name must be +specified explicitly. +@xref{ODBC administrator}. @item -If you get an error, like @code{Result length mismatch} or @code{Result -content mismatch} it means that the output of the test didn't match -exactly the expected output. This could be a bug in @strong{MySQL} or -that your mysqld version produces slight different results under some -circumstances. +The @code{passwd} parameter contains the password for @code{user}. If +@code{passwd} is @code{NULL}, only entries in the @code{user} table for the +user that have a blank (empty) password field will be checked for a match. This +allows the database administrator to set up the @strong{MySQL} privilege +system in such a way that users get different privileges depending on whether +or not they have specified a password. -Failed test results are put in a file with the same base name as the -result file with the @code{.reject} extension. If your test case is -failing, you should do a diff on the two files. If you cannot see how -they are different, examine both with @code{od -c} and also check their -lengths. +NOTE: Do not attempt to encrypt the password before calling +@code{mysql_real_connect()}; password encryption is handled automatically by +the client API. @item -If a test fails totally, you should check the logs file in the -@code{mysql-test/var/log} directory for hints of what went wrong. +@code{db} is the database name. +If @code{db} is not @code{NULL}, the connection will set the default +database to this value. @item -If you have compiled @strong{MySQL} with debugging you can try to debug this -by running @code{mysql-test-run} with the @code{--gdb} and/or @code{--debug} -options. -@xref{Making trace files}. - -If you have not compiled @strong{MySQL} for debugging you should probably -do that. Just specify the @code{--with-debug} options to @code{configure}! -@xref{Installing source}. -@end itemize - -@page -@cindex environment variables, list of -@node Environment variables, Users, MySQL internals, Top -@appendix Environment Variables - -Here is a list of all the environment variables that are used directly or -indirectly by @strong{MySQL}. Most of these can also be found in other -places in this manual. - -Note that any options on the command line will take precedence over -values specified in configuration files and environment variables, and -values in configuration files take precedence over values in environment -variables. +If @code{port} is not 0, the value will be used as the port number +for the TCP/IP connection. Note that the @code{host} parameter +determines the type of the connection. -In many cases it's preferable to use a configure file instead of environment -variables to modify the behavior of @strong{MySQL}. @xref{Option files}. +@item +If @code{unix_socket} is not @code{NULL}, the string specifies the +socket or named pipe that should be used. Note that the @code{host} +parameter determines the type of the connection. -@tindex CCX environment variable -@tindex Environment variable, CCX -@tindex CC environment variable -@tindex Environment variable, CC -@tindex CFLAGS environment variable -@tindex Environment variable, CFLAGS -@tindex CXXFLAGS environment variable -@tindex Environment variable, CXXFLAGS -@tindex DBI_USER environment variable -@tindex Environment variable, DBI_USER -@tindex DBI_TRACE environment variable -@tindex Environment variable, DBI_TRACE -@tindex HOME environment variable -@tindex Environment variable, HOME -@tindex LD_RUN_PATH environment variable -@tindex Environment variable, LD_RUN_PATH -@tindex MYSQL_DEBUG environment variable -@tindex Environment variable, MYSQL_DEBUG -@tindex MYSQL_HISTFILE environment variable -@tindex Environment variable, MYSQL_HISTFILE -@tindex MYSQL_HOST environment variable -@tindex Environment variable, MYSQL_HOST -@tindex MYSQL_PWD environment variable -@tindex Environment variable, MYSQL_PWD -@tindex MYSQL_TCP_PORT environment variable -@tindex Environment variable, MYSQL_TCP_PORT -@tindex MYSQL_UNIX_PORT environment variable -@tindex Environment variable, MYSQL_UNIX_PORT -@tindex PATH environment variable -@tindex Environment variable, PATH -@tindex TMPDIR environment variable -@tindex Environment variable, TMPDIR -@tindex TZ environment variable -@tindex Environment variable, TZ -@tindex UMASK_DIR environment variable -@tindex Environment variable, UMASK_DIR -@tindex UMASK environment variable -@tindex Environment variable, UMASK -@tindex USER environment variable -@tindex Environment variable, USER +@item +The value of @code{client_flag} is usually 0, but can be set to a combination +of the following flags in very special circumstances: -@multitable @columnfractions .2 .8 -@item @code{CCX} @tab Set this to your C++ compiler when running configure. -@item @code{CC} @tab Set this to your C compiler when running configure. -@item @code{CFLAGS} @tab Flags for your C compiler when running configure. -@item @code{CXXFLAGS} @tab Flags for your C++ compiler when running configure. -@item @code{DBI_USER} @tab The default user name for Perl DBI. -@item @code{DBI_TRACE} @tab Used when tracing Perl DBI. -@item @code{HOME} @tab The default path for the @code{mysql} history file is @file{$HOME/.mysql_history}. -@item @code{LD_RUN_PATH} @tab Used to specify where your @code{libmysqlclient.so} is. -@item @code{MYSQL_DEBUG} @tab Debug-trace options when debugging. -@item @code{MYSQL_HISTFILE} @tab The path to the @code{mysql} history file. -@item @code{MYSQL_HOST} @tab Default host name used by the @code{mysql} command-line prompt. -@item @code{MYSQL_PWD} @tab The default password when connecting to @code{mysqld}. Note that use of this is insecure! -@item @code{MYSQL_TCP_PORT} @tab The default TCP/IP port. -@item @code{MYSQL_UNIX_PORT} @tab The default socket; used for connections to @code{localhost}. -@item @code{PATH} @tab Used by the shell to finds the @strong{MySQL} programs. -@item @code{TMPDIR} @tab The directory where temporary tables/files are created. -@item @code{TZ} @tab This should be set to your local time zone. @xref{Timezone problems}. -@item @code{UMASK_DIR} @tab The user-directory creation mask when creating directories. Note that this is ANDed with @code{UMASK}! -@item @code{UMASK} @tab The user-file creation mask when creating files. -@item @code{USER} @tab The default user on Windows to use when connecting to @code{mysqld}. +@multitable @columnfractions .25 .7 +@item @strong{Flag name} @tab @strong{Flag meaning} +@code{mysqld} to be more ODBC-friendly. +@item @code{CLIENT_COMPRESS} @tab Use compression protocol. +@item @code{CLIENT_FOUND_ROWS} @tab Return the number of found (matched) rows, not the number of affected rows. +@item @code{CLIENT_IGNORE_SPACE} @tab Allow spaces after function names. Makes all functions names reserved words. +@item @code{CLIENT_INTERACTIVE} @tab Allow @code{interactive_timeout} seconds (instead of @code{wait_timeout} seconds) of inactivity before closing the connection. +@item @code{CLIENT_NO_SCHEMA} @tab Don't allow the @code{db_name.tbl_name.col_name} syntax. This is for ODBC. It causes the parser to generate an error if you use that syntax, which is useful for trapping bugs in some ODBC programs. +@item @code{CLIENT_ODBC} @tab The client is an ODBC client. This changes +@item @code{CLIENT_SSL} @tab Use SSL (encrypted protocol). @end multitable +@end itemize -@page -@cindex users, of MySQL -@cindex news sites -@node Users, MySQL customer usage, Environment variables, Top -@appendix Some MySQL Users +@subsubheading Return Values -This appendix lists some users of @strong{MySQL} that have given us permission -to list them in our documentation. It is by far not a complete list, but -should give you a general idea of who uses @strong{MySQL} and what it can -be used for. +A @code{MYSQL*} connection handle if the connection was successful, +@code{NULL} if the connection was unsuccessful. For a successful connection, +the return value is the same as the value of the first parameter, unless you +pass @code{NULL} for that parameter. -@appendixsec General News Sites +@subsubheading Errors -@itemize @bullet +@table @code +@item CR_CONN_HOST_ERROR +Failed to connect to the @strong{MySQL} server. -@item @uref{http://www.yahoo.com/, Yahoo!} +@item CR_CONNECTION_ERROR +Failed to connect to the local @strong{MySQL} server. -@item @uref{http://slashdot.org/, Slashdot: A pro-Linux/tech news and comment/discussion site} +@item CR_IPSOCK_ERROR +Failed to create an IP socket. -@item @uref{http://www.linux.com/, All about Linux} +@item CR_OUT_OF_MEMORY +Out of memory. -@item @uref{http://www.linuxtoday.com/, Linuxtoday} +@item CR_SOCKET_CREATE_ERROR +Failed to create a Unix socket. -@item @uref{http://www.32bitsonline.com/, 32Bits Online: because there's -more than one way to compute} +@item CR_UNKNOWN_HOST +Failed to find the IP address for the hostname. -@item @uref{http://www.freshmeat.net/, Freshmeat: News about new versions of computer-related stuff} +@item CR_VERSION_ERROR +A protocol mismatch resulted from attempting to connect to a server with a +client library that uses a different protocol version. This can happen if you +use a very old client library to connect to a new server that wasn't started +with the @code{--old-protocol} option. -@end itemize +@item CR_NAMEDPIPEOPEN_ERROR +Failed to create a named pipe on Windows. -@cindex search engines, web -@cindex web search engines -@appendixsec Some Web Search Engines +@item CR_NAMEDPIPEWAIT_ERROR +Failed to wait for a named pipe on Windows. -@itemize @bullet +@item CR_NAMEDPIPESETSTATE_ERROR +Failed to get a pipe handler on Windows. -@item @uref{http://www.aaa.com.au, AAA Matilda Web Search} +@item CR_SERVER_LOST +If @code{connect_timeout} > 0 and it took longer then @code{connect_timeout} +seconds to connect to the server or if the server died while executing the +@code{init-command}. -@item @uref{http://www.whatsnu.com/, What's New} +@end table -@item @uref{http://www.aladin.de/, Aladin} +@subsubheading Example -@item @uref{http://www.columbus-finder.de/, Columbus Finder} - -@item @uref{http://www.spider.de/, Spider} +@example +MYSQL mysql; -@item @uref{http://www.blitzsuche.de/, Blitzsuche} +mysql_init(&mysql); +mysql_options(&mysql,MYSQL_READ_DEFAULT_GROUP,"your_prog_name"); +if (!mysql_real_connect(&mysql,"host","user","passwd","database",0,NULL,0)) +@{ + fprintf(stderr, "Failed to connect to database: Error: %s\n", + mysql_error(&mysql)); +@} +@end example -@item @uref{http://www.indoseek.co.id, Indoseek Indonesia} +By using @code{mysql_options()} the @strong{MySQL} library will read the +@code{[client]} and @code{your_prog_name} sections in the @code{my.cnf} +file which will ensure that your program will work, even if someone has +set up @strong{MySQL} in some non-standard way. -@item @uref{http://www.yaboo.dk/, Yaboo - Yet Another BOOkmarker} +Note that upon connection, @code{mysql_real_connect()} sets the @code{reconnect} +flag (part of the MYSQL structure) to a value of @code{1}. This flag indicates, +in the event that a query cannot be performed because of a lost connection, to +try reconnecting to the server before giving up. +@findex @code{mysql_real_escape_string()} +@node mysql_real_escape_string, mysql_real_query, mysql_real_connect, C API functions +@subsubsection @code{mysql_real_escape_string()} -@item @uref{http://www.ozsearch.com.au, OzSearch Internet Guide} +@code{unsigned int mysql_real_escape_string(MYSQL *mysql, char *to, const char *from, unsigned int length)} -@item @uref{http://www.splatsearch.com/, Splat! Search} +@subsubheading Description -@item @uref{http://osdls.library.arizona.edu/, The Open Source Digital Library System Project} -@end itemize +This function is used to create a legal SQL string that you can use in a +SQL statement. @xref{String syntax}. -@appendixsec Some Information Search Engines Concentrated on Some Area +The string in @code{from} is encoded to an escaped SQL string, taking +into account the current character set of the connection. The result is placed +in @code{to} and a terminating null byte is appended. Characters +encoded are @code{NUL} (ASCII 0), @samp{\n}, @samp{\r}, @samp{\}, +@samp{'}, @samp{"}, and Control-Z (@pxref{Literals}). -@itemize @bullet +The string pointed to by @code{from} must be @code{length} bytes long. You +must allocate the @code{to} buffer to be at least @code{length*2+1} bytes +long. (In the worse case, each character may need to be encoded as using two +bytes, and you need room for the terminating null byte.) When +@code{mysql_escape_string()} returns, the contents of @code{to} will be a +null-terminated string. The return value is the length of the encoded +string, not including the terminating null character. -@item @uref{http://www.spylog.ru/, SpyLOG ; A very popular Web counter site} +@subsubheading Example -@item @uref{http://www.tucows.com/, TuCows Network; Free Software archive} +@example +char query[1000],*end; -@item @uref{http://www.jobvertise.com,Jobvertise: Post and search for jobs} +end = strmov(query,"INSERT INTO test_table values("); +*end++ = '\''; +end += mysql_real_escape_string(&mysql, end,"What's this",11); +*end++ = '\''; +*end++ = ','; +*end++ = '\''; +end += mysql_real_escape_string(&mysql, end,"binary data: \0\r\n",16); +*end++ = '\''; +*end++ = ')'; -@item @uref{http://www.musicdatabase.com, The Music Database} +if (mysql_real_query(&mysql,query,(unsigned int) (end - query))) +@{ + fprintf(stderr, "Failed to insert row, Error: %s\n", + mysql_error(&mysql)); +@} +@end example -@item @uref{http://www.soccersearch.com, Football (Soccer) search page} +The @code{strmov()} function used in the example is included in the +@code{mysqlclient} library and works like @code{strcpy()} but returns a +pointer to the terminating null of the first parameter. -@item @uref{http://www.headrush.net/takedown, TAKEDOWN - wrestling} +@subsubheading Return Values -@item @uref{http://www.lyrics.net, The International Lyrics Network} +The length of the value placed into @code{to}, not including the +terminating null character. -@item @uref{http://TheMatrix.com/~matrix/band_search.phtml, Musicians looking for other musicians (Free Service)} +@subsubheading Errors +None. -@item @uref{http://www.addall.com/AddBooks/Stores.html,AddALL books searching and price comparison} +@findex @code{mysql_real_query()} +@node mysql_real_query, mysql_reload, mysql_real_escape_string, C API functions +@subsubsection @code{mysql_real_query()} -@item @uref{http://www.herbaria.harvard.edu/Data/Gray/gray.html,Harvard's Gray Herbarium Index of Plant Names} +@code{int mysql_real_query(MYSQL *mysql, const char *query, unsigned int length)} -@item @uref{http://www.game-developer.com/,The Game Development Search Engine} +@subsubheading Description -@item @uref{www.theinnkeeper.com, The Innkeeper Vacation Guides} +Executes the SQL query pointed to by @code{query}, which should be a string +@code{length} bytes long. The query must consist of a single SQL statement. +You should not add a terminating semicolon (@samp{;}) or @code{\g} to the +statement. -@item @uref{http://www.macgamedatabase.com/, The Mac Game Database uses PHP and MySQL} -@c From: Marc Antony Vose <suzerain@suzerain.com> +You @emph{must} use @code{mysql_real_query()} rather than +@code{mysql_query()} for queries that contain binary data, because binary data +may contain the @samp{\0} character. In addition, @code{mysql_real_query()} +is faster than @code{mysql_query()} because it does not call @code{strlen()} on +the query string. -@item @uref{http://www.csse.monash.edu.au/publications/, Research -Publications at Monash University in Australia} +If you want to know if the query should return a result set or not, you can +use @code{mysql_field_count()} to check for this. +@xref{mysql_field_count, @code{mysql_field_count}}. -@item @uref{http://www.ipielle.emr.it/bts/index.html, -Occupational Health & Safety Web site database (a project for the ECC)} -@c c.presutti@ipielle.emr.it +@subsubheading Return Values -@item @uref{http://data.mch.mcgill.ca/, Bioinformatics databases at the -Montreal Children's Hospital using MySQL} -@c saeed@www.debelle.mcgill.ca -@end itemize +Zero if the query was successful. Non-zero if an error occurred. -@cindex online magazines -@cindex magazines, online -@appendixsec Online Magazines +@subsubheading Errors -@itemize @bullet -@item @uref{http://www.spoiler.com, Spoiler Webzine}. -An online magazine featuring music, literature, arts, and design content. -@item @uref{http://www.linux-magazin.de/newsflash/, Daily news about Linux in German language} -@item @uref{http://www.betazine.com,Betazine - The Ultimate Online Beta Tester's Magazine} -@item @uref{http://www.currents.net/ccinfo/aboutcc.html,Computer Currents Magazine} -@end itemize +@table @code +@item CR_COMMANDS_OUT_OF_SYNC +Commands were executed in an improper order. +@item CR_SERVER_GONE_ERROR +The @strong{MySQL} server has gone away. +@item CR_SERVER_LOST +The connection to the server was lost during the query. +@item CR_UNKNOWN_ERROR +An unknown error occurred. +@end table -@cindex web sites -@appendixsec Web Sites that Use MySQL as a Backend +@findex @code{mysql_reload()} +@node mysql_reload, mysql_row_seek, mysql_real_query, C API functions +@subsubsection @code{mysql_reload()} -@itemize @bullet +@code{int mysql_reload(MYSQL *mysql)} -@item @uref{http://liftoff.msfc.nasa.gov, NASA} -@item @uref{http://kids.msfc.nasa.gov, NASA KIDS} -@item @uref{http://science.nasa.gov, Sience@@NASA} +@subsubheading Description -@item @uref{http://www.handy.de/, handy.de} +Asks the @strong{MySQL} server to reload the grant tables. The +connected user must have the @strong{reload} privilege. -@item @uref{http://lindev.jmc.tju.edu/qwor, Qt Widget and Object Repository} +This function is deprecated. It is preferable to use @code{mysql_query()} +to issue a SQL @code{FLUSH PRIVILEGES} statement instead. -@item @uref{http://www.samba-choro.com.br, Brazilian samba site (in Portuguese)} +@subsubheading Return Values -@item @uref{http://pgss.iss.uw.edu.pl/en_index.ISS, Polish General Social Survey} +Zero for success. Non-zero if an error occurred. -@item @uref{http://www.expo2000.com, Expo2000} World-wide distribution of -tickets for this event is implemented using @strong{MySQL} and tcl/tk. More than -5000 travel agencies all over the world have access to it. +@subsubheading Errors -@item @uref{http://www.freevote.com/, FreeVote.com is a free voting -service with millions of users.} +@table @code +@item CR_COMMANDS_OUT_OF_SYNC +Commands were executed in an improper order. +@item CR_SERVER_GONE_ERROR +The @strong{MySQL} server has gone away. +@item CR_SERVER_LOST +The connection to the server was lost during the query. +@item CR_UNKNOWN_ERROR +An unknown error occurred. +@end table -@item @uref{http://f1.tauzero.se, Forza Motorsport} +@findex @code{mysql_row_seek()} +@node mysql_row_seek, mysql_row_tell, mysql_reload, C API functions +@subsubsection @code{mysql_row_seek()} -@item @uref{http://www.dreamhost.com/, DreamHost Web Hosting} +@code{MYSQL_ROW_OFFSET mysql_row_seek(MYSQL_RES *result, MYSQL_ROW_OFFSET offset)} -@end itemize +@subsubheading Description +Sets the row cursor to an arbitrary row in a query result set. This requires +that the result set structure contains the entire result of the query, so +@code{mysql_row_seek()} may be used in conjunction only with +@code{mysql_store_result()}, not with @code{mysql_use_result()}. -@cindex services -@appendixsec Some Domain/Internet/Web and Related Services +The offset should be a value returned from a call to @code{mysql_row_tell()} +or to @code{mysql_row_seek()}. This value is not simply a row number; if you +want to seek to a row within a result set using a row number, use +@code{mysql_data_seek()} instead. -@itemize @bullet +@subsubheading Return Values -@item @uref{http://www.wix.com/mysql-hosting, Registry of Web providers that -support @strong{MySQL}} +The previous value of the row cursor. This value may be passed to a +subsequent call to @code{mysql_row_seek()}. -@item @uref{http://www.yi.org/, Dynamic DNS Services} +@subsubheading Errors +None. -@item @uref{http://www.dynodns.net/, Dynamic domain name service} +@findex @code{mysql_row_tell()} +@node mysql_row_tell, mysql_select_db, mysql_row_seek, C API functions +@subsubsection @code{mysql_row_tell()} -@item @uref{http://www.ods.org/, Open DNS Project; free dynamic DNS service} +@code{MYSQL_ROW_OFFSET mysql_row_tell(MYSQL_RES *result)} -@c @item @uref{http://dynodns.net, Free dynamic DNS implementation} -@c EMAIL: A Moore <amoore@mooresystems.com> +@subsubheading Description -@item @uref{http://www.hn.org/, Hammernode; Public DNS Servers} +Returns the current position of the row cursor for the last +@code{mysql_fetch_row()}. This value can be used as an argument to +@code{mysql_row_seek()}. -@item @uref{http://www.fdns.net/, Free 3rd level domains} +You should use @code{mysql_row_tell()} only after @code{mysql_store_result()}, +not after @code{mysql_use_result()}. -@item @uref{http://worldcommunity.com/, Online Database} +@subsubheading Return Values -@item @uref{http://www.bigbiz.com, BigBiz Internet Services} +The current offset of the row cursor. -@item @uref{http://virt.circle.net, The Virt Gazette} +@subsubheading Errors +None. -@item @uref{http://www.california.com, Global InfoNet Inc} +@findex @code{mysql_select_db()} +@node mysql_select_db, mysql_shutdown, mysql_row_tell, C API functions +@subsubsection @code{mysql_select_db()} -@item @uref{http://www.webhosters.com, WebHosters - A Guide to WWW Providers} +@code{int mysql_select_db(MYSQL *mysql, const char *db)} -@item @uref{http://online.dn.ru, Internet information server} +@subsubheading Description -@item @uref{http://www.stopbit.com, A technology news site} +Causes the database specified by @code{db} to become the default (current) +database on the connection specified by @code{mysql}. In subsequent queries, +this database is the default for table references that do not include an +explicit database specifier. -@item @uref{http://www.worldnetla.net, WorldNet Communications - An Internet Services Provider} +@code{mysql_select_db()} fails unless the connected user can be authenticated +as having permission to use the database. -@item @uref{http://www.netizen.com.au/, Netizen: Australian-based Web consultancy} +@subsubheading Return Values -@item @uref{http://www.trainingpages.co.uk, Search site for training courses in the UK} +Zero for success. Non-zero if an error occurred. -@item @uref{http://chat.nitco.com, Gannon Chat (GPL). Written in Perl and Javascript} +@subsubheading Errors -@item @uref{http://www.addurls.com/,A general links directory} +@table @code +@item CR_COMMANDS_OUT_OF_SYNC +Commands were executed in an improper order. +@item CR_SERVER_GONE_ERROR +The @strong{MySQL} server has gone away. +@item CR_SERVER_LOST +The connection to the server was lost during the query. +@item CR_UNKNOWN_ERROR +An unknown error occurred. +@end table -@item @uref{http://www.bookmarktracker.com, A Web-based bookmark management service} +@findex @code{mysql_shutdown()} +@node mysql_shutdown, mysql_stat, mysql_select_db, C API functions +@subsubsection @code{mysql_shutdown()} -@item @uref{http://www.cdrom.com,Walnut Creek CDROM} +@code{int mysql_shutdown(MYSQL *mysql)} -@item @uref{http://www.wwwthreads.org/, WWWThreads; Interactive discussion Forums} +@subsubheading Description -@item @uref{http://pvmon.portici.enea.it/Meteo, In Italian; Storage data from meteo station} +Asks the database server to shut down. The connected user must have +@strong{shutdown} privileges. -@item @uref{http://www.buysell.net/, Online "Person To Person" Auction} +@subsubheading Return Values -@item @uref{http://tips.pair.com,Tips on Web development} +Zero for success. Non-zero if an error occurred. -@item @uref{http://www.mailfriends.com, Mailfriends.com is a FREE service for -everybody who wants to find friends over the internet.} +@subsubheading Errors -@item @uref{http://www.uninova.com/cgi-bin/wctelnets?list, Web Page Telnet BBS List} +@table @code +@item CR_COMMANDS_OUT_OF_SYNC +Commands were executed in an improper order. +@item CR_SERVER_GONE_ERROR +The @strong{MySQL} server has gone away. +@item CR_SERVER_LOST +The connection to the server was lost during the query. +@item CR_UNKNOWN_ERROR +An unknown error occurred. +@end table -@item @uref{http://www.uninova.com/cnc.html,UniNova Digital Postcards} +@findex @code{mysql_stat()} +@node mysql_stat, mysql_store_result, mysql_shutdown, C API functions +@subsubsection @code{mysql_stat()} -@c @item @uref{http://cabinboy.powersurfr.com, An Internet RFC search engine} +@code{char *mysql_stat(MYSQL *mysql)} -@item @uref{http://www.dslreports.com, DSL-provider search with reviews}. -Made with @strong{MySQL} and Modperl, all pages are generated dynamically out of -the @strong{MySQL} database -@end itemize +@subsubheading Description -@cindex PHP, web sites -@appendixsec Web Sites that Use @code{PHP} and MySQL +Returns a character string containing information similar to that provided by +the @code{mysqladmin status} command. This includes uptime in seconds and +the number of running threads, questions, reloads, and open tables. -@itemize @bullet -@c @item @uref{http://www.wh200th.com, White House 200th Anniversary site} +@subsubheading Return Values -@item @uref{http://support.jgaa.com/, Jgaa's Internet - Official Support Site} +A character string describing the server status. @code{NULL} if an +error occurred. -@item @uref{http://io.incluso.com, Ionline - online publication:} @strong{MySQL}, -PHP, Java, Web programming, DB development +@subsubheading Errors -@item @uref{http://www.baboo.com, BaBoo(Browse and bookmark). Free Web-based bookmark manager and Calendar} +@table @code +@item CR_COMMANDS_OUT_OF_SYNC +Commands were executed in an improper order. +@item CR_SERVER_GONE_ERROR +The @strong{MySQL} server has gone away. +@item CR_SERVER_LOST +The connection to the server was lost during the query. +@item CR_UNKNOWN_ERROR +An unknown error occurred. +@end table -@item @uref{http://www.courses.pjc.cc.fl.us/Schedule/index.php, Course -Schedule System at Pensacola Junior College} +@findex @code{mysql_store_result()} +@node mysql_store_result, mysql_thread_id, mysql_stat, C API functions +@subsubsection @code{mysql_store_result()} -@item @uref{http://www.fccj.org, Florida Community College at Jacksonville} +@code{MYSQL_RES *mysql_store_result(MYSQL *mysql)} -@item @uref{http://www.32bit.com/, 32bit.com; An extensive shareware / freeware archive} +@subsubheading Description -@item @uref{http://www.jokes2000.com/, Jokes 2000} -@c Added 990604; EMAIL: ah@dybdahl.dk +You must call @code{mysql_store_result()} or @code{mysql_use_result()} +for every query that successfully retrieves data (@code{SELECT}, +@code{SHOW}, @code{DESCRIBE}, @code{EXPLAIN}). +You don't have to call @code{mysql_store_result()} or +@code{mysql_use_result()} for other queries, but it will not do any +harm or cause any notable performance if you call @code{mysql_store_result()} +in all cases. You can detect if the query didn't have a result set by +checking if @code{mysql_store_result()} returns 0 (more about this later one). -@item @uref{http://www.burken.nu/ , Burken.NU} Burken is a webhotel that -provides scripts, among other things, for remote users, like counters, -guestbooks etc. -@c Added 990608; EMAIL: spacedmp@SpaceDump.Burken.NU (Anders Olausson) +If you want to know if the query should return a result set or not, you can +use @code{mysql_field_count()} to check for this. +@xref{mysql_field_count, @code{mysql_field_count}}. -@item @uref{http://tips.pair.com, tips.pair.com} Contains tips on html, -javascript, 2d/3d graphics, and PHP3/MySQL. All pages are generated from -a database. -@c Added 990614; EMAIL: downey@image.dk (Rune Madsen) +@code{mysql_store_result()} reads the entire result of a query to the client, +allocates a @code{MYSQL_RES} structure, and places the result into this +structure. -@item @uref{http://www.softwarezrus.com/, Softwarezrus.com} -Ecommerce site that is selling computers. -@end itemize +@code{mysql_store_results()} returns a null pointer if the query didn't return +a result set (if the query was, for example, an @code{INSERT} statement). -@cindex consultants, list of -@appendixsec Some MySQL Consultants +@code{mysql_store_results()} also returns a null pointer if reading of the +result set failed. You can check if you got an error by checking if +@code{mysql_error()} doesn't return a null pointer, if +@code{mysql_errno()} returns <> 0, or if @code{mysql_field_count()} +returns <> 0. -@itemize @bullet +An empty result set is returned if there are no rows returned. (An empty +result set differs from a null pointer as a return value.) -@item @uref{http://www.ayni.com, Ayni AG} +Once you have called @code{mysql_store_result()} and got a result back +that isn't a null pointer, you may call @code{mysql_num_rows()} to find +out how many rows are in the result set. -@item @uref{http://worldcommunity.com/, Online Database} +You can call @code{mysql_fetch_row()} to fetch rows from the result set, +or @code{mysql_row_seek()} and @code{mysql_row_tell()} to obtain or +set the current row position within the result set. -@item @uref{http://www2.dataguard.no/,DataGuard (Uses @strong{MySQL} and PHP)} +You must call @code{mysql_free_result()} once you are done with the result +set. -@item @uref{http://wwits.net/programs/mysql.phtml, WWITS (Uses @strong{MySQL} and PHP)} +@xref{NULL mysql_store_result, , @code{NULL mysql_store_result()}}. -@item @uref{http://www.worldcommunity.com/, WCN - The World Community Network} +@subsubheading Return Values -@item @uref{http://www.chipcastle.com, Chip Castle Dot Com Inc} -@c Added 990603 EMAIL: chip@chipcastle.com (Chip Castle) +A @code{MYSQL_RES} result structure with the results. @code{NULL} if +an error occurred. -@item @uref{http://www.cyber.com.au/, Cybersource Pty. Ltd} +@subsubheading Errors -@item @uref{http://www.spring.de, Spring infotainment gmbh & co. kg} -@c added 990905 "Oliver Pischke" <opischke@spring.de> +@table @code +@item CR_COMMANDS_OUT_OF_SYNC +Commands were executed in an improper order. +@item CR_OUT_OF_MEMORY +Out of memory. +@item CR_SERVER_GONE_ERROR +The @strong{MySQL} server has gone away. +@item CR_SERVER_LOST +The connection to the server was lost during the query. +@item CR_UNKNOWN_ERROR +An unknown error occurred. +@end table -@item @uref{http://www.wamdesign.com/, Develops Web sites using MySQL} -@c Added 990905; max@wamdesign.com +@findex @code{mysql_thread_id()} +@node mysql_thread_id, mysql_use_result, mysql_store_result, C API functions +@subsubsection @code{mysql_thread_id()} -@item @uref{http://www.berkeleyconsultants.com, Berkeley Consultants Group} +@code{unsigned long mysql_thread_id(MYSQL *mysql)} -@item @uref{http://www.jammconsulting.com/, JAMM Consulting Inc.} +@subsubheading Description -@end itemize +Returns the thread ID of the current connection. This value can be used as +an argument to @code{mysql_kill()} to kill the thread. -@appendixsec Programming +If the connection is lost and you reconnect with @code{mysql_ping()}, the +thread ID will change. This means you should not get the thread ID and store +it for later. You should get it when you need it. -@cindex web pages, miscellaneous -@appendixsec Uncategorized Pages +@subsubheading Return Values -@itemize @bullet +The thread ID of the current connection. -@item @uref{http://www.feature-showcase.com/htmls/demo_mysql.sql, -AZC.COM's Feature Showcase} +@subsubheading Errors +None. -@item @uref{http://www.teach.org.uk/subjects/trainingcourse/g.html, Course Search} +@findex @code{mysql_use_result()} +@node mysql_use_result, , mysql_thread_id, C API functions +@subsubsection @code{mysql_use_result()} -@item @uref{http://www.northerbys.com, Northerbys Online Auctions} +@code{MYSQL_RES *mysql_use_result(MYSQL *mysql)} -@item @uref{http://www.schiphol.nl/flights/home.htm, Amsterdam Airport Schiphol} +@subsubheading Description -@item @uref{http://TheMatrix.com/seventhsin/query.phtml, CD database} +You must call @code{mysql_store_result()} or @code{mysql_use_result()} for +every query that successfully retrieves data (@code{SELECT}, @code{SHOW}, +@code{DESCRIBE}, @code{EXPLAIN}). -@item @uref{http://TheMatrix.com/~flmm/GEAR.html, Used Audio Gear Database} +@code{mysql_use_result()} initiates a result set retrieval but does not +actually read the result set into the client like @code{mysql_store_result()} +does. Instead, each row must be retrieved individually by making calls to +@code{mysql_fetch_row()}. This reads the result of a query directly from the +server without storing it in a temporary table or local buffer, which is +somewhat faster and uses much less memory than @code{mysql_store_result()}. +The client will only allocate memory for the current row and a communication +buffer that may grow up to @code{max_allowed_packet} bytes. -@item @uref{http://www.kiss.de/musik-mueller, Musical note-sheets} +On the other hand, you shouldn't use @code{mysql_use_result()} if you are +doing a lot of processing for each row on the client side, or if the output +is sent to a screen on which the user may type a @code{^S} (stop scroll). +This will tie up the server and prevent other threads from updating any +tables from which the data is being fetched. -@item @uref{http://www.bagism.com, Bagism - A John Lennon fan page} +When using @code{mysql_use_result()}, you must execute +@code{mysql_fetch_row()} until a @code{NULL} value is returned, otherwise the +unfetched rows will be returned as part of the result set for your next +query. The C API will give the error @code{Commands out of sync; You can't +run this command now} if you forget to do this! -@item @uref{http://www.selftaught.com/, US Folk art broker} +You may not use @code{mysql_data_seek()}, @code{mysql_row_seek()}, +@code{mysql_row_tell()}, @code{mysql_num_rows()}, or +@code{mysql_affected_rows()} with a result returned from +@code{mysql_use_result()}, nor may you issue other queries until the +@code{mysql_use_result()} has finished. (However, after you have fetched all +the rows, @code{mysql_num_rows()} will accurately return the number of rows +fetched.) -@item @uref{http://organizer.net/, Mail reading on the Web} +You must call @code{mysql_free_result()} once you are done with the result +set. -@item @uref{http://www.mypage.org/, Free home pages on www.somecoolname.mypage.org} +@subsubheading Return Values -@item @uref{http://www.schulweb.de/, Der Server f@"ur Schulen im Web (In German)} +A @code{MYSQL_RES} result structure. @code{NULL} if an error occurred. -@item @uref{http://www.ald.net/, Auldhaefen Online Services} +@subsubheading Errors -@item @uref{http://www.cary.net/, CaryNET Information Center} +@table @code +@item CR_COMMANDS_OUT_OF_SYNC +Commands were executed in an improper order. +@item CR_OUT_OF_MEMORY +Out of memory. +@item CR_SERVER_GONE_ERROR +The @strong{MySQL} server has gone away. +@item CR_SERVER_LOST +The connection to the server was lost during the query. +@item CR_UNKNOWN_ERROR +An unknown error occurred. +@end table -@item @uref{http://www.dataden.com/, Dataden Computer Systems} +@node C API problems, Thread-safe clients, C API functions, C +@subsection Common questions and problems when using the C API -@item @uref{http://andree.grm.se/, Andr@'emuseet (In Swedish)} +@tindex @code{mysql_query()} +@tindex @code{mysql_store_result()} +@menu +* NULL mysql_store_result:: +* Query results:: +* Getting unique ID:: +* C API linking problems:: +@end menu -@item @uref{http://www.him.net/, HOMESITE Internet Marketing} +@node NULL mysql_store_result, Query results, C API problems, C API problems +@subsubsection Why Is It that After @code{mysql_query()} Returns Success, @code{mysql_store_result()} Sometimes Returns @code{NULL?} -@item @uref{http://www.jade-v.com/techinfo.html, Jade-V Network Services } +It is possible for @code{mysql_store_result()} to return @code{NULL} +following a successful call to @code{mysql_query()}. When this happens, it +means one of the following conditions occurred: -@item @uref{http://ww2010.atmos.uiuc.edu/(Gl)/abt/aknw/tech.rxml, -Weather World 2010 Technical Credits} @* - -@item @uref{http://gimp.foebud.org/registry/doc/, About The Gimp plugin registry} - -@item @uref{http://www.fast-inc.com/Products/Archiver/database.html, Java tool -Archiver technical detail (Slightly optimistic about @strong{MySQL} -ANSI-92 compliance)} - -@item @uref{http://www.gamesdomain.com/cheats/usrcheat.phtml, Games Domain Cheats Database} - -@item @uref{http://www.kcilink.com/poweredby/, The "Powered By" Page (Kcilink)} - -@item @uref{http://www.netcasting.net/index.whtml, Netcasting} - -@item @uref{http://homepages.tig.com.au/~mjj/nbltips, NBL (Australian National Basketball League) tipping} - -@item @uref{http://www.cgishop.com/, CGI shop} - -@item @uref{http://www.whirlycott.com/, Whirlycott: Website Design} - -@item @uref{http://www.mtp.dk, Museum Tusculanum Press} +@itemize @bullet +@item +There was a @code{malloc()} failure (for example, if the result set was too +large). -@item @uref{http://csdgi.historie.ku.dk/biblio, Centro Siciliano di Documentazione} +@item +The data couldn't be read (an error occurred on the connection). -@item @uref{http://caribou.dyn.ml.org:8000, Quake statistics database} +@item +The query returned no data (for example, it was an @code{INSERT}, +@code{UPDATE}, or @code{DELETE}). +@end itemize -@item @uref{http://www.astroforum.ch, Astroforum: Astrologie and related things (in German)} +You can always check whether or not the statement should have produced a +non-empty result by calling @code{mysql_field_count()}. If +@code{mysql_field_count()} returns zero, the result is empty and the last +query was a statement that does not return values (for example, an +@code{INSERT} or a @code{DELETE}). If @code{mysql_field_count()} returns a +non-zero value, the statement should have produced a non-empty result. +See the description of the @code{mysql_field_count()} function for an +example. -@item @uref{http://www.opendebate.com, OpenDebate - Interactive Polls & Open Discussion} +You can test for an error by calling @code{mysql_error()} or +@code{mysql_errno()}. -@item @uref{http://vermeer.organik.uni-erlangen.de/dissertationen/, Online chemical dissertation server} +@cindex queries, C API results +@menu +* Query results:: +* Getting unique ID:: +* C API linking problems:: +@end menu -@item @uref{http://www.freschinfo.com, FreSch! The Free Scholarship Search Service} +@node Query results, Getting unique ID, NULL mysql_store_result, C API problems +@subsubsection What Results Can I Get From a Query? -@item @uref{http://www.nada.kth.se/~staffanu/pinball, Stockholm Pinball Locator} +In addition to the result set returned by a query, you can also get the +following information: -@item @uref{http://www.hek.com, HEK A construction company} +@itemize @bullet +@item +@code{mysql_affected_rows()} returns the number of rows affected by the last +query when doing an @code{INSERT}, @code{UPDATE}, or @code{DELETE}. An +exception is that if @code{DELETE} is used without a @code{WHERE} clause, the +table is re-created empty, which is much faster! In this case, +@code{mysql_affected_rows()} returns zero for the number of records +affected. -@item @uref{http://www.nbi.nl, Elsevier Bussines Information} +@item +@code{mysql_num_rows()} returns the number of rows in a result set. With +@code{mysql_store_result()}, @code{mysql_num_rows()} may be called as soon as +@code{mysql_store_result()} returns. With @code{mysql_use_result()}, +@code{mysql_num_rows()} may be called only after you have fetched all the +rows with @code{mysql_fetch_row()}. -@item @uref{http://vaccination.medicallink.se/, Medical Links (Using ColdFusion and @strong{MySQL})} +@item +@code{mysql_insert_id()} returns the ID generated by the last +query that inserted a row into a table with an @code{AUTO_INCREMENT} index. +@xref{mysql_insert_id, , @code{mysql_insert_id()}}. -@item @uref{http://www.joblink-usa.com, Search for jobs & people at JobLink-USA} +@item +Some queries (@code{LOAD DATA INFILE ...}, @code{INSERT INTO +... SELECT ...}, @code{UPDATE}) return additional information. The result is +returned by @code{mysql_info()}. See the description for @code{mysql_info()} +for the format of the string that it returns. @code{mysql_info()} returns a +@code{NULL} pointer if there is no additional information. +@end itemize -@item @uref{http://www.skydive.net/competfs, Competition Formation Skydiving} +@cindex unique ID +@cindex last row, unique ID +@cindex ID, unique +@cindex tables, unique ID for last row +@node Getting unique ID, C API linking problems, Query results, C API problems +@subsubsection How Can I Get the Unique ID for the Last Inserted Row? -@item @uref{http://www.galaxy-net.net/Galaxy-NET Telecommunications, E-commerce and internal accounting} +If you insert a record in a table containing a column that has the +@code{AUTO_INCREMENT} attribute, you can get the most recently generated +ID by calling the @code{mysql_insert_id()} function. -@item @uref{http://www.borsen.dk/, Denmark's leading business daily newspaper B@o{}rsen} +You can also retrieve the ID by using the @code{LAST_INSERT_ID()} function in +a query string that you pass to @code{mysql_query()}. -@item @uref{http://tmmm.simplenet.com/indb/, The Internet NES Database} +You can check if an @code{AUTO_INCREMENT} index is used by executing +the following code. This also checks if the query was an @code{INSERT} with +an @code{AUTO_INCREMENT} index: -@item @uref{http://www.russia.cz, Travel agency in Prague in 3 languages} +@example +if (mysql_error(&mysql)[0] == 0 && + mysql_num_fields(result) == 0 && + mysql_insert_id(&mysql) != 0) +@{ + used_id = mysql_insert_id(&mysql); +@} +@end example -@item @uref{http://www.linkstation.de, Linkstation} +The most recently generated ID is maintained in the server on a +per-connection basis. It will not be changed by another client. It will not +even be changed if you update another @code{AUTO_INCREMENT} column with a +non-magic value (that is, a value that is not @code{NULL} and not @code{0}). -@item @uref{http://www.peoplestaff.com, Searchable online database at Peoplestaff} +If you want to use the ID that was generated for one table and insert +it into a second table, you can use SQL statements like this: -@item @uref{http://www.dreamhorse.com, A searchable database system for horse classified ads} +@example +INSERT INTO foo (auto,text) + VALUES(NULL,'text'); # generate ID by inserting NULL +INSERT INTO foo2 (id,text) + VALUES(LAST_INSERT_ID(),'text'); # use ID in second table +@end example -@item @uref{http://pootpoot.com/,The Poot site} +@cindex linking, problems +@cindex C API, linking problems +@node C API linking problems, , Getting unique ID, C API problems +@subsubsection Problems Linking with the C API -@item @uref{http://grateful.net/hw_html/,"Playin' in the LAN"; a network monitoring suite} +When linking with the C API, the following errors may occur on some systems: -@c Update from Christopher Milton <cmilton@bwn.net> 1999-12-21 -@item @uref{http://www.usapa.army.mil,U.S. Army Publishing Agency} +@example +gcc -g -o client test.o -L/usr/local/lib/mysql -lmysqlclient -lsocket -lnsl -@item @uref{http://www.nekretnine.co.yu/,Realestate handling in Yugoslavia} +Undefined first referenced + symbol in file +floor /usr/local/lib/mysql/libmysqlclient.a(password.o) +ld: fatal: Symbol referencing errors. No output written to client +@end example -@item @uref{http://demo.cpsoft.com/pims/devFAQ.html, PIMS; a Patient Information Management System} +If this happens on your system, you must include the math library by +adding @code{-lm} to the end of the compile/link line. -@item @uref{http://cpsoft.com,Pilkington Software Inc} +@cindex clients, thread-safe +@cindex thread-safe clients +@node Thread-safe clients, , C API problems, C +@subsection How to Make a Thread-safe Client -@item @uref{http://www.no-quarter.org/,A Vietnam Veteran's Memorial (The Wall) database} +The client library is almost thread safe. The biggest problem is +that the subroutines in @file{net.c} that read from sockets are not +interrupt safe. This was done with the thought that you might want to +have your own alarm that can break a long read to a server. If you +install interrupt handlers for the @code{SIGPIPE} interrupt, +the socket handling should be thread safe. -@item @uref{http://www.gamers-union.com/,Gamer's Union specializes in auctions of used & out-of-print gaming material} +In the older binaries we distribute on our Web site, the client +libraries are not normally compiled with the thread-safe option (the +Windows binaries are by default compiled to be thread safe). +Newer binary distributions should have both a normal and a +thread-safe client library. -@item @uref{http://www.montereyhigh.com/office/dbul.php3, A daily bulletin at Monterey High school} +To get a really thread-safe client where you can interrupt the client +from other threads and set timeouts when talking with the @strong{MySQL} +server, you should use the @code{-lmysys}, @code{-lstring}, and @code{-ldbug} +libraries and the @code{net_serv.o} code that the server uses. -@item @uref{http://www.myEastside.com,Community-owned site serving Lake -Washington's Eastside residents and businesses} +If you don't need interrupts or timeouts, you can just compile a thread +safe client library @code{(mysqlclient_r)} and use this. @xref{C,, +MySQL C API}. In this case you don't have to worry about the +@code{net_serv.o} object file or the other @strong{MySQL} libraries. -@item @uref{http://bowling-france.net/,French bowling site} -@end itemize +When using a threaded client and you want to use timeouts and interrupts, +you can make great use of the routines in the @file{thr_alarm.c} file. +If you are using routines from the @code{mysys} library, the only thing +you must remember is to call @code{my_init()} first! -Send any additions to this list to @email{webmaster@@mysql.com}. +All functions except @code{mysql_real_connect()} are by default +thread safe. The following notes describe how to compile a thread safe +client library and use it in a thread-safe manner. (The notes below for +@code{mysql_real_connect()} actually apply to @code{mysql_connect()} as +well, but because @code{mysql_connect()} is deprecated, you should be +using @code{mysql_real_connect()} anyway.) -@page -@cindex MySQL usage -@node MySQL customer usage, Contrib, Users, Top -@appendix MySQL customer usage +To make @code{mysql_real_connect()} thread safe, you must recompile the +client library with this command: -The section 'Some MySQL Users' contains a lot of different links to -@strong{MySQL} users but doesn't provide that much information about how -they are using @strong{MySQL}. @xref{Users}. This manual section is to -give you an idea of how other @strong{MySQL} users are using -@strong{MySQL} to solve their problems. +@example +shell> ./configure --with-thread-safe-client +@end example -This manual section is very new and we plan to add more stories here -shortly. If you are interested in contributing of how you use -@strong{MySQL} in a unique environment or have success store about how you -use @strong{MySQL}, you can write to @code{docs@@mysql.com} with -subject @code{Success:}. Note that as we are very busy it may take some -time before you get some feedback for your story. +This will create a thread-safe client library @code{libmysqlclient_r}. +@code{--with-thread-safe-client}. This library is thread safe per +connection. You can let two threads share the same connection as long +as you do the following: @itemize @bullet -@item -@strong{Peter Zaitsev of Spylog.ru} writes: -I think you might be interested in my database size. The whole database -is currently on 15 servers and I think it's about 60.000 of tables -containing about 5.000.000.000 of rows. My mostly loaded server -currently holds about 10.000 of tables with 1.000.000.000 of rows in it. -Hugest tables have about 50.000.000 of rows, and this value will raise -as soon as I'll move to 2.4 kernel with large files. Currently I have to -delete much of logs for large sites to hold table sizes in 2Gb. - @item -@strong{Texas Instruments} is using @strong{MySQL} for handling tables -that contain up to 2,000 million rows in a validation regression database. +Two threads can't send a query to the @strong{MySQL} at the same time on +the same connection. In particular, you have to ensure that between a +@code{mysql_query()} and @code{mysql_store_result()} no other thread is using +the same connection. +@item +Many threads can access different result sets that are retrieved with +@code{mysql_store_result()}. +@item +If you use @code{mysql_use_result}, you have to ensure that no other thread +is asking anything on the same connection until the result set is closed. +However, it really is best for threaded clients that share the same +connection to use @code{mysql_use_result()}. +@item +If you want to use multiple threads on the same connection, you must +have a mutex lock around your @code{mysql_query()} and +@code{mysql_store_result()} call combination. Once +@code{mysql_store_result()} is ready, the lock can be released and other +threads may query the same connection. +@item +If you program with POSIX threads, you can use +@code{pthread_mutex_lock()} and @code{pthread_mutex_unlock()} to +establish and release a mutex lock. @end itemize -@page -@cindex contributed programs -@cindex programs, contributed -@node Contrib, Credits, MySQL customer usage, Top -@appendix Contributed Programs - -Many users of @strong{MySQL} have contributed @emph{very} useful support -tools and add-ons. +You may get some errors because of undefined symbols when linking your +client with @code{mysqlclient_r}. In most cases this is because you haven't +included the thread libraries on the link/compile line. -@ifclear web -A list of what is available at @uref{http://www.mysql.com/Downloads/Contrib/} -(or any mirror) is shown below. -If you want to build @strong{MySQL} support for the Perl @code{DBI}/@code{DBD} -interface, you should fetch the @code{Data-Dumper}, @code{DBI}, and -@code{Msql-Mysql-modules} files and install them. -@xref{Perl support}. -@end ifclear +@cindex APIs, Perl +@cindex Perl API +@node Perl, Eiffel, C, Clients +@section MySQL Perl API -@appendixsec APIs +This section documents the Perl @code{DBI} interface. The former interface +was called @code{mysqlperl}. @code{DBI}/@code{DBD} now is the +recommended Perl interface, so @code{mysqlperl} is obsolete and is not +documented here. -@cindex Perl, modules -@itemize @bullet -@item Perl Modules -@itemize @minus -@item @uref{http://www.mysql.com/Downloads/Contrib/Data-Dumper-2.101.tar.gz, Data-Dumper-2.101.tar.gz} -Perl @code{Data-Dumper} module. Useful with @code{DBI}/@code{DBD} support for -older Perl installations. +@menu +* DBI with DBD:: @code{DBI} with @code{DBD::mysql} +* Perl DBI Class:: The @code{DBI} interface +* DBI-info:: More @code{DBI}/@code{DBD} information +@end menu -@item @uref{http://www.mysql.com/Downloads/Contrib/DBI-1.15.tar.gz, DBI-1.15.tar.gz} -Perl @code{DBI} module. +@cindex @code{DBI} interface +@node DBI with DBD, Perl DBI Class, Perl, Perl +@subsection @code{DBI} with @code{DBD::mysql} -@item @uref{http://www.mysql.com/Downloads/Contrib/KAMXbase1.2.tar.gz,KAMXbase1.2.tar.gz} -Convert between @file{.dbf} files and @strong{MySQL} tables. Perl -module written by Pratap Pereira @email{pereira@@ee.eng.ohio-state.edu}, -extended by Kevin A. McGrail @email{kmcgrail@@digital1.peregrinehw.com}. -This converter can handle MEMO fields. +@code{DBI} is a generic interface for many databases. That means that +you can write a script that works with many different database engines +without change. You need a DataBase Driver (DBD) defined for each +database type. For @strong{MySQL}, this driver is called +@code{DBD::mysql}. -@item @uref{http://www.mysql.com/Downloads/Contrib/Msql-Mysql-modules-1.2216.tar.gz, Msql-Mysql-modules-1.2216.tar.gz} -Perl @code{DBD} module to access mSQL and @strong{MySQL} databases. +For more information on the Perl5 DBI, please visit the @code{DBI} Web +page and read the documentation: +@example +@uref{http://www.symbolstone.org/technology/perl/DBI/index.html} +@end example +For more information on Object Oriented Programming +(OOP) as defined in Perl5, see the Perl OOP page: +@example +@uref{http://language.perl.com/info/documentation.html} +@end example -@item @uref{http://www.mysql.com/Downloads/Contrib/Data-ShowTable-3.3.tar.gz, Data-ShowTable-3.3.tar.gz} -Perl @code{Data-ShowTable} module. Useful with @code{DBI}/@code{DBD} support. +Note that if you want to use transactions with Perl, you need to have +@code{Msql-Mysql-modules} version 1.2216 or newer. -@item @uref{http://www.mysql.com/Downloads/Contrib/HandySQL-1.1.tar.gz, HandySQL-1.1.tar.gz} -HandySQL is a MySQL access module. It offers a C interface embedded in Perl and is -approximately 20% faster than regular DBI. +Installation instructions for @strong{MySQL} Perl support are given in +@ref{Perl support}. -@end itemize +@cindex @code{DBI} Perl module +@node Perl DBI Class, DBI-info, DBI with DBD, Perl +@subsection The @code{DBI} Interface -@cindex JDBC -@item JDBC -@itemize @bullet -@item @uref{http://www.mysql.com/Downloads/Contrib/mm.mysql.jdbc-1.2c.tar.gz, mm.mysql.jdbc-1.2c.tar.gz} -The mm JDBC driver for @strong{MySQL}. This is a production release -and is actively developed. By Mark Matthews -(@email{mmatthew@@ecn.purdue.edu}). +@noindent +@strong{Portable DBI Methods} -@item @uref{http://www.mysql.com/Downloads/Contrib/mm.mysql.jdbc-2.0pre5.tar.gz, mm.mysql.jdbc-2.0pre5.tar.gz} -The mm JDBC driver for @strong{MySQL}. This is a pre-release beta version -and is actively developed. By Mark Matthews -(@email{mmatthew@@ecn.purdue.edu}). -The two drivers above have an LGPL -license. Please check @uref{http://www.worldserver.com/mm.mysql/} for -the latest drivers (and other JDBC information) because these drivers may be out of date. +@multitable @columnfractions .3 .7 +@item @code{connect} @tab Establishes a connection to a database server. +@item @code{disconnect} @tab Disconnects from the database server. +@item @code{prepare} @tab Prepares a SQL statement for execution. +@item @code{execute} @tab Executes prepared statements. +@item @code{do} @tab Prepares and executes a SQL statement. +@item @code{quote} @tab Quotes string or @code{BLOB} values to be inserted. +@item @code{fetchrow_array} @tab Fetches the next row as an array of fields. +@item @code{fetchrow_arrayref} @tab Fetches next row as a reference array of fields. +@item @code{fetchrow_hashref} @tab Fetches next row as a reference to a hashtable. +@item @code{fetchall_arrayref} @tab Fetches all data as an array of arrays. +@item @code{finish} @tab Finishes a statement and lets the system free resources. +@item @code{rows} @tab Returns the number of rows affected. +@item @code{data_sources} @tab Returns an array of databases available on localhost. +@item @code{ChopBlanks} @tab Controls whether @code{fetchrow_*} methods trim spaces. +@item @code{NUM_OF_PARAMS} @tab The number of placeholders in the prepared statement. +@item @code{NULLABLE} @tab Which columns can be @code{NULL}. +@item @code{trace} @tab Perform tracing for debugging. +@end multitable -@item @uref{http://www.caucho.com/projects/jdbc-mysql/index.xtp} -The Resin commercial JDBC driver, which is released under open source. -It claims to be faster than the mm driver, but we haven't received that much -information about this yet. +@noindent +@strong{MySQL-specific Methods} -@item @uref{http://www.mysql.com/Downloads/Contrib/twz1jdbcForMysql-1.0.4-GA.tar.gz, twz1jdbcForMysql-1.0.4-GA.tar.gz} -The twz driver: A type 4 JDBC driver by Terrence W. Zellers -@email{zellert@@voicenet.com}. This is commercial but is free for -private and educational use. (Not supported anymore.) -@c no answer from server 990830 -@c You can always find the latest driver at @uref{http://www.voicenet.com/~zellert/tjFM/}. -@item @uref{http://www.mysql.com/Downloads/Contrib/pmdamysql.tgz,pmdamysql.tgz} -A @strong{MySQL} PMDA. Provides @strong{MySQL} server status and configuration -variables. -@end itemize +@multitable @columnfractions .3 .7 +@item @code{insertid} @tab The latest @code{AUTO_INCREMENT} value. +@item @code{is_blob} @tab Which columns are @code{BLOB} values. +@item @code{is_key} @tab Which columns are keys. +@item @code{is_num} @tab Which columns are numeric. +@item @code{is_pri_key} @tab Which columns are primary keys. +@item @code{is_not_null} @tab Which columns CANNOT be @code{NULL}. See @code{NULLABLE}. +@item @code{length} @tab Maximum possible column sizes. +@item @code{max_length} @tab Maximum column sizes actually present in result. +@item @code{NAME} @tab Column names. +@item @code{NUM_OF_FIELDS} @tab Number of fields returned. +@item @code{table} @tab Table names in returned set. +@item @code{type} @tab All column types. +@end multitable -@cindex OLEDB -@item OLEDB -@itemize @bullet -@item @uref{http://www.mysql.com/Downloads/Win32/MyOLEDB.exe, MyOLEDB.exe} -OLEDB handler for @strong{MySQL}. By SWsoft. -@item @uref{http://www.mysql.com/Downloads/Win32/MySamples.zip, MySamples.zip} -Examples and documentation for MyOLEDB. By SWsoft. -@item @uref{http://www.mysql.com/Downloads/Win32/Myoledb.zip, Myoledb.zip} -Source for MyOLEDB. By SWsoft. -@item @uref{http://www.mysql.com/Downloads/Win32/MyOLEDB.chm, MyOLEDB.chm} -Help files for MyOLEDB. -@item @uref{http://www.mysql.com/Downloads/Win32/libmyodbc.zip, libmyodbc.zip} -Static MyODBC library used for build MyOLEDB. Based on MyODBC code. -@end itemize +The Perl methods are described in more detail in the following sections. +Variables used for method return values have these meanings: -@cindex C++ -@item C++ -@itemize @bullet +@table @code +@item $dbh +Database handle -@item @uref{http://www.mysql.com/Downloads/Contrib/mysql-c++-0.02.tar.gz, mysql-c++-0.02.tar.gz} -@strong{MySQL} C++ wrapper library. By Roland Haenel, -@email{rh@@ginster.net}. +@item $sth +Statement handle -@item @uref{http://www.mysql.com/Downloads/Contrib/MyDAO.tar.gz, MyDAO} -@strong{MySQL} C++ API. By Satish @email{spitfire@@pn3.vsnl.net.in}. Inspired -by Roland Haenel's C++ API and Ed Carp's MyC library. +@item $rc +Return code (often a status) -@item @uref{http://www.mysql.com/download_mysql++.html, mysql++} -@strong{MySQL} C++ API (More than just a wrapper library.) Originally by -@email{kevina@@clark.net}. Nowadays maintained by Sinisa at @strong{MySQL AB}. +@item $rv +Return value (often a row count) +@end table -@item @uref{http://nelsonjr.homepage.com/NJrAPI,NJrAPI} -A C++ database independent library that supports @strong{MySQL}. -@end itemize +@noindent +@strong{Portable DBI Methods} -@cindex Delphi -@item Delphi -@itemize @bullet -@item @uref{http://www.mysql.com/Downloads/Contrib/delphi-interface.gz, delphi-interface.gz} -Delphi interface to @code{libmysql.dll}, by Blestan Tabakov, -@email{root@@tdg.bis.bg}. +@table @code -@item @uref{http://www.mysql.com/Downloads/Contrib/DelphiMySQL2.zip, DelphiMySQL2.zip} -Delphi interface to @code{libmysql.dll}, by @email{bsilva@@umesd.k12.or.us}. +@findex DBI->connect() +@findex connect() DBI method +@item connect($data_source, $username, $password) +Use the @code{connect} method to make a database connection to the data +source. The @code{$data_source} value should begin with +@code{DBI:driver_name:}. +Example uses of @code{connect} with the @code{DBD::mysql} driver: +@example +$dbh = DBI->connect("DBI:mysql:$database", $user, $password); +$dbh = DBI->connect("DBI:mysql:$database:$hostname", + $user, $password); +$dbh = DBI->connect("DBI:mysql:$database:$hostname:$port", + $user, $password); +@end example +If the user name and/or password are undefined, @code{DBI} uses the +values of the @code{DBI_USER} and @code{DBI_PASS} environment variables, +respectively. If you don't specify a hostname, it defaults to +@code{'localhost'}. If you don't specify a port number, it defaults to the +default @strong{MySQL} port (@value{default_port}). -@item @uref{http://www.mysql.com/Downloads/Contrib/Udmysel.pas, Udmysql.pas} -A wrapper for libmysql.dll for usage in Delphi. By Reiner Sombrowsky. +As of @code{Msql-Mysql-modules} Version 1.2009, +the @code{$data_source} value allows certain modifiers: -@item @uref{http://www.fichtner.net/delphi/mysql.delphi.phtml, A Delphi interface to @strong{MySQL}.} -With source code. By Matthias Fichtner. +@table @code +@item mysql_read_default_file=file_name +Read @file{filename} as an option file. For information on option files, +see @ref{Option files}. -@item @uref{http://www.productivity.org/projects/mysql/, @strong{TmySQL} -A library to use @strong{MySQL} with Delphi}. +@item mysql_read_default_group=group_name +The default group when reading an option file is normally the +@code{[client]} group. By specifying the @code{mysql_read_default_group} +option, the default group becomes the @code{[group_name]} group. -@item @uref{http://www.geocities.com/CapeCanaveral/2064/mysql.html, Delphi TDataset-component}. -@item -@item @uref{http://www.mysql.com/Downloads/Contrib/Win32/SBMySQL50Share.exe, Delphi 5 Shareware MySQL Dataset Components} -@end itemize +@item mysql_compression=1 +Use compressed communication between the client and server (@strong{MySQL} +Version 3.22.3 or later). -@item @uref{http://www.mysql.com/Downloads/Contrib/mysql-ruby-2.2.0.tar.gz, mysql-ruby-2.2.0.tar.gz} -@strong{MySQL} Ruby module. By TOMITA Masahiro @email{tommy@@tmtm.org} -@uref{http://www.netlab.co.jp/ruby/. Ruby} is an Object-Oriented Interpreter Language. +@item mysql_socket=/path/to/socket +Specify the pathname of the Unix socket that is used to connect +to the server (@strong{MySQL} Version 3.21.15 or later). +@end table -@item @uref{http://www.mysql.com/Downloads/Contrib/JdmMysqlDriver-0.1.0.tar.gz,JdmMysqlDriver-0.1.0.tar.gz} -A VisualWorks 3.0 Smalltalk driver for @strong{MySQL}. By -@email{joshmiller@@earthlink.net}. +Multiple modifiers may be given; each must be preceded by a semicolon. -@item @uref{http://www.mysql.com/Downloads/Contrib/Db.py, Db.py} -Python module with caching. By @email{gandalf@@rosmail.com}. +For example, if you want to avoid hardcoding the user name and password into +a @code{DBI} script, you can take them from the user's @file{~/.my.cnf} +option file instead by writing your @code{connect} call like this: -@item @uref{http://www.mysql.com/Downloads/Contrib/MySQLmodule-1.4.tar.gz, MySQLmodule-1.4.tar.gz} -Python interface for @strong{MySQL}. By Joseph Skinner @email{joe@@earthlight.co.nz}. Modified by Joerg Senekowitsch @email{senekow@@ibm.net}. +@example +$dbh = DBI->connect("DBI:mysql:$database" + . ";mysql_read_default_file=$ENV@{HOME@}/.my.cnf", + $user, $password); +@end example -@item @uref{http://www.mysql.com/Downloads/Contrib/MySQL-python-0.3.0.tar.gz, MySQL-python-0.3.0.tar.gz} -MySQLdb Python is an DB-API v2.0-compliant interface to MySQL. Transactions -are supported if the server and tables support them. It is -thread-safe, and contains a compatibility module for older code -written for the no-longer-maintained MySQLmodule interface. +This call will read options defined for the @code{[client]} group in the +option file. If you wanted to do the same thing but use options specified +for the @code{[perl]} group as well, you could use this: -@item @uref{http://www.mysql.com/Downloads/Contrib/mysql_mex_12.tar.gz, mysql_mex_1_12.tar.gz} -An interface program for the Matlab program by MathWorks. The interface -is done by Kimmo Uutela and John Fisher (not by Mathworks). -Check @uref{http://boojum.hut.fi/~kuutela/mysqlmex.html,mysqlmex.html} -for more information. +@example +$dbh = DBI->connect("DBI:mysql:$database" + . ";mysql_read_default_file=$ENV@{HOME@}/.my.cnf" + . ";mysql_read_default_group=perl", + $user, $password); +@end example -@item @uref{http://www.mysql.com/Downloads/Contrib/mysqltcl-1.53.tar.gz, mysqltcl-1.53.tar.gz} -Tcl interface for @strong{MySQL}. Based on @file{msqltcl-1.50.tar.gz}. -Updated by Tobias Ritzau, @email{tobri@@ida.liu.se}. +@findex DBI->disconnect +@findex disconnect DBI method +@item disconnect +The @code{disconnect} method disconnects the database handle from the database. +This is typically called right before you exit from the program. +Example: +@example +$rc = $dbh->disconnect; +@end example -@item @uref{http://www.mysql.com/Downloads/Contrib/MyC-0.1.tar.gz, MyC-0.1.tar.gz} -A Visual Basic-like API, by Ed Carp. +@findex DBI->prepare() +@findex prepare() DBI method +@item prepare($statement) +Prepares a SQL statement for execution by the database engine +and returns a statement handle @code{($sth)}, which you can use to invoke +the @code{execute} method. +Typically you handle @code{SELECT} statements (and @code{SELECT}-like statements +such as @code{SHOW}, @code{DESCRIBE}, and @code{EXPLAIN}) by means of +@code{prepare} and @code{execute}. +Example: +@example +$sth = $dbh->prepare($statement) + or die "Can't prepare $statement: $dbh->errstr\n"; +@end example -@item @uref{http://www.mysql.com/Downloads/Contrib/sqlscreens-1.0.1.tar.gz, sqlscreens-1.0.1.tar.gz} -Tcl/Tk code to generate database screens. By Jean-Francois Dockes. +@findex DBI->execute +@findex execute DBI method +@item execute +The @code{execute} method executes a prepared statement. For +non-@code{SELECT} statements, @code{execute} returns the number of rows +affected. If no rows are affected, @code{execute} returns @code{"0E0"}, +which Perl treats as zero but regards as true. If an error occurs, +@code{execute} returns @code{undef}. For @code{SELECT} statements, +@code{execute} only starts the SQL query in the database; you need to use one +of the @code{fetch_*} methods described below to retrieve the data. +Example: +@example +$rv = $sth->execute + or die "can't execute the query: $sth->errstr; +@end example -@item @uref{http://www.mysql.com/Downloads/Contrib/Vdb-dflts-2.1.tar.gz, Vdb-dflts-2.1.tar.gz} -This is a new version of a set of library utilities intended -to provide a generic interface to SQL database engines such that your -application becomes a 3-tiered application. The advantage is that you -can easily switch between and move to other database engines by -implementing one file for the new backend without making any -changes to your applications. By @email{damian@@cablenet.net}. +@findex DBI->do() +@findex do() DBI method +@item do($statement) +The @code{do} method prepares and executes a SQL statement and returns the +number of rows affected. If no rows are affected, @code{do} returns +@code{"0E0"}, which Perl treats as zero but regards as true. This method is +generally used for non-@code{SELECT} statements that cannot be prepared in +advance (due to driver limitations) or that do not need to be executed more +than once (inserts, deletes, etc.). Example: +@example +$rv = $dbh->do($statement) + or die "Can't execute $statement: $dbh- >errstr\n"; +@end example -@item @uref{http://www.mysql.com/Downloads/Contrib/DbFramework-1.10.tar.gz, DbFramework-1.10.tar.gz} -DbFramework is a collection of classes for manipulating @strong{MySQL} -databases. The classes are loosely based on the CDIF Data Model -Subject Area. By Paul Sharpe @email{paul@@miraclefish.com}. +Generally the 'do' statement is MUCH faster (and is preferable) +than prepare/execute for statements that don't contain parameters. -@item @uref{http://www.mysql.com/Downloads/Contrib/pike-mysql-1.4.tar.gz, pike-mysql-1.4.tar.gz} -@strong{MySQL} module for pike. For use with the Roxen web server. +@findex DBI->quote() +@findex quote() DBI method +@cindex quoting strings +@cindex strings, quoting +@item quote($string) +The @code{quote} method is used to "escape" any special characters contained in +the string and to add the required outer quotation marks. +Example: +@example +$sql = $dbh->quote($string) +@end example -@item @uref{http://www.mysql.com/Downloads/Contrib/squile.tar.gz, squile.tar.gz} -Module for @code{guile} that allows @code{guile} to interact with SQL -databases. By Hal Roberts. +@findex DBI->fetchrow_array +@findex fetchrow_array DBI method +@item fetchrow_array +This method fetches the next row of data and returns it as an array of +field values. Example: +@example +while(@@row = $sth->fetchrow_array) @{ + print qw($row[0]\t$row[1]\t$row[2]\n); +@} +@end example -@item @uref{http://www.mysql.com/Downloads/Contrib/stk-mysql.tar.gz, stk-mysql.tar.gz} -Interface for Stk. Stk is the Tk widgets with Scheme underneath instead of Tcl. -By Terry Jones. +@findex DBI->fetchrow_arrayref +@findex fetchrow_arrayref DBI method +@item fetchrow_arrayref +This method fetches the next row of data and returns it as a reference +to an array of field values. Example: +@example +while($row_ref = $sth->fetchrow_arrayref) @{ + print qw($row_ref->[0]\t$row_ref->[1]\t$row_ref->[2]\n); +@} +@end example -@item @uref{http://www.mysql.com/Downloads/Contrib/eiffel-wrapper-1.0.tar.gz,eiffel-wrapper-1.0.tar.gz} -Eiffel wrapper by Michael Ravits. +@findex DBI->fetchrow_hashref +@findex fetchrow_hashref DBI method +@item fetchrow_hashref +This method fetches a row of data and returns a reference to a hash +table containing field name/value pairs. This method is not nearly as +efficient as using array references as demonstrated above. Example: +@example +while($hash_ref = $sth->fetchrow_hashref) @{ + print qw($hash_ref->@{firstname@}\t$hash_ref->@{lastname@}\t\ + $hash_ref- > title@}\n); +@} +@end example -@item @uref{http://www.mysql.com/Downloads/Contrib/SQLmy0.06.tgz,SQLmy0.06.tgz} -FlagShip Replaceable Database Driver (RDD) for MySQL. By Alejandro -Fernandez Herrero. -@uref{http://www.fship.com/rdds.html, Flagship RDD home page} +@findex DBI->fetchall_arrayref +@findex fetchall_arrayref DBI method +@item fetchall_arrayref +This method is used to get all the data (rows) to be returned from the +SQL statement. It returns a reference to an array of references to arrays +for each row. You access or print the data by using a nested +loop. Example: +@example +my $table = $sth->fetchall_arrayref + or die "$sth->errstr\n"; +my($i, $j); +for $i ( 0 .. $#@{$table@} ) @{ + for $j ( 0 .. $#@{$table->[$i]@} ) @{ + print "$table->[$i][$j]\t"; + @} + print "\n"; +@} +@end example -@item @uref{http://www.mysql.com/Downloads/Contrib/mydsn-1.0.zip,mydsn-1.0.zip} -Binary and source for @code{mydsn.dll}. mydsn should be used to build -and remove the DSN registry file for the MyODBC driver in Coldfusion -applications. By Miguel Angel Solórzano. +@findex DBI->finish +@findex finish DBI method +@item finish +Indicates that no more data will be fetched from this statement +handle. You call this method to free up the statement handle and any +system resources associated with it. Example: +@example +$rc = $sth->finish; +@end example -@item @uref{http://www.mysql.com/Downloads/Contrib/MySQL-ADA95_API.zip, MySQL-ADA95_API.zip} -An ADA95 interface to the @strong{MySQL} API. By Francois Fabien. +@findex DBI->rows +@findex rows DBI method +@item rows +Returns the number of rows changed (updated, deleted, etc.) by the last +command. This is usually used after a non-@code{SELECT} @code{execute} +statement. Example: +@example +$rv = $sth->rows; +@end example -@item @uref{http://www.mysql.com/Downloads/Contrib/MyTool-DLL_for_VB_and_MySQL.zip, MyTool-DLL_for_VB_and_MySQL.zip} -A DLL with @strong{MySQL} C API for Visual Basic. -By Ken Menzel @email{kenm@@icarz.com}. +@findex DBI->@{NULLABLE@} +@findex NULLABLE DBI method +@item NULLABLE +Returns a reference to an array of boolean values; for each element of +the array, a value of TRUE indicates that this +column may contain @code{NULL} values. +Example: +@example +$null_possible = $sth->@{NULLABLE@}; +@end example -@item @uref{http://www.mysql.com/Downloads/Contrib/MYSQLX.EXE, MYSQL.EXE} -@strong{MySQL} ActiveX Object for directly accessing your @strong{MySQL} -servers from IIS/ASP, VB, VC++ skipping the slower ODBC methods. Fully -updatable, multithreaded with full support for all @strong{MySQL} fieldtypes -(version 2001.1.1). By SciBit @uref{http://www.scibit.com/}. +@findex DBI->@{NUM_OF_FIELDS@} +@findex NUM_OF_FIELDS DBI method +@item NUM_OF_FIELDS +This attribute indicates +the number of fields returned by a @code{SELECT} or @code{SHOW FIELDS} +statement. You may use this for checking whether a statement returned a +result: A zero value indicates a non-@code{SELECT} statement like +@code{INSERT}, @code{DELETE}, or @code{UPDATE}. +Example: +@example +$nr_of_fields = $sth->@{NUM_OF_FIELDS@}; +@end example -@item @uref{http://www.fastflow.it/mylua/, MyLUA home page} -How to use the LUA language to write @strong{MySQL} @code{PROCEDURE} that can -be loaded runtime. +@findex DBI->data_sources() +@findex data_sources() DBI method +@item data_sources($driver_name) +This method returns an array containing names of databases available to the +@strong{MySQL} server on the host @code{'localhost'}. +Example: +@example +@@dbs = DBI->data_sources("mysql"); +@end example -@itemize @bullet -@item @uref{http://www.mysql.com/Downloads/Contrib/lua-4.0.tar.gz, Lua 4.0} -LUA 4.0 -@item @uref{http://www.mysql.com/Downloads/Contrib/mylua-3.23.32.1.tar.gz, mylua-3.23.32.1.tar.gz} -Patch for @strong{MySQL} 3.23.32 to use LUA 4.0. By Cristian Giussani. -@end itemize +@findex DBI->@{ChopBlanks@} +@findex ChopBlanks DBI method +@item ChopBlanks +This attribute determines whether the @code{fetchrow_*} methods will chop +leading and trailing blanks from the returned values. +Example: +@example +$sth->@{'ChopBlanks'@} =1; +@end example -@item @uref{http://www.mysql.com/Downloads/Contrib/patched_myodbc.zip, patched_myodbc.zip} -Patch (for Omniform 4.0 support) to the MyODBC driver. -By Thomas Thaele @email{tthaele@@papenmeier.de} +@findex DBI->trace +@findex trace DBI method +@item trace($trace_level) +@itemx trace($trace_level, $trace_filename) +The @code{trace} method enables or disables tracing. When invoked as a +@code{DBI} class method, it affects tracing for all handles. When invoked as +a database or statement handle method, it affects tracing for the given +handle (and any future children of the handle). Setting @code{$trace_level} +to 2 provides detailed trace information. Setting @code{$trace_level} to 0 +disables tracing. Trace output goes to the standard error output by +default. If @code{$trace_filename} is specified, the file is opened in +append mode and output for @emph{all} traced handles is written to that +file. Example: +@example +DBI->trace(2); # trace everything +DBI->trace(2,"/tmp/dbi.out"); # trace everything to + # /tmp/dbi.out +$dth->trace(2); # trace this database handle +$sth->trace(2); # trace this statement handle +@end example -@end itemize +@tindex DBI_TRACE environment variable +@tindex Environment variable, DBI_TRACE +You can also enable @code{DBI} tracing by setting the @code{DBI_TRACE} +environment variable. Setting it to a numeric value is equivalent to calling +@code{DBI->(value)}. Setting it to a pathname is equivalent to calling +@code{DBI->(2,value)}. -@appendixsec Clients +@end table -@itemize @bullet -@item Graphical clients -@itemize @minus -@item @uref{http://www.ideit.com/products/dbvis/, DbVisualizer}. -Freeware JDBC client to graphically visualize the data and structure -of several databases simultaneously. By Innovative-IT Development AB. +@noindent +@strong{MySQL-specific Methods} -@item @uref{http://www.mysql.com/downloads/gui-clients.html, MySQLGUI} -The @strong{MySQL} GUI client homepage. By Sinisa at @strong{MySQL AB}. +The methods shown below are @strong{MySQL}-specific and not part of the +@code{DBI} standard. Several of them are now deprecated: +@code{is_blob}, @code{is_key}, @code{is_num}, @code{is_pri_key}, +@code{is_not_null}, @code{length}, @code{max_length}, and @code{table}. +Where @code{DBI}-standard alternatives exist, they are noted below: -@item @uref{http://www.mysql.com/Downloads/Contrib/mysql_navigator_0.9.0.tar.gz, MySQL navigator 0.9} -MySQL Navigator is a @strong{MySQL} database server GUI client program. The purpose -of MySQL Navigator is to provide a useful client interface to @strong{MySQL} -database servers, whilst supporting multiple operating systems and -languages. You can currently import/export database, enter queries, get -result sets, edit scripts, run scripts, add, alter, and delete users, -and retrieve client and server information. Uses QT 2.2. GPL -@uref{http://sql.kldp.org/mysql, Home page for MySQL Navigator}. +@table @code +@findex DBI->@{insertid@} +@findex insertid DBI method +@tindex AUTO_INCREMENT, using with DBI +@item insertid +If you use the @code{AUTO_INCREMENT} feature of @strong{MySQL}, the new +auto-incremented values will be stored here. +Example: +@example +$new_id = $sth->@{insertid@}; +@end example -@item @uref{http://www.mysql.com/Downloads/Win32/secman.zip, MySQL Security GUI} -A user and security management GUI for @strong{MySQL} on Windows. -By Martin Jeremic. -@uref{http://jsoft.webjump.com/, Home page for MySQL Security GUI}. +As an alternative, you can use @code{$dbh->@{'mysql_insertid'@}}. -@item @uref{http://www.mysql.com/Downloads/Contrib/kmysqladmin-0.4.1.tar.gz, kmysqladmin-0.4.1.tar.gz}. -@item @uref{http://www.mysql.com/Downloads/Contrib/kmysqladmin-0.4.1-1.src.rpm, kmysqladmin-0.4.1-1.src.rpm}. -@item @uref{http://www.mysql.com/Downloads/Contrib/kmysqladmin-0.4.1-1.i386.rpm, kmysqladmin-0.4.1-1.i386.rpm} -An administration tool for the @strong{MySQL} server using QT / KDE. Tested -only on Linux. +@findex DBI->@{is_blob@} +@findex is_blob DBI method +@item is_blob +Returns a reference to an array of boolean values; for each element of the +array, a value of TRUE indicates that the +respective column is a @code{BLOB}. +Example: +@example +$keys = $sth->@{is_blob@}; +@end example -@item @uref{http://www.mysql.com/Downloads/Contrib/mysql-admin-using-java+swing.tar.gz, Java client using Swing} By Fredy Fischer, @email{se-afs@@dial.eunet.ch}. -You can always find the latest version -@uref{http://www.trash.net/~ffischer/admin/index.html, here}. +@findex DBI->@{is_key@} +@findex is_key DBI method +@item is_key +Returns a reference to an array of boolean values; for each element of the +array, a value of TRUE indicates that the +respective column is a key. +Example: +@example +$keys = $sth->@{is_key@}; +@end example -@item @uref{http://www.mysql.com/Downloads/Win32/MySQL-Maker-1.0.zip,MySQL-Maker 1.0}. -Shareware @strong{MySQL} client for Windows. It's a WYSIWYG tool which allows -you to create, change and delete databases and tables. -You can change field - structure and add, change and delete data in -these tables directly without ODBC-driver. -@uref{http://www.presult.de/presult/frames/fs_mysqlmaker.html, MySQL Maker homepage} +@findex DBI->@{is_num@} +@findex is_num DBI method +@item is_num +Returns a reference to an array of boolean values; for each element of the +array, a value of TRUE indicates that the +respective column contains numeric values. +Example: +@example +$nums = $sth->@{is_num@}; +@end example -@item @uref{http://www.mysql.com/Downloads/Contrib/mysqlwinadmn.zip, mysqlwinadmn.zip} -Windows GUI (binary only) to administrate a database, by David B. Mansel, -@email{david@@zhadum.org}. +@findex DBI->@{is_pri_key@} +@findex is_pri_key DBI method +@item is_pri_key +Returns a reference to an array of boolean values; for each element of the +array, a value of TRUE indicates that the respective column is a primary key. +Example: +@example +$pri_keys = $sth->@{is_pri_key@}; +@end example -@item @uref{http://home.online.no/~runeberg/myqa, MyQA} -is a Linux-based query client for the @strong{MySQL} database server. MyQA -lets you enter SQL queries, execute them, and view the results, all in a -graphical user interface. The GUI is roughly similar to that of the -'Query Analyzer' client that comes with MS SQL Server. +@findex DBI->@{is_not_null@} +@findex is_not_null DBI method +@item is_not_null +Returns a reference to an array of boolean values; for each element of the +array, a value of FALSE indicates that this column may contain @code{NULL} +values. +Example: +@example +$not_nulls = $sth->@{is_not_null@}; +@end example -@item @uref{http://members.xoom.com/_opex_/mysqlmanager/index.html, MySQL Manager} -a graphical @strong{MySQL} server manager for @strong{MySQL} server written in Java, for Windows +@code{is_not_null} is deprecated; it is preferable to use the +@code{NULLABLE} attribute (described above), because that is a DBI standard. +@findex DBI->@{length@} +@findex length DBI method +@findex DBI->@{max_length@} +@findex max_length DBI method +@item length +@itemx max_length +Each of these methods returns a reference to an array of column sizes. The +@code{length} array indicates the maximum possible sizes that each column may +be (as declared in the table description). The @code{max_length} array +indicates the maximum sizes actually present in the result table. Example: -@item @uref{http://www.mysql.com/Downloads/Win32/netadmin.zip, netadmin.zip} -An administrator tool for @strong{MySQL} on Windows 95/98 and Windows NT -4.0. Only tested with @strong{MySQL} Versions 3.23.5 - 3.23.7. Written -using the Tmysql components. +@example +$lengths = $sth->@{length@}; +$max_lengths = $sth->@{max_length@}; +@end example -You can write queries and show tables, indexes, table syntax, and -administrate user, host, and database and so on. This is beta and -still has some bugs. You can test the program with all features. Please -send bugs and hints to Marco Suess @email{ms@@it-netservice.de}. Original -URL @url{http://www.it-netservice.de/pages/software/index.html}. +@findex DBI->@{NAME@} +@findex NAME DBI method +@item NAME +Returns a reference to an array of column names. +Example: +@example +$names = $sth->@{NAME@}; +@end example -@item @uref{http://www.mysql.com/Downloads/Win32/netadmin2.zip, netadmin2.zip} -New version of netadmin. See above for details. +@findex DBI->@{table@} +@findex table DBI method +@item table +Returns a reference to an array of table names. +Example: +@example +$tables = $sth->@{table@}; +@end example -@item @uref{http://www.mysql.com/Downloads/Win32/ARTADMIN203.EXE,Atronic's @strong{MySQL} client for Windows 2.0.3.0}. -Home page for this can be found at: @uref{http://www.artronic.hr}. +@findex DBI->@{type@} +@findex type DBI method +@item type +Returns a reference to an array of column types. +Example: +@example +$types = $sth->@{type@}; +@end example -@item @uref{http://www.mysql.com/Downloads/Win32/mysqlfront.zip, mysqlfront} -Home page: @uref{http://www.mysqlfront.de/}. -Win32-Client for accessing and managing dbs, tables, table-data, indexes, -import-/export-files. (Freeware). By Ansgar Becker. +@end table -@item @uref{http://www.mysql.com/Downloads/Win32/W9xstop.zip,Utility from Artronic to stop MySQL on win9x}. +@cindex @code{DBI/DBD} +@node DBI-info, , Perl DBI Class, Perl +@subsection More @code{DBI}/@code{DBD} Information -@item @uref{http://bardo.hyperlink.cz/mysqlmon,a light weight GUI client for Windows}. +You can use the @code{perldoc} command to get more information about +@code{DBI}. -@item @uref{http://dbtools.vila.bol.com.br/, Dbtools} -A tool to manage @strong{MySQL} databases. Currently only for Windows. -Some features: -@itemize @bullet -@item Manage servers, databases, tables, columns, indexes, and users -@item Import wizard to import structure and data from MS Access, MS Excel, Dbase, FoxPro, Paradox, and ODBC Databases. +@example +perldoc DBI +perldoc DBI::FAQ +perldoc DBD::mysql +@end example -@item @uref{http://www.mysql.com/Downloads/Contrib/KMYENG113.zip,KMYENG113.zip} -An administrator GUI for @strong{MySQL}. Works only on windows, no source. -Available in English and Japanese. By Mitunobu Kaneko. -Home page: @uref{http://sql.jnts.ne.jp/} -@end itemize +You can also use the @code{pod2man}, @code{pod2html}, etc., tools to +translate to other formats. -@item @uref{http://www.mysql.com/Downloads/Contrib/xmysqladmin-1.0.tar.gz, xmysqladmin-1.0.tar.gz} -An X-based front end to the @strong{MySQL} database engine. It allows reloads, -status check, process control, myisamchk, grant/revoke privileges, -creating databases, dropping databases, create, alter, browse, and drop -tables. Originally by Gilbert Therrien, @email{gilbert@@ican.net} but -now in public domain and supported by @strong{MySQL AB}. +You can find the latest @code{DBI} information at +the @code{DBI} Web page: +@example +@uref{http://www.symbolstone.org/technology/perl/DBI/index.html} +@end example -@item @uref{http://www.mysql.com/Downloads/Contrib/xmysql-1.9.tar.gz, xmysql-1.9.tar.gz}. -@item @uref{http://web.wt.net/~dblhack, xmysql home page} -A front end to the @strong{MySQL} database engine. It allows for simple -queries and table maintenance, as well as batch queries. By Rick -Mehalick, @email{dblhack@@wt.net}. -Requires @uref{http://bragg.phys.uwm.edu/xforms,xforms 0.88} to work. -@item @uref{http://www.tamos.net/sw/dbMetrix,dbMetrix} -An open source client for exploring databases and executing SQL. Supports -@strong{MySQL}, Oracle, PostgreSQL, and mSQL. -@item @uref{http://www.multimania.com/bbrox/GtkSQL,GtkSQL} -A query tool for @strong{MySQL} and PostgreSQL. -@item @uref{http://dbman.linux.cz/,dbMan} -A query tool written in Perl. Uses DBI and Tk. -@item @uref{http://www.mysql.com/Downloads/Win32/Msc201.EXE, Mascon 202} -@item @uref{http://www.mysql.com/Downloads/Win32/FrMsc202.EXE, Free Mascon 202} -Mascon is a powerful Win32 GUI for the administering @strong{MySQL} server -databases. Mascon's features include visual table design, connections to -multiple servers, data and blob editing of tables, security setting, SQL -color coding, dump functionality and much more. -@uref{http://www.scibit.com/Products/Software/Utils/Mascon.asp,Mascon home page}. -@item @uref{http://www.virtualbeer.net/dbui/,DBUI} -DBUI is a Gtk graphical database editor. -@item @uref{http://www.rtlabs.com/, MacSQL} -GUI for MySQL, ODBC, and JDBC databases for the Mac OS. -@item @uref{http://www.caleb.com.au/, JRetriever} -JRetriever is a generic database front-end tool for JDBC compliant -databases written with Java 2. JRetriever displays database -tables/views in a Windows explorer-like front end. Users can retrieve -data either by clicking on the table folder or by composing their own SQL -statements with our built-in SQL editor. The tool has been tested with -Oracle 8 and @strong{MySQL} as the back-end databases. It requires JDK 1.3 from -JavaSoft. -@item @uref{http://www.jetools.com/products/databrowser/, DataBrowser} -The DataBrowser is a cross-database, cross-platform data access tool. It is more -user friendly than tools like SQL Plus, psql (command line based tools). It is more -flexible than TOAD, ISQL, PGAccess which are GUI's that are limitied to a single -platform or database. -@item @uref{http://www.intrex.net/amit/software/, SQLC} -The SQL Console is a standalone java application that allows you to connect to a - SQL database system and issue SQL queries and updates. It has an easy-to use -graphical user interface. The SQL Console uses JDBC to connect to the database -systems and, therefore, with proper JDBC drivers, you can use this utility to -connect to some of the most popular database systems. -@item @uref{http://www.mysql.com/Downloads/Contrib/mysql_mmc.zip, MySQL MMC} -MySQL MMC is a GUI Management Tool developed using kdevelop -with a very good interface completely like Microsoft -Enterprise Tool (for SQL Server) or Sybase Central. We -can use it to manage server, database, table, index, -users and to edit table data in grid or execute Sql -by Query Analysis. -@end itemize - -@cindex Web clients -@item Web Clients -@itemize @minus -@item @uref{http://www.mysql.com/Downloads/Contrib/mysqladmin-atif-1.0.tar.gz, mysqladmin-atif-1.0.tar.gz} -WWW @strong{MySQL} administrator for the @code{user,} @code{db} and -@code{host} tables. By Tim Sailer, modified by Atif Ghaffar -@email{aghaffar@@artemedia.ch}. - -@item @uref{http://www.mysql.com/Downloads/Contrib/mysql-webadmin-1.0a8-rz.tar.gz, mysql-webadmin-1.0a8-rz.tar.gz} -A tool written in PHP-FI to administrate @strong{MySQL} databases -remotely over the web within a Web-Browser. By Peter Kuppelwieser, -@email{peter.kuppelwieser@@kantea.it}. Updated by Wim Bonis, -@email{bonis@@kiss.de}. Not maintained anymore! +@cindex Eiffel Wrapper +@cindex wrappers, Eiffel +@node Eiffel, Java, Perl, Clients +@section MySQL Eiffel wrapper -@item @uref{http://www.mysql.com/Downloads/Contrib/mysqladm.tar.gz, mysqladm.tar.gz} -@strong{MySQL} Web Database Administration written in Perl. By Tim Sailer. +The @strong{MySQL} @uref{http://www.mysql.com/Downloads/Contrib/,Contrib directory} +contains an Eiffel wrapper written by Michael Ravits. -@item @uref{http://www.mysql.com/Downloads/Contrib/mysqladm-2.tar.gz, mysqladm-2.tar.gz} -Updated version of @file{mysqladm.tar.gz}, by High Tide. +You can also find this at: +@url{http://www.netpedia.net/hosting/newplayer/} -@item @uref{http://www.mysql.com/Downloads/Contrib/billowmysql.zip, billowmysql.zip} -Updated version of @file{mysqladm.tar.gz}, by Ying Gao. You can get the -newest version from @uref{http://civeng.com/sqldemo/, the home site}. +@cindex Java connectivity +@cindex JDBC +@node Java, PHP, Eiffel, Clients +@section MySQL Java Connectivity (JDBC) -@item @uref{http://www.mysql.com/Downloads/Contrib/myadmin-0.4.tar.gz, myadmin-0.4.tar.gz}. -@item @uref{http://myadmin.cheapnet.net/, MyAdmin home page} -A Web-based @strong{MySQL} administrator by Mike Machado. +There are 2 supported JDBC drivers for @strong{MySQL} (the mm driver and +the Reisin JDBC driver). You can find a copy of the mm driver at +@uref{http://mmmysql.sourceforge.net/} or +@uref{http://www.mysql.com/Downloads/Contrib/} and the Reisin driver at +@uref{http://www.caucho.com/projects/jdbc-mysql/index.xtp} For +documentation consult any JDBC documentation and the driver's own +documentation for @strong{MySQL}-specific features. -@item @uref{http://www.mysql.com/Downloads/Contrib/phpMyAdmin_2.0.1.tar.gz,phpMyAdmin_2.0.1.tar.gz} -A set of PHP3-scripts to adminstrate @strong{MySQL} over the WWW. +@cindex PHP API +@node PHP, Cplusplus, Java, Clients +@section MySQL PHP API -@item @uref{http://www.phpwizard.net/projects/phpMyAdmin/, phpMyAdmin home page} -A PHP3 tool in the spirit of mysql-webadmin, by Tobias Ratschiller, tobias@@dnet.it. +PHP is a server-side, HTML-embedded scripting language that may be used to +create dynamic Web pages. It contains support for accessing several +databases, including @strong{MySQL}. PHP may be run as a separate program +or compiled as a module for use with the Apache Web server. -@item @uref{http://www.mysql.com/Downloads/Contrib/useradm.tar.gz, useradm.tar.gz} -@strong{MySQL} administrator in PHP. By Ofni Thomas -@email{othomas@@vaidsystems.com}. +The distribution and documentation are available at the +@uref{http://www.php.net/, PHP web site}. -@item @uref{http://gossamer-threads.com/perl/mysqlman/mysql.cgi, MySQLMan} -Similar functionality as phpmyadmin, but written with Perl and using -html templates. By Alex Krohn. -@end itemize +@menu +* PHP problems:: Common problems with MySQL and PHP +@end menu -@item @uref{http://www.mysql.com/Downloads/Contrib/mysql-editor.tar.gz,mysql-editor.tar.gz} -This cgi scripts in Perl enables you to edit content of Mysql -database. By Tomas Zeman. -@item -@uref{http://worldcommunity.com/opensource/futuresql, FutureSQL Web Database Administration Tool}. -FutureSQL by Peter F. Brown, is a free, open source rapid application -development Web database administration tool, written in Perl, -using @strong{MySQL}. It uses @code{DBI:DBD} and @code{CGI.pm}. +@node PHP problems, , PHP, PHP +@subsection Common Problems with MySQL and PHP -FutureSQL allows one to easily set up config files to view, edit, delete, -and otherwise process records from a @strong{MySQL} database. It uses a data -dictionary, configuration files and templates, and allows "pre-processing" -and "post-processing" on both fields, records, and operations. +@itemize @bullet +@item Error: "Maximum Execution Time Exceeded" +This is a PHP limit; Go into the @file{php3.ini} file and set the maximum +execution time up from 30 seconds to something higher, as needed. +It is also not a bad idea to double the ram allowed per script to 16MB instead of +8 MB. +@item Error: "Fatal error: Call to unsupported or undefined function mysql_connect() in .." +This means that your PHP version isn't compiled with @strong{MySQL} support. +You can either compile a dynamic @strong{MySQL} module and load it into PHP or +recompile PHP with built-in @strong{MySQL} support. This is described in +detail in the PHP manual. +@item Error: "undefined reference to `uncompress'" +This means that the client library is compiled with support for a compressed +client/server protocol. The fix is to add @code{-lz} last when linking +with @code{-lmysqlclient}. @end itemize -@cindex web tools -@cindex tools,, web -@appendixsec Web Tools +@cindex C++ APIs +@node Cplusplus, Python, PHP, Clients +@section MySQL C++ APIs -@itemize @bullet +Two APIs are available in the @strong{MySQL} +@uref{http://www.mysql.com/Downloads/Contrib/,Contrib directory}. -@item @uref{http://www.mysql.com/Downloads/Contrib/mod_mysql_include_1.0.tar.gz, mod_mysql_include_1.0.tar.gz} -Apache module to include HTML from @strong{MySQL} queries into your pages, -and run update queries. Originally written to implement a simple fast -low-overhead banner-rotation system. By Sasha Pachev. +@cindex Python APIs +@node Python, Tcl, Cplusplus, Clients +@section MySQL Python APIs -@item @uref{http://htcheck.sourceforge.net, htCheck} - URL checker with -@strong{MySQL} backend. Spidered URLs can later be queried using SQL to retrieve -various kinds of information, eg. broken links. Written by Gabriele Bartolini. +The @strong{MySQL} @uref{http://www.mysql.com/Downloads/Contrib/,Contrib directory} +contains a Python interface written by Joseph Skinner. -@item @uref{http://www.odbsoft.com/cook/sources.htm} -This package has various functions for generating html code from a SQL -table structure and for generating SQL statements (Select, Insert, -Update, Delete) from an html form. You can build a complete forms -interface to a SQL database (query, add, update, delete) without any -programming! By Marc Beneteau, @email{marc@@odbsoft.com}. +You can also use the Python interface to iODBC to access a +@strong{MySQL} server. +@uref{http://starship.skyport.net/~lemburg/,mxODBC} -@item @uref{http://www.mysql.com/Downloads/Contrib/sqlhtml.tar.gz, sqlhtml.tar.gz} -SQL/HTML is an HTML database manager for @strong{MySQL} using @code{DBI} 1.06. +@cindex Tcl APIs +@node Tcl, , Python, Clients +@section MySQL Tcl APIs -@item @uref{http://www.mysql.com/Downloads/Contrib/udmsearch-3.0.23.tar.gz, UdmSearch 3.0.23 (stable version)}. -@item @uref{http://www.mysql.com/Downloads/Contrib/mnogosearch-3.1.12.tar.gz, mnogosearch 3.1.12 (development but recommended version)}. -@item @uref{http://search.mnoGo.ru, UdmSearch home page} -A SQL-based search engine for Internet. By -Alexander I. Barkov @email{bar@@izhcom.ru}. +@uref{http://www.binevolve.com/~tdarugar/tcl-sql/, Tcl at binevolve}. +The +@uref{http://www.mysql.com/Downloads/Contrib,Contrib directory} contains a Tcl +interface that is based on msqltcl 1.50. -@item @uref{http://www.mysql.com/Downloads/Contrib/wmtcl.doc, wmtcl.doc}. -@item @uref{http://www.mysql.com/Downloads/Contrib/wmtcl.lex, wmtcl.lex} -With this you can write HTML files with inclusions of Tcl code. By -@email{vvs@@scil.npi.msu.su}. -@item @uref{http://www.mysql.com/Downloads/Contrib/www-sql-0.5.7.lsm, www-sql-0.5.7.lsm}. -@item @uref{http://www.mysql.com/Downloads/Contrib/www-sql-0.5.7.tar.gz, www-sql-0.5.7.tar.gz} -A CGI program that parses an HTML file containing special tags, parses -them, and inserts data from a @strong{MySQL} database. -@item @uref{http://www.mysql.com/Downloads/Contrib/genquery.zip, genquery.zip} -Perl SQL database interface package for html. +@cindex internals +@cindex threads +@node MySQL internals, Environment variables, Clients, Top +@chapter MySQL Internals -@item @uref{http://www.mysql.com/Downloads/Contrib/cgi++-0.8.tar.gz, cgi++-0.8.tar.gz} -A macro-processor to simply writing CGI/Database programs in C++ by Sasha Pachev. +This chapter describes a lot of things that you need to know when +working on the @strong{MySQL} code. If you plan to contribute to MySQL +development, want to have access to the bleeding-edge in-between +versions code, or just want to keep track of development, follow the +instructions in @xref{Installing source tree}. If you are interested in MySQL +internals, you should also subscribe to @email{internals@@lists.mysql.com}. +This is a relatively low traffic list, in comparison with +@email{mysql@@lists.mysql.com}. -@item @uref{http://www.mysql.com/Downloads/Contrib/webboard-1.0.zip, WebBoard 1.0} -EU-Industries Internet-Message-Board. +@menu +* MySQL threads:: MySQL threads +* MySQL test suite:: MySQL test suite +@end menu -@item @uref{http://www.mysql.com/Downloads/Contrib/DBIx-TextIndex-0.02.tar.gz, DBIx-TextIndex-0.02.tar.gz} -Full-text searching with Perl on @code{BLOB}/@code{TEXT} columns by Daniel Koch. -@end itemize +@node MySQL threads, MySQL test suite, MySQL internals, MySQL internals +@section MySQL Threads -@cindex tools, benchmarking -@cindex benchmarking, tools -@appendixsec Performance Benchmarking Tools +The @strong{MySQL} server creates the following threads: @itemize @bullet -@item @uref{http://www.mysql.com/Downloads/super-smack/super-smack-1.0.tar.gz, - super-smack} -Multi-threaded benchmarking tool for @strong{MySQL} and -@strong{PostgreSQL}. Written in C++. Easy to extend to support other -databases that have C/C++ client libraries. By Sasha Pachev. -@end itemize - -@cindex tools, authentication -@cindex authentication tools -@appendixsec Authentication Tools -@itemize @bullet -@item @uref{http://www.mysql.com/Downloads/Contrib/ascend-radius-mysql-0.7.2.patch.gz,ascend-radius-mysql-0.7.2.patch.gz} -This is an authentication and logging patch using @strong{MySQL} for -Ascend-Radius. By @email{takeshi@@SoftAgency.co.jp}. +@item +The TCP/IP connection thread handles all connection requests and +creates a new dedicated thread to handle the authentication and +and SQL query processing for each connection. -@item @uref{http://www.mysql.com/Downloads/Contrib/icradius-0.10.tar.gz, icradius 0.10} -@uref{http://www.mysql.com/Downloads/Contrib/icradius.README, icradius readme file}. +@item +On Windows NT there is a named pipe handler thread that does the same work as +the TCP/IP connection thread on named pipe connect requests. -@item @uref{http://www.mysql.com/Downloads/Contrib/checkpassword-0.81-mysql-0.6.6.patch.gz, -checkpassword-0.81-mysql-0.6.6.patch.gz} -@strong{MySQL} authentication patch for QMAIL and checkpassword. These are -useful for management user (mail, pop account) by @strong{MySQL}. -By @email{takeshi@@SoftAgency.co.jp}. +@item +The signal thread handles all signals. This thread also normally handles +alarms and calls @code{process_alarm()} to force timeouts on connections +that have been idle too long. -@item @uref{http://www.mysql.com/Downloads/Contrib/jradius-diff.gz, jradius-diff.gz} -@strong{MySQL} support for Livingston's Radius 2.01. Authentication and -Accounting. By Jose de Leon, @email{jdl@@thevision.net}. +@item +If @code{mysqld} is compiled with @code{-DUSE_ALARM_THREAD}, a dedicated +thread that handles alarms is created. This is only used on some systems where +there are problems with @code{sigwait()} or if one wants to use the +@code{thr_alarm()} code in ones application without a dedicated signal +handling thread. -@item @uref{http://www.mysql.com/Downloads/Contrib/mod_auth_mysql-2.20.tar.gz, mod_auth_mysql-2.20.tar.gz} -Apache authentication module for @strong{MySQL}. By Zeev Suraski, -@email{bourbon@@netvision.net.il}. +@item +If one uses the @code{--flush_time=#} option, a dedicated thread is created +to flush all tables at the given interval. -@c @strong{Please} register this module at: -@c @url{http://bourbon.netvision.net.il/mysql/mod_auth_mysql/register.html}. The -@c registering information is only used for statistical purposes and will -@c encourage further development of this module! +@item +Every connection has its own thread. -@item @uref{http://www.mysql.com/Downloads/Contrib/mod_log_mysql-1.05.tar.gz, mod_log_mysql-1.05.tar.gz} -@strong{MySQL} logging module for Apache. By Zeev Suraski, -@email{bourbon@@netvision.net.il}. +@item +Every different table on which one uses @code{INSERT DELAYED} gets its +own thread. -@item @uref{http://www.mysql.com/Downloads/Contrib/mypasswd-2.0.tar.gz, mypasswd-2.0.tar.gz} -Extra for @code{mod_auth_mysql}. This is a little tool that allows you -to add/change user records storing group and/or password entries in -@strong{MySQL} tables. By Harry Brueckner, @email{brueckner@@respublica.de}. +@item +If you use @code{--master-host}, a slave replication thread will be +started to read and apply updates from the master. +@end itemize -@item @uref{http://www.mysql.com/Downloads/Contrib/mysql-passwd.README, mysql-passwd.README}. -@item @uref{http://www.mysql.com/Downloads/Contrib/mysql-passwd-1.2.tar.gz, mysql-passwd-1.2.tar.gz} -Extra for @code{mod_auth_mysql}. This is a two-part system for use with -@code{mod_auth_mysql}. +@code{mysqladmin processlist} only shows the connection, @code{INSERT DELAYED}, +and replication threads. -@item @uref{http://www.mysql.com/Downloads/Contrib/pam_mysql.tar.gz, pam_mysql.tar.gz} -This module authenticates users via @code{pam}, using @strong{MySQL}. +@cindex mysqltest, MySQL Test Suite +@cindex testing mysqld, mysqltest +@node MySQL test suite, , MySQL threads, MySQL internals +@section MySQL Test Suite -@item @uref{http://www.mysql.com/Downloads/Contrib/nsapi_auth_mysql.tar, nsapi_auth_mysql.tar} -Netscape Web Server API (NSAPI) functions to authenticate (BASIC) users -against @strong{MySQL} tables. By Yuan John Jiang. +Until recently, our main full-coverage test suite was based on proprietary +customer data and for that reason has not been publicly available. The only +publicly available part of our testing process consisted of the @code{crash-me} +test, a Perl DBI/DBD benchmark found in the @code{sql-bench} directory, and +miscellaneous tests located in @code{tests} directory. The lack of a +standardized publicly available test suite has made it difficult for our users, +as well developers, to do regression tests on the @strong{MySQL} code. To +address this problem, we have created a new test system that is included in +the source and binary distributions starting in Version 3.23.29. -@item @uref{http://www.mysql.com/Downloads/Contrib/qmail-1.03-mysql-0.6.6.patch.gz,qmail-1.03-mysql-0.6.6.patch.gz} -Patch for qmail to authenticate users from a @strong{MySQL} table. -By @email{takeshi@@SoftAgency.co.jp}. +The current set of test cases doesn't test everything in @strong{MySQL}, but it +should catch most obvious bugs in the SQL processing code, OS/library +issues, and is quite thorough in testing replication. Our eventual goal +is to have the tests cover 100% of the code. We welcome contributions +to our test suite. You may especially want to contribute tests that +examine the functionality critical to your system, as this will ensure +that all future @strong{MySQL} releases will work well with your +applications. -@item @uref{http://www.mysql.com/Downloads/Contrib/proftpd-1.2.0rc2-fix-mysql.patch, proftpd-1.2.0rc2-fix-mysql.patch} -Patch for proftpd1.2.0rc2. By @email{takeshi@@SoftAgency.co.jp}. +@menu +* running mysqltest:: +* extending mysqltest:: +* Reporting mysqltest bugs:: +@end menu -@item @uref{http://www.mysql.com/Downloads/Contrib/pwcheck_mysql-0.1.tar.gz,pwcheck_mysql-0.1.tar.gz} -An authentication module for the Cyrus IMAP server. By Aaron Newsome. +@node running mysqltest, extending mysqltest, MySQL test suite, MySQL test suite +@subsection Running the MySQL Test Suite +The test system consist of a test language interpreter +(@code{mysqltest}), a shell script to run all +tests(@code{mysql-test-run}), the actual test cases written in a special +test language, and their expected results. To run the test suite on +your system after a build, type @code{make test} or +@code{mysql-test/mysql-test-run} from the source root. If you have +installed a binary distribution, @code{cd} to the install root +(eg. @code{/usr/local/mysql}), and do @code{scripts/mysql-test-run}. +All tests should succeed. If not, you should try to find out why and +report the problem if this is a bug in @strong{MySQL}. +@xref{Reporting mysqltest bugs}. -@end itemize +If you have a copy of @code{mysqld} running on the machine where you want to +run the test suite you do not have to stop it, as long as it is not using +ports @code{9306} and @code{9307}. If one of those ports is taken, you should +edit @code{mysql-test-run} and change the values of the master and/or slave +port to one that is available. -@cindex converters -@appendixsec Converters +You can run one individual test case with +@code{mysql-test/mysql-test-run test_name}. -@itemize @bullet -@item @uref{http://www.mysql.com/Downloads/Contrib/mssql2mysql.txt, mssql2mysql.txt} -Converter from MS-SQL to MySQL. By Michael Kofler. -@uref{http://www.kofler.cc/mysql/mssql2mysql.html, mssql2mysql home page}. +If one test fails, you should test running @code{mysql-test-run} with +the @code{--force} option to check if any other tests fails. -@item @uref{http://www.mysql.com/Downloads/Contrib/dbf2mysql-1.14.tar.gz, dbf2mysql-1.14.tar.gz} -Convert between @file{.dbf} files and @strong{MySQL} tables. By Maarten -Boekhold (@email{boekhold@@cindy.et.tudelft.nl}), William Volkman, and -Michael Widenius. This converter includes rudimentary read-only support -for MEMO fields. +@node extending mysqltest, Reporting mysqltest bugs, running mysqltest, MySQL test suite +@subsection Extending the MySQL Test Suite -@item @uref{http://www.mysql.com/Downloads/Contrib/dbf2mysql-1.13.tgz, dbf2mysql-1.13.tgz} -Convert between @file{.dbf} files and @strong{MySQL} tables. By Maarten -Boekhold, @email{boekhold@@cindy.et.tudelft.nl}, and Michael Widenius. -This converter can't handle MEMO fields. +You can use the @code{mysqltest} language to write your own test cases. +Unfortunately, we have not yet written full documentation for it - we plan to +do this shortly. You can, however, look at our current test cases and use +them as an example. The following points should help you get started: -@item @uref{http://www.mysql.com/Downloads/Contrib/dbf2mysql.zip, dbf2mysql.zip} -Convert between FoxPro @file{.dbf} files and @strong{MySQL} tables on Windows. -By Alexander Eltsyn, @email{ae@@nica.ru} or @email{ae@@usa.net}. +@itemize +@item +The tests are located in @code{mysql-test/t/*.test} -@item @uref{http://www.mysql.com/Downloads/Contrib/dbf2sql.zip, dbf2sql.zip} -Short and simple prg that can help you transport your data from foxpro -table into @strong{MySQL} table. By Danko Josic. +@item +A test case consists of @code{;} terminated statements and is similar to the +input of @code{mysql} command line client. A statement by default is a query +to be sent to @strong{MySQL} server, unless it is recognized as internal +command ( eg. @code{sleep} ). -@item @uref{http://www.mysql.com/Downloads/Contrib/dump2h-1.20.gz, dump2h-1.20.gz} -Convert from @code{mysqldump} output to a C header file. By Harry Brueckner, -@email{brueckner@@mail.respublica.de}. +@item +All queries that produce results, e.g. @code{SELECT}, @code{SHOW}, +@code{EXPLAIN}, etc., must be preceded with @code{@@/path/to/result/file}. The +file must contain the expected results. An easy way to generate the result +file is to run @code{mysqltest -r < t/test-case-name.test} from +@code{mysql-test} directory, and then edit the generated result files, if +needed, to adjust them to the expected output. In that case, be very careful +about not adding or deleting any invisible characters - make sure to only +change the text and/or delete lines. If you have to insert a line, make sure +the fields are separated with a hard tab, and there is a hard tab at the end. +You may want to use @code{od -c} to make sure your text editor has not messed +anything up during edit. We, of course, hope that you will never have to edit +the output of @code{mysqltest -r} as you only have to do it when you find a +bug. -@item @uref{http://www.mysql.com/Downloads/Contrib/exportsql.txt, exportsql.txt} -A script that is similar to @code{access_to_mysql.txt}, except that this -one is fully configurable, has better type conversion (including -detection of @code{TIMESTAMP} fields), provides warnings and suggestions -while converting, quotes @strong{all} special characters in text and -binary data, and so on. It will also convert to @code{mSQL} v1 and v2, -and is free of charge for anyone. See -@uref{http://www.cynergi.net/exportsql/} for the latest version. By -Pedro Freire, @email{support@@cynergi.net}. NOTE: Doesn't work with -Access2! -@item @uref{http://www.mysql.com/Downloads/Contrib/access_to_mysql.txt, access_to_mysql.txt} -Paste this function into an Access module of a database that has the -tables you want to export. See also @code{exportsql}. By Brian Andrews. -NOTE: Doesn't work with Access2! +@item +To be consistent with our setup, you should put your result files in +@code{mysql-test/r} directory and name them @code{test_name.result}. If the +test produces more than one result, you should use @code{test_name.a.result}, +@code{test_name.b.result}, etc. -@item @uref{http://www.mysql.com/Downloads/Contrib/importsql.txt, importsql.txt} -A script that does the exact reverse of @code{exportsql.txt}. That is, -it imports data from @strong{MySQL} into an Access database via -ODBC. This is very handy when combined with exportsql, because it lets you -use Access for all DB design and administration, and synchronize with -your actual @strong{MySQL} server either way. Free of charge. See -@uref{http://www.netdive.com/freebies/importsql/} for any updates. -Created by Laurent Bossavit of NetDIVE. -@strong{NOTE:} Doesn't work with Access2! +@item +If a statement returns an error, you should on the line before the statement +specify with the @code{--error error-number}. The error number can be +a list of possible error numbers separated with @code{','}. -@item @uref{http://www.mysql.com/Downloads/Contrib/mdb2sql.bas, mdb2sql.bas} -Converter from Access97 to @strong{MySQL} by Moshe Gurvich. +@item +If you are writing a replication test case, you should on the first line of +the test file, put @code{source include/master-slave.inc;}. To switch between +master and slave, use @code{connection master;} and @code{connection slave;}. +If you need to do something on an alternate connection, you can do +@code{connection master1;} for the master, and @code{connection slave1;} for +the slave. -@item @uref{http://www.mysql.com/Downloads/Contrib/msql2mysqlWrapper-1.0.tgz, msql2mysqlWrapper 1.0} -A C wrapper from @code{mSQL} to @strong{MySQL}. By @email{alfred@@sb.net} +@item +If you need to do something in a loop, you can use something like this: +@example +let $1=1000; +while ($1) +@{ + # do your queries here + dec $1; +@} +@end example -@item @uref{http://www.mysql.com/Downloads/Contrib/sqlconv.pl, sqlconv.pl} -A simple script that can be used to copy fields from one @strong{MySQL} table -to another in bulk. Basically, you can run @code{mysqldump} and pipe it to -the @code{sqlconv.pl} script. The script will parse through the -@code{mysqldump} output and will rearrange the fields so they can be -inserted into a new table. An example is when you want to create a new -table for a different site you are working on, but the table is just a -bit different (that is - fields in different order, etc.). -By Steve Shreeve. -@item @uref{http://www.mysql.com/Downloads/Contrib/oracledump, oracledump} -Perl program to convert Oracle databases to @strong{MySQL}. Has same -output format as mysqldump. By Johan Andersson. +@item +To sleep between queries, use the @code{sleep} command. It supports fractions +of a second, so you can do @code{sleep 1.3;}, for example, to sleep 1.3 +seconds. -@item @uref{http://www.mysql.com/Downloads/Contrib/excel2mysql, excel2mysql} -Perl program to import Excel spreadsheets into a @strong{MySQL} database. By Stephen Hurd @email{shurd@@sk.sympatico.ca} +@item +To run the slave with additional options for your test case, put them +in the command-line format in @code{mysql-test/t/test_name-slave.opt}. For +the master, put them in @code{mysql-test/t/test_name-master.opt}. -@item @uref{http://www.mysql.com/Downloads/Contrib/T2S_100.ZIP, T2S_100.ZIP}. -Windows program to convert text files to @strong{MySQL} databases. By -Asaf Azulay. +@item +If you have a question about the test suite, or have a test case to contribute, +e-mail to @email{internals@@lists.mysql.com}. As the list does not accept +attachments, you should ftp all the relevant files to: +@url{ftp://support.mysql.com/pub/mysql/Incoming} @end itemize -@appendixsec Using MySQL with Other Products - -@itemize @bullet -@item @uref{http://www.mysql.com/Downloads/Contrib/emacs-sql-mode.tar.gz, emacs-sql-mode.tar.gz} -Raw port of a SQL mode for XEmacs. Supports completion. Original by -Peter D. Pezaris @email{pez@@atlantic2.sbi.com} and partial -@strong{MySQL} port by David Axmark. +@node Reporting mysqltest bugs, , extending mysqltest, MySQL test suite +@subsection Reporting bugs in the MySQL Test Suite -@item @uref{http://www.mysql.com/Downloads/Win32/myaccess97_1_4.zip, MyAccess97 1.4}. -@item @uref{http://www.mysql.com/Downloads/Win32/myaccess2000_1_4.zip, MyAccess2000 1.4}. +If your @strong{MySQL} version doesn't pass the test suite you should +do the following: -MyAccess is an AddIn for MS Access 97/2000 that allows you to manage -@strong{MySQL} databases from within Access. Main functions are: -@itemize @minus -@item Create/Modify Tables -@item Execute Queries against MySQL -@item Extract ''Create Table-Scripts'' from MySQL -@item Import/Export tables from Access to MySQL and vice versa -@item Log Changes -@item Show a "Database Definition Report -@end itemize +@itemize @bullet +@item +Don't send a bug report before you have found out as much as possible of +what when wrong! When you do it, please use the @code{mysqlbug} script +so that we can get information about your system and @code{MySQL} +version. @xref{Bug reports}. +@item +Make sure to include the output of @code{mysql-test-run}, as well as +contents of all @code{.reject} files in @code{mysql-test/r} directory. +@item +If a test in the test suite fails, check if the test fails also when run +by its own: -Written by Hubertus Hiden. @uref{http://www.accessmysql.com, MyAccess homepage}. +@example +cd mysql-test +mysql-test-run --local test-name +@end example -@item @uref{http://www.mysql.com/Downloads/Contrib/radius-0.3.tar.gz, radius-0.3.tar.gz} -Patches for @code{radiusd} to make it support @strong{MySQL}. By Wim Bonis, -@email{bonis@@kiss.de}. -@end itemize +If this fails, then you should configure @strong{MySQL} with +@code{--with-debug} and run @code{mysql-test-run} with the +@code{--debug} option. If this also fails send the trace file +@file{var/tmp/master.trace} to ftp://support.mysql.com/pub/mysql/secret +so that we can examine it. Please remember to also include a full +description of your system, the version of the mysqld binary and how you +compiled it. -@cindex tools, useful -@appendixsec Useful Tools +@item +Try also to run @code{mysql-test-run} with the @code{--force} option to +see if there is any other test that fails. -@itemize @bullet -@item @uref{http://worldcommunity.com/opensource/utilities/mysql_backup.html, MySQL Backup}. +@item +If you have compiled @strong{MySQL} yourself, check our manual for how +to compile @strong{MySQL} on your platform or, preferable, use one of +the binaries we have compiled for you at +@uref{http://www.mysql.com/downloads/}. All our standard binaries should +pass the test suite ! -A backup script for MySQL. By Peter F. Brown. +@item +If you get an error, like @code{Result length mismatch} or @code{Result +content mismatch} it means that the output of the test didn't match +exactly the expected output. This could be a bug in @strong{MySQL} or +that your mysqld version produces slight different results under some +circumstances. -@item @uref{http://www.mysql.com/Downloads/Contrib/mytop, mytop} -@item @uref{http://public.yahoo.com/~jzawodn/mytop/, mytop home page} -mytop is a Perl program that allows you to monitor @strong{MySQL} servers by -viewing active threads, queries, and overall server performance -numbers. By Jeremy D. Zawodny. +Failed test results are put in a file with the same base name as the +result file with the @code{.reject} extension. If your test case is +failing, you should do a diff on the two files. If you cannot see how +they are different, examine both with @code{od -c} and also check their +lengths. -@item @uref{http://www.mysql.com/Downloads/Contrib/mysql_watchdog.pl, mysql_watchdog.pl} -Monitor the @strong{MySQL} daemon for possible lockups. By Yermo Lamers, -@email{yml@@yml.com}. +@item +If a test fails totally, you should check the logs file in the +@code{mysql-test/var/log} directory for hints of what went wrong. -@item @uref{http://www.mysql.com/Downloads/Contrib/mysqltop.tar.gz, mysqltop.tar.gz} -Sends a query in a fixed time interval to the server and shows the -resulting table. By Thomas Wana. -@item @uref{http://www.mysql.com/Downloads/Contrib/mysql_structure_dumper.tar.gz,mysql_structure_dumper.tar.gz} -Prints out the structure of the all tables in a database. By Thomas Wana. +@item +If you have compiled @strong{MySQL} with debugging you can try to debug this +by running @code{mysql-test-run} with the @code{--gdb} and/or @code{--debug} +options. +@xref{Making trace files}. -@item @uref{http://www.mysql.com/Downloads/Contrib/mysql_structure_dumper.tgz, structure_dumper.tgz} -Prints the structure of every table in a database. By Thomas Wana. +If you have not compiled @strong{MySQL} for debugging you should probably +do that. Just specify the @code{--with-debug} options to @code{configure}! +@xref{Installing source}. +@end itemize -@item @uref{http://www.mysql.com/Downloads/Contrib/mysqlsync, mysqlsync-1.0-alpha.tar.gz}. -A Perl script to keep remote copies of a @strong{MySQL} database in sync with a -central master copy. By Mark Jeftovic. @email{markjr@@easydns.com}. +@page +@cindex environment variables, list of +@node Environment variables, Users, MySQL internals, Top +@appendix Environment Variables -@item @uref{http://www.mysql.com/Downloads/Contrib/MySQLTutor-0.2.tar.gz, MySQLTutor}. -MySQLTutor. A @strong{MySQL} tutorial for beginners. +Here is a list of all the environment variables that are used directly or +indirectly by @strong{MySQL}. Most of these can also be found in other +places in this manual. -@item @uref{http://www.mysql.com/Downloads/Contrib/MySQLDB.zip, MySQLDB.zip} -A COM library for @strong{MySQL} by Alok Singh. -@item @uref{http://www.mysql.com/Downloads/Contrib/MySQLDB-readme.html, MySQLDB-readme.html}. +Note that any options on the command line will take precedence over +values specified in configuration files and environment variables, and +values in configuration files take precedence over values in environment +variables. -@item @uref{http://www.mysql.com/Downloads/Contrib/mysql_replicate.pl, mysql_replicate.pl} -Perl program that handles replication. By @email{elble@@icculus.nsg.nwu.edu} +In many cases it's preferable to use a configure file instead of environment +variables to modify the behavior of @strong{MySQL}. @xref{Option files}. -@item @uref{http://www.mysql.com/Downloads/Contrib/DBIx-TextIndex-0.02.tar.gz, DBIx-TextIndex-0.02.tar.gz} -Perl script that uses reverse indexing to handle text searching. -By Daniel Koch. +@tindex CCX environment variable +@tindex Environment variable, CCX +@tindex CC environment variable +@tindex Environment variable, CC +@tindex CFLAGS environment variable +@tindex Environment variable, CFLAGS +@tindex CXXFLAGS environment variable +@tindex Environment variable, CXXFLAGS +@tindex DBI_USER environment variable +@tindex Environment variable, DBI_USER +@tindex DBI_TRACE environment variable +@tindex Environment variable, DBI_TRACE +@tindex HOME environment variable +@tindex Environment variable, HOME +@tindex LD_RUN_PATH environment variable +@tindex Environment variable, LD_RUN_PATH +@tindex MYSQL_DEBUG environment variable +@tindex Environment variable, MYSQL_DEBUG +@tindex MYSQL_HISTFILE environment variable +@tindex Environment variable, MYSQL_HISTFILE +@tindex MYSQL_HOST environment variable +@tindex Environment variable, MYSQL_HOST +@tindex MYSQL_PWD environment variable +@tindex Environment variable, MYSQL_PWD +@tindex MYSQL_TCP_PORT environment variable +@tindex Environment variable, MYSQL_TCP_PORT +@tindex MYSQL_UNIX_PORT environment variable +@tindex Environment variable, MYSQL_UNIX_PORT +@tindex PATH environment variable +@tindex Environment variable, PATH +@tindex TMPDIR environment variable +@tindex Environment variable, TMPDIR +@tindex TZ environment variable +@tindex Environment variable, TZ +@tindex UMASK_DIR environment variable +@tindex Environment variable, UMASK_DIR +@tindex UMASK environment variable +@tindex Environment variable, UMASK +@tindex USER environment variable +@tindex Environment variable, USER -@item @uref{http://www.mysql.com/Downloads/Contrib/dbcheck, dbcheck} -Perl script that takes a backup of tables before running isamchk on them. -By Elizabeth. +@multitable @columnfractions .2 .8 +@item @code{CCX} @tab Set this to your C++ compiler when running configure. +@item @code{CC} @tab Set this to your C compiler when running configure. +@item @code{CFLAGS} @tab Flags for your C compiler when running configure. +@item @code{CXXFLAGS} @tab Flags for your C++ compiler when running configure. +@item @code{DBI_USER} @tab The default user name for Perl DBI. +@item @code{DBI_TRACE} @tab Used when tracing Perl DBI. +@item @code{HOME} @tab The default path for the @code{mysql} history file is @file{$HOME/.mysql_history}. +@item @code{LD_RUN_PATH} @tab Used to specify where your @code{libmysqlclient.so} is. +@item @code{MYSQL_DEBUG} @tab Debug-trace options when debugging. +@item @code{MYSQL_HISTFILE} @tab The path to the @code{mysql} history file. +@item @code{MYSQL_HOST} @tab Default host name used by the @code{mysql} command-line prompt. +@item @code{MYSQL_PWD} @tab The default password when connecting to @code{mysqld}. Note that use of this is insecure! +@item @code{MYSQL_TCP_PORT} @tab The default TCP/IP port. +@item @code{MYSQL_UNIX_PORT} @tab The default socket; used for connections to @code{localhost}. +@item @code{PATH} @tab Used by the shell to finds the @strong{MySQL} programs. +@item @code{TMPDIR} @tab The directory where temporary tables/files are created. +@item @code{TZ} @tab This should be set to your local time zone. @xref{Timezone problems}. +@item @code{UMASK_DIR} @tab The user-directory creation mask when creating directories. Note that this is ANDed with @code{UMASK}! +@item @code{UMASK} @tab The user-file creation mask when creating files. +@item @code{USER} @tab The default user on Windows to use when connecting to @code{mysqld}. +@end multitable -@item @uref{http://www.mysql.com/Downloads/Contrib/mybackup}. -@item @uref{http://www.mswanson.com/mybackup, mybackup home page} -Wrapper for mysqldump to backup all databases. By Marc Swanson. +@page +@cindex users, of MySQL +@cindex news sites +@node Users, MySQL customer usage, Environment variables, Top +@appendix Some MySQL Users -@item @uref{http://www.mysql.com/Downloads/Contrib/mdu.pl.gz,mdu.pl.gz} -Prints the storage usage of a @strong{MySQL} database. -@end itemize +This appendix lists some users of @strong{MySQL} that have given us permission +to list them in our documentation. It is by far not a complete list, but +should give you a general idea of who uses @strong{MySQL} and what it can +be used for. -@cindex RPMs, for common tools -@cindex tools, RPMs for -@appendixsec RPMs for Common Tools (Most Are for RedHat 6.1) +@appendixsec General News Sites @itemize @bullet -@item @uref{http://www.mysql.com/Downloads/Contrib/perl-Data-ShowTable-3.3-2.i386.rpm,perl-Data-ShowTable-3.3-2.i386.rpm} -@item @uref{http://www.mysql.com/Downloads/Contrib/perl-Msql-Mysql-modules-1.2210-2.i386.rpm,perl-Msql-Mysql-modules-1.2210-2.i386.rpm} -@item @uref{http://www.mysql.com/Downloads/Contrib/php-pg-3.0.13-1.i386.rpm,php-pg-3.0.13-1.i386.rpm} -@item @uref{http://www.mysql.com/Downloads/Contrib/php-pg-manual-3.0.13-1.i386.rpm,php-pg-manual-3.0.13-1.i386.rpm} -@item @uref{http://www.mysql.com/Downloads/Contrib/php-pg-mysql-3.0.13-1.i386.rpm,php-pg-mysql-3.0.13-1.i386.rpm} -@item @uref{http://www.mysql.com/Downloads/Contrib/phpMyAdmin-2.0.5-1.noarch.rpm,phpMyAdmin-2.0.5-1.noarch.rpm} -@end itemize -@cindex functions, useful -@appendixsec Useful Functions -@itemize @bullet -@item @uref{http://www.mysql.com/Downloads/Contrib/mysnprintf.c,mysnprintf.c} -sprintf() function for SQL queries that can escape blobs. By Chunhua Liu. +@item @uref{http://www.yahoo.com/, Yahoo!} + +@item @uref{http://slashdot.org/, Slashdot: A pro-Linux/tech news and comment/discussion site} + +@item @uref{http://www.linux.com/, All about Linux} + +@item @uref{http://www.linuxtoday.com/, Linuxtoday} + +@item @uref{http://www.32bitsonline.com/, 32Bits Online: because there's +more than one way to compute} + +@item @uref{http://www.freshmeat.net/, Freshmeat: News about new versions of computer-related stuff} + @end itemize -@appendixsec Windows programs +@cindex search engines, web +@cindex web search engines +@appendixsec Some Web Search Engines + @itemize @bullet -@item @uref{http://www.mysql.com/Downloads/Contrib/LaunchMySQL.zip, LaunchMySQL.zip} -The program launches the @strong{MySQL} server, shuts it down, and -display status information. By Bill Thompson + +@item @uref{http://www.aaa.com.au, AAA Matilda Web Search} + +@item @uref{http://www.whatsnu.com/, What's New} + +@item @uref{http://www.aladin.de/, Aladin} + +@item @uref{http://www.columbus-finder.de/, Columbus Finder} + +@item @uref{http://www.spider.de/, Spider} + +@item @uref{http://www.blitzsuche.de/, Blitzsuche} + +@item @uref{http://www.indoseek.co.id, Indoseek Indonesia} + +@item @uref{http://www.yaboo.dk/, Yaboo - Yet Another BOOkmarker} + + +@item @uref{http://www.ozsearch.com.au, OzSearch Internet Guide} + +@item @uref{http://www.splatsearch.com/, Splat! Search} + +@item @uref{http://osdls.library.arizona.edu/, The Open Source Digital Library System Project} @end itemize -@appendixsec Uncategorized +@appendixsec Some Information Search Engines Concentrated on Some Area @itemize @bullet -@item @uref{http://www.mysql.com/Downloads/Contrib/findres.pl, findres.pl} -Find reserved words in tables. By Nem W Schlecht. +@item @uref{http://www.spylog.ru/, SpyLOG ; A very popular Web counter site} -@item @uref{http://www.mysql.com/Downloads/Contrib/handicap.tar.gz, handicap.tar.gz} -Performance handicapping system for yachts. Uses PHP. By -@email{rhill@@stobyn.ml.org}. +@item @uref{http://www.tucows.com/, TuCows Network; Free Software archive} -@item @uref{http://www.mysql.com/Downloads/Contrib/hylalog-1.0.tar.gz, hylalog-1.0.tar.gz} -Store @code{hylafax} outgoing faxes in a @strong{MySQL} database. By Sinisa -Milivojevic, @email{sinisa@@mysql.com}. +@item @uref{http://www.jobvertise.com,Jobvertise: Post and search for jobs} -@item @uref{http://www.mysql.com/Downloads/Contrib/mrtg-mysql-1.0.tar.gz, mrtg-mysql-1.0.tar.gz} -@strong{MySQL} status plotting with MRTG, by Luuk de Boer, @email{luuk@@wxs.nl}. +@item @uref{http://www.musicdatabase.com, The Music Database} -@item @uref{http://www.mysql.com/Downloads/Contrib/wuftpd-2.4.2.18-mysql_support.2.tar.gz, wuftpd-2.4.2.18-mysql_support.2.tar.gz} -Patches to add logging to @strong{MySQL} for WU-ftpd. By Zeev Suraski, -@email{bourbon@@netvision.net.il}. +@item @uref{http://www.soccersearch.com, Football (Soccer) search page} -@item @uref{http://www.mysql.com/Downloads/Contrib/wu-ftpd-2.6.0-mysql.4.tar.gz,wu-ftpd-2.6.0-mysql.4.tar.gz} -Patches to add logging to @strong{MySQL} for WU-ftpd 2.6.0. By -@email{takeshi@@SoftAgency.co.jp}, based on Zeev Suraski wuftpd patches. +@item @uref{http://www.headrush.net/takedown, TAKEDOWN - wrestling} -@item @uref{http://www.mysql.com/Downloads/Contrib/Old-Versions, Old-Versions} -Previous versions of things found here that you probably won't be -interested in. -@end itemize +@item @uref{http://www.lyrics.net, The International Lyrics Network} -@page -@node Credits, News, Contrib, Top -@appendix Credits +@item @uref{http://TheMatrix.com/~matrix/band_search.phtml, Musicians looking for other musicians (Free Service)} -This appendix lists the developers, contributors, and supporters that have -helped to make @strong{MySQL} what it is today. +@item @uref{http://www.addall.com/AddBooks/Stores.html,AddALL books searching and price comparison} -@cindex developers, list of -@menu -* Developers:: -* Contributors:: -* Supporters:: -@end menu +@item @uref{http://www.herbaria.harvard.edu/Data/Gray/gray.html,Harvard's Gray Herbarium Index of Plant Names} -@node Developers, Contributors, Credits, Credits -@appendixsubsec Developers at MySQL AB +@item @uref{http://www.game-developer.com/,The Game Development Search Engine} -These are the developers that are or have been employed by @strong{MySQL AB} -to work on @strong{MySQL}, roughly in the order they started to work with us. -Following each developer is a small list of the tasks that the developer is -responsible for, or the accomplishments they have made. +@item @uref{www.theinnkeeper.com, The Innkeeper Vacation Guides} -@table @asis -@item Michael (Monty) Widenius -Has written the following parts of @strong{MySQL}: -@itemize @bullet -@item -All the main code in @code{mysqld}. -@item -New functions for the string library. -@item -Most of the @code{mysys} library. -@item -The @code{ISAM} and @code{MyISAM} libraries (B-tree index file -handlers with index compression and different record formats). -@item -The @code{HEAP} library. A memory table system with our superior full dynamic -hashing. In use since 1981 and published around 1984. -@item -The @code{replace} program (look into it, it's COOL!). -@item -@strong{MyODBC}, the ODBC driver for Windows95. -@item -Fixing bugs in MIT-pthreads to get it to work for @strong{MySQL}. And -also Unireg, a curses-based application tool with many utilities. -@item -Porting of @code{mSQL} tools like @code{msqlperl}, @code{DBD}/@code{DBI}, and -@code{DB2mysql}. -@item -Most of crash-me and the foundation for the @strong{MySQL} benchmarks. -@end itemize +@item @uref{http://www.macgamedatabase.com/, The Mac Game Database uses PHP and MySQL} +@c From: Marc Antony Vose <suzerain@suzerain.com> -@item David Axmark -@itemize @bullet -@item -Coordinator and initial main writer of the @strong{Reference Manual}, -including enhancements to @code{texi2html}. -@item -Automatic Web site updating from the manual. -@item -Initial Autoconf, Automake, and Libtool support. -@item -The licensing stuff. -@item -Parts of all the text files. (Nowadays only the @file{README} is -left. The rest ended up in the manual.) -@item -Lots of testing of new features. -@item -Our in-house ``free'' software lawyer. -@item -Mailing list maintainer (who never has the time to do it right...) -@item -Our original portability code (more than 10 years old now). Nowadays -only some parts of @code{mysys} are left. -@item -Someone for Monty to call in the middle of the night when he just got -that new feature to work. -@end itemize +@item @uref{http://www.csse.monash.edu.au/publications/, Research +Publications at Monash University in Australia} -@item Jani Tolonen -@itemize @bullet -@item -@code{mysqlimport} -@item -A lot of extensions to the @code{mysql} client. -@item -@code{procedure analyse()} -@end itemize +@item @uref{http://www.ipielle.emr.it/bts/index.html, +Occupational Health & Safety Web site database (a project for the ECC)} +@c c.presutti@ipielle.emr.it -@item Sinisa Milivojevic -@itemize @bullet -@item -Compression (with @code{zlib}) in the client/server protocol. -@item -Perfect hashing for the lexical analyzer phase. -@item -The MySQLGUI client. -@item -Maintainer of mysql++. +@item @uref{http://data.mch.mcgill.ca/, Bioinformatics databases at the +Montreal Children's Hospital using MySQL} +@c saeed@www.debelle.mcgill.ca @end itemize -@item Tonu Samuel -@itemize @bullet -@item -Our security expert. -@item -Vio interface (The foundation for the encrypted client/server protocol). -@item -@strong{MySQL} Filesystem (A way to use @strong{MySQL} databases as files -and directories). -@item -The CASE Expression. -@item -The MD5() and COALESCE() functions. -@item -@code{RAID} support for @code{MyISAM} tables. -@end itemize +@cindex online magazines +@cindex magazines, online +@appendixsec Online Magazines -@item Sasha Pachev @itemize @bullet -@item -Replication. -@item -@code{SHOW CREATE TABLE}. -@item -mod_mysql_include -@item -cgi++ -@item -mysql-bench +@item @uref{http://www.spoiler.com, Spoiler Webzine}. +An online magazine featuring music, literature, arts, and design content. +@item @uref{http://www.linux-magazin.de/newsflash/, Daily news about Linux in German language} +@item @uref{http://www.betazine.com,Betazine - The Ultimate Online Beta Tester's Magazine} +@item @uref{http://www.currents.net/ccinfo/aboutcc.html,Computer Currents Magazine} @end itemize -@item Matt Wagner -@itemize @bullet -@item -@strong{MySQL} test suite. -@item -Our webmaster. -@end itemize +@cindex web sites +@appendixsec Web Sites that Use MySQL as a Backend -@item Miguel Solorzano @itemize @bullet -@item -Winmysqladmin. -@end itemize -@item Timothy Smith -@itemize @bullet -@item -Dynamic character support. -@item -Responsible for @strong{MySQL} configure. -@end itemize +@item @uref{http://liftoff.msfc.nasa.gov, NASA} +@item @uref{http://kids.msfc.nasa.gov, NASA KIDS} +@item @uref{http://science.nasa.gov, Sience@@NASA} -@item Sergei Golubchik -@itemize @bullet -@item -Full-text search. -@item -Added keys to the @code{MERGE} library. -@end itemize +@item @uref{http://www.handy.de/, handy.de} -@item Jeremy Cole -@itemize @bullet -@item -Proofreading and editing this fine manual. -@item -@code{ALTER TABLE ... ORDER BY ...}. -@item -@code{UPDATE ... ORDER BY ...}. -@item -@code{DELETE ... ORDER BY ...}. -@end itemize +@item @uref{http://lindev.jmc.tju.edu/qwor, Qt Widget and Object Repository} + +@item @uref{http://www.samba-choro.com.br, Brazilian samba site (in Portuguese)} + +@item @uref{http://pgss.iss.uw.edu.pl/en_index.ISS, Polish General Social Survey} + +@item @uref{http://www.expo2000.com, Expo2000} World-wide distribution of +tickets for this event is implemented using @strong{MySQL} and tcl/tk. More than +5000 travel agencies all over the world have access to it. + +@item @uref{http://www.freevote.com/, FreeVote.com is a free voting +service with millions of users.} + +@item @uref{http://f1.tauzero.se, Forza Motorsport} + +@item @uref{http://www.dreamhost.com/, DreamHost Web Hosting} -@item John Dean -@itemize @bullet -@item -The new @strong{MySQL} GUI client. @end itemize -@item Indrek Siitan +@cindex services +@appendixsec Some Domain/Internet/Web and Related Services + @itemize @bullet -@item -Designer/programmer of our web interface. -@end itemize -@end table -The following non-developers are also working in/with @strong{MySQL AB}: +@item @uref{http://www.wix.com/mysql-hosting, Registry of Web providers that +support @strong{MySQL}} -@table @asis -@item -Hans Kierkegaard - Responsible for @strong{MySQL} license handling. -@item -Antti Halonen - Sales manager. -@item -Jonas Norrman - Handles licensing questions sent to @email{info@@mysql.com}. -@item -Erik Granberg - Handles @strong{MySQL} partners (and a lot of other stuff). -@item -Allan Larsson (The BOSS for TCX DataKonsult AB). -@end table +@item @uref{http://www.yi.org/, Dynamic DNS Services} -@cindex contributors, list of -@node Contributors, Supporters, Developers, Credits -@appendixsubsec Contributors to MySQL +@item @uref{http://www.dynodns.net/, Dynamic domain name service} -Contributors to the @strong{MySQL} distribution are listed here, in -somewhat random order: +@item @uref{http://www.ods.org/, Open DNS Project; free dynamic DNS service} -@table @asis -@item Paul DuBois -Help with making the Reference Manual correct and understandable. That -includes rewriting Monty's and David's attempts at English into English -as other people know it. -@item Gianmassimo Vigazzola @email{qwerg@@mbox.vol.it} or @email{qwerg@@tin.it} -The initial port to Win32/NT. -@item Kim Aldale -Helped to rewrite Monty's and David's early attempts at English into -English. -@item Per Eric Olsson -For more or less constructive criticism and real testing of the dynamic -record format. -@item Irena Pancirov @email{irena@@mail.yacc.it} -Win32 port with Borland compiler. @code{mysqlshutdown.exe} and -@code{mysqlwatch.exe} -@item David J. Hughes -For the effort to make a shareware SQL database. We at TcX started with -@code{mSQL}, but found that it couldn't satisfy our purposes so instead we -wrote a SQL interface to our application builder Unireg. @code{mysqladmin} -and @code{mysql} are programs that were largely influenced by their -@code{mSQL} counterparts. We have put a lot of effort into making the -@strong{MySQL} syntax a superset of @code{mSQL}. Many of the API's ideas are -borrowed from @code{mSQL} to make it easy to port free @code{mSQL} programs -to @strong{MySQL}. @strong{MySQL} doesn't contain any code from @code{mSQL}. -Two files in the distribution (@file{client/insert_test.c} and -@file{client/select_test.c}) are based on the corresponding (non-copyrighted) -files in the @code{mSQL} distribution, but are modified as examples showing -the changes necessary to convert code from @code{mSQL} to @strong{MySQL}. -(@code{mSQL} is copyrighted David J. Hughes.) -@item Fred Fish -For his excellent C debugging and trace library. Monty has made a number -of smaller improvements to the library (speed and additional options). -@item Richard A. O'Keefe -For his public domain string library. -@item Henry Spencer -For his regex library, used in @code{WHERE column REGEXP regexp}. -@item Free Software Foundation -From whom we got an excellent compiler (@code{gcc}), the @code{libc} library -(from which we have borrowed @file{strto.c} to get some code working in Linux), -and the @code{readline} library (for the @code{mysql} client). -@item Free Software Foundation & The XEmacs development team -For a really great editor/environment used by almost everybody at -TcX/MySQL AB/detron. -@item Patrick Lynch -For helping us acquire @code{http://www.mysql.com/}. -@item Fred Lindberg -For setting up qmail to handle the @strong{MySQL} mailing list and for the -incredible help we got in managing the @strong{MySQL} mailing lists. -@item Igor Romanenko @email{igor@@frog.kiev.ua} -@code{mysqldump} (previously @code{msqldump}, but ported and enhanced by -Monty). -@item Tim Bunce, Alligator Descartes -For the @code{DBD} (Perl) interface. -@item Tim Bunce -Author of @code{mysqlhotcopy}. -@item Andreas Koenig @email{a.koenig@@mind.de} -For the Perl interface to @strong{MySQL}. -@item Eugene Chan @email{eugene@@acenet.com.sg} -For porting PHP to @strong{MySQL}. -@item Michael J. Miller Jr. @email{mke@@terrapin.turbolift.com} -For the first @strong{MySQL} manual. And a lot of spelling/language -fixes for the FAQ (that turned into the @strong{MySQL} manual a long -time ago). -@item Yan Cailin -First translator of the @strong{MySQL} Reference Manual into simplified -chinese in early 2000 on which the Big5 and HK coded -(@uref{http://mysql.hitstar.com, mysql.hitstar.com}) versions were -based. @uref{http://linuxdb.yeah.net, Personal home page at -linuxdb.yeah.net}. -@item Giovanni Maruzzelli @email{maruzz@@matrice.it} -For porting iODBC (Unix ODBC). -@item Chris Provenzano -Portable user level pthreads. From the copyright: This product includes -software developed by Chris Provenzano, the University of California, -Berkeley, and contributors. We are currently using version 1_60_beta6 -patched by Monty (see @file{mit-pthreads/Changes-mysql}). -@item Xavier Leroy @email{Xavier.Leroy@@inria.fr} -The author of LinuxThreads (used by @strong{MySQL} on Linux). -@item Zarko Mocnik @email{zarko.mocnik@@dem.si} -Sorting for Slovenian language and the @file{cset.tar.gz} module that makes -it easier to add other character sets. -@item "TAMITO" @email{tommy@@valley.ne.jp} -The @code{_MB} character set macros and the ujis and sjis character sets. -@item Joshua Chamas @email{joshua@@chamas.com} -Base for concurrent insert, extended date syntax, debugging on NT, and -answering on the @strong{MySQL} mailing list. -@item Yves Carlier @email{Yves.Carlier@@rug.ac.be} -@code{mysqlaccess}, a program to show the access rights for a user. -@item Rhys Jones @email{rhys@@wales.com} (And GWE Technologies Limited) -For the JDBC, a module to extract data from @strong{MySQL} with a Java -client. -@item Dr Xiaokun Kelvin ZHU @email{X.Zhu@@brad.ac.uk} -Further development of the JDBC driver and other @strong{MySQL}-related -Java tools. -@item James Cooper @email{pixel@@organic.com} -For setting up a searchable mailing list archive at his site. -@item Rick Mehalick @email{Rick_Mehalick@@i-o.com} -For @code{xmysql}, a graphical X client for @strong{MySQL}. -@item Doug Sisk @email{sisk@@wix.com} -For providing RPM packages of @strong{MySQL} for RedHat Linux. -@item Diemand Alexander V. @email{axeld@@vial.ethz.ch} -For providing RPM packages of @strong{MySQL} for RedHat Linux-Alpha. -@item Antoni Pamies Olive @email{toni@@readysoft.es} -For providing RPM versions of a lot of @strong{MySQL} clients for Intel -and SPARC. -@item Jay Bloodworth @email{jay@@pathways.sde.state.sc.us} -For providing RPM versions for @strong{MySQL} Version 3.21. -@item Jochen Wiedmann @email{wiedmann@@neckar-alb.de} -For maintaining the Perl @code{DBD::mysql} module. -@item Therrien Gilbert @email{gilbert@@ican.net}, Jean-Marc Pouyot @email{jmp@@scalaire.fr} -French error messages. -@item Petr snajdr, @email{snajdr@@pvt.net} -Czech error messages. -@item Jaroslaw Lewandowski @email{jotel@@itnet.com.pl} -Polish error messages. -@item Miguel Angel Fernandez Roiz -Spanish error messages. -@item Roy-Magne Mo @email{rmo@@www.hivolda.no} -Norwegian error messages and testing of Version 3.21.#. -@item Timur I. Bakeyev @email{root@@timur.tatarstan.ru} -Russian error messages. -@item @email{brenno@@dewinter.com} && Filippo Grassilli @email{phil@@hyppo.com} -Italian error messages. -@item Dirk Munzinger @email{dirk@@trinity.saar.de} -German error messages. -@item Billik Stefan @email{billik@@sun.uniag.sk} -Slovak error messages. -@item Stefan Saroiu @email{tzoompy@@cs.washington.edu} -Romanian error messages. -@item Peter Feher -Hungarian error messages. -@item Roberto M. Serqueira -Portugise error messages. -@item David Sacerdote @email{davids@@secnet.com} -Ideas for secure checking of DNS hostnames. -@item Wei-Jou Chen @email{jou@@nematic.ieo.nctu.edu.tw} -Some support for Chinese(BIG5) characters. -@item Wei He @email{hewei@@mail.ied.ac.cn} -A lot of functionality for the Chinese(GBK) character set. -@item Zeev Suraski @email{bourbon@@netvision.net.il} -@code{FROM_UNIXTIME()} time formatting, @code{ENCRYPT()} functions, and -@code{bison} advisor. -Active mailing list member. -@item Luuk de Boer @email{luuk@@wxs.nl} -Ported (and extended) the benchmark suite to @code{DBI}/@code{DBD}. Have -been of great help with @code{crash-me} and running benchmarks. Some new -date functions. The mysql_setpermissions script. -@item Jay Flaherty @email{fty@@mediapulse.com} -Big parts of the Perl @code{DBI}/@code{DBD} section in the manual. -@item Paul Southworth @email{pauls@@etext.org}, Ray Loyzaga @email{yar@@cs.su.oz.au} -Proof-reading of the Reference Manual. -@item Alexis Mikhailov @email{root@@medinf.chuvashia.su} -User-definable functions (UDFs); @code{CREATE FUNCTION} and -@code{DROP FUNCTION}. -@item Andreas F. Bobak @email{bobak@@relog.ch} -The @code{AGGREGATE} extension to UDF functions. -@item Ross Wakelin @email{R.Wakelin@@march.co.uk} -Help to set up InstallShield for @strong{MySQL}-Win32. -@item Jethro Wright III @email{jetman@@li.net} -The @file{libmysql.dll} library. -@item James Pereria @email{jpereira@@iafrica.com} -Mysqlmanager, a Win32 GUI tool for administrating @strong{MySQL}. -@item Curt Sampson @email{cjs@@portal.ca} -Porting of MIT-pthreads to NetBSD/Alpha and NetBSD 1.3/i386. -@item Antony T. Curtis @email{antony.curtis@@olcs.net} -Porting of @strong{MySQL} to OS/2. -@item Martin Ramsch @email{m.ramsch@@computer.org} -Examples in the @strong{MySQL} Tutorial. -@item Steve Harvey -For making @code{mysqlaccess} more secure. -@item Konark IA-64 Centre of Persistent Systems Private Limited -@uref{http://www.pspl.co.in/konark/}. Help with the Win64 port of the -@strong{MySQL} server. -@item Albert Chin-A-Young. -Configure updates for Tru64, large file support and better TCP wrappers -support. -@item Valueclick Inc. -For sponsoring the optimize section in this manual. -@end table +@c @item @uref{http://dynodns.net, Free dynamic DNS implementation} +@c EMAIL: A Moore <amoore@mooresystems.com> + +@item @uref{http://www.hn.org/, Hammernode; Public DNS Servers} + +@item @uref{http://www.fdns.net/, Free 3rd level domains} + +@item @uref{http://worldcommunity.com/, Online Database} + +@item @uref{http://www.bigbiz.com, BigBiz Internet Services} + +@item @uref{http://virt.circle.net, The Virt Gazette} + +@item @uref{http://www.california.com, Global InfoNet Inc} + +@item @uref{http://www.webhosters.com, WebHosters - A Guide to WWW Providers} + +@item @uref{http://online.dn.ru, Internet information server} + +@item @uref{http://www.stopbit.com, A technology news site} + +@item @uref{http://www.worldnetla.net, WorldNet Communications - An Internet Services Provider} + +@item @uref{http://www.netizen.com.au/, Netizen: Australian-based Web consultancy} + +@item @uref{http://www.trainingpages.co.uk, Search site for training courses in the UK} + +@item @uref{http://chat.nitco.com, Gannon Chat (GPL). Written in Perl and Javascript} + +@item @uref{http://www.addurls.com/,A general links directory} + +@item @uref{http://www.bookmarktracker.com, A Web-based bookmark management service} + +@item @uref{http://www.cdrom.com,Walnut Creek CDROM} + +@item @uref{http://www.wwwthreads.org/, WWWThreads; Interactive discussion Forums} + +@item @uref{http://pvmon.portici.enea.it/Meteo, In Italian; Storage data from meteo station} + +@item @uref{http://www.buysell.net/, Online "Person To Person" Auction} + +@item @uref{http://tips.pair.com,Tips on Web development} + +@item @uref{http://www.mailfriends.com, Mailfriends.com is a FREE service for +everybody who wants to find friends over the internet.} + +@item @uref{http://www.uninova.com/cgi-bin/wctelnets?list, Web Page Telnet BBS List} + +@item @uref{http://www.uninova.com/cnc.html,UniNova Digital Postcards} + +@c @item @uref{http://cabinboy.powersurfr.com, An Internet RFC search engine} + +@item @uref{http://www.dslreports.com, DSL-provider search with reviews}. +Made with @strong{MySQL} and Modperl, all pages are generated dynamically out of +the @strong{MySQL} database +@end itemize + +@cindex PHP, web sites +@appendixsec Web Sites that Use @code{PHP} and MySQL + +@itemize @bullet +@c @item @uref{http://www.wh200th.com, White House 200th Anniversary site} + +@item @uref{http://support.jgaa.com/, Jgaa's Internet - Official Support Site} + +@item @uref{http://io.incluso.com, Ionline - online publication:} @strong{MySQL}, +PHP, Java, Web programming, DB development + +@item @uref{http://www.baboo.com, BaBoo(Browse and bookmark). Free Web-based bookmark manager and Calendar} + +@item @uref{http://www.courses.pjc.cc.fl.us/Schedule/index.php, Course +Schedule System at Pensacola Junior College} + +@item @uref{http://www.fccj.org, Florida Community College at Jacksonville} + +@item @uref{http://www.32bit.com/, 32bit.com; An extensive shareware / freeware archive} + +@item @uref{http://www.jokes2000.com/, Jokes 2000} +@c Added 990604; EMAIL: ah@dybdahl.dk + + +@item @uref{http://www.burken.nu/ , Burken.NU} Burken is a webhotel that +provides scripts, among other things, for remote users, like counters, +guestbooks etc. +@c Added 990608; EMAIL: spacedmp@SpaceDump.Burken.NU (Anders Olausson) + +@item @uref{http://tips.pair.com, tips.pair.com} Contains tips on html, +javascript, 2d/3d graphics, and PHP3/MySQL. All pages are generated from +a database. +@c Added 990614; EMAIL: downey@image.dk (Rune Madsen) + +@item @uref{http://www.softwarezrus.com/, Softwarezrus.com} +Ecommerce site that is selling computers. +@end itemize + +@cindex consultants, list of +@appendixsec Some MySQL Consultants + +@itemize @bullet + +@item @uref{http://www.ayni.com, Ayni AG} + +@item @uref{http://worldcommunity.com/, Online Database} + +@item @uref{http://www2.dataguard.no/,DataGuard (Uses @strong{MySQL} and PHP)} + +@item @uref{http://wwits.net/programs/mysql.phtml, WWITS (Uses @strong{MySQL} and PHP)} + +@item @uref{http://www.worldcommunity.com/, WCN - The World Community Network} + +@item @uref{http://www.chipcastle.com, Chip Castle Dot Com Inc} +@c Added 990603 EMAIL: chip@chipcastle.com (Chip Castle) + +@item @uref{http://www.cyber.com.au/, Cybersource Pty. Ltd} + +@item @uref{http://www.spring.de, Spring infotainment gmbh & co. kg} +@c added 990905 "Oliver Pischke" <opischke@spring.de> + +@item @uref{http://www.wamdesign.com/, Develops Web sites using MySQL} +@c Added 990905; max@wamdesign.com + +@item @uref{http://www.berkeleyconsultants.com, Berkeley Consultants Group} + +@item @uref{http://www.jammconsulting.com/, JAMM Consulting Inc.} + +@end itemize + +@appendixsec Programming + +@cindex web pages, miscellaneous +@appendixsec Uncategorized Pages + +@itemize @bullet + +@item @uref{http://www.feature-showcase.com/htmls/demo_mysql.sql, +AZC.COM's Feature Showcase} + +@item @uref{http://www.teach.org.uk/subjects/trainingcourse/g.html, Course Search} + +@item @uref{http://www.northerbys.com, Northerbys Online Auctions} + +@item @uref{http://www.schiphol.nl/flights/home.htm, Amsterdam Airport Schiphol} + +@item @uref{http://TheMatrix.com/seventhsin/query.phtml, CD database} + +@item @uref{http://TheMatrix.com/~flmm/GEAR.html, Used Audio Gear Database} + +@item @uref{http://www.kiss.de/musik-mueller, Musical note-sheets} + +@item @uref{http://www.bagism.com, Bagism - A John Lennon fan page} + +@item @uref{http://www.selftaught.com/, US Folk art broker} + +@item @uref{http://organizer.net/, Mail reading on the Web} + +@item @uref{http://www.mypage.org/, Free home pages on www.somecoolname.mypage.org} + +@item @uref{http://www.schulweb.de/, Der Server f@"ur Schulen im Web (In German)} + +@item @uref{http://www.ald.net/, Auldhaefen Online Services} + +@item @uref{http://www.cary.net/, CaryNET Information Center} + +@item @uref{http://www.dataden.com/, Dataden Computer Systems} + +@item @uref{http://andree.grm.se/, Andr@'emuseet (In Swedish)} + +@item @uref{http://www.him.net/, HOMESITE Internet Marketing} + +@item @uref{http://www.jade-v.com/techinfo.html, Jade-V Network Services } + +@item @uref{http://ww2010.atmos.uiuc.edu/(Gl)/abt/aknw/tech.rxml, +Weather World 2010 Technical Credits} @* + +@item @uref{http://gimp.foebud.org/registry/doc/, About The Gimp plugin registry} + +@item @uref{http://www.fast-inc.com/Products/Archiver/database.html, Java tool +Archiver technical detail (Slightly optimistic about @strong{MySQL} +ANSI-92 compliance)} + +@item @uref{http://www.gamesdomain.com/cheats/usrcheat.phtml, Games Domain Cheats Database} + +@item @uref{http://www.kcilink.com/poweredby/, The "Powered By" Page (Kcilink)} + +@item @uref{http://www.netcasting.net/index.whtml, Netcasting} + +@item @uref{http://homepages.tig.com.au/~mjj/nbltips, NBL (Australian National Basketball League) tipping} + +@item @uref{http://www.cgishop.com/, CGI shop} + +@item @uref{http://www.whirlycott.com/, Whirlycott: Website Design} + +@item @uref{http://www.mtp.dk, Museum Tusculanum Press} + +@item @uref{http://csdgi.historie.ku.dk/biblio, Centro Siciliano di Documentazione} + +@item @uref{http://caribou.dyn.ml.org:8000, Quake statistics database} + +@item @uref{http://www.astroforum.ch, Astroforum: Astrologie and related things (in German)} + +@item @uref{http://www.opendebate.com, OpenDebate - Interactive Polls & Open Discussion} + +@item @uref{http://vermeer.organik.uni-erlangen.de/dissertationen/, Online chemical dissertation server} + +@item @uref{http://www.freschinfo.com, FreSch! The Free Scholarship Search Service} + +@item @uref{http://www.nada.kth.se/~staffanu/pinball, Stockholm Pinball Locator} + +@item @uref{http://www.hek.com, HEK A construction company} + +@item @uref{http://www.nbi.nl, Elsevier Bussines Information} + +@item @uref{http://vaccination.medicallink.se/, Medical Links (Using ColdFusion and @strong{MySQL})} + +@item @uref{http://www.joblink-usa.com, Search for jobs & people at JobLink-USA} + +@item @uref{http://www.skydive.net/competfs, Competition Formation Skydiving} + +@item @uref{http://www.galaxy-net.net/Galaxy-NET Telecommunications, E-commerce and internal accounting} + +@item @uref{http://www.borsen.dk/, Denmark's leading business daily newspaper B@o{}rsen} + +@item @uref{http://tmmm.simplenet.com/indb/, The Internet NES Database} + +@item @uref{http://www.russia.cz, Travel agency in Prague in 3 languages} + +@item @uref{http://www.linkstation.de, Linkstation} + +@item @uref{http://www.peoplestaff.com, Searchable online database at Peoplestaff} + +@item @uref{http://www.dreamhorse.com, A searchable database system for horse classified ads} + +@item @uref{http://pootpoot.com/,The Poot site} + +@item @uref{http://grateful.net/hw_html/,"Playin' in the LAN"; a network monitoring suite} + +@c Update from Christopher Milton <cmilton@bwn.net> 1999-12-21 +@item @uref{http://www.usapa.army.mil,U.S. Army Publishing Agency} + +@item @uref{http://www.nekretnine.co.yu/,Realestate handling in Yugoslavia} + +@item @uref{http://demo.cpsoft.com/pims/devFAQ.html, PIMS; a Patient Information Management System} + +@item @uref{http://cpsoft.com,Pilkington Software Inc} + +@item @uref{http://www.no-quarter.org/,A Vietnam Veteran's Memorial (The Wall) database} + +@item @uref{http://www.gamers-union.com/,Gamer's Union specializes in auctions of used & out-of-print gaming material} + +@item @uref{http://www.montereyhigh.com/office/dbul.php3, A daily bulletin at Monterey High school} + +@item @uref{http://www.myEastside.com,Community-owned site serving Lake +Washington's Eastside residents and businesses} + +@item @uref{http://bowling-france.net/,French bowling site} +@end itemize + +Send any additions to this list to @email{webmaster@@mysql.com}. + +@page +@cindex MySQL usage +@node MySQL customer usage, Contrib, Users, Top +@appendix MySQL customer usage + +The section 'Some MySQL Users' contains a lot of different links to +@strong{MySQL} users but doesn't provide that much information about how +they are using @strong{MySQL}. @xref{Users}. This manual section is to +give you an idea of how other @strong{MySQL} users are using +@strong{MySQL} to solve their problems. + +This manual section is very new and we plan to add more stories here +shortly. If you are interested in contributing of how you use +@strong{MySQL} in a unique environment or have success store about how you +use @strong{MySQL}, you can write to @code{docs@@mysql.com} with +subject @code{Success:}. Note that as we are very busy it may take some +time before you get some feedback for your story. + +@itemize @bullet +@item +@strong{Peter Zaitsev of Spylog.ru} writes: +I think you might be interested in my database size. The whole database +is currently on 15 servers and I think it's about 60.000 of tables +containing about 5.000.000.000 of rows. My mostly loaded server +currently holds about 10.000 of tables with 1.000.000.000 of rows in it. +Hugest tables have about 50.000.000 of rows, and this value will raise +as soon as I'll move to 2.4 kernel with large files. Currently I have to +delete much of logs for large sites to hold table sizes in 2Gb. + +@item +@strong{Texas Instruments} is using @strong{MySQL} for handling tables +that contain up to 2,000 million rows in a validation regression database. +@end itemize + +@page +@cindex contributed programs +@cindex programs, contributed +@node Contrib, Credits, MySQL customer usage, Top +@appendix Contributed Programs + +Many users of @strong{MySQL} have contributed @emph{very} useful support +tools and add-ons. + +@ifclear web +A list of what is available at @uref{http://www.mysql.com/Downloads/Contrib/} +(or any mirror) is shown below. +If you want to build @strong{MySQL} support for the Perl @code{DBI}/@code{DBD} +interface, you should fetch the @code{Data-Dumper}, @code{DBI}, and +@code{Msql-Mysql-modules} files and install them. +@xref{Perl support}. +@end ifclear + +@appendixsec APIs + +@cindex Perl, modules +@itemize @bullet +@item Perl Modules +@itemize @minus +@item @uref{http://www.mysql.com/Downloads/Contrib/Data-Dumper-2.101.tar.gz, Data-Dumper-2.101.tar.gz} +Perl @code{Data-Dumper} module. Useful with @code{DBI}/@code{DBD} support for +older Perl installations. + +@item @uref{http://www.mysql.com/Downloads/Contrib/DBI-1.15.tar.gz, DBI-1.15.tar.gz} +Perl @code{DBI} module. + +@item @uref{http://www.mysql.com/Downloads/Contrib/KAMXbase1.2.tar.gz,KAMXbase1.2.tar.gz} +Convert between @file{.dbf} files and @strong{MySQL} tables. Perl +module written by Pratap Pereira @email{pereira@@ee.eng.ohio-state.edu}, +extended by Kevin A. McGrail @email{kmcgrail@@digital1.peregrinehw.com}. +This converter can handle MEMO fields. + +@item @uref{http://www.mysql.com/Downloads/Contrib/Msql-Mysql-modules-1.2216.tar.gz, Msql-Mysql-modules-1.2216.tar.gz} +Perl @code{DBD} module to access mSQL and @strong{MySQL} databases. + +@item @uref{http://www.mysql.com/Downloads/Contrib/Data-ShowTable-3.3.tar.gz, Data-ShowTable-3.3.tar.gz} +Perl @code{Data-ShowTable} module. Useful with @code{DBI}/@code{DBD} support. + +@item @uref{http://www.mysql.com/Downloads/Contrib/HandySQL-1.1.tar.gz, HandySQL-1.1.tar.gz} +HandySQL is a MySQL access module. It offers a C interface embedded in Perl and is +approximately 20% faster than regular DBI. + +@end itemize + +@cindex JDBC +@item JDBC +@itemize @bullet +@item @uref{http://www.mysql.com/Downloads/Contrib/mm.mysql.jdbc-1.2c.tar.gz, mm.mysql.jdbc-1.2c.tar.gz} +The mm JDBC driver for @strong{MySQL}. This is a production release +and is actively developed. By Mark Matthews +(@email{mmatthew@@ecn.purdue.edu}). + +@item @uref{http://www.mysql.com/Downloads/Contrib/mm.mysql.jdbc-2.0pre5.tar.gz, mm.mysql.jdbc-2.0pre5.tar.gz} +The mm JDBC driver for @strong{MySQL}. This is a pre-release beta version +and is actively developed. By Mark Matthews +(@email{mmatthew@@ecn.purdue.edu}). +The two drivers above have an LGPL +license. Please check @uref{http://www.worldserver.com/mm.mysql/} for +the latest drivers (and other JDBC information) because these drivers may be out of date. + +@item @uref{http://www.caucho.com/projects/jdbc-mysql/index.xtp} +The Resin commercial JDBC driver, which is released under open source. +It claims to be faster than the mm driver, but we haven't received that much +information about this yet. + +@item @uref{http://www.mysql.com/Downloads/Contrib/twz1jdbcForMysql-1.0.4-GA.tar.gz, twz1jdbcForMysql-1.0.4-GA.tar.gz} +The twz driver: A type 4 JDBC driver by Terrence W. Zellers +@email{zellert@@voicenet.com}. This is commercial but is free for +private and educational use. (Not supported anymore.) +@c no answer from server 990830 +@c You can always find the latest driver at @uref{http://www.voicenet.com/~zellert/tjFM/}. +@item @uref{http://www.mysql.com/Downloads/Contrib/pmdamysql.tgz,pmdamysql.tgz} +A @strong{MySQL} PMDA. Provides @strong{MySQL} server status and configuration +variables. +@end itemize + +@cindex OLEDB +@item OLEDB +@itemize @bullet +@item @uref{http://www.mysql.com/Downloads/Win32/MyOLEDB.exe, MyOLEDB.exe} +OLEDB handler for @strong{MySQL}. By SWsoft. +@item @uref{http://www.mysql.com/Downloads/Win32/MySamples.zip, MySamples.zip} +Examples and documentation for MyOLEDB. By SWsoft. +@item @uref{http://www.mysql.com/Downloads/Win32/Myoledb.zip, Myoledb.zip} +Source for MyOLEDB. By SWsoft. +@item @uref{http://www.mysql.com/Downloads/Win32/MyOLEDB.chm, MyOLEDB.chm} +Help files for MyOLEDB. +@item @uref{http://www.mysql.com/Downloads/Win32/libmyodbc.zip, libmyodbc.zip} +Static MyODBC library used for build MyOLEDB. Based on MyODBC code. +@end itemize + +@cindex C++ +@item C++ +@itemize @bullet + +@item @uref{http://www.mysql.com/Downloads/Contrib/mysql-c++-0.02.tar.gz, mysql-c++-0.02.tar.gz} +@strong{MySQL} C++ wrapper library. By Roland Haenel, +@email{rh@@ginster.net}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/MyDAO.tar.gz, MyDAO} +@strong{MySQL} C++ API. By Satish @email{spitfire@@pn3.vsnl.net.in}. Inspired +by Roland Haenel's C++ API and Ed Carp's MyC library. + +@item @uref{http://www.mysql.com/download_mysql++.html, mysql++} +@strong{MySQL} C++ API (More than just a wrapper library.) Originally by +@email{kevina@@clark.net}. Nowadays maintained by Sinisa at @strong{MySQL AB}. + +@item @uref{http://nelsonjr.homepage.com/NJrAPI,NJrAPI} +A C++ database independent library that supports @strong{MySQL}. +@end itemize + +@cindex Delphi +@item Delphi +@itemize @bullet +@item @uref{http://www.mysql.com/Downloads/Contrib/delphi-interface.gz, delphi-interface.gz} +Delphi interface to @code{libmysql.dll}, by Blestan Tabakov, +@email{root@@tdg.bis.bg}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/DelphiMySQL2.zip, DelphiMySQL2.zip} +Delphi interface to @code{libmysql.dll}, by @email{bsilva@@umesd.k12.or.us}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/Udmysel.pas, Udmysql.pas} +A wrapper for libmysql.dll for usage in Delphi. By Reiner Sombrowsky. + +@item @uref{http://www.fichtner.net/delphi/mysql.delphi.phtml, A Delphi interface to @strong{MySQL}.} +With source code. By Matthias Fichtner. + +@item @uref{http://www.productivity.org/projects/mysql/, @strong{TmySQL} +A library to use @strong{MySQL} with Delphi}. + +@item @uref{http://www.geocities.com/CapeCanaveral/2064/mysql.html, Delphi TDataset-component}. +@item +@item @uref{http://www.mysql.com/Downloads/Contrib/Win32/SBMySQL50Share.exe, Delphi 5 Shareware MySQL Dataset Components} +@end itemize + +@item @uref{http://www.mysql.com/Downloads/Contrib/mysql-ruby-2.2.0.tar.gz, mysql-ruby-2.2.0.tar.gz} +@strong{MySQL} Ruby module. By TOMITA Masahiro @email{tommy@@tmtm.org} +@uref{http://www.netlab.co.jp/ruby/. Ruby} is an Object-Oriented Interpreter Language. + +@item @uref{http://www.mysql.com/Downloads/Contrib/JdmMysqlDriver-0.1.0.tar.gz,JdmMysqlDriver-0.1.0.tar.gz} +A VisualWorks 3.0 Smalltalk driver for @strong{MySQL}. By +@email{joshmiller@@earthlink.net}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/Db.py, Db.py} +Python module with caching. By @email{gandalf@@rosmail.com}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/MySQLmodule-1.4.tar.gz, MySQLmodule-1.4.tar.gz} +Python interface for @strong{MySQL}. By Joseph Skinner @email{joe@@earthlight.co.nz}. Modified by Joerg Senekowitsch @email{senekow@@ibm.net}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/MySQL-python-0.3.0.tar.gz, MySQL-python-0.3.0.tar.gz} +MySQLdb Python is an DB-API v2.0-compliant interface to MySQL. Transactions +are supported if the server and tables support them. It is +thread-safe, and contains a compatibility module for older code +written for the no-longer-maintained MySQLmodule interface. + +@item @uref{http://www.mysql.com/Downloads/Contrib/mysql_mex_12.tar.gz, mysql_mex_1_12.tar.gz} +An interface program for the Matlab program by MathWorks. The interface +is done by Kimmo Uutela and John Fisher (not by Mathworks). +Check @uref{http://boojum.hut.fi/~kuutela/mysqlmex.html,mysqlmex.html} +for more information. + +@item @uref{http://www.mysql.com/Downloads/Contrib/mysqltcl-1.53.tar.gz, mysqltcl-1.53.tar.gz} +Tcl interface for @strong{MySQL}. Based on @file{msqltcl-1.50.tar.gz}. +Updated by Tobias Ritzau, @email{tobri@@ida.liu.se}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/MyC-0.1.tar.gz, MyC-0.1.tar.gz} +A Visual Basic-like API, by Ed Carp. + +@item @uref{http://www.mysql.com/Downloads/Contrib/sqlscreens-1.0.1.tar.gz, sqlscreens-1.0.1.tar.gz} +Tcl/Tk code to generate database screens. By Jean-Francois Dockes. + +@item @uref{http://www.mysql.com/Downloads/Contrib/Vdb-dflts-2.1.tar.gz, Vdb-dflts-2.1.tar.gz} +This is a new version of a set of library utilities intended +to provide a generic interface to SQL database engines such that your +application becomes a 3-tiered application. The advantage is that you +can easily switch between and move to other database engines by +implementing one file for the new backend without making any +changes to your applications. By @email{damian@@cablenet.net}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/DbFramework-1.10.tar.gz, DbFramework-1.10.tar.gz} +DbFramework is a collection of classes for manipulating @strong{MySQL} +databases. The classes are loosely based on the CDIF Data Model +Subject Area. By Paul Sharpe @email{paul@@miraclefish.com}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/pike-mysql-1.4.tar.gz, pike-mysql-1.4.tar.gz} +@strong{MySQL} module for pike. For use with the Roxen web server. + +@item @uref{http://www.mysql.com/Downloads/Contrib/squile.tar.gz, squile.tar.gz} +Module for @code{guile} that allows @code{guile} to interact with SQL +databases. By Hal Roberts. + +@item @uref{http://www.mysql.com/Downloads/Contrib/stk-mysql.tar.gz, stk-mysql.tar.gz} +Interface for Stk. Stk is the Tk widgets with Scheme underneath instead of Tcl. +By Terry Jones. + +@item @uref{http://www.mysql.com/Downloads/Contrib/eiffel-wrapper-1.0.tar.gz,eiffel-wrapper-1.0.tar.gz} +Eiffel wrapper by Michael Ravits. + +@item @uref{http://www.mysql.com/Downloads/Contrib/SQLmy0.06.tgz,SQLmy0.06.tgz} +FlagShip Replaceable Database Driver (RDD) for MySQL. By Alejandro +Fernandez Herrero. +@uref{http://www.fship.com/rdds.html, Flagship RDD home page} + +@item @uref{http://www.mysql.com/Downloads/Contrib/mydsn-1.0.zip,mydsn-1.0.zip} +Binary and source for @code{mydsn.dll}. mydsn should be used to build +and remove the DSN registry file for the MyODBC driver in Coldfusion +applications. By Miguel Angel Solórzano. + +@item @uref{http://www.mysql.com/Downloads/Contrib/MySQL-ADA95_API.zip, MySQL-ADA95_API.zip} +An ADA95 interface to the @strong{MySQL} API. By Francois Fabien. + +@item @uref{http://www.mysql.com/Downloads/Contrib/MyTool-DLL_for_VB_and_MySQL.zip, MyTool-DLL_for_VB_and_MySQL.zip} +A DLL with @strong{MySQL} C API for Visual Basic. +By Ken Menzel @email{kenm@@icarz.com}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/MYSQLX.EXE, MYSQL.EXE} +@strong{MySQL} ActiveX Object for directly accessing your @strong{MySQL} +servers from IIS/ASP, VB, VC++ skipping the slower ODBC methods. Fully +updatable, multithreaded with full support for all @strong{MySQL} fieldtypes +(version 2001.1.1). By SciBit @uref{http://www.scibit.com/}. + +@item @uref{http://www.fastflow.it/mylua/, MyLUA home page} +How to use the LUA language to write @strong{MySQL} @code{PROCEDURE} that can +be loaded runtime. -Other contributors, bugfinders, and testers: James H. Thompson, Maurizio -Menghini, Wojciech Tryc, Luca Berra, Zarko Mocnik, Wim Bonis, Elmar -Haneke, @email{jehamby@@lightside}, @email{psmith@@BayNetworks.com}, -@email{duane@@connect.com.au}, Ted Deppner @email{ted@@psyber.com}, -Mike Simons, Jaakko Hyvatti. +@itemize @bullet +@item @uref{http://www.mysql.com/Downloads/Contrib/lua-4.0.tar.gz, Lua 4.0} +LUA 4.0 +@item @uref{http://www.mysql.com/Downloads/Contrib/mylua-3.23.32.1.tar.gz, mylua-3.23.32.1.tar.gz} +Patch for @strong{MySQL} 3.23.32 to use LUA 4.0. By Cristian Giussani. +@end itemize -And lots of bug report/patches from the folks on the mailing list. +@item @uref{http://www.mysql.com/Downloads/Contrib/patched_myodbc.zip, patched_myodbc.zip} +Patch (for Omniform 4.0 support) to the MyODBC driver. +By Thomas Thaele @email{tthaele@@papenmeier.de} -A big tribute goes to those that help us answer questions on the -@code{mysql@@lists.mysql.com} mailing list: +@end itemize -@table @asis -@item Daniel Koch @email{dkoch@@amcity.com} -Irix setup. -@item Luuk de Boer @email{luuk@@wxs.nl} -Benchmark questions. -@item Tim Sailer @email{tps@@users.buoy.com} -@code{DBD-mysql} questions. -@item Boyd Lynn Gerber @email{gerberb@@zenez.com} -SCO-related questions. -@item Richard Mehalick @email{RM186061@@shellus.com} -@code{xmysql}-related questions and basic installation questions. -@item Zeev Suraski @email{bourbon@@netvision.net.il} -Apache module configuration questions (log & auth), PHP-related -questions, SQL syntax-related questions and other general questions. -@item Francesc Guasch @email{frankie@@citel.upc.es} -General questions. -@item Jonathan J Smith @email{jsmith@@wtp.net} -Questions pertaining to OS-specifics with Linux, SQL syntax, and other -things that might need some work. -@item David Sklar @email{sklar@@student.net} -Using @strong{MySQL} from PHP and Perl. -@item Alistair MacDonald @email{A.MacDonald@@uel.ac.uk} -Not yet specified, but is flexible and can handle Linux and maybe HP-UX. -Will try to get user to use @code{mysqlbug}. -@item John Lyon @email{jlyon@@imag.net} -Questions about installing @strong{MySQL} on Linux systems, using either -@file{.rpm} files or compiling from source. -@item Lorvid Ltd. @email{lorvid@@WOLFENET.com} -Simple billing/license/support/copyright issues. -@item Patrick Sherrill @email{patrick@@coconet.com} -ODBC and VisualC++ interface questions. -@item Randy Harmon @email{rjharmon@@uptimecomputers.com} -@code{DBD}, Linux, some SQL syntax questions. -@end table +@appendixsec Clients -@cindex contributing companies, list of -@node Supporters, , Contributors, Credits -@appendixsubsec Supporters to MySQL +@itemize @bullet +@item Graphical clients +@itemize @minus +@item @uref{http://www.ideit.com/products/dbvis/, DbVisualizer}. +Freeware JDBC client to graphically visualize the data and structure +of several databases simultaneously. By Innovative-IT Development AB. -The following companies has helped us finance development of -@strong{MySQL} by either paying us for developing a new feature, -developed a @strong{MySQL} feature themselves or by giving us hardware for -@strong{MySQL} development. +@item @uref{http://www.mysql.com/downloads/gui-clients.html, MySQLGUI} +The @strong{MySQL} GUI client homepage. By Sinisa at @strong{MySQL AB}. -@table @asis -@item VA Linux / Andover.net -Funded replication. -@item NuSphere -Editing of the @strong{MySQL} manual. -@item Stork Design studio -The @strong{MySQL} web site in use between 1998-2000. -@item Intel -Contributed to development on Windows and Linux platforms. -@item Compaq -Contributed to Development on Linux/Alpha. -@item SWSoft -Development on the embedded @code{mysqld} version. -@item FutureQuest -@code{--skip-show-variables} -@end table +@item @uref{http://www.mysql.com/Downloads/Contrib/mysql_navigator_0.9.0.tar.gz, MySQL navigator 0.9} +MySQL Navigator is a @strong{MySQL} database server GUI client program. The purpose +of MySQL Navigator is to provide a useful client interface to @strong{MySQL} +database servers, whilst supporting multiple operating systems and +languages. You can currently import/export database, enter queries, get +result sets, edit scripts, run scripts, add, alter, and delete users, +and retrieve client and server information. Uses QT 2.2. GPL +@uref{http://sql.kldp.org/mysql, Home page for MySQL Navigator}. -@node News, Bugs, Credits, Top -@appendix MySQL change history +@item @uref{http://www.mysql.com/Downloads/Win32/secman.zip, MySQL Security GUI} +A user and security management GUI for @strong{MySQL} on Windows. +By Martin Jeremic. +@uref{http://jsoft.webjump.com/, Home page for MySQL Security GUI}. -This appendix lists the changes from version to version in the @strong{MySQL} -source code. +@item @uref{http://www.mysql.com/Downloads/Contrib/kmysqladmin-0.4.1.tar.gz, kmysqladmin-0.4.1.tar.gz}. +@item @uref{http://www.mysql.com/Downloads/Contrib/kmysqladmin-0.4.1-1.src.rpm, kmysqladmin-0.4.1-1.src.rpm}. +@item @uref{http://www.mysql.com/Downloads/Contrib/kmysqladmin-0.4.1-1.i386.rpm, kmysqladmin-0.4.1-1.i386.rpm} +An administration tool for the @strong{MySQL} server using QT / KDE. Tested +only on Linux. + +@item @uref{http://www.mysql.com/Downloads/Contrib/mysql-admin-using-java+swing.tar.gz, Java client using Swing} By Fredy Fischer, @email{se-afs@@dial.eunet.ch}. +You can always find the latest version +@uref{http://www.trash.net/~ffischer/admin/index.html, here}. + +@item @uref{http://www.mysql.com/Downloads/Win32/MySQL-Maker-1.0.zip,MySQL-Maker 1.0}. +Shareware @strong{MySQL} client for Windows. It's a WYSIWYG tool which allows +you to create, change and delete databases and tables. +You can change field - structure and add, change and delete data in +these tables directly without ODBC-driver. +@uref{http://www.presult.de/presult/frames/fs_mysqlmaker.html, MySQL Maker homepage} + +@item @uref{http://www.mysql.com/Downloads/Contrib/mysqlwinadmn.zip, mysqlwinadmn.zip} +Windows GUI (binary only) to administrate a database, by David B. Mansel, +@email{david@@zhadum.org}. + +@item @uref{http://home.online.no/~runeberg/myqa, MyQA} +is a Linux-based query client for the @strong{MySQL} database server. MyQA +lets you enter SQL queries, execute them, and view the results, all in a +graphical user interface. The GUI is roughly similar to that of the +'Query Analyzer' client that comes with MS SQL Server. + +@item @uref{http://members.xoom.com/_opex_/mysqlmanager/index.html, MySQL Manager} +a graphical @strong{MySQL} server manager for @strong{MySQL} server written in Java, for Windows + + +@item @uref{http://www.mysql.com/Downloads/Win32/netadmin.zip, netadmin.zip} +An administrator tool for @strong{MySQL} on Windows 95/98 and Windows NT +4.0. Only tested with @strong{MySQL} Versions 3.23.5 - 3.23.7. Written +using the Tmysql components. + +You can write queries and show tables, indexes, table syntax, and +administrate user, host, and database and so on. This is beta and +still has some bugs. You can test the program with all features. Please +send bugs and hints to Marco Suess @email{ms@@it-netservice.de}. Original +URL @url{http://www.it-netservice.de/pages/software/index.html}. + +@item @uref{http://www.mysql.com/Downloads/Win32/netadmin2.zip, netadmin2.zip} +New version of netadmin. See above for details. + +@item @uref{http://www.mysql.com/Downloads/Win32/ARTADMIN203.EXE,Atronic's @strong{MySQL} client for Windows 2.0.3.0}. +Home page for this can be found at: @uref{http://www.artronic.hr}. + +@item @uref{http://www.mysql.com/Downloads/Win32/mysqlfront.zip, mysqlfront} +Home page: @uref{http://www.mysqlfront.de/}. +Win32-Client for accessing and managing dbs, tables, table-data, indexes, +import-/export-files. (Freeware). By Ansgar Becker. + +@item @uref{http://www.mysql.com/Downloads/Win32/W9xstop.zip,Utility from Artronic to stop MySQL on win9x}. + +@item @uref{http://bardo.hyperlink.cz/mysqlmon,a light weight GUI client for Windows}. + +@item @uref{http://dbtools.vila.bol.com.br/, Dbtools} +A tool to manage @strong{MySQL} databases. Currently only for Windows. +Some features: +@itemize @bullet +@item Manage servers, databases, tables, columns, indexes, and users +@item Import wizard to import structure and data from MS Access, MS Excel, Dbase, FoxPro, Paradox, and ODBC Databases. + +@item @uref{http://www.mysql.com/Downloads/Contrib/KMYENG113.zip,KMYENG113.zip} +An administrator GUI for @strong{MySQL}. Works only on windows, no source. +Available in English and Japanese. By Mitunobu Kaneko. +Home page: @uref{http://sql.jnts.ne.jp/} +@end itemize + +@item @uref{http://www.mysql.com/Downloads/Contrib/xmysqladmin-1.0.tar.gz, xmysqladmin-1.0.tar.gz} +An X-based front end to the @strong{MySQL} database engine. It allows reloads, +status check, process control, myisamchk, grant/revoke privileges, +creating databases, dropping databases, create, alter, browse, and drop +tables. Originally by Gilbert Therrien, @email{gilbert@@ican.net} but +now in public domain and supported by @strong{MySQL AB}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/xmysql-1.9.tar.gz, xmysql-1.9.tar.gz}. +@item @uref{http://web.wt.net/~dblhack, xmysql home page} +A front end to the @strong{MySQL} database engine. It allows for simple +queries and table maintenance, as well as batch queries. By Rick +Mehalick, @email{dblhack@@wt.net}. +Requires @uref{http://bragg.phys.uwm.edu/xforms,xforms 0.88} to work. +@item @uref{http://www.tamos.net/sw/dbMetrix,dbMetrix} +An open source client for exploring databases and executing SQL. Supports +@strong{MySQL}, Oracle, PostgreSQL, and mSQL. +@item @uref{http://www.multimania.com/bbrox/GtkSQL,GtkSQL} +A query tool for @strong{MySQL} and PostgreSQL. +@item @uref{http://dbman.linux.cz/,dbMan} +A query tool written in Perl. Uses DBI and Tk. +@item @uref{http://www.mysql.com/Downloads/Win32/Msc201.EXE, Mascon 202} +@item @uref{http://www.mysql.com/Downloads/Win32/FrMsc202.EXE, Free Mascon 202} +Mascon is a powerful Win32 GUI for the administering @strong{MySQL} server +databases. Mascon's features include visual table design, connections to +multiple servers, data and blob editing of tables, security setting, SQL +color coding, dump functionality and much more. +@uref{http://www.scibit.com/Products/Software/Utils/Mascon.asp,Mascon home page}. +@item @uref{http://www.virtualbeer.net/dbui/,DBUI} +DBUI is a Gtk graphical database editor. +@item @uref{http://www.rtlabs.com/, MacSQL} +GUI for MySQL, ODBC, and JDBC databases for the Mac OS. +@item @uref{http://www.caleb.com.au/, JRetriever} +JRetriever is a generic database front-end tool for JDBC compliant +databases written with Java 2. JRetriever displays database +tables/views in a Windows explorer-like front end. Users can retrieve +data either by clicking on the table folder or by composing their own SQL +statements with our built-in SQL editor. The tool has been tested with +Oracle 8 and @strong{MySQL} as the back-end databases. It requires JDK 1.3 from +JavaSoft. +@item @uref{http://www.jetools.com/products/databrowser/, DataBrowser} +The DataBrowser is a cross-database, cross-platform data access tool. It is more +user friendly than tools like SQL Plus, psql (command line based tools). It is more +flexible than TOAD, ISQL, PGAccess which are GUI's that are limitied to a single +platform or database. +@item @uref{http://www.intrex.net/amit/software/, SQLC} +The SQL Console is a standalone java application that allows you to connect to a + SQL database system and issue SQL queries and updates. It has an easy-to use +graphical user interface. The SQL Console uses JDBC to connect to the database +systems and, therefore, with proper JDBC drivers, you can use this utility to +connect to some of the most popular database systems. +@item @uref{http://www.mysql.com/Downloads/Contrib/mysql_mmc.zip, MySQL MMC} +MySQL MMC is a GUI Management Tool developed using kdevelop +with a very good interface completely like Microsoft +Enterprise Tool (for SQL Server) or Sybase Central. We +can use it to manage server, database, table, index, +users and to edit table data in grid or execute Sql +by Query Analysis. +@end itemize + +@cindex Web clients +@item Web Clients +@itemize @minus +@item @uref{http://www.mysql.com/Downloads/Contrib/mysqladmin-atif-1.0.tar.gz, mysqladmin-atif-1.0.tar.gz} +WWW @strong{MySQL} administrator for the @code{user,} @code{db} and +@code{host} tables. By Tim Sailer, modified by Atif Ghaffar +@email{aghaffar@@artemedia.ch}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/mysql-webadmin-1.0a8-rz.tar.gz, mysql-webadmin-1.0a8-rz.tar.gz} +A tool written in PHP-FI to administrate @strong{MySQL} databases +remotely over the web within a Web-Browser. By Peter Kuppelwieser, +@email{peter.kuppelwieser@@kantea.it}. Updated by Wim Bonis, +@email{bonis@@kiss.de}. Not maintained anymore! + +@item @uref{http://www.mysql.com/Downloads/Contrib/mysqladm.tar.gz, mysqladm.tar.gz} +@strong{MySQL} Web Database Administration written in Perl. By Tim Sailer. + +@item @uref{http://www.mysql.com/Downloads/Contrib/mysqladm-2.tar.gz, mysqladm-2.tar.gz} +Updated version of @file{mysqladm.tar.gz}, by High Tide. + +@item @uref{http://www.mysql.com/Downloads/Contrib/billowmysql.zip, billowmysql.zip} +Updated version of @file{mysqladm.tar.gz}, by Ying Gao. You can get the +newest version from @uref{http://civeng.com/sqldemo/, the home site}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/myadmin-0.4.tar.gz, myadmin-0.4.tar.gz}. +@item @uref{http://myadmin.cheapnet.net/, MyAdmin home page} +A Web-based @strong{MySQL} administrator by Mike Machado. + +@item @uref{http://www.mysql.com/Downloads/Contrib/phpMyAdmin_2.0.1.tar.gz,phpMyAdmin_2.0.1.tar.gz} +A set of PHP3-scripts to adminstrate @strong{MySQL} over the WWW. + +@item @uref{http://www.phpwizard.net/projects/phpMyAdmin/, phpMyAdmin home page} +A PHP3 tool in the spirit of mysql-webadmin, by Tobias Ratschiller, tobias@@dnet.it. + +@item @uref{http://www.mysql.com/Downloads/Contrib/useradm.tar.gz, useradm.tar.gz} +@strong{MySQL} administrator in PHP. By Ofni Thomas +@email{othomas@@vaidsystems.com}. + +@item @uref{http://gossamer-threads.com/perl/mysqlman/mysql.cgi, MySQLMan} +Similar functionality as phpmyadmin, but written with Perl and using +html templates. By Alex Krohn. +@end itemize + +@item @uref{http://www.mysql.com/Downloads/Contrib/mysql-editor.tar.gz,mysql-editor.tar.gz} +This cgi scripts in Perl enables you to edit content of Mysql +database. By Tomas Zeman. +@item +@uref{http://worldcommunity.com/opensource/futuresql, FutureSQL Web Database Administration Tool}. +FutureSQL by Peter F. Brown, is a free, open source rapid application +development Web database administration tool, written in Perl, +using @strong{MySQL}. It uses @code{DBI:DBD} and @code{CGI.pm}. + +FutureSQL allows one to easily set up config files to view, edit, delete, +and otherwise process records from a @strong{MySQL} database. It uses a data +dictionary, configuration files and templates, and allows "pre-processing" +and "post-processing" on both fields, records, and operations. +@end itemize + +@cindex web tools +@cindex tools,, web +@appendixsec Web Tools + +@itemize @bullet + +@item @uref{http://www.mysql.com/Downloads/Contrib/mod_mysql_include_1.0.tar.gz, mod_mysql_include_1.0.tar.gz} +Apache module to include HTML from @strong{MySQL} queries into your pages, +and run update queries. Originally written to implement a simple fast +low-overhead banner-rotation system. By Sasha Pachev. + +@item @uref{http://htcheck.sourceforge.net, htCheck} - URL checker with +@strong{MySQL} backend. Spidered URLs can later be queried using SQL to retrieve +various kinds of information, eg. broken links. Written by Gabriele Bartolini. + +@item @uref{http://www.odbsoft.com/cook/sources.htm} +This package has various functions for generating html code from a SQL +table structure and for generating SQL statements (Select, Insert, +Update, Delete) from an html form. You can build a complete forms +interface to a SQL database (query, add, update, delete) without any +programming! By Marc Beneteau, @email{marc@@odbsoft.com}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/sqlhtml.tar.gz, sqlhtml.tar.gz} +SQL/HTML is an HTML database manager for @strong{MySQL} using @code{DBI} 1.06. + +@item @uref{http://www.mysql.com/Downloads/Contrib/udmsearch-3.0.23.tar.gz, UdmSearch 3.0.23 (stable version)}. +@item @uref{http://www.mysql.com/Downloads/Contrib/mnogosearch-3.1.12.tar.gz, mnogosearch 3.1.12 (development but recommended version)}. +@item @uref{http://search.mnoGo.ru, UdmSearch home page} +A SQL-based search engine for Internet. By +Alexander I. Barkov @email{bar@@izhcom.ru}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/wmtcl.doc, wmtcl.doc}. +@item @uref{http://www.mysql.com/Downloads/Contrib/wmtcl.lex, wmtcl.lex} +With this you can write HTML files with inclusions of Tcl code. By +@email{vvs@@scil.npi.msu.su}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/www-sql-0.5.7.lsm, www-sql-0.5.7.lsm}. +@item @uref{http://www.mysql.com/Downloads/Contrib/www-sql-0.5.7.tar.gz, www-sql-0.5.7.tar.gz} +A CGI program that parses an HTML file containing special tags, parses +them, and inserts data from a @strong{MySQL} database. + +@item @uref{http://www.mysql.com/Downloads/Contrib/genquery.zip, genquery.zip} +Perl SQL database interface package for html. + +@item @uref{http://www.mysql.com/Downloads/Contrib/cgi++-0.8.tar.gz, cgi++-0.8.tar.gz} +A macro-processor to simply writing CGI/Database programs in C++ by Sasha Pachev. + +@item @uref{http://www.mysql.com/Downloads/Contrib/webboard-1.0.zip, WebBoard 1.0} +EU-Industries Internet-Message-Board. + +@item @uref{http://www.mysql.com/Downloads/Contrib/DBIx-TextIndex-0.02.tar.gz, DBIx-TextIndex-0.02.tar.gz} +Full-text searching with Perl on @code{BLOB}/@code{TEXT} columns by Daniel Koch. +@end itemize + +@cindex tools, benchmarking +@cindex benchmarking, tools +@appendixsec Performance Benchmarking Tools + +@itemize @bullet +@item @uref{http://www.mysql.com/Downloads/super-smack/super-smack-1.0.tar.gz, + super-smack} +Multi-threaded benchmarking tool for @strong{MySQL} and +@strong{PostgreSQL}. Written in C++. Easy to extend to support other +databases that have C/C++ client libraries. By Sasha Pachev. +@end itemize + +@cindex tools, authentication +@cindex authentication tools +@appendixsec Authentication Tools + +@itemize @bullet +@item @uref{http://www.mysql.com/Downloads/Contrib/ascend-radius-mysql-0.7.2.patch.gz,ascend-radius-mysql-0.7.2.patch.gz} +This is an authentication and logging patch using @strong{MySQL} for +Ascend-Radius. By @email{takeshi@@SoftAgency.co.jp}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/icradius-0.10.tar.gz, icradius 0.10} +@uref{http://www.mysql.com/Downloads/Contrib/icradius.README, icradius readme file}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/checkpassword-0.81-mysql-0.6.6.patch.gz, +checkpassword-0.81-mysql-0.6.6.patch.gz} +@strong{MySQL} authentication patch for QMAIL and checkpassword. These are +useful for management user (mail, pop account) by @strong{MySQL}. +By @email{takeshi@@SoftAgency.co.jp}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/jradius-diff.gz, jradius-diff.gz} +@strong{MySQL} support for Livingston's Radius 2.01. Authentication and +Accounting. By Jose de Leon, @email{jdl@@thevision.net}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/mod_auth_mysql-2.20.tar.gz, mod_auth_mysql-2.20.tar.gz} +Apache authentication module for @strong{MySQL}. By Zeev Suraski, +@email{bourbon@@netvision.net.il}. + +@c @strong{Please} register this module at: +@c @url{http://bourbon.netvision.net.il/mysql/mod_auth_mysql/register.html}. The +@c registering information is only used for statistical purposes and will +@c encourage further development of this module! + +@item @uref{http://www.mysql.com/Downloads/Contrib/mod_log_mysql-1.05.tar.gz, mod_log_mysql-1.05.tar.gz} +@strong{MySQL} logging module for Apache. By Zeev Suraski, +@email{bourbon@@netvision.net.il}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/mypasswd-2.0.tar.gz, mypasswd-2.0.tar.gz} +Extra for @code{mod_auth_mysql}. This is a little tool that allows you +to add/change user records storing group and/or password entries in +@strong{MySQL} tables. By Harry Brueckner, @email{brueckner@@respublica.de}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/mysql-passwd.README, mysql-passwd.README}. +@item @uref{http://www.mysql.com/Downloads/Contrib/mysql-passwd-1.2.tar.gz, mysql-passwd-1.2.tar.gz} +Extra for @code{mod_auth_mysql}. This is a two-part system for use with +@code{mod_auth_mysql}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/pam_mysql.tar.gz, pam_mysql.tar.gz} +This module authenticates users via @code{pam}, using @strong{MySQL}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/nsapi_auth_mysql.tar, nsapi_auth_mysql.tar} +Netscape Web Server API (NSAPI) functions to authenticate (BASIC) users +against @strong{MySQL} tables. By Yuan John Jiang. + +@item @uref{http://www.mysql.com/Downloads/Contrib/qmail-1.03-mysql-0.6.6.patch.gz,qmail-1.03-mysql-0.6.6.patch.gz} +Patch for qmail to authenticate users from a @strong{MySQL} table. +By @email{takeshi@@SoftAgency.co.jp}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/proftpd-1.2.0rc2-fix-mysql.patch, proftpd-1.2.0rc2-fix-mysql.patch} +Patch for proftpd1.2.0rc2. By @email{takeshi@@SoftAgency.co.jp}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/pwcheck_mysql-0.1.tar.gz,pwcheck_mysql-0.1.tar.gz} +An authentication module for the Cyrus IMAP server. By Aaron Newsome. + + +@end itemize + +@cindex converters +@appendixsec Converters + +@itemize @bullet +@item @uref{http://www.mysql.com/Downloads/Contrib/mssql2mysql.txt, mssql2mysql.txt} +Converter from MS-SQL to MySQL. By Michael Kofler. +@uref{http://www.kofler.cc/mysql/mssql2mysql.html, mssql2mysql home page}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/dbf2mysql-1.14.tar.gz, dbf2mysql-1.14.tar.gz} +Convert between @file{.dbf} files and @strong{MySQL} tables. By Maarten +Boekhold (@email{boekhold@@cindy.et.tudelft.nl}), William Volkman, and +Michael Widenius. This converter includes rudimentary read-only support +for MEMO fields. + +@item @uref{http://www.mysql.com/Downloads/Contrib/dbf2mysql-1.13.tgz, dbf2mysql-1.13.tgz} +Convert between @file{.dbf} files and @strong{MySQL} tables. By Maarten +Boekhold, @email{boekhold@@cindy.et.tudelft.nl}, and Michael Widenius. +This converter can't handle MEMO fields. + +@item @uref{http://www.mysql.com/Downloads/Contrib/dbf2mysql.zip, dbf2mysql.zip} +Convert between FoxPro @file{.dbf} files and @strong{MySQL} tables on Windows. +By Alexander Eltsyn, @email{ae@@nica.ru} or @email{ae@@usa.net}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/dbf2sql.zip, dbf2sql.zip} +Short and simple prg that can help you transport your data from foxpro +table into @strong{MySQL} table. By Danko Josic. + +@item @uref{http://www.mysql.com/Downloads/Contrib/dump2h-1.20.gz, dump2h-1.20.gz} +Convert from @code{mysqldump} output to a C header file. By Harry Brueckner, +@email{brueckner@@mail.respublica.de}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/exportsql.txt, exportsql.txt} +A script that is similar to @code{access_to_mysql.txt}, except that this +one is fully configurable, has better type conversion (including +detection of @code{TIMESTAMP} fields), provides warnings and suggestions +while converting, quotes @strong{all} special characters in text and +binary data, and so on. It will also convert to @code{mSQL} v1 and v2, +and is free of charge for anyone. See +@uref{http://www.cynergi.net/exportsql/} for the latest version. By +Pedro Freire, @email{support@@cynergi.net}. NOTE: Doesn't work with +Access2! +@item @uref{http://www.mysql.com/Downloads/Contrib/access_to_mysql.txt, access_to_mysql.txt} +Paste this function into an Access module of a database that has the +tables you want to export. See also @code{exportsql}. By Brian Andrews. +NOTE: Doesn't work with Access2! + +@item @uref{http://www.mysql.com/Downloads/Contrib/importsql.txt, importsql.txt} +A script that does the exact reverse of @code{exportsql.txt}. That is, +it imports data from @strong{MySQL} into an Access database via +ODBC. This is very handy when combined with exportsql, because it lets you +use Access for all DB design and administration, and synchronize with +your actual @strong{MySQL} server either way. Free of charge. See +@uref{http://www.netdive.com/freebies/importsql/} for any updates. +Created by Laurent Bossavit of NetDIVE. +@strong{NOTE:} Doesn't work with Access2! + +@item @uref{http://www.mysql.com/Downloads/Contrib/mdb2sql.bas, mdb2sql.bas} +Converter from Access97 to @strong{MySQL} by Moshe Gurvich. + +@item @uref{http://www.mysql.com/Downloads/Contrib/msql2mysqlWrapper-1.0.tgz, msql2mysqlWrapper 1.0} +A C wrapper from @code{mSQL} to @strong{MySQL}. By @email{alfred@@sb.net} + +@item @uref{http://www.mysql.com/Downloads/Contrib/sqlconv.pl, sqlconv.pl} +A simple script that can be used to copy fields from one @strong{MySQL} table +to another in bulk. Basically, you can run @code{mysqldump} and pipe it to +the @code{sqlconv.pl} script. The script will parse through the +@code{mysqldump} output and will rearrange the fields so they can be +inserted into a new table. An example is when you want to create a new +table for a different site you are working on, but the table is just a +bit different (that is - fields in different order, etc.). +By Steve Shreeve. +@item @uref{http://www.mysql.com/Downloads/Contrib/oracledump, oracledump} +Perl program to convert Oracle databases to @strong{MySQL}. Has same +output format as mysqldump. By Johan Andersson. + +@item @uref{http://www.mysql.com/Downloads/Contrib/excel2mysql, excel2mysql} +Perl program to import Excel spreadsheets into a @strong{MySQL} database. By Stephen Hurd @email{shurd@@sk.sympatico.ca} + +@item @uref{http://www.mysql.com/Downloads/Contrib/T2S_100.ZIP, T2S_100.ZIP}. +Windows program to convert text files to @strong{MySQL} databases. By +Asaf Azulay. + +@end itemize + +@appendixsec Using MySQL with Other Products + +@itemize @bullet +@item @uref{http://www.mysql.com/Downloads/Contrib/emacs-sql-mode.tar.gz, emacs-sql-mode.tar.gz} +Raw port of a SQL mode for XEmacs. Supports completion. Original by +Peter D. Pezaris @email{pez@@atlantic2.sbi.com} and partial +@strong{MySQL} port by David Axmark. + +@item @uref{http://www.mysql.com/Downloads/Win32/myaccess97_1_4.zip, MyAccess97 1.4}. +@item @uref{http://www.mysql.com/Downloads/Win32/myaccess2000_1_4.zip, MyAccess2000 1.4}. -Note that we tend to update the manual at the same time we make changes to -@strong{MySQL}. If you find a version listed below that you can't -find on the -@uref{http://www.mysql.com/downloads/,@strong{MySQL} download page}, -this means that the version has not yet been released! +MyAccess is an AddIn for MS Access 97/2000 that allows you to manage +@strong{MySQL} databases from within Access. Main functions are: +@itemize @minus +@item Create/Modify Tables +@item Execute Queries against MySQL +@item Extract ''Create Table-Scripts'' from MySQL +@item Import/Export tables from Access to MySQL and vice versa +@item Log Changes +@item Show a "Database Definition Report +@end itemize -@menu -* News-4.0.x:: Changes in release 4.0.x (Development; Alpha) -* News-3.23.x:: Changes in release 3.23.x (Stable) -* News-3.22.x:: Changes in release 3.22.x (Older; Still supported) -* News-3.21.x:: Changes in release 3.21.x -* News-3.20.x:: Changes in release 3.20.x -* News-3.19.x:: Changes in release 3.19.x -@end menu +Written by Hubertus Hiden. @uref{http://www.accessmysql.com, MyAccess homepage}. -@node News-4.0.x, News-3.23.x, News, News -@appendixsec Changes in release 4.0.x (Development; Alpha) +@item @uref{http://www.mysql.com/Downloads/Contrib/radius-0.3.tar.gz, radius-0.3.tar.gz} +Patches for @code{radiusd} to make it support @strong{MySQL}. By Wim Bonis, +@email{bonis@@kiss.de}. +@end itemize -We have now started to work on @strong{MySQL} 4.0. We will update this -section as we add new features, so that others can follow our development. +@cindex tools, useful +@appendixsec Useful Tools -Our TODO section contains what we plan to have in 4.0. @xref{TODO MySQL 4.0}. +@itemize @bullet +@item @uref{http://worldcommunity.com/opensource/utilities/mysql_backup.html, MySQL Backup}. -@menu -* News-4.0.0:: Changes in release 4.0.0 -@end menu +A backup script for MySQL. By Peter F. Brown. -@node News-4.0.0, , News-4.0.x, News-4.0.x -@appendixsubsec Changes in release 4.0.0 +@item @uref{http://www.mysql.com/Downloads/Contrib/mytop, mytop} +@item @uref{http://public.yahoo.com/~jzawodn/mytop/, mytop home page} +mytop is a Perl program that allows you to monitor @strong{MySQL} servers by +viewing active threads, queries, and overall server performance +numbers. By Jeremy D. Zawodny. -@itemize @bullet -@item -Multi-table @code{DELETE}. -@item -Don't support old client protocols prior to @strong{MySQL} 3.21 any more. -@item -Don't include the old C API functions @code{mysql_drop_db}, -@code{mysql_create_db} and @code{mysql_connect}, unless compiled with -@code{USE_OLD_FUNCTIONS}. -@item -Renamed @code{safe_mysqld} to @code{mysqld_safe}. -@item -Allow @code{IN} as a synonym for @code{FROM} in @code{SHOW} commands. -@item -@code{SHOW INDEXES} is now a synonym for @code{SHOW INDEX}. -@item -Added support for symbolic links to @code{MyISAM} tables. Symlink handling is -now enabled by default for Windows. -@item -@code{LOAD DATA FROM MASTER} ``auto-magically'' sets up a slave. -@item -A new @code{HANDLER} interface to @code{MyISAM} tables. -@item -@code{COUNT(DISTINCT)} is about 30% faster. -@item -@code{FULLTEXT} index creation now is much faster. -@item -Searching on packed (@code{CHAR}/@code{VARCHAR}) keys now is much faster. -@item -Added @code{SQL_CALC_FOUND_ROWS} and @code{FOUND_ROWS()}. This makes it -possible to know how many rows a query would have returned -without a @code{LIMIT} clause. -@item -Changed output format of @code{SHOW OPEN TABLES}. -@item -Allow @code{SELECT expression LIMIT ...}. -@item -Added @code{IDENTITY} as a synonym for @code{AUTO_INCREMENT} (like Sybase). -@item -Added @code{ORDER BY} syntax to @code{UPDATE} and @code{DELETE}. -@item -Optimized queries of type: -@code{SELECT DISTINCT * from table_name ORDER by key_part1 LIMIT #} -@end itemize +@item @uref{http://www.mysql.com/Downloads/Contrib/mysql_watchdog.pl, mysql_watchdog.pl} +Monitor the @strong{MySQL} daemon for possible lockups. By Yermo Lamers, +@email{yml@@yml.com}. -@node News-3.23.x, News-3.22.x, News-4.0.x, News -@appendixsec Changes in release 3.23.x (Stable) +@item @uref{http://www.mysql.com/Downloads/Contrib/mysqltop.tar.gz, mysqltop.tar.gz} +Sends a query in a fixed time interval to the server and shows the +resulting table. By Thomas Wana. +@item @uref{http://www.mysql.com/Downloads/Contrib/mysql_structure_dumper.tar.gz,mysql_structure_dumper.tar.gz} +Prints out the structure of the all tables in a database. By Thomas Wana. -The 3.23 release has several major features that are not -present in previous versions. We have added three new -table types: +@item @uref{http://www.mysql.com/Downloads/Contrib/mysql_structure_dumper.tgz, structure_dumper.tgz} +Prints the structure of every table in a database. By Thomas Wana. -@table @asis -@item @strong{MyISAM} -A new ISAM library which is tuned for SQL and supports large files. -@item @strong{BerkeleyDB} or @strong{BDB} -Uses the Berkeley DB library from Sleepycat Software to implement -transaction-safe tables. -@item @strong{InnoDB} -A transaction-safe table handler that supports row level locking, and many -Oracle-like features. -@end table +@item @uref{http://www.mysql.com/Downloads/Contrib/mysqlsync, mysqlsync-1.0-alpha.tar.gz}. +A Perl script to keep remote copies of a @strong{MySQL} database in sync with a +central master copy. By Mark Jeftovic. @email{markjr@@easydns.com}. -Note that only MyISAM is available in the standard binary distribution. +@item @uref{http://www.mysql.com/Downloads/Contrib/MySQLTutor-0.2.tar.gz, MySQLTutor}. +MySQLTutor. A @strong{MySQL} tutorial for beginners. -The 3.23 release also includes support for database replication between -a master and many slaves, full-text indexing, and much more. +@item @uref{http://www.mysql.com/Downloads/Contrib/MySQLDB.zip, MySQLDB.zip} +A COM library for @strong{MySQL} by Alok Singh. +@item @uref{http://www.mysql.com/Downloads/Contrib/MySQLDB-readme.html, MySQLDB-readme.html}. -All new features are being developed in the 4.0 version. Only -bug fixes and minor enhancements to existing features will be -added to 3.23. +@item @uref{http://www.mysql.com/Downloads/Contrib/mysql_replicate.pl, mysql_replicate.pl} +Perl program that handles replication. By @email{elble@@icculus.nsg.nwu.edu} -The replication code and BerkeleyDB code is still not as tested and as -the rest of the code, so we will probably need to do a couple of future -releases of 3.23 with small fixes for this part of the code. As long as -you don't use these features, you should be quite safe with -@strong{MySQL} 3.23! +@item @uref{http://www.mysql.com/Downloads/Contrib/DBIx-TextIndex-0.02.tar.gz, DBIx-TextIndex-0.02.tar.gz} +Perl script that uses reverse indexing to handle text searching. +By Daniel Koch. -Note that the above doesn't mean that replication or Berkeley DB don't -work; We have done a lot of testing of all code, including replication -and BDB without finding any problems. It only means that not as many -users use this code as the rest of the code and because of this we are -not yet 100% confident in this code. +@item @uref{http://www.mysql.com/Downloads/Contrib/dbcheck, dbcheck} +Perl script that takes a backup of tables before running isamchk on them. +By Elizabeth. -@menu -* News-3.23.40:: Changes in release 3.23.40 -* News-3.23.39:: Changes in release 3.23.39 -* News-3.23.38:: Changes in release 3.23.38 -* News-3.23.37:: Changes in release 3.23.37 -* News-3.23.36:: Changes in release 3.23.36 -* News-3.23.35:: Changes in release 3.23.35 -* News-3.23.34a:: Changes in release 3.23.34a -* News-3.23.34:: Changes in release 3.23.34 -* News-3.23.33:: Changes in release 3.23.33 -* News-3.23.32:: Changes in release 3.23.32 -* News-3.23.31:: Changes in release 3.23.31 -* News-3.23.30:: Changes in release 3.23.30 -* News-3.23.29:: Changes in release 3.23.29 -* News-3.23.28:: Changes in release 3.23.28 -* News-3.23.27:: Changes in release 3.23.27 -* News-3.23.26:: Changes in release 3.23.26 -* News-3.23.25:: Changes in release 3.23.25 -* News-3.23.24:: Changes in release 3.23.24 -* News-3.23.23:: Changes in release 3.23.23 -* News-3.23.22:: Changes in release 3.23.22 -* News-3.23.21:: Changes in release 3.23.21 -* News-3.23.20:: Changes in release 3.23.20 -* News-3.23.19:: Changes in release 3.23.19 -* News-3.23.18:: Changes in release 3.23.18 -* News-3.23.17:: Changes in release 3.23.17 -* News-3.23.16:: Changes in release 3.23.16 -* News-3.23.15:: Changes in release 3.23.15 -* News-3.23.14:: Changes in release 3.23.14 -* News-3.23.13:: Changes in release 3.23.13 -* News-3.23.12:: Changes in release 3.23.12 -* News-3.23.11:: Changes in release 3.23.11 -* News-3.23.10:: Changes in release 3.23.10 -* News-3.23.9:: Changes in release 3.23.9 -* News-3.23.8:: Changes in release 3.23.8 -* News-3.23.7:: Changes in release 3.23.7 -* News-3.23.6:: Changes in release 3.23.6 -* News-3.23.5:: Changes in release 3.23.5 -* News-3.23.4:: Changes in release 3.23.4 -* News-3.23.3:: Changes in release 3.23.3 -* News-3.23.2:: Changes in release 3.23.2 -* News-3.23.1:: Changes in release 3.23.1 -* News-3.23.0:: Changes in release 3.23.0 -@end menu +@item @uref{http://www.mysql.com/Downloads/Contrib/mybackup}. +@item @uref{http://www.mswanson.com/mybackup, mybackup home page} +Wrapper for mysqldump to backup all databases. By Marc Swanson. -@node News-3.23.40, News-3.23.39, News-3.23.x, News-3.23.x -@appendixsubsec Changes in release 3.23.40 -@itemize @bullet -@item -Fixed problem with @code{--low-priority-updates} and @code{INSERT}'s. -@item -Fixed bug in slave thread when under some rare circumstances it could -get 22 bytes ahead on the offset in the master. -@item -Added @code{slave_wait_timeout} for replication. -@item -Fixed problem with @code{UPDATE} and @code{BDB} tables. -@item -Fixed hard bug in @code{BDB} tables when using key parts. -@item -Fixed problem when using the @code{GRANT FILE ON database.* ...}; Previously -we added the @code{DROP} privilege for the database. -@item -Fixed @code{DELETE FROM table_name ... LIMIT 0} and -@code{UPDATE FROM table_name ... LIMIT 0} acted as though the @code{LIMIT} -clause was not present (they deleted or updated all selected rows). -@item -@code{CHECK TABLE} now checks if an @code{AUTO_INCREMENT} column contains -the value 0. -@item -Sending a @code{SIGHUP} to @code{mysqld} will now only flush the logs, -not reset the replication. -@item -Fixed parser to allow floats of type @code{1.0e1} (no sign after @code{e}). -@item -Option @code{--force} to @code{myisamchk} now also updates states. -@item -Added option @code{--warnings} to @code{mysqld}. Now @code{mysqld} -only prints the error @code{Aborted connection} if this option is used. -@item -Fixed problem with @code{SHOW CREATE TABLE} when you didn't have a -@code{PRIMARY KEY}. -@item -Properly fixed the rename of @code{innodb_unix_file_flush_method} to -@code{innodb_flush_method}. -@item -Fixed bug when converting @code{UNSIGNED BIGINT} to @code{DOUBLE}. This caused -a problem when doing comparisons with @code{BIGINT} values outside of the -signed range. -@item -Fixed bug in @code{BDB} tables when querying empty tables. -@item -Fixed a bug when using @code{COUNT(DISTINCT)} with @code{LEFT JOIN} and -there wasn't any matching rows. -@item -Removed all documentation referring to the @code{GEMINI} table -type. @code{GEMINI} is not released under an Open Source license. +@item @uref{http://www.mysql.com/Downloads/Contrib/mdu.pl.gz,mdu.pl.gz} +Prints the storage usage of a @strong{MySQL} database. @end itemize -@node News-3.23.39, News-3.23.38, News-3.23.40, News-3.23.x -@appendixsubsec Changes in release 3.23.39 -@itemize @bullet -@item -The @code{AUTO_INCREMENT} sequence wasn't reset when dropping -and adding an @code{AUTO_INCREMENT} column. -@item -@code{CREATE ... SELECT} now creates non-unique indexes delayed. -@item -Fixed problem where @code{LOCK TABLES table_name READ} followed by -@code{FLUSH TABLES} put an exclusive lock on the table. -@item -@code{REAL} @@variables with was represented with 2 digits when -converted to strings. -@item -Fixed problem that client 'hung' when @code{LOAD TABLE FROM MASTER} failed. -@item -Running @code{myisamchk --fast --force} will no longer repair tables -that only had the open count wrong. -@item -Added functions to handle symbolic links to make life easier in 4.0. -@item -We are now using the @code{-lcma} thread library on HP-UX 10.20 so -that @strong{MySQL} will be more stable on HP-UX. -@item -Fixed problem with @code{IF()} and number of decimals in the result. -@item -Fixed date-part extraction functions to work with dates where day -and/or month is 0. -@item -Extended argument length in option files from 256 to 512 chars. -@item -Fixed problem with shutdown when @code{INSERT DELAYED} was waiting for -a @code{LOCK TABLE}. -@item -Fixed coredump bug in InnoDB when tablespace was full. -@item -Fixed problem with @code{MERGE} tables and big tables (> 4G) when using -@code{ORDER BY}. -@end itemize +@cindex RPMs, for common tools +@cindex tools, RPMs for +@appendixsec RPMs for Common Tools (Most Are for RedHat 6.1) -@node News-3.23.38, News-3.23.37, News-3.23.39, News-3.23.x -@appendixsubsec Changes in release 3.23.38 @itemize @bullet -@item -Fixed a bug when @code{SELECT} from @code{MERGE} table -sometimes results in incorrectly ordered rows. -@item -Fixed a bug in @code{REPLACE()} when using the ujis character set. -@item -Applied Sleepycat BDB patches 3.2.9.1 and 3.2.9.2. -@item -Added option @code{--skip-stack-trace} to @code{mysqld}. -@item -@code{CREATE TEMPORARY} now works with @code{InnoDB} tables. -@item -@code{InnoDB} now promotes sub keys to whole keys. -@item -Added option @code{CONCURRENT} to @code{LOAD DATA}. -@item -Better error message when slave @code{max_allowed_packet} is too low to -read a very long log event from the master. -@item -Fixed bug when too many rows where removed when using -@code{SELECT DISTINCT ... HAVING}. -@item -@code{SHOW CREATE TABLE} now returns @code{TEMPORARY} for temporary tables. -@item -Added @code{Rows_examined} to slow query log. -@item -Fixed problems with function returning empty string when using -together with a group functions and a @code{WHERE} that didn't match any rows. -@item -New program @code{mysqlcheck}. -@item -Added database name to output for administrative commands like @code{CHECK}, -@code{REPAIR}, @code{OPTIMIZE}. -@item -Lots of portability fixes for InnoDB. -@item -Changed optimizer so that queries like -@code{SELECT * FROM table_name,table_name2 ... ORDER BY key_part1 LIMIT #} -will use index on @code{key_part1} instead of @code{filesort}. -@item -Fixed bug when doing -@code{LOCK TABLE to_table WRITE,...; INSERT INTO to_table... SELECT ...} -when @code{to_table} was empty. -@item -Fixed bug with @code{LOCK TABLE} and BDB tables. +@item @uref{http://www.mysql.com/Downloads/Contrib/perl-Data-ShowTable-3.3-2.i386.rpm,perl-Data-ShowTable-3.3-2.i386.rpm} +@item @uref{http://www.mysql.com/Downloads/Contrib/perl-Msql-Mysql-modules-1.2210-2.i386.rpm,perl-Msql-Mysql-modules-1.2210-2.i386.rpm} +@item @uref{http://www.mysql.com/Downloads/Contrib/php-pg-3.0.13-1.i386.rpm,php-pg-3.0.13-1.i386.rpm} +@item @uref{http://www.mysql.com/Downloads/Contrib/php-pg-manual-3.0.13-1.i386.rpm,php-pg-manual-3.0.13-1.i386.rpm} +@item @uref{http://www.mysql.com/Downloads/Contrib/php-pg-mysql-3.0.13-1.i386.rpm,php-pg-mysql-3.0.13-1.i386.rpm} +@item @uref{http://www.mysql.com/Downloads/Contrib/phpMyAdmin-2.0.5-1.noarch.rpm,phpMyAdmin-2.0.5-1.noarch.rpm} @end itemize -@node News-3.23.37, News-3.23.36, News-3.23.38, News-3.23.x -@appendixsubsec Changes in release 3.23.37 +@cindex functions, useful +@appendixsec Useful Functions @itemize @bullet -@item -Fixed a bug when using @code{MATCH} in @code{HAVING} clause. -@item -Fixed a bug when using @code{HEAP} tables with @code{LIKE}. -@item -Added @code{--mysql-version} to @code{safe_mysqld} -@item -Changed @code{INNOBASE} to @code{InnoDB} (because the @code{INNOBASE} -name was already used). All @code{configure} options and @code{mysqld} -start options are now using @code{innodb} instead of @code{innobase}. This -means that you have to change any configuration files where you have used -@code{innobase} options before upgrading to this version! -@item -Fixed bug when using indexes on @code{CHAR(255) NULL} columns. -@item -Slave thread will now be started even if @code{master-host} is not set, as -long as @code{server-id} is set and valid @code{master.info} is present -@item -Partial updates (terminated with kill) are now logged with a special error -code to the binary log. Slave will refuse to execute them if the error code -indicates the update was terminated abnormally, and will have to be recovered -with @code{SET SQL_SLAVE_SKIP_COUNTER=1; SLAVE START} after a manual sanity -check/correction of data integrity. -@item -Fixed bug that erroneously logged a drop of internal temporary table -on thread termination to the binary log - bug affected replication. -@item -Fixed a bug in @code{REGEXP()} on 64-bit machines. -@item -@code{UPDATE} and @code{DELETE} with @code{WHERE unique_key_part IS NULL} -didn't update/delete all rows. -@item -Disabled @code{INSERT DELAYED} for tables that support transactions. -@item -Fixed bug when using date functions on @code{TEXT}/@code{BLOB} column -with wrong date format. -@item -UDFs now also work on Windows. (Patch by Ralph Mason) -@item -Fixed bug in @code{ALTER TABLE} and @code{LOAD DATA INFILE} that disabled -key-sorting. These commands should now be faster in most cases. -@item -Fixed performance bug where reopened tables (tables that had been -waiting for @code{FLUSH} or @code{REPAIR}) would not use indexes for the -next query. -@item -Fixed problem with @code{ALTER TABLE} to Innobase tables on FreeBSD. -@item -Added @code{mysqld} variables @code{myisam_max_sort_file_size} and -@code{myisam_max_extra_sort_file_size}. -@item -Initialize signals early to avoid problem with signals in Innobase. -@item -Applied patch for the @code{tis620} character set to make comparisons -case-independent and to fix a bug in @code{LIKE} for this character set. -@strong{NOTE}: All tables that uses the @code{tis620} character set must be -fixed with @code{myisamchk -r} or @code{REPAIR TABLE} ! -@item -Added @code{--skip-safemalloc} option to @code{mysqld}. +@item @uref{http://www.mysql.com/Downloads/Contrib/mysnprintf.c,mysnprintf.c} +sprintf() function for SQL queries that can escape blobs. By Chunhua Liu. @end itemize -@node News-3.23.36, News-3.23.35, News-3.23.37, News-3.23.x -@appendixsubsec Changes in release 3.23.36 +@appendixsec Windows programs @itemize @bullet -@item -Fixed a bug that allowed you to use database names containing a @samp{.} -character. This fixes a serious security issue when @code{mysqld} is run -as root. -@item -Fixed bug when thread creation failed (could happen when doing a LOT -of connections in a short time). -@item -Fixed some problems with @code{FLUSH TABLES} and @code{TEMPORARY} tables. -(Problem with freeing the key cache and error @code{Can't reopen table...}). -@item -Fixed a problem in Innobase with other character sets than @code{latin1} -and another problem when using many columns. -@item -Fixed bug that caused a core dump when using a very complex query involving -@code{DISTINCT} and summary functions. -@item -Added @code{SET TRANSACTION ISOLATION LEVEL ...} -@item -Added @code{SELECT ... FOR UPDATE}. -@item -Fixed bug where the number of affected rows was not returned when @code{MySQL} -was compiled without transaction support. -@item -Fixed a bug in @code{UPDATE} where keys weren't always used to find the -rows to be updated. -@item -Fixed a bug in @code{CONCAT_WS()} where it returned wrong results. -@item -Changed @code{CREATE ... INSERT} and @code{INSERT ... SELECT} to not -allow concurrent inserts as this could make the binary log hard to repeat. -(Concurrent inserts are enabled if you are not using the binary or update log). -@item -Changed some macros to be able to use fast mutex with glibc 2.2. +@item @uref{http://www.mysql.com/Downloads/Contrib/LaunchMySQL.zip, LaunchMySQL.zip} +The program launches the @strong{MySQL} server, shuts it down, and +display status information. By Bill Thompson @end itemize -@node News-3.23.35, News-3.23.34a, News-3.23.36, News-3.23.x -@appendixsubsec Changes in release 3.23.35 +@appendixsec Uncategorized + @itemize @bullet -@item -Fixed newly introduced bug in @code{ORDER BY}. -@item -Fixed wrong define @code{CLIENT_TRANSACTIONS}. -@item -Fixed bug in @code{SHOW VARIABLES} when using @code{INNOBASE} tables. -@item -Setting and using user variables in @code{SELECT DISTINCT} didn't work. -@item -Tuned @code{SHOW ANALYZE} for small tables. -@item -Fixed handling of arguments in the benchmark script @file{run-all-tests}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/findres.pl, findres.pl} +Find reserved words in tables. By Nem W Schlecht. + +@item @uref{http://www.mysql.com/Downloads/Contrib/handicap.tar.gz, handicap.tar.gz} +Performance handicapping system for yachts. Uses PHP. By +@email{rhill@@stobyn.ml.org}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/hylalog-1.0.tar.gz, hylalog-1.0.tar.gz} +Store @code{hylafax} outgoing faxes in a @strong{MySQL} database. By Sinisa +Milivojevic, @email{sinisa@@mysql.com}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/mrtg-mysql-1.0.tar.gz, mrtg-mysql-1.0.tar.gz} +@strong{MySQL} status plotting with MRTG, by Luuk de Boer, @email{luuk@@wxs.nl}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/wuftpd-2.4.2.18-mysql_support.2.tar.gz, wuftpd-2.4.2.18-mysql_support.2.tar.gz} +Patches to add logging to @strong{MySQL} for WU-ftpd. By Zeev Suraski, +@email{bourbon@@netvision.net.il}. + +@item @uref{http://www.mysql.com/Downloads/Contrib/wu-ftpd-2.6.0-mysql.4.tar.gz,wu-ftpd-2.6.0-mysql.4.tar.gz} +Patches to add logging to @strong{MySQL} for WU-ftpd 2.6.0. By +@email{takeshi@@SoftAgency.co.jp}, based on Zeev Suraski wuftpd patches. + +@item @uref{http://www.mysql.com/Downloads/Contrib/Old-Versions, Old-Versions} +Previous versions of things found here that you probably won't be +interested in. @end itemize -@node News-3.23.34a, News-3.23.34, News-3.23.35, News-3.23.x -@appendixsubsec Changes in release 3.23.34a +@page +@node Credits, News, Contrib, Top +@appendix Credits + +This appendix lists the developers, contributors, and supporters that have +helped to make @strong{MySQL} what it is today. + +@cindex developers, list of +@menu +* Developers:: +* Contributors:: +* Supporters:: +@end menu + +@node Developers, Contributors, Credits, Credits +@appendixsubsec Developers at MySQL AB + +These are the developers that are or have been employed by @strong{MySQL AB} +to work on @strong{MySQL}, roughly in the order they started to work with us. +Following each developer is a small list of the tasks that the developer is +responsible for, or the accomplishments they have made. + +@table @asis +@item Michael (Monty) Widenius +Has written the following parts of @strong{MySQL}: @itemize @bullet @item -Added extra files to the distribution to allow @code{INNOBASE} support -to be compiled. -@end itemize - -@node News-3.23.34, News-3.23.33, News-3.23.34a, News-3.23.x -@appendixsubsec Changes in release 3.23.34 -@itemize @bullet +All the main code in @code{mysqld}. @item -Added the @code{INNOBASE} table handler and the @code{BDB} table handler -to the @strong{MySQL} source distribution. +New functions for the string library. @item -Updated the documentation about @code{GEMINI} tables. +Most of the @code{mysys} library. @item -Fixed a bug in @code{INSERT DELAYED} that caused threads to hang when -inserting @code{NULL} into an @code{AUTO_INCREMENT} column. +The @code{ISAM} and @code{MyISAM} libraries (B-tree index file +handlers with index compression and different record formats). @item -Fixed a bug in @code{CHECK TABLE} / @code{REPAIR TABLE} that could cause -a thread to hang. +The @code{HEAP} library. A memory table system with our superior full dynamic +hashing. In use since 1981 and published around 1984. @item -@code{REPLACE} will not replace a row that conflicts with an -@code{auto_increment} generated key. +The @code{replace} program (look into it, it's COOL!). @item -@code{mysqld} now only sets @code{CLIENT_TRANSACTIONS} in -@code{mysql->server_capabilities} if the server supports a -transaction-safe handler. +@strong{MyODBC}, the ODBC driver for Windows95. @item -Fixed @code{LOAD DATA INFILE} to allow numeric values to be read into -@code{ENUM} and @code{SET} columns. +Fixing bugs in MIT-pthreads to get it to work for @strong{MySQL}. And +also Unireg, a curses-based application tool with many utilities. @item -Improved error diagnostic for slave thread exit. +Porting of @code{mSQL} tools like @code{msqlperl}, @code{DBD}/@code{DBI}, and +@code{DB2mysql}. @item -Fixed bug in @code{ALTER TABLE ... ORDER BY}. +Most of crash-me and the foundation for the @strong{MySQL} benchmarks. +@end itemize + +@item David Axmark +@itemize @bullet @item -Added option @code{max_user_connections} to @code{mysqld}. +Coordinator and initial main writer of the @strong{Reference Manual}, +including enhancements to @code{texi2html}. @item -Limit query length for replication by @code{max_allowed_packet}, not the -arbitrary limit of 4 MB. +Automatic Web site updating from the manual. @item -Allow space around @code{=} in argument to @code{--set-variable}. +Initial Autoconf, Automake, and Libtool support. @item -Fixed problem in automatic repair that could leave some threads in state -@code{Waiting for table}. +The licensing stuff. @item -@code{SHOW CREATE TABLE} now dumps the @code{UNION()} for @code{MERGE} tables. +Parts of all the text files. (Nowadays only the @file{README} is +left. The rest ended up in the manual.) @item -@code{ALTER TABLE} now remembers the old @code{UNION()} definition. +Lots of testing of new features. @item -Fixed bug when replicating timestamps. +Our in-house ``free'' software lawyer. @item -Fixed bug in bidirectional replication. +Mailing list maintainer (who never has the time to do it right...) @item -Fixed bug in the @code{BDB} table handler that occurred when using an index -on multi-part key where a key part may be @code{NULL}. +Our original portability code (more than 10 years old now). Nowadays +only some parts of @code{mysys} are left. @item -Fixed @code{MAX()} optimization on sub-key for @code{BDB} tables. +Someone for Monty to call in the middle of the night when he just got +that new feature to work. +@end itemize + +@item Jani Tolonen +@itemize @bullet @item -Fixed problem where garbage results were returned when using @code{BDB} -tables and @code{BLOB} or @code{TEXT} fields when joining many tables. +@code{mysqlimport} @item -Fixed a problem with @code{BDB} tables and @code{TEXT} columns. +A lot of extensions to the @code{mysql} client. @item -Fixed bug when using a @code{BLOB} key where a const row wasn't found. +@code{procedure analyse()} +@end itemize + +@item Sinisa Milivojevic +@itemize @bullet @item -Fixed that @code{mysqlbinlog} writes the timestamp value for each query. -This ensures that one gets same values for date functions like @code{NOW()} -when using @code{mysqlbinlog} to pipe the queries to another server. +Compression (with @code{zlib}) in the client/server protocol. @item -Allow one to use @code{--skip-gemini}, @code{--skip-bdb} and -@code{--skip-innobase} to @code{mysqld} even if these databases are not -compiled in @code{mysqld}. +Perfect hashing for the lexical analyzer phase. @item -One can now do @code{GROUP BY ... DESC}. +The MySQLGUI client. @item -Fixed a deadlock in the @code{SET} code, when one ran @code{SET @@foo=bar}, -where @code{bar} is a column reference, an error was not properly generated. +Maintainer of mysql++. @end itemize -@node News-3.23.33, News-3.23.32, News-3.23.34, News-3.23.x -@appendixsubsec Changes in release 3.23.33 +@item Tonu Samuel @itemize @bullet @item -Fixed that DNS lookups are not using the same mutex as the hostname -cache. This will enable known hosts to be quickly resolved even if a -DNS lookup takes a long time. +Our security expert. @item -Added @code{--character-sets-dir} to @code{myisampack}. +Vio interface (The foundation for the encrypted client/server protocol). @item -Removed warnings when running @code{REPAIR TABLE ... EXTENDED}. +@strong{MySQL} Filesystem (A way to use @strong{MySQL} databases as files +and directories). @item -Fixed a bug that caused a core dump when using @code{GROUP BY} on an alias, -where the alias was the same as an existing column name. +The CASE Expression. @item -Added @code{SEQUENCE()} as an example UDF function. +The MD5() and COALESCE() functions. @item -Changed @code{mysql_install_db} to use @code{BINARY} for @code{CHAR} -columns in the privilege tables. +@code{RAID} support for @code{MyISAM} tables. +@end itemize + +@item Sasha Pachev +@itemize @bullet @item -Changed @code{TRUNCATE table_name} to @code{TRUNCATE TABLE table_name} -to use the same syntax as Oracle. Until 4.0 we will also allow -@code{TRUNCATE table_name} to not crash old code. +Replication. @item -Fixed 'no found rows' bug in @code{MyISAM} tables when a @code{BLOB} was -first part of a multi-part key. +@code{SHOW CREATE TABLE}. @item -Fixed bug where @code{CASE} didn't work with @code{GROUP BY}. +mod_mysql_include @item -Added option @code{--sort-recover} to @code{myisamchk}. +cgi++ @item -@code{myisamchk -S} and @code{OPTIMIZE TABLE} now work on Windows. +mysql-bench +@end itemize + +@item Matt Wagner +@itemize @bullet @item -Fixed bug when using @code{DISTINCT} on results from functions that referred -to a group function, like: -@example -SELECT a, DISTINCT SEC_TO_TIME(sum(a)) from table_name GROUP BY a, b; -@end example +@strong{MySQL} test suite. @item -Fixed buffer overrun in @code{libmysqlclient} library. -Fixed bug in handling @code{STOP} event after @code{ROTATE} event in -replication. +Our webmaster. +@end itemize + +@item Miguel Solorzano +@itemize @bullet @item -Fixed another buffer overrun in @code{DROP DATABASE}. +Winmysqladmin. +@end itemize + +@item Timothy Smith +@itemize @bullet @item -Added @code{Table_locks_immediate} and @code{Table_locks_waited} status -variables. +Dynamic character support. @item -Fixed bug in replication that broke slave server start with existing -@code{master.info}. This fixes a bug introduced in 3.23.32. +Responsible for @strong{MySQL} configure. +@end itemize + +@item Sergei Golubchik +@itemize @bullet @item -Added @code{SET SQL_SLAVE_SKIP_COUNTER=n} command to recover from -replication glitches without a full database copy. +Full-text search. @item -Added @code{max_binlog_size} variable; the binary log will be rotated -automatically when the size crosses the limit. +Added keys to the @code{MERGE} library. +@end itemize + +@item Jeremy Cole +@itemize @bullet @item -Added @code{Last_error}, @code{Last_errno}, and @code{Slave_skip_counter} to -@code{SHOW SLAVE STATUS}. +Proofreading and editing this fine manual. @item -Fixed bug in @code{MASTER_POS_WAIT()} function. +@code{ALTER TABLE ... ORDER BY ...}. @item -Execute coredump handler on @code{SIGILL}, and @code{SIGBUS} in addition to -@code{SIGSEGV}. +@code{UPDATE ... ORDER BY ...}. @item -On x86 Linux, print the current query and thread (connection) id, if -available, in the coredump handler. +@code{DELETE ... ORDER BY ...}. +@end itemize + +@item John Dean +@itemize @bullet @item -Fixed several timing bugs in the test suite. +The new @strong{MySQL} GUI client. +@end itemize + +@item Indrek Siitan +@itemize @bullet @item -Extended @code{mysqltest} to take care of the timing issues in the test -suite. +Designer/programmer of our web interface. +@end itemize +@end table + +The following non-developers are also working in/with @strong{MySQL AB}: + +@table @asis @item -@code{ALTER TABLE} can now be used to change the definition for a -@code{MERGE} table. +Hans Kierkegaard - Responsible for @strong{MySQL} license handling. @item -Fixed creation of @code{MERGE} tables on Windows. +Antti Halonen - Sales manager. @item -Portability fixes for OpenBSD and OS2. +Jonas Norrman - Handles licensing questions sent to @email{info@@mysql.com}. @item -Added @code{--temp-pool} option to @code{mysqld}. Using this option -will cause most temporary files created to use a small set of names, -rather than a unique name for each new file. This is to work around a -problem in the Linux kernel dealing with creating a bunch of new files -with different names. With the old behavior, Linux seems to "leak" -memory, as it's being allocated to the directory entry cache instead of -the disk cache. -@end itemize +Erik Granberg - Handles @strong{MySQL} partners (and a lot of other stuff). +@item +Allan Larsson (The BOSS for TCX DataKonsult AB). +@end table + +@cindex contributors, list of +@node Contributors, Supporters, Developers, Credits +@appendixsubsec Contributors to MySQL + +Contributors to the @strong{MySQL} distribution are listed here, in +somewhat random order: + +@table @asis +@item Paul DuBois +Help with making the Reference Manual correct and understandable. That +includes rewriting Monty's and David's attempts at English into English +as other people know it. +@item Gianmassimo Vigazzola @email{qwerg@@mbox.vol.it} or @email{qwerg@@tin.it} +The initial port to Win32/NT. +@item Kim Aldale +Helped to rewrite Monty's and David's early attempts at English into +English. +@item Per Eric Olsson +For more or less constructive criticism and real testing of the dynamic +record format. +@item Irena Pancirov @email{irena@@mail.yacc.it} +Win32 port with Borland compiler. @code{mysqlshutdown.exe} and +@code{mysqlwatch.exe} +@item David J. Hughes +For the effort to make a shareware SQL database. We at TcX started with +@code{mSQL}, but found that it couldn't satisfy our purposes so instead we +wrote a SQL interface to our application builder Unireg. @code{mysqladmin} +and @code{mysql} are programs that were largely influenced by their +@code{mSQL} counterparts. We have put a lot of effort into making the +@strong{MySQL} syntax a superset of @code{mSQL}. Many of the API's ideas are +borrowed from @code{mSQL} to make it easy to port free @code{mSQL} programs +to @strong{MySQL}. @strong{MySQL} doesn't contain any code from @code{mSQL}. +Two files in the distribution (@file{client/insert_test.c} and +@file{client/select_test.c}) are based on the corresponding (non-copyrighted) +files in the @code{mSQL} distribution, but are modified as examples showing +the changes necessary to convert code from @code{mSQL} to @strong{MySQL}. +(@code{mSQL} is copyrighted David J. Hughes.) +@item Fred Fish +For his excellent C debugging and trace library. Monty has made a number +of smaller improvements to the library (speed and additional options). +@item Richard A. O'Keefe +For his public domain string library. +@item Henry Spencer +For his regex library, used in @code{WHERE column REGEXP regexp}. +@item Free Software Foundation +From whom we got an excellent compiler (@code{gcc}), the @code{libc} library +(from which we have borrowed @file{strto.c} to get some code working in Linux), +and the @code{readline} library (for the @code{mysql} client). +@item Free Software Foundation & The XEmacs development team +For a really great editor/environment used by almost everybody at +TcX/MySQL AB/detron. +@item Patrick Lynch +For helping us acquire @code{http://www.mysql.com/}. +@item Fred Lindberg +For setting up qmail to handle the @strong{MySQL} mailing list and for the +incredible help we got in managing the @strong{MySQL} mailing lists. +@item Igor Romanenko @email{igor@@frog.kiev.ua} +@code{mysqldump} (previously @code{msqldump}, but ported and enhanced by +Monty). +@item Tim Bunce, Alligator Descartes +For the @code{DBD} (Perl) interface. +@item Tim Bunce +Author of @code{mysqlhotcopy}. +@item Andreas Koenig @email{a.koenig@@mind.de} +For the Perl interface to @strong{MySQL}. +@item Eugene Chan @email{eugene@@acenet.com.sg} +For porting PHP to @strong{MySQL}. +@item Michael J. Miller Jr. @email{mke@@terrapin.turbolift.com} +For the first @strong{MySQL} manual. And a lot of spelling/language +fixes for the FAQ (that turned into the @strong{MySQL} manual a long +time ago). +@item Yan Cailin +First translator of the @strong{MySQL} Reference Manual into simplified +chinese in early 2000 on which the Big5 and HK coded +(@uref{http://mysql.hitstar.com, mysql.hitstar.com}) versions were +based. @uref{http://linuxdb.yeah.net, Personal home page at +linuxdb.yeah.net}. +@item Giovanni Maruzzelli @email{maruzz@@matrice.it} +For porting iODBC (Unix ODBC). +@item Chris Provenzano +Portable user level pthreads. From the copyright: This product includes +software developed by Chris Provenzano, the University of California, +Berkeley, and contributors. We are currently using version 1_60_beta6 +patched by Monty (see @file{mit-pthreads/Changes-mysql}). +@item Xavier Leroy @email{Xavier.Leroy@@inria.fr} +The author of LinuxThreads (used by @strong{MySQL} on Linux). +@item Zarko Mocnik @email{zarko.mocnik@@dem.si} +Sorting for Slovenian language and the @file{cset.tar.gz} module that makes +it easier to add other character sets. +@item "TAMITO" @email{tommy@@valley.ne.jp} +The @code{_MB} character set macros and the ujis and sjis character sets. +@item Joshua Chamas @email{joshua@@chamas.com} +Base for concurrent insert, extended date syntax, debugging on NT, and +answering on the @strong{MySQL} mailing list. +@item Yves Carlier @email{Yves.Carlier@@rug.ac.be} +@code{mysqlaccess}, a program to show the access rights for a user. +@item Rhys Jones @email{rhys@@wales.com} (And GWE Technologies Limited) +For the JDBC, a module to extract data from @strong{MySQL} with a Java +client. +@item Dr Xiaokun Kelvin ZHU @email{X.Zhu@@brad.ac.uk} +Further development of the JDBC driver and other @strong{MySQL}-related +Java tools. +@item James Cooper @email{pixel@@organic.com} +For setting up a searchable mailing list archive at his site. +@item Rick Mehalick @email{Rick_Mehalick@@i-o.com} +For @code{xmysql}, a graphical X client for @strong{MySQL}. +@item Doug Sisk @email{sisk@@wix.com} +For providing RPM packages of @strong{MySQL} for RedHat Linux. +@item Diemand Alexander V. @email{axeld@@vial.ethz.ch} +For providing RPM packages of @strong{MySQL} for RedHat Linux-Alpha. +@item Antoni Pamies Olive @email{toni@@readysoft.es} +For providing RPM versions of a lot of @strong{MySQL} clients for Intel +and SPARC. +@item Jay Bloodworth @email{jay@@pathways.sde.state.sc.us} +For providing RPM versions for @strong{MySQL} Version 3.21. +@item Jochen Wiedmann @email{wiedmann@@neckar-alb.de} +For maintaining the Perl @code{DBD::mysql} module. +@item Therrien Gilbert @email{gilbert@@ican.net}, Jean-Marc Pouyot @email{jmp@@scalaire.fr} +French error messages. +@item Petr snajdr, @email{snajdr@@pvt.net} +Czech error messages. +@item Jaroslaw Lewandowski @email{jotel@@itnet.com.pl} +Polish error messages. +@item Miguel Angel Fernandez Roiz +Spanish error messages. +@item Roy-Magne Mo @email{rmo@@www.hivolda.no} +Norwegian error messages and testing of Version 3.21.#. +@item Timur I. Bakeyev @email{root@@timur.tatarstan.ru} +Russian error messages. +@item @email{brenno@@dewinter.com} && Filippo Grassilli @email{phil@@hyppo.com} +Italian error messages. +@item Dirk Munzinger @email{dirk@@trinity.saar.de} +German error messages. +@item Billik Stefan @email{billik@@sun.uniag.sk} +Slovak error messages. +@item Stefan Saroiu @email{tzoompy@@cs.washington.edu} +Romanian error messages. +@item Peter Feher +Hungarian error messages. +@item Roberto M. Serqueira +Portugise error messages. +@item David Sacerdote @email{davids@@secnet.com} +Ideas for secure checking of DNS hostnames. +@item Wei-Jou Chen @email{jou@@nematic.ieo.nctu.edu.tw} +Some support for Chinese(BIG5) characters. +@item Wei He @email{hewei@@mail.ied.ac.cn} +A lot of functionality for the Chinese(GBK) character set. +@item Zeev Suraski @email{bourbon@@netvision.net.il} +@code{FROM_UNIXTIME()} time formatting, @code{ENCRYPT()} functions, and +@code{bison} advisor. +Active mailing list member. +@item Luuk de Boer @email{luuk@@wxs.nl} +Ported (and extended) the benchmark suite to @code{DBI}/@code{DBD}. Have +been of great help with @code{crash-me} and running benchmarks. Some new +date functions. The mysql_setpermissions script. +@item Jay Flaherty @email{fty@@mediapulse.com} +Big parts of the Perl @code{DBI}/@code{DBD} section in the manual. +@item Paul Southworth @email{pauls@@etext.org}, Ray Loyzaga @email{yar@@cs.su.oz.au} +Proof-reading of the Reference Manual. +@item Alexis Mikhailov @email{root@@medinf.chuvashia.su} +User-definable functions (UDFs); @code{CREATE FUNCTION} and +@code{DROP FUNCTION}. +@item Andreas F. Bobak @email{bobak@@relog.ch} +The @code{AGGREGATE} extension to UDF functions. +@item Ross Wakelin @email{R.Wakelin@@march.co.uk} +Help to set up InstallShield for @strong{MySQL}-Win32. +@item Jethro Wright III @email{jetman@@li.net} +The @file{libmysql.dll} library. +@item James Pereria @email{jpereira@@iafrica.com} +Mysqlmanager, a Win32 GUI tool for administrating @strong{MySQL}. +@item Curt Sampson @email{cjs@@portal.ca} +Porting of MIT-pthreads to NetBSD/Alpha and NetBSD 1.3/i386. +@item Antony T. Curtis @email{antony.curtis@@olcs.net} +Porting of @strong{MySQL} to OS/2. +@item Martin Ramsch @email{m.ramsch@@computer.org} +Examples in the @strong{MySQL} Tutorial. +@item Steve Harvey +For making @code{mysqlaccess} more secure. +@item Konark IA-64 Centre of Persistent Systems Private Limited +@uref{http://www.pspl.co.in/konark/}. Help with the Win64 port of the +@strong{MySQL} server. +@item Albert Chin-A-Young. +Configure updates for Tru64, large file support and better TCP wrappers +support. +@item Valueclick Inc. +For sponsoring the optimize section in this manual. +@end table + +Other contributors, bugfinders, and testers: James H. Thompson, Maurizio +Menghini, Wojciech Tryc, Luca Berra, Zarko Mocnik, Wim Bonis, Elmar +Haneke, @email{jehamby@@lightside}, @email{psmith@@BayNetworks.com}, +@email{duane@@connect.com.au}, Ted Deppner @email{ted@@psyber.com}, +Mike Simons, Jaakko Hyvatti. + +And lots of bug report/patches from the folks on the mailing list. + +A big tribute goes to those that help us answer questions on the +@code{mysql@@lists.mysql.com} mailing list: + +@table @asis +@item Daniel Koch @email{dkoch@@amcity.com} +Irix setup. +@item Luuk de Boer @email{luuk@@wxs.nl} +Benchmark questions. +@item Tim Sailer @email{tps@@users.buoy.com} +@code{DBD-mysql} questions. +@item Boyd Lynn Gerber @email{gerberb@@zenez.com} +SCO-related questions. +@item Richard Mehalick @email{RM186061@@shellus.com} +@code{xmysql}-related questions and basic installation questions. +@item Zeev Suraski @email{bourbon@@netvision.net.il} +Apache module configuration questions (log & auth), PHP-related +questions, SQL syntax-related questions and other general questions. +@item Francesc Guasch @email{frankie@@citel.upc.es} +General questions. +@item Jonathan J Smith @email{jsmith@@wtp.net} +Questions pertaining to OS-specifics with Linux, SQL syntax, and other +things that might need some work. +@item David Sklar @email{sklar@@student.net} +Using @strong{MySQL} from PHP and Perl. +@item Alistair MacDonald @email{A.MacDonald@@uel.ac.uk} +Not yet specified, but is flexible and can handle Linux and maybe HP-UX. +Will try to get user to use @code{mysqlbug}. +@item John Lyon @email{jlyon@@imag.net} +Questions about installing @strong{MySQL} on Linux systems, using either +@file{.rpm} files or compiling from source. +@item Lorvid Ltd. @email{lorvid@@WOLFENET.com} +Simple billing/license/support/copyright issues. +@item Patrick Sherrill @email{patrick@@coconet.com} +ODBC and VisualC++ interface questions. +@item Randy Harmon @email{rjharmon@@uptimecomputers.com} +@code{DBD}, Linux, some SQL syntax questions. +@end table + +@cindex contributing companies, list of +@node Supporters, , Contributors, Credits +@appendixsubsec Supporters to MySQL + +The following companies has helped us finance development of +@strong{MySQL} by either paying us for developing a new feature, +developed a @strong{MySQL} feature themselves or by giving us hardware for +@strong{MySQL} development. + +@table @asis +@item VA Linux / Andover.net +Funded replication. +@item NuSphere +Editing of the @strong{MySQL} manual. +@item Stork Design studio +The @strong{MySQL} web site in use between 1998-2000. +@item Intel +Contributed to development on Windows and Linux platforms. +@item Compaq +Contributed to Development on Linux/Alpha. +@item SWSoft +Development on the embedded @code{mysqld} version. +@item FutureQuest +@code{--skip-show-variables} +@end table + +@node News, Porting, Credits, Top +@appendix MySQL change history + +This appendix lists the changes from version to version in the @strong{MySQL} +source code. + +Note that we tend to update the manual at the same time we make changes to +@strong{MySQL}. If you find a version listed below that you can't +find on the +@uref{http://www.mysql.com/downloads/,@strong{MySQL} download page}, +this means that the version has not yet been released! + +@menu +* News-4.0.x:: Changes in release 4.0.x (Development; Alpha) +* News-3.23.x:: Changes in release 3.23.x (Stable) +* News-3.22.x:: Changes in release 3.22.x (Older; Still supported) +* News-3.21.x:: Changes in release 3.21.x +* News-3.20.x:: Changes in release 3.20.x +* News-3.19.x:: Changes in release 3.19.x +@end menu -@node News-3.23.32, News-3.23.31, News-3.23.33, News-3.23.x -@appendixsubsec Changes in release 3.23.32 -@itemize @bullet -@item -Changed code to get around compiler bug in Compaq C++ on OSF1, that broke -@code{BACKUP}, @code{RESTORE}, @code{CHECK}, @code{REPAIR}, and -@code{ANALYZE TABLE}. -@item -Added option @code{FULL} to @code{SHOW COLUMNS}. Now we show the -privilege list for the columns only if this option is given. -@item -Fixed bug in @code{SHOW LOGS} when there weren't any BDB logs. -@item -Fixed a timing problem in replication that could delay sending an update -to the client until a new update was done. -@item -Don't convert field names when using @code{mysql_list_fields()}. This is -to keep this code compatible with @code{SHOW FIELDS}. -@item -@code{MERGE} tables didn't work on Windows. -@item -Fixed problem with @code{SET PASSWORD=...} on Windows. -@item -Added missing @file{my_config.h} to RPM distribution. -@item -@code{TRIM("foo" from "foo")} didn't return an empty string. -@item -Added @code{--with-version-suffix} to @code{configure}. -@item -Fixed coredump when client aborted connection without @code{mysql_close()}. -@item -Fixed a bug in @code{RESTORE TABLE} when trying to restore from a non-existent -directory. -@item -Fixed a bug which caused a core dump on the slave when replicating -@code{SET PASSWORD}. -@item -Added @code{MASTER_POS_WAIT()}. -@end itemize +@node News-4.0.x, News-3.23.x, News, News +@appendixsec Changes in release 4.0.x (Development; Alpha) + +We have now started to work on @strong{MySQL} 4.0. We will update this +section as we add new features, so that others can follow our development. + +Our TODO section contains what we plan to have in 4.0. @xref{TODO MySQL 4.0}. + +@menu +* News-4.0.0:: Changes in release 4.0.0 +@end menu + +@node News-4.0.0, , News-4.0.x, News-4.0.x +@appendixsubsec Changes in release 4.0.0 -@node News-3.23.31, News-3.23.30, News-3.23.32, News-3.23.x -@appendixsubsec Changes in release 3.23.31 @itemize @bullet @item -The test suite now tests all reachable BDB interface code. During -testing we found and fixed many errors in the interface code. +Multi-table @code{DELETE}. @item -Using @code{HAVING} on an empty table could produce one result row when -it shouldn't. +Don't support old client protocols prior to @strong{MySQL} 3.21 any more. @item -Fixed that the @strong{MySQL} RPM is not dependent on Perl5 anymore. +Don't include the old C API functions @code{mysql_drop_db}, +@code{mysql_create_db} and @code{mysql_connect}, unless compiled with +@code{USE_OLD_FUNCTIONS}. @item -Fixed some problems with @code{HEAP} tables on Windows. +Renamed @code{safe_mysqld} to @code{mysqld_safe}. @item -@code{SHOW TABLE STATUS} didn't show correct average row length for tables -larger than 4G. +Allow @code{IN} as a synonym for @code{FROM} in @code{SHOW} commands. @item -@code{CHECK TABLE ... EXTENDED} didn't check row links for fixed size tables. +@code{SHOW INDEXES} is now a synonym for @code{SHOW INDEX}. @item -Added option @code{MEDIUM} to @code{CHECK TABLE}. +Added support for symbolic links to @code{MyISAM} tables. Symlink handling is +now enabled by default for Windows. @item -Fixed problem when using @code{DECIMAL()} keys on negative numbers. +@code{LOAD DATA FROM MASTER} ``auto-magically'' sets up a slave. @item -@code{HOUR()} (and some other @code{TIME} functions) on a @code{CHAR} column -always returned @code{NULL}. +A new @code{HANDLER} interface to @code{MyISAM} tables. @item -Fixed security bug in something (please upgrade if you are using a earlier -@strong{MySQL} 3.23 version). +@code{COUNT(DISTINCT)} is about 30% faster. @item -Fixed buffer overflow bug when writing a certain error message. +@code{FULLTEXT} index creation now is much faster. @item -Added usage of @code{setrlimit()} on Linux to get -@code{-O --open-files-limit=#} to work on Linux. +Searching on packed (@code{CHAR}/@code{VARCHAR}) keys now is much faster. @item -Added new @code{mysqld} variable: @code{bdb_version}. +Added @code{SQL_CALC_FOUND_ROWS} and @code{FOUND_ROWS()}. This makes it +possible to know how many rows a query would have returned +without a @code{LIMIT} clause. @item -Fixed bug when using expression of type: -@example -SELECT ... FROM t1 LEFT JOIN t2 ON (t1.a=t2.a) WHERE t1.a=t2.a -@end example -In this case the test in the @code{WHERE} clause was wrongly optimized away. +Changed output format of @code{SHOW OPEN TABLES}. @item -Fixed bug in @code{MyISAM} when deleting keys with possible @code{NULL} -values, but the first key-column was not a prefix-compressed text column. +Allow @code{SELECT expression LIMIT ...}. @item -Fixed @code{mysql.server} to read the @code{mysql.server} option section -instead of @code{mysql_server}. +Added @code{IDENTITY} as a synonym for @code{AUTO_INCREMENT} (like Sybase). @item -Fixed @code{safe_mysqld} and @code{mysql.server} to also read the -@code{server} option section. +Added @code{ORDER BY} syntax to @code{UPDATE} and @code{DELETE}. @item -Added @code{Threads_created} status variable to @code{mysqld}. +Optimized queries of type: +@code{SELECT DISTINCT * from table_name ORDER by key_part1 LIMIT #} @end itemize -@node News-3.23.30, News-3.23.29, News-3.23.31, News-3.23.x -@appendixsubsec Changes in release 3.23.30 +@node News-3.23.x, News-3.22.x, News-4.0.x, News +@appendixsec Changes in release 3.23.x (Stable) + +The 3.23 release has several major features that are not +present in previous versions. We have added three new +table types: + +@table @asis +@item @strong{MyISAM} +A new ISAM library which is tuned for SQL and supports large files. +@item @strong{BerkeleyDB} or @strong{BDB} +Uses the Berkeley DB library from Sleepycat Software to implement +transaction-safe tables. +@item @strong{InnoDB} +A transaction-safe table handler that supports row level locking, and many +Oracle-like features. +@end table + +Note that only MyISAM is available in the standard binary distribution. + +The 3.23 release also includes support for database replication between +a master and many slaves, full-text indexing, and much more. + +All new features are being developed in the 4.0 version. Only +bug fixes and minor enhancements to existing features will be +added to 3.23. + +The replication code and BerkeleyDB code is still not as tested and as +the rest of the code, so we will probably need to do a couple of future +releases of 3.23 with small fixes for this part of the code. As long as +you don't use these features, you should be quite safe with +@strong{MySQL} 3.23! + +Note that the above doesn't mean that replication or Berkeley DB don't +work; We have done a lot of testing of all code, including replication +and BDB without finding any problems. It only means that not as many +users use this code as the rest of the code and because of this we are +not yet 100% confident in this code. + +@menu +* News-3.23.40:: Changes in release 3.23.40 +* News-3.23.39:: Changes in release 3.23.39 +* News-3.23.38:: Changes in release 3.23.38 +* News-3.23.37:: Changes in release 3.23.37 +* News-3.23.36:: Changes in release 3.23.36 +* News-3.23.35:: Changes in release 3.23.35 +* News-3.23.34a:: Changes in release 3.23.34a +* News-3.23.34:: Changes in release 3.23.34 +* News-3.23.33:: Changes in release 3.23.33 +* News-3.23.32:: Changes in release 3.23.32 +* News-3.23.31:: Changes in release 3.23.31 +* News-3.23.30:: Changes in release 3.23.30 +* News-3.23.29:: Changes in release 3.23.29 +* News-3.23.28:: Changes in release 3.23.28 +* News-3.23.27:: Changes in release 3.23.27 +* News-3.23.26:: Changes in release 3.23.26 +* News-3.23.25:: Changes in release 3.23.25 +* News-3.23.24:: Changes in release 3.23.24 +* News-3.23.23:: Changes in release 3.23.23 +* News-3.23.22:: Changes in release 3.23.22 +* News-3.23.21:: Changes in release 3.23.21 +* News-3.23.20:: Changes in release 3.23.20 +* News-3.23.19:: Changes in release 3.23.19 +* News-3.23.18:: Changes in release 3.23.18 +* News-3.23.17:: Changes in release 3.23.17 +* News-3.23.16:: Changes in release 3.23.16 +* News-3.23.15:: Changes in release 3.23.15 +* News-3.23.14:: Changes in release 3.23.14 +* News-3.23.13:: Changes in release 3.23.13 +* News-3.23.12:: Changes in release 3.23.12 +* News-3.23.11:: Changes in release 3.23.11 +* News-3.23.10:: Changes in release 3.23.10 +* News-3.23.9:: Changes in release 3.23.9 +* News-3.23.8:: Changes in release 3.23.8 +* News-3.23.7:: Changes in release 3.23.7 +* News-3.23.6:: Changes in release 3.23.6 +* News-3.23.5:: Changes in release 3.23.5 +* News-3.23.4:: Changes in release 3.23.4 +* News-3.23.3:: Changes in release 3.23.3 +* News-3.23.2:: Changes in release 3.23.2 +* News-3.23.1:: Changes in release 3.23.1 +* News-3.23.0:: Changes in release 3.23.0 +@end menu + +@node News-3.23.40, News-3.23.39, News-3.23.x, News-3.23.x +@appendixsubsec Changes in release 3.23.40 @itemize @bullet @item -Added @code{SHOW OPEN TABLES} command. -@item -Fixed that @code{myisamdump} works against old @code{mysqld} servers. -@item -Fixed @code{myisamchk -k#} so that it works again. -@item -Fixed a problem with replication when the binary log file went over 2G -on 32-bit systems. -@item -@code{LOCK TABLES} will now automatically start a new transaction. +Fixed problem with @code{--low-priority-updates} and @code{INSERT}'s. @item -Changed @code{BDB} tables to not use internal subtransactions and reuse -open files to get more speed. +Fixed bug in slave thread when under some rare circumstances it could +get 22 bytes ahead on the offset in the master. @item -Added option @code{--mysqld=#} to @code{safe_mysqld}. +Added @code{slave_wait_timeout} for replication. @item -Allow hex constants in the @code{--fields-*-by} and -@code{--lines-terminated-by} options to @code{mysqldump} and -@code{mysqlimport}. By Paul DuBois. +Fixed problem with @code{UPDATE} and @code{BDB} tables. @item -Added option @code{--safe-show-database} to @code{mysqld}. +Fixed hard bug in @code{BDB} tables when using key parts. @item -Added @code{have_bdb}, @code{have_gemini}, @code{have_innobase}, -@code{have_raid} and @code{have_ssl} to @code{SHOW VARIABLES} to make it -easy to test for supported extensions. +Fixed problem when using the @code{GRANT FILE ON database.* ...}; Previously +we added the @code{DROP} privilege for the database. @item -Added option @code{--open-files-limit} to @code{mysqld}. +Fixed @code{DELETE FROM table_name ... LIMIT 0} and +@code{UPDATE FROM table_name ... LIMIT 0} acted as though the @code{LIMIT} +clause was not present (they deleted or updated all selected rows). @item -Changed option @code{--open-files} to @code{--open-files-limit} in -@code{safe_mysqld}. +@code{CHECK TABLE} now checks if an @code{AUTO_INCREMENT} column contains +the value 0. @item -Fixed a bug where some rows were not found with @code{HEAP} tables -that had many keys. +Sending a @code{SIGHUP} to @code{mysqld} will now only flush the logs, +not reset the replication. @item -Fixed that @code{--bdb-no-sync} works. +Fixed parser to allow floats of type @code{1.0e1} (no sign after @code{e}). @item -Changed @code{--bdb-recover} to @code{--bdb-no-recover} as recover should -be on by default. +Option @code{--force} to @code{myisamchk} now also updates states. @item -Changed the default number of BDB locks to 10000. +Added option @code{--warnings} to @code{mysqld}. Now @code{mysqld} +only prints the error @code{Aborted connection} if this option is used. @item -Fixed a bug from 3.23.29 when allocating the shared structure needed -for BDB tables. +Fixed problem with @code{SHOW CREATE TABLE} when you didn't have a +@code{PRIMARY KEY}. @item -Changed @file{mysqld_multi.sh} to use configure variables. Patch by -Christopher McCrory. +Properly fixed the rename of @code{innodb_unix_file_flush_method} to +@code{innodb_flush_method}. @item -Added fixing of include files for Solaris 2.8. +Fixed bug when converting @code{UNSIGNED BIGINT} to @code{DOUBLE}. This caused +a problem when doing comparisons with @code{BIGINT} values outside of the +signed range. @item -Fixed bug with @code{--skip-networking} on Debian Linux. +Fixed bug in @code{BDB} tables when querying empty tables. @item -Fixed problem that some temporary files where reported as having the -name @code{UNOPENED} in error messages. +Fixed a bug when using @code{COUNT(DISTINCT)} with @code{LEFT JOIN} and +there wasn't any matching rows. @item -Fixed bug when running two simultaneous @code{SHOW LOGS} queries. +Removed all documentation referring to the @code{GEMINI} table +type. @code{GEMINI} is not released under an Open Source license. @end itemize -@node News-3.23.29, News-3.23.28, News-3.23.30, News-3.23.x -@appendixsubsec Changes in release 3.23.29 +@node News-3.23.39, News-3.23.38, News-3.23.40, News-3.23.x +@appendixsubsec Changes in release 3.23.39 @itemize @bullet @item -Configure updates for Tru64, large file support, and better TCP wrapper -support. By Albert Chin-A-Young. -@item -Fixed bug in @code{<=>} operator. -@item -Fixed bug in @code{REPLACE} with BDB tables. -@item -@code{LPAD()} and @code{RPAD()} will shorten the result string if it's longer -than the length argument. -@item -Added @code{SHOW LOGS} command. -@item -Remove not used BDB logs on shutdown. -@item -When creating a table, put @code{PRIMARY} keys first, followed by -@code{UNIQUE} keys. -@item -Fixed a bug in @code{UPDATE} involving multi-part keys where one -specified all key parts both in the update and the @code{WHERE} part. In -this case @strong{MySQL} could try to update a record that didn't match -the whole @code{WHERE} part. -@item -Changed drop table to first drop the tables and then the @file{.frm} file. -@item -Fixed a bug in the hostname cache which caused @code{mysqld} to report the -hostname as @code{''} in some error messages. -@item -Fixed a bug with @code{HEAP} type tables; the variable -@code{max_heap_table_size} wasn't used. Now either @code{MAX_ROWS} or -@code{max_heap_table_size} can be used to limit the size of a @code{HEAP} -type table. -@item -Changed the default server-id to 1 for masters and 2 for slaves -to make it easier to use the binary log. -@item -Renamed variable @code{bdb_lock_max} to @code{bdb_max_lock}. -@item -Added support for @code{auto_increment} on sub-fields for BDB tables. -@item -Added @code{ANALYZE} of BDB tables. -@item -In BDB tables, we now store the number of rows; This helps to optimize queries -when we need an approximation of the number of rows. -@item -If we get an error in a multi-row statement, we now only rollback the -last statement, not the entire transaction. -@item -If you do a @code{ROLLBACK} when you have updated a non-transactional table -you will get an error as a warning. -@item -Added option @code{--bdb-shared-data} to @code{mysqld}. -@item -Added status variable @code{Slave_open_temp_tables}. -@item -Added variables @code{binlog_cache_size} and @code{max_binlog_cache_size} to -@code{mysqld}. -@item -@code{DROP TABLE}, @code{RENAME TABLE}, @code{CREATE INDEX} and -@code{DROP INDEX} are now transaction endpoints. -@item -If you do a @code{DROP DATABASE} on a symbolic linked database, both -the link and the original database is deleted. -@item -Fixed that @code{DROP DATABASE} works on OS/2. -@item -Fixed bug when doing a @code{SELECT DISTINCT ... table1 LEFT JOIN -table2 ...} when table2 was empty. -@item -Added @code{--abort-slave-event-count} and -@code{--disconnect-slave-event-count} options to @code{mysqld} for -debugging and testing of replication. -@item -Fixed replication of temporary tables. Handles everything except -slave server restart. -@item -@code{SHOW KEYS} now shows whether or not key is @code{FULLTEXT}. -@item -New script @file{mysqld_multi}. @xref{mysqld_multi, , @code{mysqld_multi}}. -@item -Added new script, @file{mysql-multi.server.sh}. Thanks to -Tim Bunce @email{Tim.Bunce@@ig.co.uk} for modifying @file{mysql.server} to -easily handle hosts running many @code{mysqld} processes. -@item -@file{safe_mysqld}, @file{mysql.server}, and @file{mysql_install_db} have -been modified to use @code{mysql_print_defaults} instead of various hacks -to read the @file{my.cnf} files. In addition, the handling of various -paths has been made more consistent with how @code{mysqld} handles them -by default. +The @code{AUTO_INCREMENT} sequence wasn't reset when dropping +and adding an @code{AUTO_INCREMENT} column. @item -Automatically remove Berkeley DB transaction logs that no longer are in -use. +@code{CREATE ... SELECT} now creates non-unique indexes delayed. @item -Fixed bug with several @code{FULLTEXT} indexes in one table. +Fixed problem where @code{LOCK TABLES table_name READ} followed by +@code{FLUSH TABLES} put an exclusive lock on the table. @item -Added a warning if number of rows changes on @code{REPAIR}/@code{OPTIMIZE}. +@code{REAL} @@variables with was represented with 2 digits when +converted to strings. @item -Applied patches for OS/2 by @code{Yuri Dario}. +Fixed problem that client 'hung' when @code{LOAD TABLE FROM MASTER} failed. @item -@code{FLUSH TABLES table_name} didn't always flush the index tree -to disk properly. +Running @code{myisamchk --fast --force} will no longer repair tables +that only had the open count wrong. @item -@code{--bootstrap} is now run in a separate thread. This fixes a problem -that caused @code{mysql_install_db} to core dump on some Linux machines. +Added functions to handle symbolic links to make life easier in 4.0. @item -Changed @code{mi_create()} to use less stack space. +We are now using the @code{-lcma} thread library on HP-UX 10.20 so +that @strong{MySQL} will be more stable on HP-UX. @item -Fixed bug with optimizer trying to over-optimize @code{MATCH} when used -with @code{UNIQUE} key. +Fixed problem with @code{IF()} and number of decimals in the result. @item -Changed @code{crash-me} and the @strong{MySQL} benchmarks to also work -with FrontBase. +Fixed date-part extraction functions to work with dates where day +and/or month is 0. @item -Allow @code{RESTRICT} and @code{CASCADE} after @code{DROP TABLE} to make -porting easier. +Extended argument length in option files from 256 to 512 chars. @item -Reset status variable which could cause problem if one used @code{--slow-log}. +Fixed problem with shutdown when @code{INSERT DELAYED} was waiting for +a @code{LOCK TABLE}. @item -Added variable @code{connect_timeout} to @code{mysql} and @code{mysqladmin}. +Fixed coredump bug in InnoDB when tablespace was full. @item -Added @code{connect_timeout} as an alias for @code{timeout} for option files -read by @code{mysql_options()}. +Fixed problem with @code{MERGE} tables and big tables (> 4G) when using +@code{ORDER BY}. @end itemize -@node News-3.23.28, News-3.23.27, News-3.23.29, News-3.23.x -@appendixsubsec Changes in release 3.23.28 +@node News-3.23.38, News-3.23.37, News-3.23.39, News-3.23.x +@appendixsubsec Changes in release 3.23.38 @itemize @bullet @item -Added new options @code{--pager[=...]}, @code{--no-pager}, -@code{--tee=...} and @code{--no-tee} to the @code{mysql} client. The -new corresponding interactive commands are @code{pager}, @code{nopager}, -@code{tee} and @code{notee}. @xref{mysql, , @code{mysql}}, @code{mysql --help} -and the interactive help for more information. -@item -Fixed crash when automatic repair of @code{MyISAM} table failed. -@item -Fixed a major performance bug in the table locking code when one -constantly had a LOT of @code{SELECT}, @code{UPDATE} and @code{INSERT} -statements running. The symptom was that the @code{UPDATE} and -@code{INSERT} queries were locked for a long time while new @code{SELECT} -statements were executed before the updates. -@item -When reading @code{options_files} with @code{mysql_options()} the -@code{return-found-rows} option was ignored. -@item -One can now specify @code{interactive-timeout} in the option file that -is read by @code{mysql_options()}. This makes it possible to force -programs that run for a long time (like @code{mysqlhotcopy}) to use -@code{interactive_timeout} instead of @code{wait_timeout}. -@item -Added to the slow query log the time and the user name for each logged -query. If you are using @code{--log-long-format} then also queries that -do not use an index are logged, even if the query takes less than -@code{long_query_time} seconds. -@item -Fixed a problem in @code{LEFT JOIN} which caused all columns in a reference -table to be @code{NULL}. -@item -Fixed a problem when using @code{NATURAL JOIN} without keys. -@item -Fixed a bug when using a multi-part keys where the first part was of type -@code{TEXT} or @code{BLOB}. -@item -@code{DROP} of temporary tables wasn't stored in the update/binary log. -@item -Fixed a bug where @code{SELECT DISTINCT * ... LIMIT #} only returned one row. -@item -Fixed a bug in the assembler code in @code{strstr()} for sparc and cleaned up -the @file{global.h} header file to avoid a problem with bad aliasing with -the compiler submitted with RedHat 7.0. (Reported by Trond Eivind Glomsrød) -@item -The option @code{--skip-networking} now works properly on NT. -@item -Fixed a long outstanding bug in the @code{ISAM} tables when a row with a length -of more than 65K was shortened by a single byte. -@item -Fixed a bug in @code{MyISAM} when running multiple updating processes on -the same table. -@item -Allow one to use @code{FLUSH TABLE tablename}. -@item -Added @code{--replicate-ignore-table}, @code{--replicate-do-table}, - @code{--replicate-wild-ignore-table}, @code{--replicate-wild-do-table}. -@item -Changed all log files to use our own @code{IO_CACHE} mechanism instead of -@code{FILE} to avoid OS problems when there are many files open. -@item -Added options @code{--open-files} and @code{--timezone} to @code{safe_mysqld}. -@item -Fixed a fatal bug in @code{CREATE TEMPORARY TABLE ... SELECT ...}. -@item -Fixed a problem with @code{CREATE TABLE ... SELECT NULL}. -@item -Added variables @code{large_file_support},@code{net_read_timeout}, -@code{net_write_timeout} and @code{query_buffer_size} to @code{SHOW VARIABLES}. -@item -Added status variables @code{created_tmp_files} and @code{sort_merge_passes} -to @code{SHOW STATUS}. +Fixed a bug when @code{SELECT} from @code{MERGE} table +sometimes results in incorrectly ordered rows. @item -Fixed a bug where we didn't allow an index name after the -@code{FOREIGN KEY} definition. +Fixed a bug in @code{REPLACE()} when using the ujis character set. @item -Added @code{TRUNCATE table_name} as a synonym for -@code{DELETE FROM table_name}. +Applied Sleepycat BDB patches 3.2.9.1 and 3.2.9.2. @item -Fixed a bug in a BDB key compare function when comparing part keys. +Added option @code{--skip-stack-trace} to @code{mysqld}. @item -Added variable @code{bdb_lock_max} to @code{mysqld}. +@code{CREATE TEMPORARY} now works with @code{InnoDB} tables. @item -Added more tests to the benchmark suite. +@code{InnoDB} now promotes sub keys to whole keys. @item -Fixed an overflow bug in the client code when using overly long database names. +Added option @code{CONCURRENT} to @code{LOAD DATA}. @item -@code{mysql_connect()} now aborts on Linux if the server doesn't answer in -@code{timeout} seconds. +Better error message when slave @code{max_allowed_packet} is too low to +read a very long log event from the master. @item -@code{SLAVE START} did not work if you started with -@code{--skip-slave-start} and had not explicitly run @code{CHANGE MASTER TO}. +Fixed bug when too many rows where removed when using +@code{SELECT DISTINCT ... HAVING}. @item -Fixed the output of @code{SHOW MASTER STATUS} to be consistent with -@code{SHOW SLAVE STATUS}. (It now has no directory in the log name.) +@code{SHOW CREATE TABLE} now returns @code{TEMPORARY} for temporary tables. @item -Added @code{PURGE MASTER LOGS TO}. +Added @code{Rows_examined} to slow query log. @item -Added @code{SHOW MASTER LOGS}. +Fixed problems with function returning empty string when using +together with a group functions and a @code{WHERE} that didn't match any rows. @item -Added @code{--safemalloc-mem-limit} option to @code{mysqld} to simulate memory -shortage when compiled @code{--with-debug=full}. +New program @code{mysqlcheck}. @item -Fixed several coredumps in out-of-memory conditions. +Added database name to output for administrative commands like @code{CHECK}, +@code{REPAIR}, @code{OPTIMIZE}. @item -@code{SHOW SLAVE STATUS} was using an uninitialized mutex if the slave had -not been started yet. +Lots of portability fixes for InnoDB. @item -Fixed bug in @code{ELT()} and @code{MAKE_SET()} when the query used -a temporary table. +Changed optimizer so that queries like +@code{SELECT * FROM table_name,table_name2 ... ORDER BY key_part1 LIMIT #} +will use index on @code{key_part1} instead of @code{filesort}. @item -@code{CHANGE MASTER TO} without specifying @code{MASTER_LOG_POS} would -set it to 0 instead of 4 and hit the magic number in the master binlog. +Fixed bug when doing +@code{LOCK TABLE to_table WRITE,...; INSERT INTO to_table... SELECT ...} +when @code{to_table} was empty. @item -@code{ALTER TABLE ... ORDER BY ...} syntax added. This will create the -new table with the rows in a specific order. +Fixed bug with @code{LOCK TABLE} and BDB tables. @end itemize -@node News-3.23.27, News-3.23.26, News-3.23.28, News-3.23.x -@appendixsubsec Changes in release 3.23.27 +@node News-3.23.37, News-3.23.36, News-3.23.38, News-3.23.x +@appendixsubsec Changes in release 3.23.37 @itemize @bullet @item -Fixed a bug where the automatic repair of MyISAM tables sometimes failed -when the data file was corrupt. +Fixed a bug when using @code{MATCH} in @code{HAVING} clause. @item -Fixed a bug in @code{SHOW CREATE} when using @code{AUTO_INCREMENT} columns. +Fixed a bug when using @code{HEAP} tables with @code{LIKE}. @item -Changed BDB tables to use new compare function in Berkeley DB 3.2.3. +Added @code{--mysql-version} to @code{safe_mysqld} @item -You can now use Unix sockets with @code{mit-pthreads}. +Changed @code{INNOBASE} to @code{InnoDB} (because the @code{INNOBASE} +name was already used). All @code{configure} options and @code{mysqld} +start options are now using @code{innodb} instead of @code{innobase}. This +means that you have to change any configuration files where you have used +@code{innobase} options before upgrading to this version! @item -Added the latin5 (turkish) character set. +Fixed bug when using indexes on @code{CHAR(255) NULL} columns. @item -Small portability fixes. -@end itemize - -@node News-3.23.26, News-3.23.25, News-3.23.27, News-3.23.x -@appendixsubsec Changes in release 3.23.26 -@itemize @bullet +Slave thread will now be started even if @code{master-host} is not set, as +long as @code{server-id} is set and valid @code{master.info} is present @item -Fixed @code{<>} to work properly with @code{NULL}. +Partial updates (terminated with kill) are now logged with a special error +code to the binary log. Slave will refuse to execute them if the error code +indicates the update was terminated abnormally, and will have to be recovered +with @code{SET SQL_SLAVE_SKIP_COUNTER=1; SLAVE START} after a manual sanity +check/correction of data integrity. @item -Fixed a problem with @code{SUBSTRING_INDEX()} and @code{REPLACE()}. -(Patch by Alexander Igonitchev) +Fixed bug that erroneously logged a drop of internal temporary table +on thread termination to the binary log - bug affected replication. @item -Fix @code{CREATE TEMPORARY TABLE IF NOT EXISTS} not to give an error -if the table exists. +Fixed a bug in @code{REGEXP()} on 64-bit machines. @item -If you don't create a @code{PRIMARY KEY} in a BDB table, a hidden -@code{PRIMARY KEY} will be created. +@code{UPDATE} and @code{DELETE} with @code{WHERE unique_key_part IS NULL} +didn't update/delete all rows. @item -Added read-only-key optimization to BDB tables. +Disabled @code{INSERT DELAYED} for tables that support transactions. @item -@code{LEFT JOIN} in some cases preferred a full table scan when there was -no @code{WHERE} clause. +Fixed bug when using date functions on @code{TEXT}/@code{BLOB} column +with wrong date format. @item -When using @code{--log-slow-queries}, don't count the time waiting for a lock. +UDFs now also work on Windows. (Patch by Ralph Mason) @item -Fixed bug in lock code on Windows which could cause the key cache -to report that the key file was crashed even if it was okay. +Fixed bug in @code{ALTER TABLE} and @code{LOAD DATA INFILE} that disabled +key-sorting. These commands should now be faster in most cases. @item -Automatic repair of @code{MyISAM} tables if you start @code{mysqld} with -@code{--myisam-recover}. +Fixed performance bug where reopened tables (tables that had been +waiting for @code{FLUSH} or @code{REPAIR}) would not use indexes for the +next query. @item -Removed the @code{TYPE=} keyword from @code{CHECK} and -@code{REPAIR}. Allow @code{CHECK} options to be combined. (You can still -use @code{TYPE=} but this usage is deprecated.) +Fixed problem with @code{ALTER TABLE} to Innobase tables on FreeBSD. @item -Fixed mutex bug in the binary replication log - long update queries could -be read only in part by the slave if it did it at the wrong time, which -was not fatal, but resulted in a performance-degrading reconnect and -a scary message in the error log. +Added @code{mysqld} variables @code{myisam_max_sort_file_size} and +@code{myisam_max_extra_sort_file_size}. @item -Changed the format of the binary log - added magic number, server -version, binlog version. Added server id and query error code for each -query event. +Initialize signals early to avoid problem with signals in Innobase. @item -Replication thread from the slave now will kill all the stale threads -from the same server. +Applied patch for the @code{tis620} character set to make comparisons +case-independent and to fix a bug in @code{LIKE} for this character set. +@strong{NOTE}: All tables that uses the @code{tis620} character set must be +fixed with @code{myisamchk -r} or @code{REPAIR TABLE} ! @item -Long replication user names were not being handled properly. +Added @code{--skip-safemalloc} option to @code{mysqld}. +@end itemize + +@node News-3.23.36, News-3.23.35, News-3.23.37, News-3.23.x +@appendixsubsec Changes in release 3.23.36 +@itemize @bullet @item -Added @code{--replicate-rewrite-db} option to @code{mysqld}. +Fixed a bug that allowed you to use database names containing a @samp{.} +character. This fixes a serious security issue when @code{mysqld} is run +as root. @item -Added @code{--skip-slave-start} option to @code{mysqld}. +Fixed bug when thread creation failed (could happen when doing a LOT +of connections in a short time). @item -Updates that generated an error code (such as @code{INSERT INTO foo(some_key) -values (1),(1)}) erroneously terminated the slave thread. +Fixed some problems with @code{FLUSH TABLES} and @code{TEMPORARY} tables. +(Problem with freeing the key cache and error @code{Can't reopen table...}). @item -Added optimization of queries where @code{DISTINCT} is only used on columns -from some of the tables. +Fixed a problem in Innobase with other character sets than @code{latin1} +and another problem when using many columns. @item -Allow floating-point numbers where there is no sign after the exponent -(like 1e1). +Fixed bug that caused a core dump when using a very complex query involving +@code{DISTINCT} and summary functions. @item -@code{SHOW GRANTS} didn't always show all column grants. +Added @code{SET TRANSACTION ISOLATION LEVEL ...} @item -Added @code{--default-extra-file=#} to all @strong{MySQL} clients. +Added @code{SELECT ... FOR UPDATE}. @item -Columns referenced in @code{INSERT} statements now are initialized properly. +Fixed bug where the number of affected rows was not returned when @code{MySQL} +was compiled without transaction support. @item -@code{UPDATE} didn't always work when used with a range on a timestamp that -was part of the key that was used to find rows. +Fixed a bug in @code{UPDATE} where keys weren't always used to find the +rows to be updated. @item -Fixed a bug in @code{FULLTEXT} index when inserting a @code{NULL} column. +Fixed a bug in @code{CONCAT_WS()} where it returned wrong results. @item -Changed to use @code{mkstemp()} instead of @code{tempnam()}. Based -on a patch from John Jones. +Changed @code{CREATE ... INSERT} and @code{INSERT ... SELECT} to not +allow concurrent inserts as this could make the binary log hard to repeat. +(Concurrent inserts are enabled if you are not using the binary or update log). +@item +Changed some macros to be able to use fast mutex with glibc 2.2. @end itemize -@node News-3.23.25, News-3.23.24, News-3.23.26, News-3.23.x -@appendixsubsec Changes in release 3.23.25 +@node News-3.23.35, News-3.23.34a, News-3.23.36, News-3.23.x +@appendixsubsec Changes in release 3.23.35 @itemize @bullet @item -Fixed that @code{databasename} works as second argument to @code{mysqlhotcopy}. +Fixed newly introduced bug in @code{ORDER BY}. @item -@code{UMASK} and @code{UMASK_DIR} can now be specified in octal. +Fixed wrong define @code{CLIENT_TRANSACTIONS}. @item -Added @code{RIGHT JOIN}. This makes @code{RIGHT} a reserved word. +Fixed bug in @code{SHOW VARIABLES} when using @code{INNOBASE} tables. @item -Added @code{@@@@IDENTITY} as a synonym for @code{LAST_INSERT_ID()}. -(This is for Visual Basic compatibility.) +Setting and using user variables in @code{SELECT DISTINCT} didn't work. @item -Fixed a bug in @code{myisamchk} and @code{REPAIR} when using @code{FULLTEXT} -index. +Tuned @code{SHOW ANALYZE} for small tables. @item -@code{LOAD DATA INFILE} now works with FIFOs. -(Patch by Toni L. Harbaugh-Blackford.) +Fixed handling of arguments in the benchmark script @file{run-all-tests}. +@end itemize + +@node News-3.23.34a, News-3.23.34, News-3.23.35, News-3.23.x +@appendixsubsec Changes in release 3.23.34a +@itemize @bullet @item -@code{FLUSH LOGS} broke replication if you specified a log name with an -explicit extension as the value of the @code{log-bin} option. +Added extra files to the distribution to allow @code{INNOBASE} support +to be compiled. +@end itemize + +@node News-3.23.34, News-3.23.33, News-3.23.34a, News-3.23.x +@appendixsubsec Changes in release 3.23.34 +@itemize @bullet @item -Fixed a bug in @code{MyISAM} with packed multi-part keys. +Added the @code{INNOBASE} table handler and the @code{BDB} table handler +to the @strong{MySQL} source distribution. @item -Fixed crash when using @code{CHECK TABLE} on Windows. +Updated the documentation about @code{GEMINI} tables. @item -Fixed a bug where @code{FULLTEXT} index always used the koi8_ukr -character set. +Fixed a bug in @code{INSERT DELAYED} that caused threads to hang when +inserting @code{NULL} into an @code{AUTO_INCREMENT} column. @item -Fixed privilege checking for @code{CHECK TABLE}. +Fixed a bug in @code{CHECK TABLE} / @code{REPAIR TABLE} that could cause +a thread to hang. @item -The @code{MyISAM} repair/reindex code didn't use the @code{--tempdir} -option for its temporary files. +@code{REPLACE} will not replace a row that conflicts with an +@code{auto_increment} generated key. @item -Added @code{BACKUP TABLE/RESTORE TABLE}. +@code{mysqld} now only sets @code{CLIENT_TRANSACTIONS} in +@code{mysql->server_capabilities} if the server supports a +transaction-safe handler. @item -Fixed coredump on @code{CHANGE MASTER TO} when the slave did not have -the master to start with. +Fixed @code{LOAD DATA INFILE} to allow numeric values to be read into +@code{ENUM} and @code{SET} columns. @item -Fixed incorrect @code{Time} in the processlist for @code{Connect} of -the slave thread. +Improved error diagnostic for slave thread exit. @item -The slave now logs when it connects to the master. +Fixed bug in @code{ALTER TABLE ... ORDER BY}. @item -Fixed a core dump bug when doing @code{FLUSH MASTER} if you didn't specify -a filename argument to @code{--log-bin}. +Added option @code{max_user_connections} to @code{mysqld}. @item -Added missing @file{ha_berkeley.x} files to the @strong{MySQL} Windows +Limit query length for replication by @code{max_allowed_packet}, not the +arbitrary limit of 4 MB. @item -Fixed some mutex bugs in the log code that could cause thread blocks if new -log files couldn't be created. +Allow space around @code{=} in argument to @code{--set-variable}. @item -Added lock time and number of selected processed rows to slow query log. +Fixed problem in automatic repair that could leave some threads in state +@code{Waiting for table}. @item -Added @code{--memlock} option to @code{mysqld} to lock @code{mysqld} -in memory on systems with the @code{mlockall()} call (like in Solaris). +@code{SHOW CREATE TABLE} now dumps the @code{UNION()} for @code{MERGE} tables. @item -@code{HEAP} tables didn't use keys properly. (Bug from 3.23.23.) +@code{ALTER TABLE} now remembers the old @code{UNION()} definition. @item -Added better support for @code{MERGE} tables (keys, mapping, creation, -documentation...). @xref{MERGE}. +Fixed bug when replicating timestamps. @item -Fixed bug in @code{mysqldump} from 3.23 which caused some @code{CHAR} columns -not to be quoted. +Fixed bug in bidirectional replication. @item -Merged @code{analyze}, @code{check}, @code{optimize} and repair code. +Fixed bug in the @code{BDB} table handler that occurred when using an index +on multi-part key where a key part may be @code{NULL}. @item -@code{OPTIMIZE TABLE} is now mapped to @code{REPAIR} with statistics and -sorting of the index tree. This means that for the moment it only -works on @code{MyISAM} tables. +Fixed @code{MAX()} optimization on sub-key for @code{BDB} tables. @item -Added a pre-alloced block to root_malloc to get fewer mallocs. +Fixed problem where garbage results were returned when using @code{BDB} +tables and @code{BLOB} or @code{TEXT} fields when joining many tables. @item -Added a lot of new statistics variables. +Fixed a problem with @code{BDB} tables and @code{TEXT} columns. @item -Fixed @code{ORDER BY} bug with BDB tables. +Fixed bug when using a @code{BLOB} key where a const row wasn't found. @item -Removed warning that @code{mysqld} couldn't remove the @file{.pid} file -under Windows. +Fixed that @code{mysqlbinlog} writes the timestamp value for each query. +This ensures that one gets same values for date functions like @code{NOW()} +when using @code{mysqlbinlog} to pipe the queries to another server. @item -Changed @code{--log-isam} to log @strong{MyISAM} tables instead of isam -tables. +Allow one to use @code{--skip-gemini}, @code{--skip-bdb} and +@code{--skip-innobase} to @code{mysqld} even if these databases are not +compiled in @code{mysqld}. @item -Fixed @code{CHECK TABLE} to work on Windows. +One can now do @code{GROUP BY ... DESC}. @item -Added file mutexes to make @code{pwrite()} safe on Windows. +Fixed a deadlock in the @code{SET} code, when one ran @code{SET @@foo=bar}, +where @code{bar} is a column reference, an error was not properly generated. @end itemize -@node News-3.23.24, News-3.23.23, News-3.23.25, News-3.23.x -@appendixsubsec Changes in release 3.23.24 +@node News-3.23.33, News-3.23.32, News-3.23.34, News-3.23.x +@appendixsubsec Changes in release 3.23.33 @itemize @bullet @item -Added @code{mysqld} variable @code{created_tmp_disk_tables}. +Fixed that DNS lookups are not using the same mutex as the hostname +cache. This will enable known hosts to be quickly resolved even if a +DNS lookup takes a long time. @item -To make it possible to reliably dump and restore tables with -@code{TIMESTAMP(X)} columns, @strong{MySQL} now reports columns with @code{X} -other than 14 or 8 to be strings. +Added @code{--character-sets-dir} to @code{myisampack}. @item -Changed sort order for latin1 as it was before @strong{MySQL} Version 3.23.23. -Any table with @code{CHAR} columns that may have characters with ASCII -values greater than 128 that was created or modified with 3.23.22 must -be repaired! +Removed warnings when running @code{REPAIR TABLE ... EXTENDED}. @item -Fixed small memory leak introduced from 3.23.22 when creating a -temporary table. +Fixed a bug that caused a core dump when using @code{GROUP BY} on an alias, +where the alias was the same as an existing column name. @item -Fixed problem with BDB tables and reading on a unique (not primary) key. +Added @code{SEQUENCE()} as an example UDF function. @item -Restored the win1251 character set (it's now only marked deprecated). -@end itemize - -@node News-3.23.23, News-3.23.22, News-3.23.24, News-3.23.x -@appendixsubsec Changes in release 3.23.23 -@itemize @bullet +Changed @code{mysql_install_db} to use @code{BINARY} for @code{CHAR} +columns in the privilege tables. @item -Changed sort order for 'German'; All tables created with 'German' sortorder -must be repaired with @code{REPAIR TABLE} or @code{myisamchk} before use! +Changed @code{TRUNCATE table_name} to @code{TRUNCATE TABLE table_name} +to use the same syntax as Oracle. Until 4.0 we will also allow +@code{TRUNCATE table_name} to not crash old code. @item -Added option @code{--core-file} to @code{mysqld} to get a core file on -Linux if @code{mysqld} dies on the SIGSEGV signal. +Fixed 'no found rows' bug in @code{MyISAM} tables when a @code{BLOB} was +first part of a multi-part key. @item -@strong{MySQL} client @code{mysql} now starts with option -@code{--no-named-commands} (@code{-g}) by default. This option can be -disabled with @code{--enable-named-commands} (@code{-G}). This may cause -incompatibility problems in some cases, for example, in SQL scripts that -use named commands without a semicolon, etc. ! Long format commands -still work from the first line. +Fixed bug where @code{CASE} didn't work with @code{GROUP BY}. @item -Fixed a problem when using many pending @code{DROP TABLE} statements at -the same time. +Added option @code{--sort-recover} to @code{myisamchk}. @item -Optimizer didn't use keys properly when using @code{LEFT JOIN} on an -empty table. +@code{myisamchk -S} and @code{OPTIMIZE TABLE} now work on Windows. @item -Added shorter help text when invoking @code{mysqld} with incorrect options. +Fixed bug when using @code{DISTINCT} on results from functions that referred +to a group function, like: +@example +SELECT a, DISTINCT SEC_TO_TIME(sum(a)) from table_name GROUP BY a, b; +@end example @item -Fixed non-fatal @code{free()} bug in @code{mysqlimport}. +Fixed buffer overrun in @code{libmysqlclient} library. +Fixed bug in handling @code{STOP} event after @code{ROTATE} event in +replication. @item -Fixed bug in @code{MyISAM} index handling of -@code{DECIMAL}/@code{NUMERIC} keys. +Fixed another buffer overrun in @code{DROP DATABASE}. @item -Fixed a bug in concurrent insert in @code{MyISAM} tables; In some contexts, -usage of @code{MIN(key_part)} or @code{MAX(key_part)} returned an empty set. +Added @code{Table_locks_immediate} and @code{Table_locks_waited} status +variables. @item -Updated @code{mysqlhotcopy} to use the new @code{FLUSH TABLES table_list} -syntax. Only tables which are being backed up are flushed now. +Fixed bug in replication that broke slave server start with existing +@code{master.info}. This fixes a bug introduced in 3.23.32. @item -Changed behavior of @code{--enable-thread-safe-client} so -that both non-threaded (@code{-lmysqlclient}) and threaded -(@code{-lmysqlclient_r}) libraries are built. Users who linked -against a threaded @code{-lmysqlclient} will need to link against -@code{libmysqlclient_r} now. +Added @code{SET SQL_SLAVE_SKIP_COUNTER=n} command to recover from +replication glitches without a full database copy. @item -Added atomic @code{RENAME} command. +Added @code{max_binlog_size} variable; the binary log will be rotated +automatically when the size crosses the limit. @item -Don't count entries with @code{NULL} in @code{COUNT(DISTINCT ...)}. +Added @code{Last_error}, @code{Last_errno}, and @code{Slave_skip_counter} to +@code{SHOW SLAVE STATUS}. @item -Changed @code{ALTER TABLE}, @code{LOAD DATA INFILE} on empty tables and -@code{INSERT ... SELECT ...} on empty tables to create non-unique indexes -in a separate batch with sorting. This will make the above calls much -faster when you have many indexes. +Fixed bug in @code{MASTER_POS_WAIT()} function. @item -@code{ALTER TABLE} now logs the first used insert_id correctly. +Execute coredump handler on @code{SIGILL}, and @code{SIGBUS} in addition to +@code{SIGSEGV}. @item -Fixed crash when adding a default value to a @code{BLOB} column. +On x86 Linux, print the current query and thread (connection) id, if +available, in the coredump handler. @item -Fixed a bug with @code{DATE_ADD/DATE_SUB} where it returned a datetime instead -of a date. +Fixed several timing bugs in the test suite. @item -Fixed a problem with the thread cache which made some threads show up as -@code{***DEAD***} in @code{SHOW PROCESSLIST}. +Extended @code{mysqltest} to take care of the timing issues in the test +suite. @item -Fixed a lock in our thr_rwlock code, which could make selects that run -at the same time as concurrent inserts crash. This only affects systems -that don't have the @code{pthread_rwlock_rdlock} code. +@code{ALTER TABLE} can now be used to change the definition for a +@code{MERGE} table. @item -When deleting rows with a non-unique key in a HEAP table, all rows weren't -always deleted. +Fixed creation of @code{MERGE} tables on Windows. @item -Fixed bug in range optimizer for HEAP tables for searches on a part index. +Portability fixes for OpenBSD and OS2. @item -Fixed that @code{SELECT} on part keys works with BDB tables. +Added @code{--temp-pool} option to @code{mysqld}. Using this option +will cause most temporary files created to use a small set of names, +rather than a unique name for each new file. This is to work around a +problem in the Linux kernel dealing with creating a bunch of new files +with different names. With the old behavior, Linux seems to "leak" +memory, as it's being allocated to the directory entry cache instead of +the disk cache. +@end itemize + +@node News-3.23.32, News-3.23.31, News-3.23.33, News-3.23.x +@appendixsubsec Changes in release 3.23.32 +@itemize @bullet @item -Fixed @code{INSERT INTO bdb_table ... SELECT} to work with BDB tables. +Changed code to get around compiler bug in Compaq C++ on OSF1, that broke +@code{BACKUP}, @code{RESTORE}, @code{CHECK}, @code{REPAIR}, and +@code{ANALYZE TABLE}. @item -@code{CHECK TABLE} now updates key statistics for the table. +Added option @code{FULL} to @code{SHOW COLUMNS}. Now we show the +privilege list for the columns only if this option is given. @item -@code{ANALYZE TABLE} will now only update tables that have been changed -since thee last @code{ANALYZE}. Note that this is a new feature and tables -will not be marked to be analyzed until they are updated in any way with -3.23.23 or newer. For older tables, you have to do @code{CHECK TABLE} -to update the key distribution. +Fixed bug in @code{SHOW LOGS} when there weren't any BDB logs. @item -Fixed some minor privilege problems with @code{CHECK}, @code{ANALYZE}, -@code{REPAIR} and @code{SHOW CREATE} commands. +Fixed a timing problem in replication that could delay sending an update +to the client until a new update was done. @item -Added @code{CHANGE MASTER TO} command. +Don't convert field names when using @code{mysql_list_fields()}. This is +to keep this code compatible with @code{SHOW FIELDS}. @item -Added @code{FAST}, @code{QUICK} @code{EXTENDED} check types to -@code{CHECK TABLES}. +@code{MERGE} tables didn't work on Windows. @item -Changed @code{myisamchk} so that @code{--fast} and -@code{--check-changed-tables} are also honored with @code{--sort-index} and -@code{--analyze}. +Fixed problem with @code{SET PASSWORD=...} on Windows. @item -Fixed fatal bug in @code{LOAD TABLE FROM MASTER} that did not lock the -table during index re-build. +Added missing @file{my_config.h} to RPM distribution. @item -@code{LOAD DATA INFILE} broke replication if the database was excluded -from replication. +@code{TRIM("foo" from "foo")} didn't return an empty string. @item -More variables in @code{SHOW SLAVE STATUS} and @code{SHOW MASTER STATUS}. +Added @code{--with-version-suffix} to @code{configure}. @item -@code{SLAVE STOP} now will not return until the slave thread actually exits. +Fixed coredump when client aborted connection without @code{mysql_close()}. @item -Full text search via the @code{MATCH} function and @code{FULLTEXT} index type. -(For MyISAM files). This makes @code{FULLTEXT} a reserved word. +Fixed a bug in @code{RESTORE TABLE} when trying to restore from a non-existent +directory. +@item +Fixed a bug which caused a core dump on the slave when replicating +@code{SET PASSWORD}. +@item +Added @code{MASTER_POS_WAIT()}. @end itemize -@node News-3.23.22, News-3.23.21, News-3.23.23, News-3.23.x -@appendixsubsec Changes in release 3.23.22 +@node News-3.23.31, News-3.23.30, News-3.23.32, News-3.23.x +@appendixsubsec Changes in release 3.23.31 @itemize @bullet @item -Fixed that @code{lex_hash.h} is created properly for each @strong{MySQL} -distribution. +The test suite now tests all reachable BDB interface code. During +testing we found and fixed many errors in the interface code. @item -Fixed that @code{MASTER} and @code{COLLECTION} are not reserved words. +Using @code{HAVING} on an empty table could produce one result row when +it shouldn't. @item -The log generated by @code{--slow-query-log} didn't contain the whole queries. +Fixed that the @strong{MySQL} RPM is not dependent on Perl5 anymore. @item -Fixed that open transactions in BDB tables are rolled back if the -connection is closed unexpectedly. +Fixed some problems with @code{HEAP} tables on Windows. @item -Added workaround for a bug in @code{gcc} 2.96 (intel) and @code{gcc} 2.9 -(Ia64) in @code{gen_lex_hash.c}. +@code{SHOW TABLE STATUS} didn't show correct average row length for tables +larger than 4G. @item -Fixed memory leak in the client library when using @code{host=} in the -@code{my.cnf} file. +@code{CHECK TABLE ... EXTENDED} didn't check row links for fixed size tables. @item -Optimized functions that manipulate the hours/minutes/seconds. +Added option @code{MEDIUM} to @code{CHECK TABLE}. @item -Fixed bug when comparing the result of @code{DATE_ADD()}/@code{DATE_SUB()} -against a number. +Fixed problem when using @code{DECIMAL()} keys on negative numbers. @item -Changed the meaning of @code{-F, --fast} for @code{myisamchk}. Added option -@code{-C, --check-only-changed} to @code{myisamchk}. +@code{HOUR()} (and some other @code{TIME} functions) on a @code{CHAR} column +always returned @code{NULL}. @item -Added @code{ANALYZE table_name} to update key statistics for tables. +Fixed security bug in something (please upgrade if you are using a earlier +@strong{MySQL} 3.23 version). @item -Changed binary items @code{0x...} to be regarded as integers by default. +Fixed buffer overflow bug when writing a certain error message. @item -Fix for SCO and @code{SHOW PROCESSLIST}. +Added usage of @code{setrlimit()} on Linux to get +@code{-O --open-files-limit=#} to work on Linux. @item -Added @code{auto-rehash} on reconnect for the @code{mysql} client. +Added new @code{mysqld} variable: @code{bdb_version}. @item -Fixed a newly introduced bug in @code{MyISAM}, where the index file couldn't -get bigger than 64M. +Fixed bug when using expression of type: +@example +SELECT ... FROM t1 LEFT JOIN t2 ON (t1.a=t2.a) WHERE t1.a=t2.a +@end example +In this case the test in the @code{WHERE} clause was wrongly optimized away. @item -Added @code{SHOW MASTER STATUS} and @code{SHOW SLAVE STATUS}. +Fixed bug in @code{MyISAM} when deleting keys with possible @code{NULL} +values, but the first key-column was not a prefix-compressed text column. +@item +Fixed @code{mysql.server} to read the @code{mysql.server} option section +instead of @code{mysql_server}. +@item +Fixed @code{safe_mysqld} and @code{mysql.server} to also read the +@code{server} option section. +@item +Added @code{Threads_created} status variable to @code{mysqld}. @end itemize -@node News-3.23.21, News-3.23.20, News-3.23.22, News-3.23.x -@appendixsubsec Changes in release 3.23.21 +@node News-3.23.30, News-3.23.29, News-3.23.31, News-3.23.x +@appendixsubsec Changes in release 3.23.30 @itemize @bullet @item -Added @code{mysql_character_set_name(MYSQL *mysql)} function to the -@strong{MySQL} C API. +Added @code{SHOW OPEN TABLES} command. +@item +Fixed that @code{myisamdump} works against old @code{mysqld} servers. +@item +Fixed @code{myisamchk -k#} so that it works again. +@item +Fixed a problem with replication when the binary log file went over 2G +on 32-bit systems. +@item +@code{LOCK TABLES} will now automatically start a new transaction. +@item +Changed @code{BDB} tables to not use internal subtransactions and reuse +open files to get more speed. @item -Made the update log @code{ASCII 0} safe. +Added option @code{--mysqld=#} to @code{safe_mysqld}. @item -Added the @code{mysql_config} script. +Allow hex constants in the @code{--fields-*-by} and +@code{--lines-terminated-by} options to @code{mysqldump} and +@code{mysqlimport}. By Paul DuBois. @item -Fixed problem when using @code{<} or @code{>} with a char column that was only -partly indexed. +Added option @code{--safe-show-database} to @code{mysqld}. @item -One would get a core dump if the log file was not readable by the MySQL user. +Added @code{have_bdb}, @code{have_gemini}, @code{have_innobase}, +@code{have_raid} and @code{have_ssl} to @code{SHOW VARIABLES} to make it +easy to test for supported extensions. @item -Changed @code{mysqladmin} to use the @code{CREATE DATABASE}/@code{DROP -DATABASE} commands instead of the old deprecated API calls. +Added option @code{--open-files-limit} to @code{mysqld}. @item -Fixed @code{chown} warning in @code{safe_mysqld}. +Changed option @code{--open-files} to @code{--open-files-limit} in +@code{safe_mysqld}. @item -Fixed a bug in @code{ORDER BY} that was introduced in 3.23.19. +Fixed a bug where some rows were not found with @code{HEAP} tables +that had many keys. @item -Only optimize the @code{DELETE FROM tbl_name} to do a drop+create of -the table if we are in @code{AUTOCOMMIT} mode. (Needed for BDB tables). +Fixed that @code{--bdb-no-sync} works. @item -Added extra checks to avoid index corruption when the @code{ISAM}/@code{MyISAM} -index files gets full during an @code{INSERT}/@code{UPDATE}. +Changed @code{--bdb-recover} to @code{--bdb-no-recover} as recover should +be on by default. @item -@code{myisamchk} didn't correctly update row checksum when used with -@code{-ro} (This only gave an warning in subsequent runs). +Changed the default number of BDB locks to 10000. @item -Fixed bug in @code{REPAIR TABLE} so that it works with tables without indexes. +Fixed a bug from 3.23.29 when allocating the shared structure needed +for BDB tables. @item -Fixed buffer overrun in @code{DROP DATABASE} +Changed @file{mysqld_multi.sh} to use configure variables. Patch by +Christopher McCrory. @item -@code{LOAD TABLE FROM MASTER} is sufficiently bug-free to announce it as -a feature. +Added fixing of include files for Solaris 2.8. @item -@code{MATCH} and @code{AGAINST} are now reserved words. -@end itemize - -@node News-3.23.20, News-3.23.19, News-3.23.21, News-3.23.x -@appendixsubsec Changes in release 3.23.20 -@itemize @bullet +Fixed bug with @code{--skip-networking} on Debian Linux. @item -Fixed bug in 3.23.19; @code{DELETE FROM tbl_name} removed the .frm file. +Fixed problem that some temporary files where reported as having the +name @code{UNOPENED} in error messages. @item -Added @code{SHOW CREATE TABLE}. +Fixed bug when running two simultaneous @code{SHOW LOGS} queries. @end itemize -@node News-3.23.19, News-3.23.18, News-3.23.20, News-3.23.x -@appendixsubsec Changes in release 3.23.19 +@node News-3.23.29, News-3.23.28, News-3.23.30, News-3.23.x +@appendixsubsec Changes in release 3.23.29 @itemize @bullet @item -Changed copyright for all files to GPL for the server code and utilities and -LGPL for the client libraries. -@item -Fixed bug where all rows matching weren't updated on a @code{MyISAM} table -when doing update based on key on a table with many keys and some key changed -values. -@item -The Linux @strong{MySQL} RPM's and binaries are now statically linked with -a linuxthread version that has faster mutex handling when used with -@strong{MySQL}. -@item -@code{ORDER BY} can now uses @code{REF} keys to find subset the rows -that needs to be sorted. -@item -Changed name of @code{print_defaults} to @code{my_print_defaults} to avoid -name confusion. +Configure updates for Tru64, large file support, and better TCP wrapper +support. By Albert Chin-A-Young. @item -Fixed @code{NULLIF()} to work according to ANSI SQL99. +Fixed bug in @code{<=>} operator. @item -Added @code{net_read_timeout} and @code{net_write_timeout} as startup -parameters to @code{mysqld}. +Fixed bug in @code{REPLACE} with BDB tables. @item -Fixed bug that destroyed index when doing @code{myisamchk --sort-records} -on a table with prefix compressed index. +@code{LPAD()} and @code{RPAD()} will shorten the result string if it's longer +than the length argument. @item -Added pack_isam and myisampack to the standard @strong{MySQL} distribution. +Added @code{SHOW LOGS} command. @item -Added the syntax @code{BEGIN WORK} (the same as @code{BEGIN}). +Remove not used BDB logs on shutdown. @item -Fixed core dump bug when using @code{ORDER BY} on a @code{CONV()} expression. -@item Added @code{LOAD TABLE FROM MASTER} -@item Added @code{FLUSH MASTER} and @code{FLUSH SLAVE} -@item Fixed big/little endian problem in the replication -@end itemize - -@node News-3.23.18, News-3.23.17, News-3.23.19, News-3.23.x -@appendixsubsec Changes in release 3.23.18 -@itemize @bullet +When creating a table, put @code{PRIMARY} keys first, followed by +@code{UNIQUE} keys. @item -Fixed a problem from 3.23.17 when choosing character set on the client side. +Fixed a bug in @code{UPDATE} involving multi-part keys where one +specified all key parts both in the update and the @code{WHERE} part. In +this case @strong{MySQL} could try to update a record that didn't match +the whole @code{WHERE} part. @item -Added @code{FLUSH TABLES WITH READ LOCK} to make a global lock suitable to -make a copy of @strong{MySQL} data files. +Changed drop table to first drop the tables and then the @file{.frm} file. @item -@code{CREATE TABLE ... SELECT ... PROCEDURE} now works. +Fixed a bug in the hostname cache which caused @code{mysqld} to report the +hostname as @code{''} in some error messages. @item -Internal temporary tables will now uses compressed index when using -@code{GROUP BY} on @code{VARCHAR/CHAR} columns. +Fixed a bug with @code{HEAP} type tables; the variable +@code{max_heap_table_size} wasn't used. Now either @code{MAX_ROWS} or +@code{max_heap_table_size} can be used to limit the size of a @code{HEAP} +type table. @item -Fixed a problem when locking the same table with both a @code{READ} and a -@code{WRITE} lock. +Changed the default server-id to 1 for masters and 2 for slaves +to make it easier to use the binary log. @item -Fixed problem with myisamchk and @code{RAID} tables. -@end itemize - -@node News-3.23.17, News-3.23.16, News-3.23.18, News-3.23.x -@appendixsubsec Changes in release 3.23.17 -@itemize @bullet +Renamed variable @code{bdb_lock_max} to @code{bdb_max_lock}. @item -Fixed a bug in @code{find_in_set()} when the first argument was @code{NULL}. +Added support for @code{auto_increment} on sub-fields for BDB tables. @item -Added table locks to Berkeley DB. +Added @code{ANALYZE} of BDB tables. @item -Fixed a bug with @code{LEFT JOIN} and @code{ORDER BY} where the first -table had only one matching row. +In BDB tables, we now store the number of rows; This helps to optimize queries +when we need an approximation of the number of rows. @item -Added 4 sample @code{my.cnf} example files in the @file{support-files} -directory. +If we get an error in a multi-row statement, we now only rollback the +last statement, not the entire transaction. @item -Fixed @code{duplicated key} problem when doing big @code{GROUP BY}'s. -(This bug was probably introduced in 3.23.15). +If you do a @code{ROLLBACK} when you have updated a non-transactional table +you will get an error as a warning. @item -Changed syntax for @code{INNER JOIN} to match ANSI SQL. +Added option @code{--bdb-shared-data} to @code{mysqld}. @item -Added @code{NATURAL JOIN} syntax. +Added status variable @code{Slave_open_temp_tables}. @item -A lot of fixes in the @code{BDB} interface. +Added variables @code{binlog_cache_size} and @code{max_binlog_cache_size} to +@code{mysqld}. @item -Added handling of @code{--no-defaults} and @code{--defaults-file} to -@code{safe_mysqld.sh} and @code{mysql_install_db.sh}. +@code{DROP TABLE}, @code{RENAME TABLE}, @code{CREATE INDEX} and +@code{DROP INDEX} are now transaction endpoints. @item -Fixed bug in reading compressed tables with many threads. +If you do a @code{DROP DATABASE} on a symbolic linked database, both +the link and the original database is deleted. @item -Fixed that @code{USE INDEX} works with @code{PRIMARY} keys. +Fixed that @code{DROP DATABASE} works on OS/2. @item -Added @code{BEGIN} statement to start a transaction in @code{AUTOCOMMIT} mode. +Fixed bug when doing a @code{SELECT DISTINCT ... table1 LEFT JOIN +table2 ...} when table2 was empty. @item -Added symbolic links support for Windows. +Added @code{--abort-slave-event-count} and +@code{--disconnect-slave-event-count} options to @code{mysqld} for +debugging and testing of replication. @item -Changed protocol to let client know if the server is in AUTOCOMMIT mode -and if there is a pending transaction. If there is a pending transaction -the client library will give an error before reconnecting to the server to -let the client know that the server did a rollback. -The protocol is still backward compatible with old clients +Fixed replication of temporary tables. Handles everything except +slave server restart. @item -@code{KILL} now works on a thread that is locked on a 'write' to a dead client. +@code{SHOW KEYS} now shows whether or not key is @code{FULLTEXT}. @item -Fixed memory leak in the replication slave thread. +New script @file{mysqld_multi}. @xref{mysqld_multi, , @code{mysqld_multi}}. @item -Added new option @code{log-slave-updates} to allow daisy-chaining the slaves. +Added new script, @file{mysql-multi.server.sh}. Thanks to +Tim Bunce @email{Tim.Bunce@@ig.co.uk} for modifying @file{mysql.server} to +easily handle hosts running many @code{mysqld} processes. @item -Fixed compile error on FreeBSD and other systems where @code{pthread_t} -is not the same as @code{int}. +@file{safe_mysqld}, @file{mysql.server}, and @file{mysql_install_db} have +been modified to use @code{mysql_print_defaults} instead of various hacks +to read the @file{my.cnf} files. In addition, the handling of various +paths has been made more consistent with how @code{mysqld} handles them +by default. @item -Fixed master shutdown aborting the slave thread. +Automatically remove Berkeley DB transaction logs that no longer are in +use. @item -Fixed a race condition in @code{INSERT DELAYED} code when doing -@code{ALTER TABLE}. +Fixed bug with several @code{FULLTEXT} indexes in one table. @item -Added deadlock detection sanity checks to @code{INSERT DELAYED} -@end itemize - -@node News-3.23.16, News-3.23.15, News-3.23.17, News-3.23.x -@appendixsubsec Changes in release 3.23.16 -@itemize @bullet +Added a warning if number of rows changes on @code{REPAIR}/@code{OPTIMIZE}. @item -Added option @code{TYPE=QUICK} to @code{CHECK} and @code{REPAIR}. +Applied patches for OS/2 by @code{Yuri Dario}. @item -Fixed bug in @code{REPAIR TABLE} when the table was in use by other threads. +@code{FLUSH TABLES table_name} didn't always flush the index tree +to disk properly. @item -Added a thread cache to make it possible to debug @strong{MySQL} with -@code{gdb} when one does a lot of reconnects. This will also improve -systems where you can't use persistent connections. +@code{--bootstrap} is now run in a separate thread. This fixes a problem +that caused @code{mysql_install_db} to core dump on some Linux machines. @item -Lots of fixes in the Berkeley DB interface. +Changed @code{mi_create()} to use less stack space. @item -@code{UPDATE IGNORE} will not abort if an update results in a -@code{DUPLICATE_KEY} error. +Fixed bug with optimizer trying to over-optimize @code{MATCH} when used +with @code{UNIQUE} key. @item -Put @code{CREATE TEMPORARY TABLE} commands in the update log. +Changed @code{crash-me} and the @strong{MySQL} benchmarks to also work +with FrontBase. @item -Fixed bug in handling of masked IP numbers in the privilege tables. +Allow @code{RESTRICT} and @code{CASCADE} after @code{DROP TABLE} to make +porting easier. @item -Fixed bug with @code{delayed_key_writes} tables and @code{CHECK TABLE}. +Reset status variable which could cause problem if one used @code{--slow-log}. @item -Added @code{replicate-do-db} and @code{replicate-ignore-db} options to -restrict which databases get replicated +Added variable @code{connect_timeout} to @code{mysql} and @code{mysqladmin}. @item -Added @code{SQL_LOG_BIN} option +Added @code{connect_timeout} as an alias for @code{timeout} for option files +read by @code{mysql_options()}. @end itemize -@node News-3.23.15, News-3.23.14, News-3.23.16, News-3.23.x -@appendixsubsec Changes in release 3.23.15 +@node News-3.23.28, News-3.23.27, News-3.23.29, News-3.23.x +@appendixsubsec Changes in release 3.23.28 @itemize @bullet @item -To start @code{mysqld} as @code{root}, you must now use the @code{--user=root} option. +Added new options @code{--pager[=...]}, @code{--no-pager}, +@code{--tee=...} and @code{--no-tee} to the @code{mysql} client. The +new corresponding interactive commands are @code{pager}, @code{nopager}, +@code{tee} and @code{notee}. @xref{mysql, , @code{mysql}}, @code{mysql --help} +and the interactive help for more information. @item -Added interface to Berkeley DB. (This is not yet functional; Play with it at -your own risk!) +Fixed crash when automatic repair of @code{MyISAM} table failed. @item -Replication between master and slaves. +Fixed a major performance bug in the table locking code when one +constantly had a LOT of @code{SELECT}, @code{UPDATE} and @code{INSERT} +statements running. The symptom was that the @code{UPDATE} and +@code{INSERT} queries were locked for a long time while new @code{SELECT} +statements were executed before the updates. @item -Fixed bug that other threads could steal a lock when a thread had -a lock on a table and did a @code{FLUSH TABLES} command. +When reading @code{options_files} with @code{mysql_options()} the +@code{return-found-rows} option was ignored. @item -Added the @code{slow_launch_time} variable and the @code{Slow_launch_threads} -status variable to @code{mysqld}. These can be examined with -@code{mysqladmin variables} and @code{mysqladmin extended-status}. +One can now specify @code{interactive-timeout} in the option file that +is read by @code{mysql_options()}. This makes it possible to force +programs that run for a long time (like @code{mysqlhotcopy}) to use +@code{interactive_timeout} instead of @code{wait_timeout}. @item -Added functions @code{INET_NTOA()} and @code{INET_ATON()}. +Added to the slow query log the time and the user name for each logged +query. If you are using @code{--log-long-format} then also queries that +do not use an index are logged, even if the query takes less than +@code{long_query_time} seconds. @item -The default type of @code{IF()} now depends on the second and -third arguments and not only on the second argument. +Fixed a problem in @code{LEFT JOIN} which caused all columns in a reference +table to be @code{NULL}. @item -Fixed case when @code{myisamchk} could go into a loop when trying to -repair a crashed table. +Fixed a problem when using @code{NATURAL JOIN} without keys. @item -Don't write @code{INSERT DELAYED} to update log if @code{SQL_LOG_UPDATE=0}. +Fixed a bug when using a multi-part keys where the first part was of type +@code{TEXT} or @code{BLOB}. @item -Fixed problem with @code{REPLACE} on @code{HEAP} tables. +@code{DROP} of temporary tables wasn't stored in the update/binary log. @item -Added possible character sets and time zone to @code{SHOW VARIABLES}. +Fixed a bug where @code{SELECT DISTINCT * ... LIMIT #} only returned one row. @item -Fixed bug in locking code that could result it locking problems with -concurrent inserts under high load. +Fixed a bug in the assembler code in @code{strstr()} for sparc and cleaned up +the @file{global.h} header file to avoid a problem with bad aliasing with +the compiler submitted with RedHat 7.0. (Reported by Trond Eivind Glomsrød) @item -Fixed a problem with @code{DELETE} of many rows on a table with -compressed keys where @strong{MySQL} scanned the index to find the rows. +The option @code{--skip-networking} now works properly on NT. @item -Fixed problem with @code{CHECK} on table with deleted keyblocks. +Fixed a long outstanding bug in the @code{ISAM} tables when a row with a length +of more than 65K was shortened by a single byte. @item -Fixed a bug in reconnect (at the client side) where it didn't free memory -properly in some contexts. +Fixed a bug in @code{MyISAM} when running multiple updating processes on +the same table. @item -Fixed problems in update log when using @code{LAST_INSERT_ID()} to update -a table with an auto_increment key. +Allow one to use @code{FLUSH TABLE tablename}. @item -Added function @code{NULLIF()}. +Added @code{--replicate-ignore-table}, @code{--replicate-do-table}, + @code{--replicate-wild-ignore-table}, @code{--replicate-wild-do-table}. @item -Fixed bug when using @code{LOAD DATA INFILE} on a table with -@code{BLOB/TEXT} columns. +Changed all log files to use our own @code{IO_CACHE} mechanism instead of +@code{FILE} to avoid OS problems when there are many files open. @item -Optimised MyISAM to be faster when inserting keys in sorted order. +Added options @code{--open-files} and @code{--timezone} to @code{safe_mysqld}. @item -@code{EXPLAIN SELECT ...} now also prints out whether @strong{MySQL} needs to -create a temporary table or use file sorting when resolving the @code{SELECT}. +Fixed a fatal bug in @code{CREATE TEMPORARY TABLE ... SELECT ...}. @item -Added optimization to skip @code{ORDER BY} parts where the part is a -constant expression in the @code{WHERE} part. Indexes can now be used -even if the @code{ORDER BY} doesn't match the index exactly, as long as -all the not used index parts and all the extra @code{ORDER BY} -columns are constants in the @code{WHERE} clause. @xref{MySQL indexes}. +Fixed a problem with @code{CREATE TABLE ... SELECT NULL}. @item -@code{UPDATE} and @code{DELETE} on a whole unique key in the @code{WHERE} part, -is now faster than before. +Added variables @code{large_file_support},@code{net_read_timeout}, +@code{net_write_timeout} and @code{query_buffer_size} to @code{SHOW VARIABLES}. @item -Changed @code{RAID_CHUNKSIZE} to be in 1024 byte increments. +Added status variables @code{created_tmp_files} and @code{sort_merge_passes} +to @code{SHOW STATUS}. @item -Fixed coredump in LOAD_FILE(NULL). -@end itemize - -@node News-3.23.14, News-3.23.13, News-3.23.15, News-3.23.x -@appendixsubsec Changes in release 3.23.14 -@itemize @bullet +Fixed a bug where we didn't allow an index name after the +@code{FOREIGN KEY} definition. @item -Fixed a bug in @code{CONCAT()} where one of the arguments was a function -that returned a modified argument. +Added @code{TRUNCATE table_name} as a synonym for +@code{DELETE FROM table_name}. @item -Fixed a critical bug in @code{myisamchk}, where it updated the header in -the index file when one only checked the table. This confused the -@code{mysqld} daemon if it updated the same table at the same time. Now -the status in the index file is only updated if one uses -@code{--update-state}. With older @code{myisamchk} versions you should -use @code{--read-only} when only checking tables, if there is the -slightest chance that the @code{mysqld} server is working on the table at the -same time! +Fixed a bug in a BDB key compare function when comparing part keys. @item -Fixed that @code{DROP TABLE} is logged in the update log. +Added variable @code{bdb_lock_max} to @code{mysqld}. @item -Fixed problem when searching on @code{DECIMAL()} key field -where the column data contained leading zeros. +Added more tests to the benchmark suite. @item -Fix bug in @code{myisamchk} when the auto_increment isn't the first key. +Fixed an overflow bug in the client code when using overly long database names. @item -Allow @code{DATETIME} in ISO8601 format: 2000-03-12T12:00:00 +@code{mysql_connect()} now aborts on Linux if the server doesn't answer in +@code{timeout} seconds. @item -Dynamic character sets. A @code{mysqld} binary can now handle many different -character sets (you can choose which when starting @code{mysqld}). +@code{SLAVE START} did not work if you started with +@code{--skip-slave-start} and had not explicitly run @code{CHANGE MASTER TO}. @item -Added command @code{REPAIR TABLE}. +Fixed the output of @code{SHOW MASTER STATUS} to be consistent with +@code{SHOW SLAVE STATUS}. (It now has no directory in the log name.) @item -Added C API function @code{mysql_thread_safe()}. +Added @code{PURGE MASTER LOGS TO}. @item -Added the @code{UMASK_DIR} environment variable. +Added @code{SHOW MASTER LOGS}. @item -Added function @code{CONNECTION_ID()}. +Added @code{--safemalloc-mem-limit} option to @code{mysqld} to simulate memory +shortage when compiled @code{--with-debug=full}. @item -When using @code{=} on @code{BLOB} or @code{VARCHAR BINARY} keys, where -only a part of the column was indexed, the whole column of the result -row wasn't compared. +Fixed several coredumps in out-of-memory conditions. @item -Fix for sjis character set and @code{ORDER BY}. +@code{SHOW SLAVE STATUS} was using an uninitialized mutex if the slave had +not been started yet. @item -When running in ANSI mode, don't allow columns to be used that aren't in -the @code{GROUP BY} part. +Fixed bug in @code{ELT()} and @code{MAKE_SET()} when the query used +a temporary table. +@item +@code{CHANGE MASTER TO} without specifying @code{MASTER_LOG_POS} would +set it to 0 instead of 4 and hit the magic number in the master binlog. +@item +@code{ALTER TABLE ... ORDER BY ...} syntax added. This will create the +new table with the rows in a specific order. @end itemize -@node News-3.23.13, News-3.23.12, News-3.23.14, News-3.23.x -@appendixsubsec Changes in release 3.23.13 +@node News-3.23.27, News-3.23.26, News-3.23.28, News-3.23.x +@appendixsubsec Changes in release 3.23.27 @itemize @bullet @item -Fixed problem when doing locks on the same table more than 2 times in -the same @code{LOCK TABLE} command; This fixed the problem one got when running -the test-ATIS test with @code{--fast} or @code{--check-only-changed}. -@item -Added option @code{SQL_BUFFER_RESULT} to @code{SELECT}. +Fixed a bug where the automatic repair of MyISAM tables sometimes failed +when the data file was corrupt. @item -Removed end space from double/float numbers in results from temporary -tables. -Added @code{CHECK TABLE} command. +Fixed a bug in @code{SHOW CREATE} when using @code{AUTO_INCREMENT} columns. @item -Added changes for MyISAM in 3.23.12 that didn't get into the source -distribution because of CVS problems. +Changed BDB tables to use new compare function in Berkeley DB 3.2.3. @item -Fixed bug so that @code{mysqladmin shutdown} will wait for the local server -to close down. +You can now use Unix sockets with @code{mit-pthreads}. @item -Fixed a possible endless loop when calculating timestamp. +Added the latin5 (turkish) character set. @item -Added @code{print_defaults} to the @file{.rpm} files. Removed @code{mysqlbug} -from the client @file{.rpm} file. +Small portability fixes. @end itemize -@node News-3.23.12, News-3.23.11, News-3.23.13, News-3.23.x -@appendixsubsec Changes in release 3.23.12 +@node News-3.23.26, News-3.23.25, News-3.23.27, News-3.23.x +@appendixsubsec Changes in release 3.23.26 @itemize @bullet @item -Fixed bug in @code{MyISAM} involving @code{REPLACE ... SELECT ...} which could -give a corrupted table. -@item -Fixed bug in @code{myisamchk} where it wrongly reset the auto_increment value. -@item -LOTS of patches for Linux Alpha. @strong{MySQL} now appears to be relatively -stable on Alpha. -@item -Changed @code{DISTINCT} on @code{HEAP} temporary tables to use hashed -keys to quickly find duplicated rows. This mostly concerns queries of -type @code{SELECT DISTINCT ... GROUP BY ...}. This fixes a problem where -not all duplicates were removed in queries of the above type. In -addition, the new code is MUCH faster. -@item -Added patches to make @strong{MySQL} compile on Mac OS X. -@item -Added option @code{IF NOT EXISTS} to @code{CREATE DATABASE}. -@item -Added options @code{--all-databases} and @code{--databases} to @code{mysqldump} -to allow dumping of many databases at the same time. -@item -Fixed bug in compressed @code{DECIMAL()} index in @code{MyISAM} tables. -@item -Fixed bug when storing 0 into a timestamp. -@item -When doing @code{mysqladmin shutdown} on a local connection, @code{mysqladmin} now -waits until the pidfile is gone before terminating. -@item -Fixed core dump with some @code{COUNT(DISTINCT ...)} queries. +Fixed @code{<>} to work properly with @code{NULL}. @item -Fixed that @code{myisamchk} works properly with RAID:ed tables. +Fixed a problem with @code{SUBSTRING_INDEX()} and @code{REPLACE()}. +(Patch by Alexander Igonitchev) @item -Fixed problem with @code{LEFT JOIN} and @code{key_field IS NULL}. +Fix @code{CREATE TEMPORARY TABLE IF NOT EXISTS} not to give an error +if the table exists. @item -Fixed bug in @code{net_clear()} which could give the error @code{Aborted -connection} in the @strong{MySQL} clients. +If you don't create a @code{PRIMARY KEY} in a BDB table, a hidden +@code{PRIMARY KEY} will be created. @item -Added options @code{USE INDEX (key_list)} and @code{IGNORE INDEX (key_list)} as -join parameters in @code{SELECT}. +Added read-only-key optimization to BDB tables. @item -@code{DELETE} and @code{RENAME} should now work on @code{RAID} tables. -@end itemize - -@node News-3.23.11, News-3.23.10, News-3.23.12, News-3.23.x -@appendixsubsec Changes in release 3.23.11 -@itemize @bullet +@code{LEFT JOIN} in some cases preferred a full table scan when there was +no @code{WHERE} clause. @item -Allow the @code{ALTER TABLE tbl_name ADD (field_list)} syntax. +When using @code{--log-slow-queries}, don't count the time waiting for a lock. @item -Fixed problem with optimizer that could sometimes use wrong keys. +Fixed bug in lock code on Windows which could cause the key cache +to report that the key file was crashed even if it was okay. @item -Fixed that @code{GRANT/REVOKE ALL PRIVILEGES} doesn't affect -@code{GRANT OPTION}. +Automatic repair of @code{MyISAM} tables if you start @code{mysqld} with +@code{--myisam-recover}. @item -Removed extra @code{)} from the output of @code{SHOW GRANTS} +Removed the @code{TYPE=} keyword from @code{CHECK} and +@code{REPAIR}. Allow @code{CHECK} options to be combined. (You can still +use @code{TYPE=} but this usage is deprecated.) @item -Fixed problem when storing numbers in timestamps. +Fixed mutex bug in the binary replication log - long update queries could +be read only in part by the slave if it did it at the wrong time, which +was not fatal, but resulted in a performance-degrading reconnect and +a scary message in the error log. @item -Fix problem with timezones that have half hour offsets. +Changed the format of the binary log - added magic number, server +version, binlog version. Added server id and query error code for each +query event. @item -Allow the syntax @code{UNIQUE INDEX} in @code{CREATE} statements. +Replication thread from the slave now will kill all the stale threads +from the same server. @item -@code{mysqlhotcopy} - fast on-line hot-backup utility for local -@strong{MySQL} databases. By Tim Bunce. +Long replication user names were not being handled properly. @item -New more secure @code{mysqlaccess}. Thanks to Steve Harvey for this. +Added @code{--replicate-rewrite-db} option to @code{mysqld}. @item -Added options @code{--i-am-a-dummy} and @code{--safe-updates} to @code{mysql}. +Added @code{--skip-slave-start} option to @code{mysqld}. @item -Added variables @code{select_limit} and @code{max_join_size} to @code{mysql}. +Updates that generated an error code (such as @code{INSERT INTO foo(some_key) +values (1),(1)}) erroneously terminated the slave thread. @item -Added sql variables: @code{SQL_MAX_JOIN_SIZE} and @code{SQL_SAFE_UPDATES}. +Added optimization of queries where @code{DISTINCT} is only used on columns +from some of the tables. @item -Added @code{READ LOCAL} lock that doesn't lock the table for concurrent -inserts. (This is used by @code{mysqldump}). +Allow floating-point numbers where there is no sign after the exponent +(like 1e1). @item -Changed that @code{LOCK TABLES ... READ} doesn't anymore allow concurrent -inserts. +@code{SHOW GRANTS} didn't always show all column grants. @item -Added option @code{--skip-delay-key-write} to @code{mysqld}. +Added @code{--default-extra-file=#} to all @strong{MySQL} clients. @item -Fixed security problem in the protocol regarding password checking. +Columns referenced in @code{INSERT} statements now are initialized properly. @item -@code{_rowid} can now be used as an alias for an integer type unique indexed -column. +@code{UPDATE} didn't always work when used with a range on a timestamp that +was part of the key that was used to find rows. @item -Added back blocking of @code{SIGPIPE} when compiling with @code{--thread-safe-clients} -to make things safe for old clients. -@end itemize - -@node News-3.23.10, News-3.23.9, News-3.23.11, News-3.23.x -@appendixsubsec Changes in release 3.23.10 -@itemize @bullet +Fixed a bug in @code{FULLTEXT} index when inserting a @code{NULL} column. @item -Fixed bug in 3.23.9 where memory wasn't properly freed when doing -@code{LOCK TABLES}. +Changed to use @code{mkstemp()} instead of @code{tempnam()}. Based +on a patch from John Jones. @end itemize -@node News-3.23.9, News-3.23.8, News-3.23.10, News-3.23.x -@appendixsubsec Changes in release 3.23.9 +@node News-3.23.25, News-3.23.24, News-3.23.26, News-3.23.x +@appendixsubsec Changes in release 3.23.25 @itemize @bullet @item -Fixed problem that affected queries that did arithmetic on group functions. -@item -Fixed problem with timestamps and @code{INSERT DELAYED}. -@item -Fixed that @code{date_column BETWEEN const_date AND const_date} works. +Fixed that @code{databasename} works as second argument to @code{mysqlhotcopy}. @item -Fixed problem when only changing a 0 to @code{NULL} in a table with -@code{BLOB/TEXT} columns. +@code{UMASK} and @code{UMASK_DIR} can now be specified in octal. @item -Fixed bug in range optimizer when using many key parts and or on the middle -key parts: @code{WHERE K1=1 and K3=2 and (K2=2 and K4=4 or K2=3 and K4=5)} +Added @code{RIGHT JOIN}. This makes @code{RIGHT} a reserved word. @item -Added command @code{source} to @code{mysql} to allow reading of batch files -inside the @code{mysql} client. Original patch by Matthew Vanecek. +Added @code{@@@@IDENTITY} as a synonym for @code{LAST_INSERT_ID()}. +(This is for Visual Basic compatibility.) @item -Fixed critical problem with the @code{WITH GRANT OPTION} option. +Fixed a bug in @code{myisamchk} and @code{REPAIR} when using @code{FULLTEXT} +index. @item -Don't give an unnecessary @code{GRANT} error when using tables from many -databases in the same query. +@code{LOAD DATA INFILE} now works with FIFOs. +(Patch by Toni L. Harbaugh-Blackford.) @item -Added VIO wrapper (needed for SSL support ; By Andrei Errapart and -Tõnu Samuel). +@code{FLUSH LOGS} broke replication if you specified a log name with an +explicit extension as the value of the @code{log-bin} option. @item -Fixed optimizer problem on @code{SELECT} when using many overlapping indexes. -@strong{MySQL} should now be able to choose keys even better when there -is many keys to choose from. +Fixed a bug in @code{MyISAM} with packed multi-part keys. @item -Changed optimizer to prefer a range key instead of a ref key when the -range key can uses more columns than the ref key (which only can use -columns with =). For example, the following type of queries should now -be faster: @code{SELECT * from key_part_1=const and key_part_2 > const2} +Fixed crash when using @code{CHECK TABLE} on Windows. @item -Fixed bug that a change of all @code{VARCHAR} columns to @code{CHAR} columns -didn't change row type from dynamic to fixed. +Fixed a bug where @code{FULLTEXT} index always used the koi8_ukr +character set. @item -Disabled floating-point exceptions for FreeBSD to fix core dump when -doing @code{SELECT floor(pow(2,63))}. +Fixed privilege checking for @code{CHECK TABLE}. @item -Changed @code{mysqld} startup option @code{--delay-key-write} to -@code{--delay-key-write-for-all-tables} +The @code{MyISAM} repair/reindex code didn't use the @code{--tempdir} +option for its temporary files. @item -Added @code{read-next-on-key} to @code{HEAP} tables. This should fix all -problems with @code{HEAP} tables when using not @code{UNIQUE} keys. +Added @code{BACKUP TABLE/RESTORE TABLE}. @item -Added print of default arguments options to all clients. +Fixed coredump on @code{CHANGE MASTER TO} when the slave did not have +the master to start with. @item -Added @code{--log-slow-queries} to @code{mysqld} to log all queries that take a -long time to a separate log file with a time of how long the query took. +Fixed incorrect @code{Time} in the processlist for @code{Connect} of +the slave thread. @item -Fixed core dump when doing @code{WHERE key_column=RAND(...)} +The slave now logs when it connects to the master. @item -Fixed optimization bug in @code{SELECT ... LEFT JOIN ... key_column IS NULL}, -when @code{key_column} could contain @code{NULL} values. +Fixed a core dump bug when doing @code{FLUSH MASTER} if you didn't specify +a filename argument to @code{--log-bin}. @item -Fixed problem with 8-bit characters as separators in @code{LOAD DATA INFILE}. -@end itemize - -@node News-3.23.8, News-3.23.7, News-3.23.9, News-3.23.x -@appendixsubsec Changes in release 3.23.8 -@itemize @bullet +Added missing @file{ha_berkeley.x} files to the @strong{MySQL} Windows @item -Fixed problem when handling indexfiles larger than 8G. +Fixed some mutex bugs in the log code that could cause thread blocks if new +log files couldn't be created. @item -Added latest patches to mit-pthreads for NetBSD. +Added lock time and number of selected processed rows to slow query log. @item -Fixed problem with timezones that are < GMT -11. +Added @code{--memlock} option to @code{mysqld} to lock @code{mysqld} +in memory on systems with the @code{mlockall()} call (like in Solaris). @item -Fixed a bug when deleting packed keys in @code{NISAM}. +@code{HEAP} tables didn't use keys properly. (Bug from 3.23.23.) @item -Fixed problem with @code{ISAM} when doing some @code{ORDER BY ... DESC} queries. +Added better support for @code{MERGE} tables (keys, mapping, creation, +documentation...). @xref{MERGE}. @item -Fixed bug when doing a join on a text key which didn't cover the whole key. +Fixed bug in @code{mysqldump} from 3.23 which caused some @code{CHAR} columns +not to be quoted. @item -Option @code{--delay-key-write} didn't enable delayed key writing. +Merged @code{analyze}, @code{check}, @code{optimize} and repair code. @item -Fixed update of @code{TEXT} column which only involved case changes. +@code{OPTIMIZE TABLE} is now mapped to @code{REPAIR} with statistics and +sorting of the index tree. This means that for the moment it only +works on @code{MyISAM} tables. @item -Fixed that @code{INSERT DELAYED} doesn't update timestamps that are given. +Added a pre-alloced block to root_malloc to get fewer mallocs. @item -Added function @code{YEARWEEK()} and options @code{x}, @code{X}, @code{v} and -@code{V} to @code{DATE_FORMAT()}. +Added a lot of new statistics variables. @item -Fixed problem with @code{MAX(indexed_column)} and HEAP tables. +Fixed @code{ORDER BY} bug with BDB tables. @item -Fixed problem with @code{BLOB NULL} keys and @code{LIKE} "prefix%". +Removed warning that @code{mysqld} couldn't remove the @file{.pid} file +under Windows. @item -Fixed problem with @code{MyISAM} and fixed-length rows < 5 bytes. +Changed @code{--log-isam} to log @strong{MyISAM} tables instead of isam +tables. @item -Fixed problem that could cause @strong{MySQL} to touch freed memory when -doing very complicated @code{GROUP BY} queries. +Fixed @code{CHECK TABLE} to work on Windows. @item -Fixed core dump if you got a crashed table where an @code{ENUM} field value -was too big. +Added file mutexes to make @code{pwrite()} safe on Windows. @end itemize -@node News-3.23.7, News-3.23.6, News-3.23.8, News-3.23.x -@appendixsubsec Changes in release 3.23.7 +@node News-3.23.24, News-3.23.23, News-3.23.25, News-3.23.x +@appendixsubsec Changes in release 3.23.24 @itemize @bullet @item -Fixed workaround under Linux to avoid problems with @code{pthread_mutex_timedwait}, -which is used with @code{INSERT DELAYED}. @xref{Linux}. -@item -Fixed that one will get a 'disk full' error message if one gets disk full -when doing sorting (instead of waiting until we got more disk space). -@item -Fixed a bug in @code{MyISAM} with keys > 250 characters. -@item -In @code{MyISAM} one can now do an @code{INSERT} at the same time as other -threads are reading from the table. -@item -Added variable @code{max_write_lock_count} to @code{mysqld} to force a -@code{READ} lock after a certain number of @code{WRITE} locks. +Added @code{mysqld} variable @code{created_tmp_disk_tables}. @item -Inverted flag @code{delayed_key_write} on @code{show variables}. +To make it possible to reliably dump and restore tables with +@code{TIMESTAMP(X)} columns, @strong{MySQL} now reports columns with @code{X} +other than 14 or 8 to be strings. @item -Renamed variable @code{concurrency} to @code{thread_concurrency}. +Changed sort order for latin1 as it was before @strong{MySQL} Version 3.23.23. +Any table with @code{CHAR} columns that may have characters with ASCII +values greater than 128 that was created or modified with 3.23.22 must +be repaired! @item -The following functions are now multi-byte-safe: -@code{LOCATE(substr,str)}, @code{POSITION(substr IN str)}, -@code{LOCATE(substr,str,pos)}, @code{INSTR(str,substr)}, -@code{LEFT(str,len)}, @code{RIGHT(str,len)}, -@code{SUBSTRING(str,pos,len)}, @code{SUBSTRING(str FROM pos FOR len)}, -@code{MID(str,pos,len)}, @code{SUBSTRING(str,pos)}, @code{SUBSTRING(str -FROM pos)}, @code{SUBSTRING_INDEX(str,delim,count)}, @code{RTRIM(str)}, -@code{TRIM([[BOTH | TRAILING] [remstr] FROM] str)}, -@code{REPLACE(str,from_str,to_str)}, @code{REVERSE(str)}, -@code{INSERT(str,pos,len,newstr)}, @code{LCASE(str)}, @code{LOWER(str)}, -@code{UCASE(str)} and @code{UPPER(str)}; Patch by Wei He. +Fixed small memory leak introduced from 3.23.22 when creating a +temporary table. @item -Fix core dump when releasing a lock from a non-existent table. +Fixed problem with BDB tables and reading on a unique (not primary) key. @item -Remove locks on tables before starting to remove duplicates. +Restored the win1251 character set (it's now only marked deprecated). +@end itemize + +@node News-3.23.23, News-3.23.22, News-3.23.24, News-3.23.x +@appendixsubsec Changes in release 3.23.23 +@itemize @bullet @item -Added option @code{FULL} to @code{SHOW PROCESSLIST}. +Changed sort order for 'German'; All tables created with 'German' sortorder +must be repaired with @code{REPAIR TABLE} or @code{myisamchk} before use! @item -Added option @code{--verbose} to @code{mysqladmin}. +Added option @code{--core-file} to @code{mysqld} to get a core file on +Linux if @code{mysqld} dies on the SIGSEGV signal. @item -Fixed problem when automatically converting HEAP to MyISAM. +@strong{MySQL} client @code{mysql} now starts with option +@code{--no-named-commands} (@code{-g}) by default. This option can be +disabled with @code{--enable-named-commands} (@code{-G}). This may cause +incompatibility problems in some cases, for example, in SQL scripts that +use named commands without a semicolon, etc. ! Long format commands +still work from the first line. @item -Fixed bug in HEAP tables when doing insert + delete + insert + scan the -table. +Fixed a problem when using many pending @code{DROP TABLE} statements at +the same time. @item -Fixed bugs on Alpha with @code{REPLACE()} and @code{LOAD DATA INFILE}. +Optimizer didn't use keys properly when using @code{LEFT JOIN} on an +empty table. @item -Added @code{mysqld} variable @code{interactive_timeout}. +Added shorter help text when invoking @code{mysqld} with incorrect options. @item -Changed the argument to @code{mysql_data_seek()} from @code{ulong} to -@code{ulonglong}. -@end itemize - -@node News-3.23.6, News-3.23.5, News-3.23.7, News-3.23.x -@appendixsubsec Changes in release 3.23.6 - -@itemize @bullet +Fixed non-fatal @code{free()} bug in @code{mysqlimport}. @item -Added @code{mysqld} option @code{-O lower_case_table_names=@{0|1@}} to allow -users to force table names to lowercase. +Fixed bug in @code{MyISAM} index handling of +@code{DECIMAL}/@code{NUMERIC} keys. @item -Added @code{SELECT ... INTO DUMPFILE}. +Fixed a bug in concurrent insert in @code{MyISAM} tables; In some contexts, +usage of @code{MIN(key_part)} or @code{MAX(key_part)} returned an empty set. @item -Added @code{mysqld} option @code{--ansi} to make some functions @code{ANSI SQL} -compatible. +Updated @code{mysqlhotcopy} to use the new @code{FLUSH TABLES table_list} +syntax. Only tables which are being backed up are flushed now. @item -Temporary tables now starts with @code{#sql}. +Changed behavior of @code{--enable-thread-safe-client} so +that both non-threaded (@code{-lmysqlclient}) and threaded +(@code{-lmysqlclient_r}) libraries are built. Users who linked +against a threaded @code{-lmysqlclient} will need to link against +@code{libmysqlclient_r} now. @item -Added quoting of identifiers with @code{`} (@code{"} in @code{--ansi} mode). +Added atomic @code{RENAME} command. @item -Changed to use snprintf() when printing floats to avoid some buffer -overflows on FreeBSD. +Don't count entries with @code{NULL} in @code{COUNT(DISTINCT ...)}. @item -Made @code{[floor()} overflow safe on FreeBSD. +Changed @code{ALTER TABLE}, @code{LOAD DATA INFILE} on empty tables and +@code{INSERT ... SELECT ...} on empty tables to create non-unique indexes +in a separate batch with sorting. This will make the above calls much +faster when you have many indexes. @item -Added option @code{--quote-names} to @code{mysqldump} +@code{ALTER TABLE} now logs the first used insert_id correctly. @item -Fixed bug that one could make a part of a @code{PRIMARY KEY NOT NULL}. +Fixed crash when adding a default value to a @code{BLOB} column. @item -Fixed @code{encrypt()} to be thread safe and not reuse buffer. +Fixed a bug with @code{DATE_ADD/DATE_SUB} where it returned a datetime instead +of a date. @item -Added @code{mysql_odbc_escape_string()} function to support big5 characters in -MyODBC. +Fixed a problem with the thread cache which made some threads show up as +@code{***DEAD***} in @code{SHOW PROCESSLIST}. @item -Rewrote the table handler to use classes. This introduces a lot of new code, -but will make table handling faster and better. +Fixed a lock in our thr_rwlock code, which could make selects that run +at the same time as concurrent inserts crash. This only affects systems +that don't have the @code{pthread_rwlock_rdlock} code. @item -Added patch by Sasha for user-defined variables. +When deleting rows with a non-unique key in a HEAP table, all rows weren't +always deleted. @item -Changed that @code{FLOAT} and @code{DOUBLE} (without any length modifiers) are -not anymore fixed decimal point numbers. +Fixed bug in range optimizer for HEAP tables for searches on a part index. @item -Changed the meaning of @code{FLOAT(X)}: Now this is the same as @code{FLOAT} if -X <= 24 and a @code{DOUBLE} if 24 < X <= 53. +Fixed that @code{SELECT} on part keys works with BDB tables. @item -@code{DECIMAL(X)} is now an alias for @code{DECIMAL(X,0)} and @code{DECIMAL} -is now an alias for @code{DECIMAL(10,0)}. The same goes for @code{NUMERIC}. +Fixed @code{INSERT INTO bdb_table ... SELECT} to work with BDB tables. @item -Added option @code{ROW_FORMAT=@{default | dynamic | static | compressed@}} to -@code{CREATE_TABLE}. +@code{CHECK TABLE} now updates key statistics for the table. @item -@code{DELETE FROM table_name} didn't work on temporary tables. +@code{ANALYZE TABLE} will now only update tables that have been changed +since thee last @code{ANALYZE}. Note that this is a new feature and tables +will not be marked to be analyzed until they are updated in any way with +3.23.23 or newer. For older tables, you have to do @code{CHECK TABLE} +to update the key distribution. @item -Changed function @code{CHAR_LENGTH()} to be multi-byte character safe. +Fixed some minor privilege problems with @code{CHECK}, @code{ANALYZE}, +@code{REPAIR} and @code{SHOW CREATE} commands. @item -Added function @code{ORD(string)}. -@end itemize - -@node News-3.23.5, News-3.23.4, News-3.23.6, News-3.23.x -@appendixsubsec Changes in release 3.23.5 - -@itemize @bullet +Added @code{CHANGE MASTER TO} command. @item -Fixed some Y2K problems in the new date handling in 3.23. +Added @code{FAST}, @code{QUICK} @code{EXTENDED} check types to +@code{CHECK TABLES}. @item -Fixed problem with @code{SELECT DISTINCT ... ORDER BY RAND()}. +Changed @code{myisamchk} so that @code{--fast} and +@code{--check-changed-tables} are also honored with @code{--sort-index} and +@code{--analyze}. @item -Added patches by Sergei A. Golubchik for text searching on the MyISAM level. +Fixed fatal bug in @code{LOAD TABLE FROM MASTER} that did not lock the +table during index re-build. @item -Fixed cache overflow problem when using full joins without keys. +@code{LOAD DATA INFILE} broke replication if the database was excluded +from replication. @item -Fixed some configure issues. +More variables in @code{SHOW SLAVE STATUS} and @code{SHOW MASTER STATUS}. @item -Some small changes to make parsing faster. +@code{SLAVE STOP} now will not return until the slave thread actually exits. @item -@code{ALTER TABLE} + adding a column after the last field didn't work. +Full text search via the @code{MATCH} function and @code{FULLTEXT} index type. +(For MyISAM files). This makes @code{FULLTEXT} a reserved word. +@end itemize + +@node News-3.23.22, News-3.23.21, News-3.23.23, News-3.23.x +@appendixsubsec Changes in release 3.23.22 +@itemize @bullet @item -Fixed problem when using an auto_increment column in two keys +Fixed that @code{lex_hash.h} is created properly for each @strong{MySQL} +distribution. @item -One can now with MyISAM have the auto_increment part as a sub part: -@code{CREATE TABLE foo (a int not null auto_increment, b char(5), primary key (b,a))} +Fixed that @code{MASTER} and @code{COLLECTION} are not reserved words. @item -Fixed bug in MyISAM with packed char keys that could be @code{NULL}. +The log generated by @code{--slow-query-log} didn't contain the whole queries. @item -@code{AS} on fieldname with @code{CREATE TABLE table_name SELECT ...} didn't -work. +Fixed that open transactions in BDB tables are rolled back if the +connection is closed unexpectedly. @item -Allow use of @code{NATIONAL} and @code{NCHAR} when defining character columns. -This is the same as not using @code{BINARY}. +Added workaround for a bug in @code{gcc} 2.96 (intel) and @code{gcc} 2.9 +(Ia64) in @code{gen_lex_hash.c}. @item -Don't allow @code{NULL} columns in a @code{PRIMARY KEY} (only in @code{UNIQUE} -keys). +Fixed memory leak in the client library when using @code{host=} in the +@code{my.cnf} file. @item -Clear @code{LAST_INSERT_ID} if one uses this in ODBC: -@code{WHERE auto_increment_column IS NULL}. This seems to fix some problems -with Access. +Optimized functions that manipulate the hours/minutes/seconds. @item -@code{SET SQL_AUTO_IS_NULL=0|1} now turns on/off the handling of -searching after the last inserted row with @code{WHERE -auto_increment_column IS NULL}. +Fixed bug when comparing the result of @code{DATE_ADD()}/@code{DATE_SUB()} +against a number. @item -Added new @code{mysqld} variable @code{concurrency} for Solaris. +Changed the meaning of @code{-F, --fast} for @code{myisamchk}. Added option +@code{-C, --check-only-changed} to @code{myisamchk}. @item -Added option @code{--relative} to @code{mysqladmin} to make -@code{extended-status} more useful to monitor changes. +Added @code{ANALYZE table_name} to update key statistics for tables. @item -Fixed bug when using @code{COUNT(DISTINCT ...)} on an empty table. +Changed binary items @code{0x...} to be regarded as integers by default. @item -Added support for the Chinese character set GBK. +Fix for SCO and @code{SHOW PROCESSLIST}. @item -Fixed problem with @code{LOAD DATA INFILE} and @code{BLOB} columns. +Added @code{auto-rehash} on reconnect for the @code{mysql} client. @item -Added bit operator @code{~} (negation). +Fixed a newly introduced bug in @code{MyISAM}, where the index file couldn't +get bigger than 64M. @item -Fixed problem with @code{UDF} functions. +Added @code{SHOW MASTER STATUS} and @code{SHOW SLAVE STATUS}. @end itemize -@node News-3.23.4, News-3.23.3, News-3.23.5, News-3.23.x -@appendixsubsec Changes in release 3.23.4 - +@node News-3.23.21, News-3.23.20, News-3.23.22, News-3.23.x +@appendixsubsec Changes in release 3.23.21 @itemize @bullet @item -Inserting a @code{DATETIME} into a @code{TIME} column will not anymore -try to store 'days' in it. +Added @code{mysql_character_set_name(MYSQL *mysql)} function to the +@strong{MySQL} C API. @item -Fixed problem with storage of float/double on little endian machines. -(This affected @code{SUM()}.) +Made the update log @code{ASCII 0} safe. @item -Added connect timeout on TCP/IP connections. +Added the @code{mysql_config} script. @item -Fixed problem with @code{LIKE} "%" on an index that may have @code{NULL} values. +Fixed problem when using @code{<} or @code{>} with a char column that was only +partly indexed. @item -@code{REVOKE ALL PRIVILEGES} didn't revoke all privileges. +One would get a core dump if the log file was not readable by the MySQL user. @item -Allow creation of temporary tables with same name as the original table. +Changed @code{mysqladmin} to use the @code{CREATE DATABASE}/@code{DROP +DATABASE} commands instead of the old deprecated API calls. @item -When granting a user a grant option for a database, he couldn't grant -privileges to other users. +Fixed @code{chown} warning in @code{safe_mysqld}. @item -New command: @code{SHOW GRANTS FOR user} (by Sinisa). +Fixed a bug in @code{ORDER BY} that was introduced in 3.23.19. @item -New @code{date_add} syntax: @code{date/datetime + INTERVAL # interval_type}. -By Joshua Chamas. +Only optimize the @code{DELETE FROM tbl_name} to do a drop+create of +the table if we are in @code{AUTOCOMMIT} mode. (Needed for BDB tables). @item -Fixed privilege check for @code{LOAD DATA REPLACE}. +Added extra checks to avoid index corruption when the @code{ISAM}/@code{MyISAM} +index files gets full during an @code{INSERT}/@code{UPDATE}. @item -Automatic fixing of broken include files on Solaris 2.7 +@code{myisamchk} didn't correctly update row checksum when used with +@code{-ro} (This only gave an warning in subsequent runs). @item -Some configure issues to fix problems with big file system detection. +Fixed bug in @code{REPAIR TABLE} so that it works with tables without indexes. @item -@code{REGEXP} is now case insensitive if you use non-binary strings. +Fixed buffer overrun in @code{DROP DATABASE} +@item +@code{LOAD TABLE FROM MASTER} is sufficiently bug-free to announce it as +a feature. +@item +@code{MATCH} and @code{AGAINST} are now reserved words. @end itemize -@node News-3.23.3, News-3.23.2, News-3.23.4, News-3.23.x -@appendixsubsec Changes in release 3.23.3 - +@node News-3.23.20, News-3.23.19, News-3.23.21, News-3.23.x +@appendixsubsec Changes in release 3.23.20 @itemize @bullet @item -Added patches for MIT-pthreads on NetBSD. -@item -Fixed range bug in MyISAM. -@item -@code{ASC} is now the default again for @code{ORDER BY}. -@item -Added @code{LIMIT} to @code{UPDATE}. +Fixed bug in 3.23.19; @code{DELETE FROM tbl_name} removed the .frm file. @item -New client function: @code{mysql_change_user()}. +Added @code{SHOW CREATE TABLE}. +@end itemize + +@node News-3.23.19, News-3.23.18, News-3.23.20, News-3.23.x +@appendixsubsec Changes in release 3.23.19 +@itemize @bullet @item -Added character set to @code{SHOW VARIABLES}. +Changed copyright for all files to GPL for the server code and utilities and +LGPL for the client libraries. @item -Added support of @code{--[whitespace]} comments. +Fixed bug where all rows matching weren't updated on a @code{MyISAM} table +when doing update based on key on a table with many keys and some key changed +values. @item -Allow @code{INSERT into tbl_name VALUES ()}, that is, you may now specify -an empty value list to insert a row in which each column is set to its -default value. +The Linux @strong{MySQL} RPM's and binaries are now statically linked with +a linuxthread version that has faster mutex handling when used with +@strong{MySQL}. @item -Changed @code{SUBSTRING(text FROM pos)} to conform to ANSI SQL. (Before this -construct returned the rightmost 'pos' characters). +@code{ORDER BY} can now uses @code{REF} keys to find subset the rows +that needs to be sorted. @item -@code{SUM()} with @code{GROUP BY} returned 0 on some systems. +Changed name of @code{print_defaults} to @code{my_print_defaults} to avoid +name confusion. @item -Changed output for @code{SHOW TABLE STATUS}. +Fixed @code{NULLIF()} to work according to ANSI SQL99. @item -Added @code{DELAY_KEY_WRITE} option to @code{CREATE TABLE}. +Added @code{net_read_timeout} and @code{net_write_timeout} as startup +parameters to @code{mysqld}. @item -Allow @code{AUTO_INCREMENT} on any key part. +Fixed bug that destroyed index when doing @code{myisamchk --sort-records} +on a table with prefix compressed index. @item -Fixed problem with @code{YEAR(NOW())} and @code{YEAR(CURDATE())}. +Added pack_isam and myisampack to the standard @strong{MySQL} distribution. @item -Added @code{CASE} construct. +Added the syntax @code{BEGIN WORK} (the same as @code{BEGIN}). @item -New function @code{COALESCE()}. +Fixed core dump bug when using @code{ORDER BY} on a @code{CONV()} expression. +@item Added @code{LOAD TABLE FROM MASTER} +@item Added @code{FLUSH MASTER} and @code{FLUSH SLAVE} +@item Fixed big/little endian problem in the replication @end itemize -@node News-3.23.2, News-3.23.1, News-3.23.3, News-3.23.x -@appendixsubsec Changes in release 3.23.2 - +@node News-3.23.18, News-3.23.17, News-3.23.19, News-3.23.x +@appendixsubsec Changes in release 3.23.18 @itemize @bullet @item -Fixed range optimizer bug: @code{SELECT * FROM table_name WHERE -key_part1 >= const AND (key_part2 = const OR key_part2 = const)}. The -bug was that some rows could be duplicated in the result. +Fixed a problem from 3.23.17 when choosing character set on the client side. @item -Running @code{myisamchk} without @code{-a} updated the index -distribution wrong. +Added @code{FLUSH TABLES WITH READ LOCK} to make a global lock suitable to +make a copy of @strong{MySQL} data files. @item -@code{SET SQL_LOW_PRIORITY_UPDATES=1} gave parse error before. +@code{CREATE TABLE ... SELECT ... PROCEDURE} now works. @item -You can now update indexes columns that are used in the @code{WHERE} clause. -@code{UPDATE tbl_name SET KEY=KEY+1 WHERE KEY > 100} +Internal temporary tables will now uses compressed index when using +@code{GROUP BY} on @code{VARCHAR/CHAR} columns. @item -Date handling should now be a bit faster. +Fixed a problem when locking the same table with both a @code{READ} and a +@code{WRITE} lock. @item -Added handling of fuzzy dates (dates where day or month is 0): -(Like: 1999-01-00) +Fixed problem with myisamchk and @code{RAID} tables. +@end itemize + +@node News-3.23.17, News-3.23.16, News-3.23.18, News-3.23.x +@appendixsubsec Changes in release 3.23.17 +@itemize @bullet @item -Fixed optimization of @code{SELECT ... WHERE key_part1=const1 AND -key_part_2=const2 AND key_part1=const4 AND key_part2=const4} ; Indextype -should be @code{range} instead of @code{ref}. +Fixed a bug in @code{find_in_set()} when the first argument was @code{NULL}. @item -Fixed @code{egcs} 1.1.2 optimizer bug (when using @code{BLOB}s) on Linux Alpha. +Added table locks to Berkeley DB. @item -Fixed problem with @code{LOCK TABLES} combined with @code{DELETE FROM table}. +Fixed a bug with @code{LEFT JOIN} and @code{ORDER BY} where the first +table had only one matching row. @item -MyISAM tables now allow keys on @code{NULL} and @code{BLOB/TEXT} columns. +Added 4 sample @code{my.cnf} example files in the @file{support-files} +directory. @item -The following join is now much faster: -@code{SELECT ... FROM t1 LEFT JOIN t2 ON ... WHERE t2.not_null_column IS NULL}. +Fixed @code{duplicated key} problem when doing big @code{GROUP BY}'s. +(This bug was probably introduced in 3.23.15). +@item +Changed syntax for @code{INNER JOIN} to match ANSI SQL. @item -@code{ORDER BY} and @code{GROUP BY} can be done on functions. +Added @code{NATURAL JOIN} syntax. @item -Changed handling of 'const_item' to allow handling of @code{ORDER BY RAND()}. +A lot of fixes in the @code{BDB} interface. @item -Indexes are now used for @code{WHERE key_column = function}. +Added handling of @code{--no-defaults} and @code{--defaults-file} to +@code{safe_mysqld.sh} and @code{mysql_install_db.sh}. @item -Indexes are now used for @code{WHERE key_column = column_name} even if -the columns are not identically packed. +Fixed bug in reading compressed tables with many threads. @item -Indexes are now used for @code{WHERE column_name IS NULL}. +Fixed that @code{USE INDEX} works with @code{PRIMARY} keys. @item -Changed heap tables to be stored in low_byte_first order (to make it easy -to convert to MyISAM tables) +Added @code{BEGIN} statement to start a transaction in @code{AUTOCOMMIT} mode. @item -Automatic change of HEAP temporary tables to MyISAM tables in case of -'table is full' errors. +Added symbolic links support for Windows. @item -Added option @code{--init-file=file_name} to @code{mysqld}. +Changed protocol to let client know if the server is in AUTOCOMMIT mode +and if there is a pending transaction. If there is a pending transaction +the client library will give an error before reconnecting to the server to +let the client know that the server did a rollback. +The protocol is still backward compatible with old clients @item -@code{COUNT(DISTINCT value, [value, ...])} +@code{KILL} now works on a thread that is locked on a 'write' to a dead client. @item -@code{CREATE TEMPORARY TABLE} now creates a temporary table, in its own -namespace, that is automatically deleted if connection is dropped. +Fixed memory leak in the replication slave thread. @item -New reserved words (required for @code{CASE}): @code{CASE, THEN, WHEN, ELSE and END}. +Added new option @code{log-slave-updates} to allow daisy-chaining the slaves. @item -New functions @code{EXPORT_SET()} and @code{MD5()}. +Fixed compile error on FreeBSD and other systems where @code{pthread_t} +is not the same as @code{int}. @item -Support for the GB2312 Chinese character set. -@end itemize - -@node News-3.23.1, News-3.23.0, News-3.23.2, News-3.23.x -@appendixsubsec Changes in release 3.23.1 - -@itemize @bullet +Fixed master shutdown aborting the slave thread. @item -Fixed some compilation problems. +Fixed a race condition in @code{INSERT DELAYED} code when doing +@code{ALTER TABLE}. +@item +Added deadlock detection sanity checks to @code{INSERT DELAYED} @end itemize -@node News-3.23.0, , News-3.23.1, News-3.23.x -@appendixsubsec Changes in release 3.23.0 - +@node News-3.23.16, News-3.23.15, News-3.23.17, News-3.23.x +@appendixsubsec Changes in release 3.23.16 @itemize @bullet @item -A new table handler library (@code{MyISAM}) with a lot of new features. -@xref{MyISAM}. -@item -You can create in-memory @code{HEAP} tables which are extremely fast for -lookups. -@item -Support for big files (63 bit) on OSes that support big files. -@item -New function @code{LOAD_FILE(filename)} to get the contents of a file as a -string value. +Added option @code{TYPE=QUICK} to @code{CHECK} and @code{REPAIR}. @item -New operator @code{<=>} which will act as @code{=} but will return TRUE -if both arguments are @code{NULL}. This is useful for comparing changes -between tables. +Fixed bug in @code{REPAIR TABLE} when the table was in use by other threads. @item -Added the ODBC 3.0 @code{EXTRACT(interval FROM datetime)} function. +Added a thread cache to make it possible to debug @strong{MySQL} with +@code{gdb} when one does a lot of reconnects. This will also improve +systems where you can't use persistent connections. @item -Columns defined as @code{FLOAT(X)} is not rounded on storage and may be -in scientific notation (1.0 E+10) when retrieved. +Lots of fixes in the Berkeley DB interface. @item -@code{REPLACE} is now faster than before. +@code{UPDATE IGNORE} will not abort if an update results in a +@code{DUPLICATE_KEY} error. @item -Changed @code{LIKE} character comparison to behave as @code{=}; -This means that @code{'e' LIKE '@'e'} (if the line doesn't display -correctly, the latter 'e' means a French 'e' with a dot above) is now true. +Put @code{CREATE TEMPORARY TABLE} commands in the update log. @item -@code{SHOW TABLE STATUS} returns a lot of information about the tables. +Fixed bug in handling of masked IP numbers in the privilege tables. @item -Added @code{LIKE} to the @code{SHOW STATUS} command. +Fixed bug with @code{delayed_key_writes} tables and @code{CHECK TABLE}. @item -Added privilege column to @code{SHOW COLUMNS}. +Added @code{replicate-do-db} and @code{replicate-ignore-db} options to +restrict which databases get replicated @item -Added columns @code{packed} and @code{comment} to @code{SHOW INDEX}. +Added @code{SQL_LOG_BIN} option +@end itemize + +@node News-3.23.15, News-3.23.14, News-3.23.16, News-3.23.x +@appendixsubsec Changes in release 3.23.15 +@itemize @bullet @item -Added comments to tables (with @code{CREATE TABLE ... COMMENT "xxx"}). +To start @code{mysqld} as @code{root}, you must now use the @code{--user=root} option. @item -Added @code{UNIQUE}, as in -@code{CREATE TABLE table_name (col int not null UNIQUE)} +Added interface to Berkeley DB. (This is not yet functional; Play with it at +your own risk!) @item -New create syntax: @code{CREATE TABLE table_name SELECT ...} +Replication between master and slaves. @item -New create syntax: @code{CREATE TABLE IF NOT EXISTS ...} +Fixed bug that other threads could steal a lock when a thread had +a lock on a table and did a @code{FLUSH TABLES} command. @item -Allow creation of @code{CHAR(0)} columns. +Added the @code{slow_launch_time} variable and the @code{Slow_launch_threads} +status variable to @code{mysqld}. These can be examined with +@code{mysqladmin variables} and @code{mysqladmin extended-status}. @item -@code{DATE_FORMAT()} now requires @samp{%} before any format character. +Added functions @code{INET_NTOA()} and @code{INET_ATON()}. @item -@code{DELAYED} is now a reserved word (sorry about that :( ). +The default type of @code{IF()} now depends on the second and +third arguments and not only on the second argument. @item -An example procedure is added: @code{analyse}, file: @file{sql_analyse.c}. -This will describe the data in your query. Try the following: -@example -SELECT ... FROM ... WHERE ... PROCEDURE ANALYSE([max elements,[max memory]]) -@end example - -This procedure is extremely useful when you want to check the data in your -table! +Fixed case when @code{myisamchk} could go into a loop when trying to +repair a crashed table. @item -@code{BINARY} cast to force a string to be compared case sensitively. +Don't write @code{INSERT DELAYED} to update log if @code{SQL_LOG_UPDATE=0}. @item -Added option @code{--skip-show-database} to @code{mysqld}. +Fixed problem with @code{REPLACE} on @code{HEAP} tables. @item -Check if a row has changed in an @code{UPDATE} now also works with -@code{BLOB}/@code{TEXT} columns. +Added possible character sets and time zone to @code{SHOW VARIABLES}. @item -Added the @code{INNER} join syntax. @strong{NOTE}: This made @code{INNER} -a reserved word! +Fixed bug in locking code that could result it locking problems with +concurrent inserts under high load. @item -Added support for netmasks to the hostname in the @strong{MySQL} tables. -You can specify a netmask using the @code{IP/NETMASK} syntax. +Fixed a problem with @code{DELETE} of many rows on a table with +compressed keys where @strong{MySQL} scanned the index to find the rows. @item -If you compare a @code{NOT NULL DATE/DATETIME} column with @code{IS -NULL}, this is changed to a compare against @code{0} to satisfy some ODBC -applications. (By @email{shreeve@@uci.edu}). +Fixed problem with @code{CHECK} on table with deleted keyblocks. @item -@code{NULL IN (...)} now returns @code{NULL} instead of @code{0}. This will -ensure that @code{null_column NOT IN (...)} doesn't match -@code{NULL} values. +Fixed a bug in reconnect (at the client side) where it didn't free memory +properly in some contexts. @item -Fix storage of floating-point values in @code{TIME} columns. +Fixed problems in update log when using @code{LAST_INSERT_ID()} to update +a table with an auto_increment key. @item -Changed parsing of @code{TIME} strings to be more strict. Now the -fractional second part is detected (and currently skipped). The -following formats are supported: -@table @code -@item [[DAYS] [H]H:]MM:]SS[.fraction] -@item [[[[[H]H]H]H]MM]SS[.fraction] -@end table +Added function @code{NULLIF()}. @item -Detect (and ignore) second fraction part from @code{DATETIME}. +Fixed bug when using @code{LOAD DATA INFILE} on a table with +@code{BLOB/TEXT} columns. @item -Added the @code{LOW_PRIORITY} attribute to @code{LOAD DATA INFILE}. +Optimised MyISAM to be faster when inserting keys in sorted order. @item -The default index name now uses the same case as the used column name. +@code{EXPLAIN SELECT ...} now also prints out whether @strong{MySQL} needs to +create a temporary table or use file sorting when resolving the @code{SELECT}. @item -Changed default number of connections to 100. +Added optimization to skip @code{ORDER BY} parts where the part is a +constant expression in the @code{WHERE} part. Indexes can now be used +even if the @code{ORDER BY} doesn't match the index exactly, as long as +all the not used index parts and all the extra @code{ORDER BY} +columns are constants in the @code{WHERE} clause. @xref{MySQL indexes}. @item -Use bigger buffers when using @code{LOAD DATA INFILE}. +@code{UPDATE} and @code{DELETE} on a whole unique key in the @code{WHERE} part, +is now faster than before. @item -@code{DECIMAL(x,y)} now works according to ANSI SQL. +Changed @code{RAID_CHUNKSIZE} to be in 1024 byte increments. @item -Added aggregate UDF functions. Thanks to Andreas F. Bobak -@email{bobak@@relog.ch} for this! +Fixed coredump in LOAD_FILE(NULL). +@end itemize + +@node News-3.23.14, News-3.23.13, News-3.23.15, News-3.23.x +@appendixsubsec Changes in release 3.23.14 +@itemize @bullet @item -@code{LAST_INSERT_ID()} is now updated for @code{INSERT INTO ... SELECT}. +Fixed a bug in @code{CONCAT()} where one of the arguments was a function +that returned a modified argument. @item -Some small changes to the join table optimizer to make some joins faster. +Fixed a critical bug in @code{myisamchk}, where it updated the header in +the index file when one only checked the table. This confused the +@code{mysqld} daemon if it updated the same table at the same time. Now +the status in the index file is only updated if one uses +@code{--update-state}. With older @code{myisamchk} versions you should +use @code{--read-only} when only checking tables, if there is the +slightest chance that the @code{mysqld} server is working on the table at the +same time! @item -@code{SELECT DISTINCT} is much faster; It uses the new @code{UNIQUE} -functionality in @code{MyISAM}. One difference compared to @strong{MySQL} Version 3.22 -is that the output of @code{DISTINCT} is not sorted anymore. +Fixed that @code{DROP TABLE} is logged in the update log. @item -All C client API macros are now functions to make shared libraries more -reliable. Because of this, you can no longer call @code{mysql_num_fields()} on -a @code{MYSQL} object, you must use @code{mysql_field_count()} instead. +Fixed problem when searching on @code{DECIMAL()} key field +where the column data contained leading zeros. @item -Added use of @code{LIBEWRAP}; Patch by Henning P . Schmiedehausen. +Fix bug in @code{myisamchk} when the auto_increment isn't the first key. @item -Don't allow @code{AUTO_INCREMENT} for other than numerical columns. +Allow @code{DATETIME} in ISO8601 format: 2000-03-12T12:00:00 @item -Using @code{AUTO_INCREMENT} will now automatically make the column -@code{NOT NULL}. +Dynamic character sets. A @code{mysqld} binary can now handle many different +character sets (you can choose which when starting @code{mysqld}). @item -Show @code{NULL} as the default value for AUTO_INCREMENT columns. +Added command @code{REPAIR TABLE}. @item -Added @code{SQL_BIG_RESULT}; @code{SQL_SMALL_RESULT} is now default. +Added C API function @code{mysql_thread_safe()}. @item -Added a shared library RPM. This enhancement was contributed by David -Fox (dsfox@@cogsci.ucsd.edu). +Added the @code{UMASK_DIR} environment variable. @item -Added a @code{--enable-large-files/--disable-large-files} switch to -@code{configure}. See @file{configure.in} for some systems where this is -automatically turned off because of broken implementations. +Added function @code{CONNECTION_ID()}. @item -Upgraded @code{readline} to 4.0. +When using @code{=} on @code{BLOB} or @code{VARCHAR BINARY} keys, where +only a part of the column was indexed, the whole column of the result +row wasn't compared. @item -New @code{CREATE TABLE} options: @code{PACK_KEYS} and @code{CHECKSUM}. +Fix for sjis character set and @code{ORDER BY}. @item -Added @code{mysqld} option @code{--default-table-type}. +When running in ANSI mode, don't allow columns to be used that aren't in +the @code{GROUP BY} part. @end itemize -@node News-3.22.x, News-3.21.x, News-3.23.x, News -@appendixsec Changes in release 3.22.x (Older; Still supported) - -The 3.22 version has faster and safer connect code than version 3.21, as well -as a lot of new nice enhancements. The reason for not including these changes -As there aren't really any MAJOR changes, upgrading from 3.21 to 3.22 should -be very easy and painless. @xref{Upgrading-from-3.21}. - -@menu -* News-3.22.35:: Changes in release 3.22.35 -* News-3.22.34:: Changes in release 3.22.34 -* News-3.22.33:: Changes in release 3.22.33 -* News-3.22.32:: Changes in release 3.22.32 -* News-3.22.31:: Changes in release 3.22.31 -* News-3.22.30:: Changes in release 3.22.30 -* News-3.22.29:: Changes in release 3.22.29 -* News-3.22.28:: Changes in release 3.22.28 -* News-3.22.27:: Changes in release 3.22.27 -* News-3.22.26:: Changes in release 3.22.26 -* News-3.22.25:: Changes in release 3.22.25 -* News-3.22.24:: Changes in release 3.22.24 -* News-3.22.23:: Changes in release 3.22.23 -* News-3.22.22:: Changes in release 3.22.22 -* News-3.22.21:: Changes in release 3.22.21 -* News-3.22.20:: Changes in release 3.22.20 -* News-3.22.19:: Changes in release 3.22.19 -* News-3.22.18:: Changes in release 3.22.18 -* News-3.22.17:: Changes in release 3.22.17 -* News-3.22.16:: Changes in release 3.22.16 -* News-3.22.15:: Changes in release 3.22.15 -* News-3.22.14:: Changes in release 3.22.14 -* News-3.22.13:: Changes in release 3.22.13 -* News-3.22.12:: Changes in release 3.22.12 -* News-3.22.11:: Changes in release 3.22.11 -* News-3.22.10:: Changes in release 3.22.10 -* News-3.22.9:: Changes in release 3.22.9 -* News-3.22.8:: Changes in release 3.22.8 -* News-3.22.7:: Changes in release 3.22.7 -* News-3.22.6:: Changes in release 3.22.6 -* News-3.22.5:: Changes in release 3.22.5 -* News-3.22.4:: Changes in release 3.22.4 -* News-3.22.3:: Changes in release 3.22.3 -* News-3.22.2:: Changes in release 3.22.2 -* News-3.22.1:: Changes in release 3.22.1 -* News-3.22.0:: Changes in release 3.22.0 -@end menu - -@node News-3.22.35, News-3.22.34, News-3.22.x, News-3.22.x -@appendixsubsec Changes in release 3.22.35 +@node News-3.23.13, News-3.23.12, News-3.23.14, News-3.23.x +@appendixsubsec Changes in release 3.23.13 @itemize @bullet @item -Fixed problem with @code{STD()}. +Fixed problem when doing locks on the same table more than 2 times in +the same @code{LOCK TABLE} command; This fixed the problem one got when running +the test-ATIS test with @code{--fast} or @code{--check-only-changed}. @item -Merged changes from the newest @code{ISAM} library from 3.23. +Added option @code{SQL_BUFFER_RESULT} to @code{SELECT}. @item -Fixed problem with @code{INSERT DELAYED}. +Removed end space from double/float numbers in results from temporary +tables. +Added @code{CHECK TABLE} command. @item -Fixed a bug core dump when using a @code{LEFT JOIN}/@code{STRAIGHT_JOIN} -on a table with only one row. -@end itemize - -@node News-3.22.34, News-3.22.33, News-3.22.35, News-3.22.x -@appendixsubsec Changes in release 3.22.34 -@itemize @bullet +Added changes for MyISAM in 3.23.12 that didn't get into the source +distribution because of CVS problems. @item -Fixed problem with @code{GROUP BY} on @code{TINYBLOB} columns; This -caused bugzilla to not show rows in some queries. +Fixed bug so that @code{mysqladmin shutdown} will wait for the local server +to close down. @item -Had to do total recompile of the Windows binary version as VC++ didn't -compile all relevant files for 3.22.33 :( +Fixed a possible endless loop when calculating timestamp. +@item +Added @code{print_defaults} to the @file{.rpm} files. Removed @code{mysqlbug} +from the client @file{.rpm} file. @end itemize -@node News-3.22.33, News-3.22.32, News-3.22.34, News-3.22.x -@appendixsubsec Changes in release 3.22.33 +@node News-3.23.12, News-3.23.11, News-3.23.13, News-3.23.x +@appendixsubsec Changes in release 3.23.12 @itemize @bullet @item -Fixed problems in Windows when locking tables with @code{LOCK TABLE} +Fixed bug in @code{MyISAM} involving @code{REPLACE ... SELECT ...} which could +give a corrupted table. @item -Quicker kill of @code{SELECT DISTINCT} queries. -@end itemize - -@node News-3.22.32, News-3.22.31, News-3.22.33, News-3.22.x -@appendixsubsec Changes in release 3.22.32 -@itemize @bullet +Fixed bug in @code{myisamchk} where it wrongly reset the auto_increment value. @item -Fixed problem when storing numbers in timestamps. +LOTS of patches for Linux Alpha. @strong{MySQL} now appears to be relatively +stable on Alpha. @item -Fix problem with timezones that have half hour offsets. +Changed @code{DISTINCT} on @code{HEAP} temporary tables to use hashed +keys to quickly find duplicated rows. This mostly concerns queries of +type @code{SELECT DISTINCT ... GROUP BY ...}. This fixes a problem where +not all duplicates were removed in queries of the above type. In +addition, the new code is MUCH faster. @item -@code{mysqlhotcopy} - fast on-line hot-backup utility for local @strong{MySQL} -databases. By Tim Bunce. +Added patches to make @strong{MySQL} compile on Mac OS X. @item -New more secure @code{mysqlaccess}. Thanks to Steve Harvey for this. +Added option @code{IF NOT EXISTS} to @code{CREATE DATABASE}. @item -Fixed security problem in the protocol regarding password checking. +Added options @code{--all-databases} and @code{--databases} to @code{mysqldump} +to allow dumping of many databases at the same time. @item -Fixed problem that affected queries that did arithmetic on @code{GROUP} functions. +Fixed bug in compressed @code{DECIMAL()} index in @code{MyISAM} tables. @item -Fixed a bug in the @code{ISAM} code when deleting rows on tables with -packed indexes. -@end itemize - -@node News-3.22.31, News-3.22.30, News-3.22.32, News-3.22.x -@appendixsubsec Changes in release 3.22.31 - -@itemize @bullet +Fixed bug when storing 0 into a timestamp. @item -A few small fixes for the Windows version. -@end itemize - - -@node News-3.22.30, News-3.22.29, News-3.22.31, News-3.22.x -@appendixsubsec Changes in release 3.22.30 - -@itemize @bullet +When doing @code{mysqladmin shutdown} on a local connection, @code{mysqladmin} now +waits until the pidfile is gone before terminating. @item -Fixed optimizer problem on @code{SELECT} when using many overlapping indexes. +Fixed core dump with some @code{COUNT(DISTINCT ...)} queries. @item -Disabled floating-point exceptions for FreeBSD to fix core dump when -doing @code{SELECT floor(pow(2,63))}. +Fixed that @code{myisamchk} works properly with RAID:ed tables. @item -Added print of default arguments options to all clients. +Fixed problem with @code{LEFT JOIN} and @code{key_field IS NULL}. @item -Fixed critical problem with the @code{WITH GRANT OPTION} option. +Fixed bug in @code{net_clear()} which could give the error @code{Aborted +connection} in the @strong{MySQL} clients. @item -Fixed non-critical Y2K problem when writing short date to log files. +Added options @code{USE INDEX (key_list)} and @code{IGNORE INDEX (key_list)} as +join parameters in @code{SELECT}. +@item +@code{DELETE} and @code{RENAME} should now work on @code{RAID} tables. @end itemize -@node News-3.22.29, News-3.22.28, News-3.22.30, News-3.22.x -@appendixsubsec Changes in release 3.22.29 - +@node News-3.23.11, News-3.23.10, News-3.23.12, News-3.23.x +@appendixsubsec Changes in release 3.23.11 @itemize @bullet @item -Upgraded the configure and include files to match the latest 3.23 version. -This should increase portability and make it easier to build shared libraries. +Allow the @code{ALTER TABLE tbl_name ADD (field_list)} syntax. @item -Added latest patches to mit-pthreads for NetBSD. +Fixed problem with optimizer that could sometimes use wrong keys. @item -Fixed problem with timezones that are < GMT -11. +Fixed that @code{GRANT/REVOKE ALL PRIVILEGES} doesn't affect +@code{GRANT OPTION}. @item -Fixed a bug when deleting packed keys in NISAM. +Removed extra @code{)} from the output of @code{SHOW GRANTS} @item -Fixed problem that could cause @strong{MySQL} to touch freed memory when -doing very complicated @code{GROUP BY} queries. +Fixed problem when storing numbers in timestamps. @item -Fixed core dump if you got a crashed table where an @code{ENUM} field value -was too big. +Fix problem with timezones that have half hour offsets. @item -Added @code{mysqlshutdown.exe} and @code{mysqlwatch.exe} to the Windows -distribution. +Allow the syntax @code{UNIQUE INDEX} in @code{CREATE} statements. @item -Fixed problem when doing @code{ORDER BY} on a reference key. +@code{mysqlhotcopy} - fast on-line hot-backup utility for local +@strong{MySQL} databases. By Tim Bunce. @item -Fixed that @code{INSERT DELAYED} doesn't update timestamps that are given. - -@end itemize - -@node News-3.22.28, News-3.22.27, News-3.22.29, News-3.22.x -@appendixsubsec Changes in release 3.22.28 - -@itemize @bullet +New more secure @code{mysqlaccess}. Thanks to Steve Harvey for this. @item -Fixed problem with @code{LEFT JOIN} and @code{COUNT()} on a column which -was declared @code{NULL} + and it had a @code{DEFAULT} value. +Added options @code{--i-am-a-dummy} and @code{--safe-updates} to @code{mysql}. @item -Fixed core dump problem when using @code{CONCAT()} in a @code{WHERE} clause. +Added variables @code{select_limit} and @code{max_join_size} to @code{mysql}. @item -Fixed problem with @code{AVG()} and @code{STD()} with @code{NULL} values. -@end itemize - -@node News-3.22.27, News-3.22.26, News-3.22.28, News-3.22.x -@appendixsubsec Changes in release 3.22.27 - -@itemize @bullet +Added sql variables: @code{SQL_MAX_JOIN_SIZE} and @code{SQL_SAFE_UPDATES}. @item -Fixed prototype in @file{my_ctype.h} when using other character sets. +Added @code{READ LOCAL} lock that doesn't lock the table for concurrent +inserts. (This is used by @code{mysqldump}). @item -Some configure issues to fix problems with big file system detection. +Changed that @code{LOCK TABLES ... READ} doesn't anymore allow concurrent +inserts. @item -Fixed problem when sorting on big blob columns. +Added option @code{--skip-delay-key-write} to @code{mysqld}. @item -@code{ROUND()} will now work on Windows. +Fixed security problem in the protocol regarding password checking. +@item +@code{_rowid} can now be used as an alias for an integer type unique indexed +column. +@item +Added back blocking of @code{SIGPIPE} when compiling with @code{--thread-safe-clients} +to make things safe for old clients. @end itemize -@node News-3.22.26, News-3.22.25, News-3.22.27, News-3.22.x -@appendixsubsec Changes in release 3.22.26 +@node News-3.23.10, News-3.23.9, News-3.23.11, News-3.23.x +@appendixsubsec Changes in release 3.23.10 +@itemize @bullet +@item +Fixed bug in 3.23.9 where memory wasn't properly freed when doing +@code{LOCK TABLES}. +@end itemize +@node News-3.23.9, News-3.23.8, News-3.23.10, News-3.23.x +@appendixsubsec Changes in release 3.23.9 @itemize @bullet @item -Fixed core dump with empty @code{BLOB/TEXT} column to @code{REVERSE()}. +Fixed problem that affected queries that did arithmetic on group functions. @item -Extended @code{/*! */} with version numbers. +Fixed problem with timestamps and @code{INSERT DELAYED}. @item -Changed @code{SUBSTRING(text FROM pos)} to conform to ANSI SQL. (Before this -construct returned the rightmost 'pos' characters). +Fixed that @code{date_column BETWEEN const_date AND const_date} works. @item -Fixed problem with @code{LOCK TABLES} combined with @code{DELETE FROM table} +Fixed problem when only changing a 0 to @code{NULL} in a table with +@code{BLOB/TEXT} columns. @item -Fixed problem that INSERT ... SELECT didn't use SQL_BIG_TABLES. +Fixed bug in range optimizer when using many key parts and or on the middle +key parts: @code{WHERE K1=1 and K3=2 and (K2=2 and K4=4 or K2=3 and K4=5)} @item -@code{SET SQL_LOW_PRIORITY_UPDATES=#} didn't work. +Added command @code{source} to @code{mysql} to allow reading of batch files +inside the @code{mysql} client. Original patch by Matthew Vanecek. @item -Password wasn't updated correctly if privileges didn't change on: -@code{GRANT ... IDENTIFIED BY} +Fixed critical problem with the @code{WITH GRANT OPTION} option. @item -Fixed range optimizer bug in -@code{SELECT * FROM table_name WHERE key_part1 >= const AND (key_part2 = const OR key_part2 = const)} +Don't give an unnecessary @code{GRANT} error when using tables from many +databases in the same query. @item -Fixed bug in compression key handling in ISAM. -@end itemize - -@node News-3.22.25, News-3.22.24, News-3.22.26, News-3.22.x -@appendixsubsec Changes in release 3.22.25 - -@itemize @bullet +Added VIO wrapper (needed for SSL support ; By Andrei Errapart and +Tõnu Samuel). @item -Fixed some small problems with the installation. -@end itemize - -@node News-3.22.24, News-3.22.23, News-3.22.25, News-3.22.x -@appendixsubsec Changes in release 3.22.24 - -@itemize @bullet +Fixed optimizer problem on @code{SELECT} when using many overlapping indexes. +@strong{MySQL} should now be able to choose keys even better when there +is many keys to choose from. +@item +Changed optimizer to prefer a range key instead of a ref key when the +range key can uses more columns than the ref key (which only can use +columns with =). For example, the following type of queries should now +be faster: @code{SELECT * from key_part_1=const and key_part_2 > const2} +@item +Fixed bug that a change of all @code{VARCHAR} columns to @code{CHAR} columns +didn't change row type from dynamic to fixed. +@item +Disabled floating-point exceptions for FreeBSD to fix core dump when +doing @code{SELECT floor(pow(2,63))}. @item -@code{DATA} is not a reserved word anymore. +Changed @code{mysqld} startup option @code{--delay-key-write} to +@code{--delay-key-write-for-all-tables} @item -Fixed optimizer bug with tables with only one row. +Added @code{read-next-on-key} to @code{HEAP} tables. This should fix all +problems with @code{HEAP} tables when using not @code{UNIQUE} keys. @item -Fixed bug when using @code{LOCK TABLES table_name READ; FLUSH TABLES;} +Added print of default arguments options to all clients. @item -Applied some patches for HP-UX. +Added @code{--log-slow-queries} to @code{mysqld} to log all queries that take a +long time to a separate log file with a time of how long the query took. @item -@code{isamchk} should now work on Windows. +Fixed core dump when doing @code{WHERE key_column=RAND(...)} @item -Changed @file{configure} to not use big file handling on Linux as this -crashes some RedHat 6.0 systems +Fixed optimization bug in @code{SELECT ... LEFT JOIN ... key_column IS NULL}, +when @code{key_column} could contain @code{NULL} values. +@item +Fixed problem with 8-bit characters as separators in @code{LOAD DATA INFILE}. @end itemize -@node News-3.22.23, News-3.22.22, News-3.22.24, News-3.22.x -@appendixsubsec Changes in release 3.22.23 - +@node News-3.23.8, News-3.23.7, News-3.23.9, News-3.23.x +@appendixsubsec Changes in release 3.23.8 @itemize @bullet @item -Upgraded to use Autoconf 2.13, Automake 1.4 and @code{libtool} 1.3.2. -@item -Better support for SCO in @code{configure}. -@item -Added option @code{--defaults-file=###} to option file handling to force use -of only one specific option file. +Fixed problem when handling indexfiles larger than 8G. @item -Extended @code{CREATE} syntax to ignore @strong{MySQL} Version 3.23 keywords. +Added latest patches to mit-pthreads for NetBSD. @item -Fixed deadlock problem when using @code{INSERT DELAYED} on a table locked with -@code{LOCK TABLES}. +Fixed problem with timezones that are < GMT -11. @item -Fixed deadlock problem when using @code{DROP TABLE} on a table that was -locked by another thread. +Fixed a bug when deleting packed keys in @code{NISAM}. @item -Add logging of @code{GRANT/REVOKE} commands in the update log. +Fixed problem with @code{ISAM} when doing some @code{ORDER BY ... DESC} queries. @item -Fixed @code{isamchk} to detect a new error condition. +Fixed bug when doing a join on a text key which didn't cover the whole key. @item -Fixed bug in @code{NATURAL LEFT JOIN}. -@end itemize - -@node News-3.22.22, News-3.22.21, News-3.22.23, News-3.22.x -@appendixsubsec Changes in release 3.22.22 - -@itemize @bullet +Option @code{--delay-key-write} didn't enable delayed key writing. @item -Fixed problem in the C API when you called @code{mysql_close()} directly after -@code{mysql_init()}. +Fixed update of @code{TEXT} column which only involved case changes. @item -Better client error message when you can't open socket. +Fixed that @code{INSERT DELAYED} doesn't update timestamps that are given. @item -Fixed @code{delayed_insert_thread} counting when you couldn't create a new -delayed_insert thread. +Added function @code{YEARWEEK()} and options @code{x}, @code{X}, @code{v} and +@code{V} to @code{DATE_FORMAT()}. @item -Fixed bug in @code{CONCAT()} with many arguments. +Fixed problem with @code{MAX(indexed_column)} and HEAP tables. @item -Added patches for DEC 3.2 and SCO. +Fixed problem with @code{BLOB NULL} keys and @code{LIKE} "prefix%". @item -Fixed path-bug when installing @strong{MySQL} as a service on NT. +Fixed problem with @code{MyISAM} and fixed-length rows < 5 bytes. @item -The @strong{MySQL}-Windows version is now compiled with VC++ 6.0 instead of -with VC++ 5.0. +Fixed problem that could cause @strong{MySQL} to touch freed memory when +doing very complicated @code{GROUP BY} queries. @item -New installation setup for @strong{MySQL}-Windows. +Fixed core dump if you got a crashed table where an @code{ENUM} field value +was too big. @end itemize -@node News-3.22.21, News-3.22.20, News-3.22.22, News-3.22.x -@appendixsubsec Changes in release 3.22.21 - +@node News-3.23.7, News-3.23.6, News-3.23.8, News-3.23.x +@appendixsubsec Changes in release 3.23.7 @itemize @bullet @item -Fixed problem with @code{DELETE FROM TABLE} when table was locked by -another thread. -@item -Fixed bug in @code{LEFT JOIN} involving empty tables. -@item -Changed the @code{mysql.db} column from @code{char(32)} to @code{char(60)}. +Fixed workaround under Linux to avoid problems with @code{pthread_mutex_timedwait}, +which is used with @code{INSERT DELAYED}. @xref{Linux}. @item -@code{MODIFY} and @code{DELAYED} are not reserved words anymore. +Fixed that one will get a 'disk full' error message if one gets disk full +when doing sorting (instead of waiting until we got more disk space). @item -Fixed a bug when storing days in a @code{TIME} column. +Fixed a bug in @code{MyISAM} with keys > 250 characters. @item -Fixed a problem with @code{Host '...' is not allowed to connect to this MySQL -server} after one had inserted a new @strong{MySQL} user with a @code{GRANT} -command. +In @code{MyISAM} one can now do an @code{INSERT} at the same time as other +threads are reading from the table. @item -Changed to use @code{TCP_NODELAY} also on Linux (Should give faster TCP/IP -connections). -@end itemize - -@node News-3.22.20, News-3.22.19, News-3.22.21, News-3.22.x -@appendixsubsec Changes in release 3.22.20 - -@itemize @bullet +Added variable @code{max_write_lock_count} to @code{mysqld} to force a +@code{READ} lock after a certain number of @code{WRITE} locks. @item -Fixed @code{STD()} for big tables when result should be 0. +Inverted flag @code{delayed_key_write} on @code{show variables}. @item -The update log didn't have newlines on some operating systems. +Renamed variable @code{concurrency} to @code{thread_concurrency}. @item -@code{INSERT DELAYED} had some garbage at end in the update log. -@end itemize - -@node News-3.22.19, News-3.22.18, News-3.22.20, News-3.22.x -@appendixsubsec Changes in release 3.22.19 - -@itemize @bullet +The following functions are now multi-byte-safe: +@code{LOCATE(substr,str)}, @code{POSITION(substr IN str)}, +@code{LOCATE(substr,str,pos)}, @code{INSTR(str,substr)}, +@code{LEFT(str,len)}, @code{RIGHT(str,len)}, +@code{SUBSTRING(str,pos,len)}, @code{SUBSTRING(str FROM pos FOR len)}, +@code{MID(str,pos,len)}, @code{SUBSTRING(str,pos)}, @code{SUBSTRING(str +FROM pos)}, @code{SUBSTRING_INDEX(str,delim,count)}, @code{RTRIM(str)}, +@code{TRIM([[BOTH | TRAILING] [remstr] FROM] str)}, +@code{REPLACE(str,from_str,to_str)}, @code{REVERSE(str)}, +@code{INSERT(str,pos,len,newstr)}, @code{LCASE(str)}, @code{LOWER(str)}, +@code{UCASE(str)} and @code{UPPER(str)}; Patch by Wei He. @item -Fixed bug in @code{mysql_install_db} (from 3.22.17). +Fix core dump when releasing a lock from a non-existent table. @item -Changed default key cache size to 8M. +Remove locks on tables before starting to remove duplicates. @item -Fixed problem with queries that needed temporary tables with @code{BLOB} -columns. -@end itemize - -@node News-3.22.18, News-3.22.17, News-3.22.19, News-3.22.x -@appendixsubsec Changes in release 3.22.18 - -@itemize @bullet +Added option @code{FULL} to @code{SHOW PROCESSLIST}. @item -Fixes a fatal problem in 3.22.17 on Linux; After @code{shutdown} -all threads didn't die properly. +Added option @code{--verbose} to @code{mysqladmin}. @item -Added option @code{-O flush_time=#} to @code{mysqld}. This is mostly -useful on Windows and tells how often @strong{MySQL} should close all -unused tables and flush all updated tables to disk. +Fixed problem when automatically converting HEAP to MyISAM. @item -Fixed problem that a @code{VARCHAR} column compared with @code{CHAR} column -didn't use keys efficiently. -@end itemize - -@node News-3.22.17, News-3.22.16, News-3.22.18, News-3.22.x -@appendixsubsec Changes in release 3.22.17 - -@itemize @bullet +Fixed bug in HEAP tables when doing insert + delete + insert + scan the +table. @item -Fixed a core dump problem when using @code{--log-update} and connecting -without a default database. +Fixed bugs on Alpha with @code{REPLACE()} and @code{LOAD DATA INFILE}. @item -Fixed some @code{configure} and portability problems. +Added @code{mysqld} variable @code{interactive_timeout}. @item -Using @code{LEFT JOIN} on tables that had circular dependencies caused -@code{mysqld} to hang forever. +Changed the argument to @code{mysql_data_seek()} from @code{ulong} to +@code{ulonglong}. @end itemize -@node News-3.22.16, News-3.22.15, News-3.22.17, News-3.22.x -@appendixsubsec Changes in release 3.22.16 +@node News-3.23.6, News-3.23.5, News-3.23.7, News-3.23.x +@appendixsubsec Changes in release 3.23.6 @itemize @bullet @item -@code{mysqladmin processlist} could kill the server if a new user logged in. +Added @code{mysqld} option @code{-O lower_case_table_names=@{0|1@}} to allow +users to force table names to lowercase. @item -@code{DELETE FROM tbl_name WHERE key_column=col_name} didn't find any matching -rows. Fixed. +Added @code{SELECT ... INTO DUMPFILE}. @item -@code{DATE_ADD(column, ...)} didn't work. +Added @code{mysqld} option @code{--ansi} to make some functions @code{ANSI SQL} +compatible. @item -@code{INSERT DELAYED} could deadlock with status 'upgrading lock' +Temporary tables now starts with @code{#sql}. @item -Extended @code{ENCRYPT()} to take longer salt strings than 2 characters. +Added quoting of identifiers with @code{`} (@code{"} in @code{--ansi} mode). @item -@code{longlong2str} is now much faster than before. For @code{Intel x86} -platforms, this function is written in optimized assembler. +Changed to use snprintf() when printing floats to avoid some buffer +overflows on FreeBSD. @item -Added the @code{MODIFY} keyword to @code{ALTER TABLE}. -@end itemize - -@node News-3.22.15, News-3.22.14, News-3.22.16, News-3.22.x -@appendixsubsec Changes in release 3.22.15 - -@itemize @bullet +Made @code{[floor()} overflow safe on FreeBSD. @item -@code{GRANT} used with @code{IDENTIFIED BY} didn't take effect until privileges -were flushed. +Added option @code{--quote-names} to @code{mysqldump} @item -Name change of some variables in @code{SHOW STATUS}. +Fixed bug that one could make a part of a @code{PRIMARY KEY NOT NULL}. @item -Fixed problem with @code{ORDER BY} with 'only index' optimization when there -were multiple key definitions for a used column. +Fixed @code{encrypt()} to be thread safe and not reuse buffer. @item -@code{DATE} and @code{DATETIME} columns are now up to 5 times faster than -before. +Added @code{mysql_odbc_escape_string()} function to support big5 characters in +MyODBC. @item -@code{INSERT DELAYED} can be used to let the client do other things while the -server inserts rows into a table. +Rewrote the table handler to use classes. This introduces a lot of new code, +but will make table handling faster and better. @item -@code{LEFT JOIN USING (col1,col2)} didn't work if one used it with tables -from 2 different databases. +Added patch by Sasha for user-defined variables. @item -@code{LOAD DATA LOCAL INFILE} didn't work in the Unix version because of -a missing file. +Changed that @code{FLOAT} and @code{DOUBLE} (without any length modifiers) are +not anymore fixed decimal point numbers. @item -Fixed problems with @code{VARCHAR}/@code{BLOB} on very short rows (< 4 bytes); -error 127 could occur when deleting rows. +Changed the meaning of @code{FLOAT(X)}: Now this is the same as @code{FLOAT} if +X <= 24 and a @code{DOUBLE} if 24 < X <= 53. @item -Updating @code{BLOB/TEXT} through formulas didn't work for short (< 256 char) -strings. +@code{DECIMAL(X)} is now an alias for @code{DECIMAL(X,0)} and @code{DECIMAL} +is now an alias for @code{DECIMAL(10,0)}. The same goes for @code{NUMERIC}. @item -When you did a @code{GRANT} on a new host, @code{mysqld} could die on the first -connect from this host. +Added option @code{ROW_FORMAT=@{default | dynamic | static | compressed@}} to +@code{CREATE_TABLE}. @item -Fixed bug when one used @code{ORDER BY} on column name that was the same -name as an alias. +@code{DELETE FROM table_name} didn't work on temporary tables. @item -Added @code{BENCHMARK(loop_count,expression)} function to time expressions. +Changed function @code{CHAR_LENGTH()} to be multi-byte character safe. +@item +Added function @code{ORD(string)}. @end itemize -@node News-3.22.14, News-3.22.13, News-3.22.15, News-3.22.x -@appendixsubsec Changes in release 3.22.14 +@node News-3.23.5, News-3.23.4, News-3.23.6, News-3.23.x +@appendixsubsec Changes in release 3.23.5 @itemize @bullet @item -Allow empty arguments to @code{mysqld} to make it easier to start from shell -scripts. +Fixed some Y2K problems in the new date handling in 3.23. @item -Setting a @code{TIMESTAMP} column to @code{NULL} didn't record the timestamp -value in the update log. +Fixed problem with @code{SELECT DISTINCT ... ORDER BY RAND()}. @item -Fixed lock handler bug when one did -@code{INSERT INTO TABLE ... SELECT ... GROUP BY}. +Added patches by Sergei A. Golubchik for text searching on the MyISAM level. @item -Added a patch for @code{localtime_r()} on Windows so that it will not crash -anymore if your date is > 2039, but instead will return a time of all zero. +Fixed cache overflow problem when using full joins without keys. @item -Names for -user-defined functions are no longer case sensitive. +Fixed some configure issues. @item -Added escape of @code{^Z} (ASCII 26) to @code{\Z} as @code{^Z} doesn't -work with pipes on Windows. +Some small changes to make parsing faster. @item -@code{mysql_fix_privileges} adds a new column to the @code{mysql.func} to -support aggregate UDF functions in future @strong{MySQL} releases. -@end itemize - -@node News-3.22.13, News-3.22.12, News-3.22.14, News-3.22.x -@appendixsubsec Changes in release 3.22.13 - -@itemize @bullet +@code{ALTER TABLE} + adding a column after the last field didn't work. @item -Saving @code{NOW()}, @code{CURDATE()} or @code{CURTIME()} directly in a -column didn't work. +Fixed problem when using an auto_increment column in two keys @item -@code{SELECT COUNT(*) ... LEFT JOIN ...} didn't work with no @code{WHERE} part. +One can now with MyISAM have the auto_increment part as a sub part: +@code{CREATE TABLE foo (a int not null auto_increment, b char(5), primary key (b,a))} @item -Updated @file{config.guess} to allow @strong{MySQL} to configure on -UnixWare 7.0.x. +Fixed bug in MyISAM with packed char keys that could be @code{NULL}. @item -Changed the implementation of @code{pthread_cond()} on the Windows version. -@code{get_lock()} now correctly times out on Windows! -@end itemize - -@node News-3.22.12, News-3.22.11, News-3.22.13, News-3.22.x -@appendixsubsec Changes in release 3.22.12 - -@itemize @bullet +@code{AS} on fieldname with @code{CREATE TABLE table_name SELECT ...} didn't +work. @item -Fixed problem when using @code{DATE_ADD()} and @code{DATE_SUB()} in a -@code{WHERE} clause. +Allow use of @code{NATIONAL} and @code{NCHAR} when defining character columns. +This is the same as not using @code{BINARY}. @item -You can now set the password for a user with the @code{GRANT ... TO user -IDENTIFIED BY 'password'} syntax. +Don't allow @code{NULL} columns in a @code{PRIMARY KEY} (only in @code{UNIQUE} +keys). @item -Fixed bug in @code{GRANT} checking with @code{SELECT} on many tables. +Clear @code{LAST_INSERT_ID} if one uses this in ODBC: +@code{WHERE auto_increment_column IS NULL}. This seems to fix some problems +with Access. @item -Added missing file @code{mysql_fix_privilege_tables} to the RPM -distribution. This is not run by default because it relies on the client -package. +@code{SET SQL_AUTO_IS_NULL=0|1} now turns on/off the handling of +searching after the last inserted row with @code{WHERE +auto_increment_column IS NULL}. @item -Added option @code{SQL_SMALL_RESULT} to @code{SELECT} to force use of -fast temporary tables when you know that the result set will be small. +Added new @code{mysqld} variable @code{concurrency} for Solaris. @item -Allow use of negative real numbers without a decimal point. +Added option @code{--relative} to @code{mysqladmin} to make +@code{extended-status} more useful to monitor changes. @item -Day number is now adjusted to maximum days in month if the resulting month -after @code{DATE_ADD}/@code{DATE_SUB()} doesn't have enough days. +Fixed bug when using @code{COUNT(DISTINCT ...)} on an empty table. @item -Fix that @code{GRANT} compares columns in case-insensitive fashion. +Added support for the Chinese character set GBK. @item -Fixed a bug in @file{sql_list.h} that made @code{ALTER TABLE} dump core in -some contexts. +Fixed problem with @code{LOAD DATA INFILE} and @code{BLOB} columns. @item -The hostname in @code{user@@hostname} can now include @samp{.} and @samp{-} -without quotes in the context of the @code{GRANT}, @code{REVOKE} and -@code{SET PASSWORD FOR ...} statements. +Added bit operator @code{~} (negation). @item -Fix for @code{isamchk} for tables which need big temporary files. +Fixed problem with @code{UDF} functions. @end itemize -@node News-3.22.11, News-3.22.10, News-3.22.12, News-3.22.x -@appendixsubsec Changes in release 3.22.11 +@node News-3.23.4, News-3.23.3, News-3.23.5, News-3.23.x +@appendixsubsec Changes in release 3.23.4 @itemize @bullet @item -@strong{IMPORTANT}: You must run the @code{mysql_fix_privilege_tables} script -when you upgrade to this version! This is needed because of the new -@code{GRANT} system. If you don't do this, you will get @code{Access -denied} when you try to use @code{ALTER TABLE}, @code{CREATE INDEX} or -@code{DROP INDEX}. -@item -@code{GRANT} to allow/deny users table and column access. -@item -Changed @code{USER()} to return @code{user@@host} -@item -Changed the syntax for how to set @code{PASSWORD} for another user. +Inserting a @code{DATETIME} into a @code{TIME} column will not anymore +try to store 'days' in it. @item -New command @code{FLUSH STATUS} that sets most status variables to zero. +Fixed problem with storage of float/double on little endian machines. +(This affected @code{SUM()}.) @item -New status variables: @code{aborted_threads}, @code{aborted_connects}. +Added connect timeout on TCP/IP connections. @item -New option variable: @code{connection_timeout}. +Fixed problem with @code{LIKE} "%" on an index that may have @code{NULL} values. @item -Added support for Thai sorting (by Pruet Boonma -@email{pruet@@ds90.intanon.nectec.or.th}). +@code{REVOKE ALL PRIVILEGES} didn't revoke all privileges. @item -Slovak and japanese error messages. +Allow creation of temporary tables with same name as the original table. @item -Configuration and portability fixes. +When granting a user a grant option for a database, he couldn't grant +privileges to other users. @item -Added option @code{SET SQL_WARNINGS=1} to get a warning count also for simple -inserts. +New command: @code{SHOW GRANTS FOR user} (by Sinisa). @item -@strong{MySQL} now uses @code{SIGTERM} instead of @code{SIGQUIT} with -shutdown to work better on FreeBSD. +New @code{date_add} syntax: @code{date/datetime + INTERVAL # interval_type}. +By Joshua Chamas. @item -Added option @code{\G} (print vertically) to @code{mysql}. +Fixed privilege check for @code{LOAD DATA REPLACE}. @item -@code{SELECT HIGH_PRIORITY ...} killed @code{mysqld}. +Automatic fixing of broken include files on Solaris 2.7 @item -@code{IS NULL} on a @code{AUTO_INCREMENT} column in a @code{LEFT JOIN} didn't -work as expected. +Some configure issues to fix problems with big file system detection. @item -New function @code{MAKE_SET()}. +@code{REGEXP} is now case insensitive if you use non-binary strings. @end itemize -@node News-3.22.10, News-3.22.9, News-3.22.11, News-3.22.x -@appendixsubsec Changes in release 3.22.10 +@node News-3.23.3, News-3.23.2, News-3.23.4, News-3.23.x +@appendixsubsec Changes in release 3.23.3 @itemize @bullet @item -@code{mysql_install_db} no longer starts the @strong{MySQL} server! You -should start @code{mysqld} with @code{safe_mysqld} after installing it! The -@strong{MySQL} RPM will, however, start the server as before. -@item -Added @code{--bootstrap} option to @code{mysqld} and recoded -@code{mysql_install_db} to use it. This will make it easier to install -@strong{MySQL} with RPMs. +Added patches for MIT-pthreads on NetBSD. @item -Changed @code{+}, @code{-} (sign and minus), @code{*}, @code{/}, @code{%}, -@code{ABS()} and @code{MOD()} to be @code{BIGINT} aware (64-bit safe). +Fixed range bug in MyISAM. @item -Fixed a bug in @code{ALTER TABLE} that caused @code{mysqld} to crash. +@code{ASC} is now the default again for @code{ORDER BY}. @item -@strong{MySQL} now always reports the conflicting key values when a -duplicate key entry occurs. (Before this was only reported for @code{INSERT}). +Added @code{LIMIT} to @code{UPDATE}. @item -New syntax: @code{INSERT INTO tbl_name SET col_name=value, col_name=value, ...} +New client function: @code{mysql_change_user()}. @item -Most errors in the @file{.err} log are now prefixed with a time stamp. +Added character set to @code{SHOW VARIABLES}. @item -Added option @code{MYSQL_INIT_COMMAND} to @code{mysql_options()} to make -a query on connect or reconnect. +Added support of @code{--[whitespace]} comments. @item -Added option @code{MYSQL_READ_DEFAULT_FILE} and -@code{MYSQL_READ_DEFAULT_GROUP} to @code{mysql_options()} to read the -following parameters from the @strong{MySQL} option files: @code{port}, -@code{socket}, @code{compress}, @code{password}, @code{pipe}, @code{timeout}, -@code{user}, @code{init-command}, @code{host} and @code{database}. +Allow @code{INSERT into tbl_name VALUES ()}, that is, you may now specify +an empty value list to insert a row in which each column is set to its +default value. @item -Added @code{maybe_null} to the UDF structure. +Changed @code{SUBSTRING(text FROM pos)} to conform to ANSI SQL. (Before this +construct returned the rightmost 'pos' characters). @item -Added option @code{IGNORE} to @code{INSERT} statements with many rows. +@code{SUM()} with @code{GROUP BY} returned 0 on some systems. @item -Fixed some problems with sorting of the koi8 character sets; Users of koi8 -@strong{MUST} run @code{isamchk -rq} on each table that has an index on -a @code{CHAR} or @code{VARCHAR} column. +Changed output for @code{SHOW TABLE STATUS}. @item -New script @code{mysql_setpermission}, by Luuk de Boer, allows one -to easily create new users with permissions for specific databases. +Added @code{DELAY_KEY_WRITE} option to @code{CREATE TABLE}. @item -Allow use of hexadecimal strings (0x...) when specifying a constant string -(like in the column separators with @code{LOAD DATA INFILE}). +Allow @code{AUTO_INCREMENT} on any key part. @item -Ported to OS/2 (thanks to Antony T. Curtis @email{antony.curtis@@olcs.net}). +Fixed problem with @code{YEAR(NOW())} and @code{YEAR(CURDATE())}. @item -Added more variables to @code{SHOW STATUS} and changed format of output to -be like @code{SHOW VARIABLES}. +Added @code{CASE} construct. @item -Added @code{extended-status} command to @code{mysqladmin} which will show the -new status variables. +New function @code{COALESCE()}. @end itemize -@node News-3.22.9, News-3.22.8, News-3.22.10, News-3.22.x -@appendixsubsec Changes in release 3.22.9 +@node News-3.23.2, News-3.23.1, News-3.23.3, News-3.23.x +@appendixsubsec Changes in release 3.23.2 @itemize @bullet @item -@code{SET SQL_LOG_UPDATE=0} caused a lockup of the server. +Fixed range optimizer bug: @code{SELECT * FROM table_name WHERE +key_part1 >= const AND (key_part2 = const OR key_part2 = const)}. The +bug was that some rows could be duplicated in the result. @item -New SQL command: @code{FLUSH [ TABLES | HOSTS | LOGS | PRIVILEGES ] [, ...]} +Running @code{myisamchk} without @code{-a} updated the index +distribution wrong. @item -New SQL command: @code{KILL} @code{thread_id}. +@code{SET SQL_LOW_PRIORITY_UPDATES=1} gave parse error before. @item -Added casts and changed include files to make @strong{MySQL} easier to -compile on AIX and DEC OSF1 4.x +You can now update indexes columns that are used in the @code{WHERE} clause. +@code{UPDATE tbl_name SET KEY=KEY+1 WHERE KEY > 100} @item -Fixed conversion problem when using @code{ALTER TABLE} from a @code{INT} -to a short @code{CHAR()} column. +Date handling should now be a bit faster. @item -Added @code{SELECT HIGH_PRIORITY}; This will get a lock for the -@code{SELECT} even if there is a thread waiting for another -@code{SELECT} to get a @code{WRITE LOCK}. +Added handling of fuzzy dates (dates where day or month is 0): +(Like: 1999-01-00) @item -Moved wild_compare to string class to be able to use @code{LIKE} on -@code{BLOB}/@code{TEXT} columns with @code{\0}. +Fixed optimization of @code{SELECT ... WHERE key_part1=const1 AND +key_part_2=const2 AND key_part1=const4 AND key_part2=const4} ; Indextype +should be @code{range} instead of @code{ref}. @item -Added @code{ESCAPE} option to @code{LIKE}. +Fixed @code{egcs} 1.1.2 optimizer bug (when using @code{BLOB}s) on Linux Alpha. @item -Added a lot more output to @code{mysqladmin debug}. +Fixed problem with @code{LOCK TABLES} combined with @code{DELETE FROM table}. @item -You can now start @code{mysqld} on Windows with the @code{--flush} option. -This will flush all tables to disk after each update. This makes things -much safer on NT/Win98 but also @strong{MUCH} slower. -@end itemize - -@node News-3.22.8, News-3.22.7, News-3.22.9, News-3.22.x -@appendixsubsec Changes in release 3.22.8 - -@itemize @bullet +MyISAM tables now allow keys on @code{NULL} and @code{BLOB/TEXT} columns. @item -Czech character sets should now work much better. You must also install -@uref{http://www.mysql.com/Downloads/Patches/czech-3.22.8-patch}. -This patch should also be installed if you are using a character set with -uses @code{my_strcoll()}! The patch should always be safe to install (for any system), -but as this patch changes ISAM internals it's not yet in the default -distribution. +The following join is now much faster: +@code{SELECT ... FROM t1 LEFT JOIN t2 ON ... WHERE t2.not_null_column IS NULL}. @item -@code{DATE_ADD()} and @code{DATE_SUB()} didn't work with group functions. +@code{ORDER BY} and @code{GROUP BY} can be done on functions. @item -@code{mysql} will now also try to reconnect on @code{USE DATABASE} commands. +Changed handling of 'const_item' to allow handling of @code{ORDER BY RAND()}. @item -Fix problem with @code{ORDER BY} and @code{LEFT JOIN} and @code{const} tables. +Indexes are now used for @code{WHERE key_column = function}. @item -Fixed problem with @code{ORDER BY} if the first @code{ORDER BY} column -was a key and the rest of the @code{ORDER BY} columns wasn't part of the key. +Indexes are now used for @code{WHERE key_column = column_name} even if +the columns are not identically packed. @item -Fixed a big problem with @code{OPTIMIZE TABLE}. +Indexes are now used for @code{WHERE column_name IS NULL}. @item -@strong{MySQL} clients on NT will now by default first try to connect with -named pipes and after this with TCP/IP. +Changed heap tables to be stored in low_byte_first order (to make it easy +to convert to MyISAM tables) @item -Fixed a problem with @code{DROP TABLE} and @code{mysqladmin shutdown} on Windows -(a fatal bug from 3.22.6). +Automatic change of HEAP temporary tables to MyISAM tables in case of +'table is full' errors. @item -Fixed problems with @code{TIME columns} and negative strings. +Added option @code{--init-file=file_name} to @code{mysqld}. +@item +@code{COUNT(DISTINCT value, [value, ...])} +@item +@code{CREATE TEMPORARY TABLE} now creates a temporary table, in its own +namespace, that is automatically deleted if connection is dropped. +@item +New reserved words (required for @code{CASE}): @code{CASE, THEN, WHEN, ELSE and END}. @item -Added an extra thread signal loop on shutdown to avoid some error messages -from the client. +New functions @code{EXPORT_SET()} and @code{MD5()}. @item -@strong{MySQL} now uses the next available number as extension for the update -log file. +Support for the GB2312 Chinese character set. +@end itemize + +@node News-3.23.1, News-3.23.0, News-3.23.2, News-3.23.x +@appendixsubsec Changes in release 3.23.1 + +@itemize @bullet @item -Added patches for UNIXWARE 7. +Fixed some compilation problems. @end itemize -@node News-3.22.7, News-3.22.6, News-3.22.8, News-3.22.x -@appendixsubsec Changes in release 3.22.7 +@node News-3.23.0, , News-3.23.1, News-3.23.x +@appendixsubsec Changes in release 3.23.0 @itemize @bullet @item -Added @code{LIMIT} clause for the @code{DELETE} statement. +A new table handler library (@code{MyISAM}) with a lot of new features. +@xref{MyISAM}. @item -You can now use the @code{/*! ... */} syntax to hide @strong{MySQL}-specific -keywords when you write portable code. @strong{MySQL} will parse the code -inside the comments as if the surrounding @code{/*!} and @code{*/} comment -characters didn't exist. +You can create in-memory @code{HEAP} tables which are extremely fast for +lookups. @item -@code{OPTIMIZE TABLE tbl_name} can now be used to reclaim disk space -after many deletes. Currently, this uses @code{ALTER TABLE} to -regenerate the table, but in the future it will use an integrated -@code{isamchk} for more speed. +Support for big files (63 bit) on OSes that support big files. @item -Upgraded @code{libtool} to get the configure more portable. +New function @code{LOAD_FILE(filename)} to get the contents of a file as a +string value. @item -Fixed slow @code{UPDATE} and @code{DELETE} operations when using -@code{DATETIME} or @code{DATE} keys. +New operator @code{<=>} which will act as @code{=} but will return TRUE +if both arguments are @code{NULL}. This is useful for comparing changes +between tables. @item -Changed optimizer to make it better at deciding when to do a full join -and when using keys. +Added the ODBC 3.0 @code{EXTRACT(interval FROM datetime)} function. @item -You can now use @code{mysqladmin proc} to display information about your own -threads. Only users with the @strong{Process_priv} privilege can get -information about all threads. +Columns defined as @code{FLOAT(X)} is not rounded on storage and may be +in scientific notation (1.0 E+10) when retrieved. @item -Added handling of formats @code{YYMMDD}, @code{YYYYMMDD}, -@code{YYMMDDHHMMSS} for numbers when using @code{DATETIME} and -@code{TIMESTAMP} types. (Formerly these formats only worked with strings.) +@code{REPLACE} is now faster than before. @item -Added connect option @code{CLIENT_IGNORE_SPACE} to allow use of spaces -after function names and before @samp{(} (Powerbuilder requires this). -This will make all function names reserved words. +Changed @code{LIKE} character comparison to behave as @code{=}; +This means that @code{'e' LIKE '@'e'} (if the line doesn't display +correctly, the latter 'e' means a French 'e' with a dot above) is now true. @item -Added the @code{--log-long-format} option to @code{mysqld} to enable timestamps -and INSERT_ID's in the update log. +@code{SHOW TABLE STATUS} returns a lot of information about the tables. @item -Added @code{--where} option to @code{mysqldump} (patch by Jim Faucette). +Added @code{LIKE} to the @code{SHOW STATUS} command. @item -The lexical analyzer now uses ``perfect hashing'' for faster parsing of SQL -statements. -@end itemize - -@node News-3.22.6, News-3.22.5, News-3.22.7, News-3.22.x -@appendixsubsec Changes in release 3.22.6 - -@itemize @bullet +Added privilege column to @code{SHOW COLUMNS}. @item -Faster @code{mysqldump}. +Added columns @code{packed} and @code{comment} to @code{SHOW INDEX}. @item -For the @code{LOAD DATA INFILE} statement, you can now use the new @code{LOCAL} -keyword to read the file from the client. @code{mysqlimport} will -automatically use @code{LOCAL} when importing with the TCP/IP protocol. +Added comments to tables (with @code{CREATE TABLE ... COMMENT "xxx"}). @item -Fixed small optimize problem when updating keys. +Added @code{UNIQUE}, as in +@code{CREATE TABLE table_name (col int not null UNIQUE)} @item -Changed makefiles to support shared libraries. +New create syntax: @code{CREATE TABLE table_name SELECT ...} @item -@strong{MySQL}-NT can now use named pipes, which means that you can now use -@strong{MySQL}-NT without having to install TCP/IP. -@end itemize - -@node News-3.22.5, News-3.22.4, News-3.22.6, News-3.22.x -@appendixsubsec Changes in release 3.22.5 +New create syntax: @code{CREATE TABLE IF NOT EXISTS ...} +@item +Allow creation of @code{CHAR(0)} columns. +@item +@code{DATE_FORMAT()} now requires @samp{%} before any format character. +@item +@code{DELAYED} is now a reserved word (sorry about that :( ). +@item +An example procedure is added: @code{analyse}, file: @file{sql_analyse.c}. +This will describe the data in your query. Try the following: +@example +SELECT ... FROM ... WHERE ... PROCEDURE ANALYSE([max elements,[max memory]]) +@end example -@itemize @bullet +This procedure is extremely useful when you want to check the data in your +table! @item -All table lock handing is changed to avoid some very subtle -deadlocks when using @code{DROP TABLE}, @code{ALTER TABLE}, @code{DELETE FROM -TABLE} and @code{mysqladmin flush-tables} under heavy usage. -Changed locking code to get better handling of locks of different types. +@code{BINARY} cast to force a string to be compared case sensitively. @item -Updated @code{DBI} to 1.00 and @code{DBD} to 1.2.0. +Added option @code{--skip-show-database} to @code{mysqld}. @item -Added a check that the error message file contains error messages suitable for -the current version of @code{mysqld}. (To avoid errors if you accidentally -try to use an old error message file.) +Check if a row has changed in an @code{UPDATE} now also works with +@code{BLOB}/@code{TEXT} columns. @item -All count structures in the client (@code{affected_rows()}, -@code{insert_id()}, ...) are now of type @code{BIGINT} to allow 64-bit values -to be used. -This required a minor change in the @strong{MySQL} protocol which should affect -only old clients when using tables with @code{AUTO_INCREMENT} values > 16M. +Added the @code{INNER} join syntax. @strong{NOTE}: This made @code{INNER} +a reserved word! @item -The return type of @code{mysql_fetch_lengths()} has changed from @code{uint *} -to @code{ulong *}. This may give a warning for old clients but should work -on most machines. +Added support for netmasks to the hostname in the @strong{MySQL} tables. +You can specify a netmask using the @code{IP/NETMASK} syntax. @item -Change @code{mysys} and @code{dbug} libraries to allocate all thread variables -in one struct. This makes it easier to make a threaded @file{libmysql.dll} -library. +If you compare a @code{NOT NULL DATE/DATETIME} column with @code{IS +NULL}, this is changed to a compare against @code{0} to satisfy some ODBC +applications. (By @email{shreeve@@uci.edu}). @item -Use the result from @code{gethostname()} (instead of @code{uname()}) when -constructing @file{.pid} file names. +@code{NULL IN (...)} now returns @code{NULL} instead of @code{0}. This will +ensure that @code{null_column NOT IN (...)} doesn't match +@code{NULL} values. @item -New better compressed server/client protocol. +Fix storage of floating-point values in @code{TIME} columns. @item -@code{COUNT()}, @code{STD()} and @code{AVG()} are extended to handle more than -4G rows. +Changed parsing of @code{TIME} strings to be more strict. Now the +fractional second part is detected (and currently skipped). The +following formats are supported: +@table @code +@item [[DAYS] [H]H:]MM:]SS[.fraction] +@item [[[[[H]H]H]H]MM]SS[.fraction] +@end table @item -You can now store values in the range @code{-838:59:59} <= x <= -@code{838:59:59} in a @code{TIME} column. +Detect (and ignore) second fraction part from @code{DATETIME}. @item -@strong{WARNING: INCOMPATIBLE CHANGE!!} -If you set a @code{TIME} column to too short a value, @strong{MySQL} now -assumes the value is given as: @code{[[[D ]HH:]MM:]SS} instead of -@code{HH[:MM[:SS]]}. +Added the @code{LOW_PRIORITY} attribute to @code{LOAD DATA INFILE}. @item -@code{TIME_TO_SEC()} and @code{SEC_TO_TIME()} can now handle negative times -and hours up to 32767. +The default index name now uses the same case as the used column name. @item -Added new option -@code{SET OPTION SQL_LOG_UPDATE=@{0|1@}} to allow users with -the @strong{process} privilege to bypass the update log. -(Modified patch from Sergey A Mukhin @email{violet@@rosnet.net}.) +Changed default number of connections to 100. @item -Fixed fatal bug in @code{LPAD()}. +Use bigger buffers when using @code{LOAD DATA INFILE}. @item -Initialize line buffer in @file{mysql.cc} to make @code{BLOB} reading from -pipes safer. +@code{DECIMAL(x,y)} now works according to ANSI SQL. @item -Added @code{-O max_connect_errors=#} option to @code{mysqld}. -Connect errors are now reset for each correct connection. +Added aggregate UDF functions. Thanks to Andreas F. Bobak +@email{bobak@@relog.ch} for this! @item -Increased the default value of @code{max_allowed_packet} to @code{1M} in -@code{mysqld}. +@code{LAST_INSERT_ID()} is now updated for @code{INSERT INTO ... SELECT}. @item -Added @code{--low-priority-updates} option to @code{mysqld}, to give -table-modifying operations (@code{INSERT}, @code{REPLACE}, @code{UPDATE}, -@code{DELETE}) lower priority than retrievals. You can now use -@code{@{INSERT | REPLACE | UPDATE | DELETE@} LOW_PRIORITY ...} You can -also use @code{SET OPTION SQL_LOW_PRIORITY_UPDATES=@{0|1@}} to change -the priority for one thread. One side effect is that @code{LOW_PRIORITY} -is now a reserved word. :( +Some small changes to the join table optimizer to make some joins faster. @item -Add support for @code{INSERT INTO table ... VALUES(...),(...),(...)}, -to allow inserting multiple rows with a single statement. +@code{SELECT DISTINCT} is much faster; It uses the new @code{UNIQUE} +functionality in @code{MyISAM}. One difference compared to @strong{MySQL} Version 3.22 +is that the output of @code{DISTINCT} is not sorted anymore. @item -@code{INSERT INTO tbl_name} is now also cached when used with @code{LOCK TABLES}. -(Previously only @code{INSERT ... SELECT} and @code{LOAD DATA INFILE} were -cached.) +All C client API macros are now functions to make shared libraries more +reliable. Because of this, you can no longer call @code{mysql_num_fields()} on +a @code{MYSQL} object, you must use @code{mysql_field_count()} instead. @item -Allow @code{GROUP BY} functions with @code{HAVING}: -@example -mysql> SELECT col FROM table GROUP BY col HAVING COUNT(*)>0; -@end example +Added use of @code{LIBEWRAP}; Patch by Henning P . Schmiedehausen. @item -@code{mysqld} will now ignore trailing @samp{;} characters in queries. This -is to make it easier to migrate from some other SQL servers that require the -trailing @samp{;}. +Don't allow @code{AUTO_INCREMENT} for other than numerical columns. @item -Fix for corrupted fixed-format output generated by @code{SELECT INTO OUTFILE}. +Using @code{AUTO_INCREMENT} will now automatically make the column +@code{NOT NULL}. @item -@strong{WARNING: INCOMPATIBLE CHANGE!!} -Added Oracle @code{GREATEST()} and @code{LEAST()} functions. You must now use -these instead of the @code{MAX()} and @code{MIN()} functions to get the -largest/smallest value from a list of values. These can now handle @code{REAL}, -@code{BIGINT} and string (@code{CHAR} or @code{VARCHAR}) values. +Show @code{NULL} as the default value for AUTO_INCREMENT columns. @item -@strong{WARNING: INCOMPATIBLE CHANGE!!} -@code{DAYOFWEEK()} had offset 0 for Sunday. Changed the offset to 1. +Added @code{SQL_BIG_RESULT}; @code{SQL_SMALL_RESULT} is now default. @item -Give an error for queries that mix @code{GROUP BY} columns and fields when -there is no @code{GROUP BY} specification. +Added a shared library RPM. This enhancement was contributed by David +Fox (dsfox@@cogsci.ucsd.edu). @item -Added @code{--vertical} option to @code{mysql}, for printing results in -vertical mode. +Added a @code{--enable-large-files/--disable-large-files} switch to +@code{configure}. See @file{configure.in} for some systems where this is +automatically turned off because of broken implementations. @item -Index-only optimization; some queries are now resolved using -only indexes. Until @strong{MySQL} 4.0, this works only for numeric columns. -@xref{MySQL indexes, , @strong{MySQL} indexes}. +Upgraded @code{readline} to 4.0. @item -Lots of new benchmarks. +New @code{CREATE TABLE} options: @code{PACK_KEYS} and @code{CHECKSUM}. @item -A new C API chapter and lots of other improvements in the manual. +Added @code{mysqld} option @code{--default-table-type}. @end itemize -@node News-3.22.4, News-3.22.3, News-3.22.5, News-3.22.x -@appendixsubsec Changes in release 3.22.4 +@node News-3.22.x, News-3.21.x, News-3.23.x, News +@appendixsec Changes in release 3.22.x (Older; Still supported) + +The 3.22 version has faster and safer connect code than version 3.21, as well +as a lot of new nice enhancements. The reason for not including these changes +As there aren't really any MAJOR changes, upgrading from 3.21 to 3.22 should +be very easy and painless. @xref{Upgrading-from-3.21}. + +@menu +* News-3.22.35:: Changes in release 3.22.35 +* News-3.22.34:: Changes in release 3.22.34 +* News-3.22.33:: Changes in release 3.22.33 +* News-3.22.32:: Changes in release 3.22.32 +* News-3.22.31:: Changes in release 3.22.31 +* News-3.22.30:: Changes in release 3.22.30 +* News-3.22.29:: Changes in release 3.22.29 +* News-3.22.28:: Changes in release 3.22.28 +* News-3.22.27:: Changes in release 3.22.27 +* News-3.22.26:: Changes in release 3.22.26 +* News-3.22.25:: Changes in release 3.22.25 +* News-3.22.24:: Changes in release 3.22.24 +* News-3.22.23:: Changes in release 3.22.23 +* News-3.22.22:: Changes in release 3.22.22 +* News-3.22.21:: Changes in release 3.22.21 +* News-3.22.20:: Changes in release 3.22.20 +* News-3.22.19:: Changes in release 3.22.19 +* News-3.22.18:: Changes in release 3.22.18 +* News-3.22.17:: Changes in release 3.22.17 +* News-3.22.16:: Changes in release 3.22.16 +* News-3.22.15:: Changes in release 3.22.15 +* News-3.22.14:: Changes in release 3.22.14 +* News-3.22.13:: Changes in release 3.22.13 +* News-3.22.12:: Changes in release 3.22.12 +* News-3.22.11:: Changes in release 3.22.11 +* News-3.22.10:: Changes in release 3.22.10 +* News-3.22.9:: Changes in release 3.22.9 +* News-3.22.8:: Changes in release 3.22.8 +* News-3.22.7:: Changes in release 3.22.7 +* News-3.22.6:: Changes in release 3.22.6 +* News-3.22.5:: Changes in release 3.22.5 +* News-3.22.4:: Changes in release 3.22.4 +* News-3.22.3:: Changes in release 3.22.3 +* News-3.22.2:: Changes in release 3.22.2 +* News-3.22.1:: Changes in release 3.22.1 +* News-3.22.0:: Changes in release 3.22.0 +@end menu +@node News-3.22.35, News-3.22.34, News-3.22.x, News-3.22.x +@appendixsubsec Changes in release 3.22.35 @itemize @bullet @item -Added @code{--tmpdir} option to @code{mysqld}, for specifying the location -of the temporary file directory. +Fixed problem with @code{STD()}. @item -@strong{MySQL} now automatically changes a query from an ODBC client: -@example -SELECT ... FROM table WHERE auto_increment_column IS NULL -@end example -to: -@example -SELECT ... FROM table WHERE auto_increment_column == LAST_INSERT_ID() -@end example -This allows some ODBC programs (Delphi, Access) to retrieve the newly -inserted row to fetch the @code{AUTO_INCREMENT} id. +Merged changes from the newest @code{ISAM} library from 3.23. @item -@code{DROP TABLE} now waits for all users to free a table before deleting it. +Fixed problem with @code{INSERT DELAYED}. @item -Fixed small memory leak in the new connect protocol. +Fixed a bug core dump when using a @code{LEFT JOIN}/@code{STRAIGHT_JOIN} +on a table with only one row. +@end itemize + +@node News-3.22.34, News-3.22.33, News-3.22.35, News-3.22.x +@appendixsubsec Changes in release 3.22.34 +@itemize @bullet @item -New functions @code{BIN()}, @code{OCT()}, @code{HEX()} and @code{CONV()} for -converting between different number bases. +Fixed problem with @code{GROUP BY} on @code{TINYBLOB} columns; This +caused bugzilla to not show rows in some queries. @item -Added function @code{SUBSTRING()} with 2 arguments. +Had to do total recompile of the Windows binary version as VC++ didn't +compile all relevant files for 3.22.33 :( +@end itemize + +@node News-3.22.33, News-3.22.32, News-3.22.34, News-3.22.x +@appendixsubsec Changes in release 3.22.33 +@itemize @bullet @item -If you created a table with a record length smaller than 5, you couldn't -delete rows from the table. +Fixed problems in Windows when locking tables with @code{LOCK TABLE} @item -Added optimization to remove const reference tables from @code{ORDER BY} and -@code{GROUP BY}. +Quicker kill of @code{SELECT DISTINCT} queries. +@end itemize + +@node News-3.22.32, News-3.22.31, News-3.22.33, News-3.22.x +@appendixsubsec Changes in release 3.22.32 +@itemize @bullet @item -@code{mysqld} now automatically disables system locking on Linux and Windows, -and for systems that use MIT-pthreads. You can force the use of locking -with the @code{--enable-locking} option. +Fixed problem when storing numbers in timestamps. @item -Added @code{--console} option to @code{mysqld}, to force a console window -(for error messages) when using Windows. +Fix problem with timezones that have half hour offsets. @item -Fixed table locks for Windows. +@code{mysqlhotcopy} - fast on-line hot-backup utility for local @strong{MySQL} +databases. By Tim Bunce. @item -Allow @samp{$} in identifiers. +New more secure @code{mysqlaccess}. Thanks to Steve Harvey for this. @item -Changed name of user-specific configuration file from @file{my.cnf} to -@file{.my.cnf} (Unix only). +Fixed security problem in the protocol regarding password checking. @item -Added @code{DATE_ADD()} and @code{DATE_SUB()} functions. +Fixed problem that affected queries that did arithmetic on @code{GROUP} functions. +@item +Fixed a bug in the @code{ISAM} code when deleting rows on tables with +packed indexes. @end itemize -@node News-3.22.3, News-3.22.2, News-3.22.4, News-3.22.x -@appendixsubsec Changes in release 3.22.3 +@node News-3.22.31, News-3.22.30, News-3.22.32, News-3.22.x +@appendixsubsec Changes in release 3.22.31 @itemize @bullet @item -Fixed a lock problem (bug in @strong{MySQL} Version 3.22.1) when closing temporary tables. +A few small fixes for the Windows version. +@end itemize + + +@node News-3.22.30, News-3.22.29, News-3.22.31, News-3.22.x +@appendixsubsec Changes in release 3.22.30 + +@itemize @bullet @item -Added missing @code{mysql_ping()} to the client library. +Fixed optimizer problem on @code{SELECT} when using many overlapping indexes. @item -Added @code{--compress} option to all @strong{MySQL} clients. +Disabled floating-point exceptions for FreeBSD to fix core dump when +doing @code{SELECT floor(pow(2,63))}. @item -Changed @code{byte} to @code{char} in @file{mysql.h} and @file{mysql_com.h}. +Added print of default arguments options to all clients. +@item +Fixed critical problem with the @code{WITH GRANT OPTION} option. +@item +Fixed non-critical Y2K problem when writing short date to log files. @end itemize -@node News-3.22.2, News-3.22.1, News-3.22.3, News-3.22.x -@appendixsubsec Changes in release 3.22.2 +@node News-3.22.29, News-3.22.28, News-3.22.30, News-3.22.x +@appendixsubsec Changes in release 3.22.29 @itemize @bullet @item -Searching on multiple constant keys that matched more than 30% of the rows -didn't always use the best possible key. +Upgraded the configure and include files to match the latest 3.23 version. +This should increase portability and make it easier to build shared libraries. @item -New functions @code{<<}, @code{>>}, @code{RPAD()} and @code{LPAD()}. +Added latest patches to mit-pthreads for NetBSD. @item -You can now save default options (like passwords) in a configuration file -(@file{my.cnf}). +Fixed problem with timezones that are < GMT -11. @item -Lots of small changes to get @code{ORDER BY} to work when no records are found -when using fields that are not in @code{GROUP BY} (@strong{MySQL} extension). +Fixed a bug when deleting packed keys in NISAM. @item -Added @code{--chroot} option to @code{mysqld}, to start @code{mysqld} in -a chroot environment (by Nikki Chumakov @email{nikkic@@cityline.ru}). +Fixed problem that could cause @strong{MySQL} to touch freed memory when +doing very complicated @code{GROUP BY} queries. @item -Trailing spaces are now ignored when comparing case-sensitive strings; -this should fix some problems with ODBC and flag 512! +Fixed core dump if you got a crashed table where an @code{ENUM} field value +was too big. @item -Fixed a core-dump bug in the range optimizer. +Added @code{mysqlshutdown.exe} and @code{mysqlwatch.exe} to the Windows +distribution. @item -Added @code{--one-thread} option to @code{mysqld}, for debugging with -LinuxThreads (or @code{glibc}). (This replaces the @code{-T32} flag) +Fixed problem when doing @code{ORDER BY} on a reference key. @item -Added @code{DROP TABLE IF EXISTS} to prevent an error from occurring if the -table doesn't exist. +Fixed that @code{INSERT DELAYED} doesn't update timestamps that are given. + +@end itemize + +@node News-3.22.28, News-3.22.27, News-3.22.29, News-3.22.x +@appendixsubsec Changes in release 3.22.28 + +@itemize @bullet @item -@code{IF} and @code{EXISTS} are now reserved words (they would have to -be sooner or later). +Fixed problem with @code{LEFT JOIN} and @code{COUNT()} on a column which +was declared @code{NULL} + and it had a @code{DEFAULT} value. @item -Added lots of new options to @code{mysqldump}. +Fixed core dump problem when using @code{CONCAT()} in a @code{WHERE} clause. @item -Server error messages are now in @file{mysqld_error.h}. +Fixed problem with @code{AVG()} and @code{STD()} with @code{NULL} values. +@end itemize + +@node News-3.22.27, News-3.22.26, News-3.22.28, News-3.22.x +@appendixsubsec Changes in release 3.22.27 + +@itemize @bullet @item -The server/client protocol now supports compression. +Fixed prototype in @file{my_ctype.h} when using other character sets. @item -All bug fixes from @strong{MySQL} Version 3.21.32. +Some configure issues to fix problems with big file system detection. +@item +Fixed problem when sorting on big blob columns. +@item +@code{ROUND()} will now work on Windows. @end itemize -@node News-3.22.1, News-3.22.0, News-3.22.2, News-3.22.x -@appendixsubsec Changes in release 3.22.1 +@node News-3.22.26, News-3.22.25, News-3.22.27, News-3.22.x +@appendixsubsec Changes in release 3.22.26 @itemize @bullet @item -Added new C API function @code{mysql_ping()}. -@item -Added new API functions @code{mysql_init()} and @code{mysql_options()}. -You now MUST call @code{mysql_init()} before you call -@code{mysql_real_connect()}. -You don't have to call @code{mysql_init()} if you only use -@code{mysql_connect()}. +Fixed core dump with empty @code{BLOB/TEXT} column to @code{REVERSE()}. @item -Added @code{mysql_options(...,MYSQL_OPT_CONNECT_TIMEOUT,...)} so you can set a -timeout for connecting to a server. +Extended @code{/*! */} with version numbers. @item -Added @code{--timeout} option to @code{mysqladmin}, as a test of -@code{mysql_options()}. +Changed @code{SUBSTRING(text FROM pos)} to conform to ANSI SQL. (Before this +construct returned the rightmost 'pos' characters). @item -Added @code{AFTER column} and @code{FIRST} options to -@code{ALTER TABLE ... ADD columns}. -This makes it possible to add a new column at some specific location -within a row in an existing table. +Fixed problem with @code{LOCK TABLES} combined with @code{DELETE FROM table} @item -@code{WEEK()} now takes an optional argument to allow handling of weeks when -the week starts on Monday (some European countries). By default, -@code{WEEK()} assumes the week starts on Sunday. +Fixed problem that INSERT ... SELECT didn't use SQL_BIG_TABLES. @item -@code{TIME} columns weren't stored properly (bug in @strong{MySQL} Version 3.22.0). +@code{SET SQL_LOW_PRIORITY_UPDATES=#} didn't work. @item -@code{UPDATE} now returns information about how many rows were -matched and updated, and how many ``warnings'' occurred when doing the update. +Password wasn't updated correctly if privileges didn't change on: +@code{GRANT ... IDENTIFIED BY} @item -Fixed incorrect result from @code{FORMAT(-100,2)}. +Fixed range optimizer bug in +@code{SELECT * FROM table_name WHERE key_part1 >= const AND (key_part2 = const OR key_part2 = const)} @item -@code{ENUM} and @code{SET} columns were compared in binary (case-sensitive) -fashion; changed to be case insensitive. +Fixed bug in compression key handling in ISAM. @end itemize -@node News-3.22.0, , News-3.22.1, News-3.22.x -@appendixsubsec Changes in release 3.22.0 +@node News-3.22.25, News-3.22.24, News-3.22.26, News-3.22.x +@appendixsubsec Changes in release 3.22.25 @itemize @bullet @item -New (backward compatible) connect protocol that allows you to -specify the database to use when connecting, to get much faster -connections to a specific database. +Fixed some small problems with the installation. +@end itemize -The @code{mysql_real_connect()} call is changed to: +@node News-3.22.24, News-3.22.23, News-3.22.25, News-3.22.x +@appendixsubsec Changes in release 3.22.24 -@example -mysql_real_connect(MYSQL *mysql, const char *host, const char *user, - const char *passwd, const char *db, uint port, - const char *unix_socket, uint client_flag) -@end example +@itemize @bullet +@item +@code{DATA} is not a reserved word anymore. +@item +Fixed optimizer bug with tables with only one row. +@item +Fixed bug when using @code{LOCK TABLES table_name READ; FLUSH TABLES;} +@item +Applied some patches for HP-UX. +@item +@code{isamchk} should now work on Windows. +@item +Changed @file{configure} to not use big file handling on Linux as this +crashes some RedHat 6.0 systems +@end itemize + +@node News-3.22.23, News-3.22.22, News-3.22.24, News-3.22.x +@appendixsubsec Changes in release 3.22.23 + +@itemize @bullet @item -Each connection is handled by its own thread, rather than by the -master @code{accept()} thread. This fixes permanently the telnet bug -that was a topic on the mail list some time ago. +Upgraded to use Autoconf 2.13, Automake 1.4 and @code{libtool} 1.3.2. @item -All TCP/IP connections are now checked with backward resolution of -the hostname to get better security. @code{mysqld} now has a local hostname -resolver cache so connections should actually be faster than before, -even with this feature. +Better support for SCO in @code{configure}. @item -A site automatically will be blocked from future connections if someone -repeatedly connects with an ``improper header'' (like when one uses telnet). +Added option @code{--defaults-file=###} to option file handling to force use +of only one specific option file. @item -You can now refer to tables in different databases with references of the form -@code{tbl_name@@db_name} or @code{db_name.tbl_name}. This makes it possible to -give a user read access to some tables and write access to others simply by -keeping them in different databases! +Extended @code{CREATE} syntax to ignore @strong{MySQL} Version 3.23 keywords. @item -Added @code{--user} option to @code{mysqld}, to allow it to run -as another Unix user (if it is started as the Unix @code{root} user). +Fixed deadlock problem when using @code{INSERT DELAYED} on a table locked with +@code{LOCK TABLES}. @item -Added caching of users and access rights (for faster access rights checking) +Fixed deadlock problem when using @code{DROP TABLE} on a table that was +locked by another thread. @item -Normal users (not anonymous ones) can change their password with -@code{mysqladmin password 'new_password'}. This uses encrypted passwords -that are not logged in the normal @strong{MySQL} log! +Add logging of @code{GRANT/REVOKE} commands in the update log. @item -All important string functions are now coded in assembler for x86 Linux -machines. This gives a speedup of 10% in many cases. +Fixed @code{isamchk} to detect a new error condition. @item -For tables that have many columns, the column names are now hashed for -much faster column name lookup (this will speed up some benchmark -tests a lot!) +Fixed bug in @code{NATURAL LEFT JOIN}. +@end itemize + +@node News-3.22.22, News-3.22.21, News-3.22.23, News-3.22.x +@appendixsubsec Changes in release 3.22.22 + +@itemize @bullet @item -Some benchmarks are changed to get better individual timing. -(Some loops were so short that a specific test took < 2 seconds. The -loops have been changed to take about 20 seconds to make it easier -to compare different databases. A test that took 1-2 seconds before -now takes 11-24 seconds, which is much better) +Fixed problem in the C API when you called @code{mysql_close()} directly after +@code{mysql_init()}. @item -Re-arranged @code{SELECT} code to handle some very specific queries -involving group functions (like @code{COUNT(*)}) without a @code{GROUP BY} but -with @code{HAVING}. The following now works: -@example -mysql> SELECT count(*) as C FROM table HAVING C > 1; -@end example +Better client error message when you can't open socket. @item -Changed the protocol for field functions to be faster and avoid some -calls to @code{malloc()}. +Fixed @code{delayed_insert_thread} counting when you couldn't create a new +delayed_insert thread. @item -Added @code{-T32} option to @code{mysqld}, for running all queries under the -main thread. This makes it possible to debug @code{mysqld} under Linux with -@code{gdb}! +Fixed bug in @code{CONCAT()} with many arguments. @item -Added optimization of @code{not_null_column IS NULL} (needed for some Access -queries). +Added patches for DEC 3.2 and SCO. @item -Allow @code{STRAIGHT_JOIN} to be used between two tables to force the optimizer -to join them in a specific order. +Fixed path-bug when installing @strong{MySQL} as a service on NT. @item -String functions now return @code{VARCHAR} rather than @code{CHAR} and -the column type is now @code{VARCHAR} for fields saved as @code{VARCHAR}. -This should make the @strong{MyODBC} driver better, but may break some old -@strong{MySQL} clients that don't handle @code{FIELD_TYPE_VARCHAR} the same -way as @code{FIELD_TYPE_CHAR}. +The @strong{MySQL}-Windows version is now compiled with VC++ 6.0 instead of +with VC++ 5.0. @item -@code{CREATE INDEX} and @code{DROP INDEX} are now implemented through -@code{ALTER TABLE}. -@code{CREATE TABLE} is still the recommended (fast) way to create indexes. +New installation setup for @strong{MySQL}-Windows. +@end itemize + +@node News-3.22.21, News-3.22.20, News-3.22.22, News-3.22.x +@appendixsubsec Changes in release 3.22.21 + +@itemize @bullet @item -Added @code{--set-variable} option @code{wait_timeout} to @code{mysqld}. +Fixed problem with @code{DELETE FROM TABLE} when table was locked by +another thread. @item -Added time column to @code{mysqladmin processlist} to show how long a query -has taken or how long a thread has slept. +Fixed bug in @code{LEFT JOIN} involving empty tables. @item -Added lots of new variables to @code{show variables} and some new to -@code{show status}. +Changed the @code{mysql.db} column from @code{char(32)} to @code{char(60)}. @item -Added new type @code{YEAR}. @code{YEAR} is stored in 1 byte with allowable -values of 0, and 1901 to 2155. +@code{MODIFY} and @code{DELAYED} are not reserved words anymore. @item -Added new @code{DATE} type that is stored in 3 bytes rather than 4 bytes. -All new tables are created with the new date type if you don't use the -@code{--old-protocol} option to @code{mysqld}. +Fixed a bug when storing days in a @code{TIME} column. @item -Fixed bug in record caches; for some queries, you could get -@code{Error from table handler: #} on some operating systems. +Fixed a problem with @code{Host '...' is not allowed to connect to this MySQL +server} after one had inserted a new @strong{MySQL} user with a @code{GRANT} +command. @item -Added @code{--enable-assembler} option to @code{configure}, for x86 machines -(tested on Linux + @code{gcc}). This will enable assembler functions for the -most important string functions for more speed! +Changed to use @code{TCP_NODELAY} also on Linux (Should give faster TCP/IP +connections). @end itemize -@node News-3.21.x, News-3.20.x, News-3.22.x, News -@appendixsec Changes in release 3.21.x - -Version 3.21 is quite old now, and should be avoided if possible. This -information is kept here for historical purposes only. +@node News-3.22.20, News-3.22.19, News-3.22.21, News-3.22.x +@appendixsubsec Changes in release 3.22.20 -@menu -* News-3.21.33:: Changes in release 3.21.33 -* News-3.21.32:: Changes in release 3.21.32 -* News-3.21.31:: Changes in release 3.21.31 -* News-3.21.30:: Changes in release 3.21.30 -* News-3.21.29:: Changes in release 3.21.29 -* News-3.21.28:: Changes in release 3.21.28 -* News-3.21.27:: Changes in release 3.21.27 -* News-3.21.26:: Changes in release 3.21.26 -* News-3.21.25:: Changes in release 3.21.25 -* News-3.21.24:: Changes in release 3.21.24 -* News-3.21.23:: Changes in release 3.21.23 -* News-3.21.22:: Changes in release 3.21.22 -* News-3.21.21a:: Changes in release 3.21.21a -* News-3.21.21:: Changes in release 3.21.21 -* News-3.21.20:: Changes in release 3.21.20 -* News-3.21.19:: Changes in release 3.21.19 -* News-3.21.18:: Changes in release 3.21.18 -* News-3.21.17:: Changes in release 3.21.17 -* News-3.21.16:: Changes in release 3.21.16 -* News-3.21.15:: Changes in release 3.21.15 -* News-3.21.14b:: Changes in release 3.21.14b -* News-3.21.14a:: Changes in release 3.21.14a -* News-3.21.13:: Changes in release 3.21.13 -* News-3.21.12:: Changes in release 3.21.12 -* News-3.21.11:: Changes in release 3.21.11 -* News-3.21.10:: Changes in release 3.21.10 -* News-3.21.9:: Changes in release 3.21.9 -* News-3.21.8:: Changes in release 3.21.8 -* News-3.21.7:: Changes in release 3.21.7 -* News-3.21.6:: Changes in release 3.21.6 -* News-3.21.5:: Changes in release 3.21.5 -* News-3.21.4:: Changes in release 3.21.4 -* News-3.21.3:: Changes in release 3.21.3 -* News-3.21.2:: Changes in release 3.21.2 -* News-3.21.0:: Changes in release 3.21.0 -@end menu +@itemize @bullet +@item +Fixed @code{STD()} for big tables when result should be 0. +@item +The update log didn't have newlines on some operating systems. +@item +@code{INSERT DELAYED} had some garbage at end in the update log. +@end itemize -@node News-3.21.33, News-3.21.32, News-3.21.x, News-3.21.x -@appendixsubsec Changes in release 3.21.33 +@node News-3.22.19, News-3.22.18, News-3.22.20, News-3.22.x +@appendixsubsec Changes in release 3.22.19 @itemize @bullet @item -Fixed problem when sending @code{SIGHUP} to @code{mysqld}; -@code{mysqld} core dumped when starting from boot on some systems. -@item -Fixed problem with losing a little memory for some connections. +Fixed bug in @code{mysql_install_db} (from 3.22.17). @item -@code{DELETE FROM tbl_name} without a @code{WHERE} condition is now done the -long way when you use @code{LOCK TABLES} or if the table is in use, to -avoid race conditions. +Changed default key cache size to 8M. @item -@code{INSERT INTO TABLE (timestamp_column) VALUES (NULL);} didn't set timestamp. +Fixed problem with queries that needed temporary tables with @code{BLOB} +columns. @end itemize -@node News-3.21.32, News-3.21.31, News-3.21.33, News-3.21.x -@appendixsubsec Changes in release 3.21.32 +@node News-3.22.18, News-3.22.17, News-3.22.19, News-3.22.x +@appendixsubsec Changes in release 3.22.18 @itemize @bullet @item -Fixed some possible race conditions when doing many reopen/close on the same -tables under heavy load! This can happen if you execute @code{mysqladmin -refresh} often. This could in some very rare cases corrupt the header of the -index file and cause error 126 or 138. - +Fixes a fatal problem in 3.22.17 on Linux; After @code{shutdown} +all threads didn't die properly. @item -Fixed fatal bug in @code{refresh()} when running with the -@code{--skip-locking} option. There was a ``very small'' time gap after -a @code{mysqladmin refresh} when a table could be corrupted if one -thread updated a table while another thread did @code{mysqladmin -refresh} and another thread started a new update ont the same table -before the first thread had finished. A refresh (or -@code{--flush-tables}) will now not return until all used tables are -closed! +Added option @code{-O flush_time=#} to @code{mysqld}. This is mostly +useful on Windows and tells how often @strong{MySQL} should close all +unused tables and flush all updated tables to disk. @item -@code{SELECT DISTINCT} with a @code{WHERE} clause that didn't match any rows -returned a row in some contexts (bug only in 3.21.31). +Fixed problem that a @code{VARCHAR} column compared with @code{CHAR} column +didn't use keys efficiently. +@end itemize + +@node News-3.22.17, News-3.22.16, News-3.22.18, News-3.22.x +@appendixsubsec Changes in release 3.22.17 + +@itemize @bullet @item -@code{GROUP BY} + @code{ORDER BY} returned one empty row when no rows where -found. +Fixed a core dump problem when using @code{--log-update} and connecting +without a default database. @item -Fixed a bug in the range optimizer that wrote -@code{Use_count: Wrong count for ...} in the error log file. +Fixed some @code{configure} and portability problems. +@item +Using @code{LEFT JOIN} on tables that had circular dependencies caused +@code{mysqld} to hang forever. @end itemize -@node News-3.21.31, News-3.21.30, News-3.21.32, News-3.21.x -@appendixsubsec Changes in release 3.21.31 +@node News-3.22.16, News-3.22.15, News-3.22.17, News-3.22.x +@appendixsubsec Changes in release 3.22.16 + @itemize @bullet @item -Fixed a sign extension problem for the @code{TINYINT} type on Irix. +@code{mysqladmin processlist} could kill the server if a new user logged in. @item -Fixed problem with @code{LEFT("constant_string",function)}. +@code{DELETE FROM tbl_name WHERE key_column=col_name} didn't find any matching +rows. Fixed. @item -Fixed problem with @code{FIND_IN_SET()}. +@code{DATE_ADD(column, ...)} didn't work. @item -@code{LEFT JOIN} core dumped if the second table is used with a constant -@code{WHERE/ON} expression that uniquely identifies one record. +@code{INSERT DELAYED} could deadlock with status 'upgrading lock' @item -Fixed problems with @code{DATE_FORMAT()} and incorrect dates. -@code{DATE_FORMAT()} now ignores @code{'%'} to make it possible to extend -it more easily in the future. +Extended @code{ENCRYPT()} to take longer salt strings than 2 characters. +@item +@code{longlong2str} is now much faster than before. For @code{Intel x86} +platforms, this function is written in optimized assembler. +@item +Added the @code{MODIFY} keyword to @code{ALTER TABLE}. @end itemize -@node News-3.21.30, News-3.21.29, News-3.21.31, News-3.21.x -@appendixsubsec Changes in release 3.21.30 +@node News-3.22.15, News-3.22.14, News-3.22.16, News-3.22.x +@appendixsubsec Changes in release 3.22.15 + @itemize @bullet @item -@code{mysql} now returns an exit code > 0 if the query returned an error. -@item -Saving of command line history to file in @code{mysql} client. -By Tommy Larsen @email{tommy@@mix.hive.no}. -@item -Fixed problem with empty lines that were ignored in @file{mysql.cc}. -@item -Save the pid of the signal handler thread in the pid file instead -of the pid of the main thread. +@code{GRANT} used with @code{IDENTIFIED BY} didn't take effect until privileges +were flushed. @item -Added patch by @email{tommy@@valley.ne.jp} to support Japanese characters -SJIS and UJIS. +Name change of some variables in @code{SHOW STATUS}. @item -Changed @code{safe_mysqld} to redirect startup messages to -@code{'hostname'.err} instead -of @code{'hostname'.log} to reclaim file space on @code{mysqladmin refresh}. +Fixed problem with @code{ORDER BY} with 'only index' optimization when there +were multiple key definitions for a used column. @item -@code{ENUM} always had the first entry as default value. +@code{DATE} and @code{DATETIME} columns are now up to 5 times faster than +before. @item -@code{ALTER TABLE} wrote two entries to the update log. +@code{INSERT DELAYED} can be used to let the client do other things while the +server inserts rows into a table. @item -@code{sql_acc()} now closes the @code{mysql} grant tables after a reload to -save table space and memory. +@code{LEFT JOIN USING (col1,col2)} didn't work if one used it with tables +from 2 different databases. @item -Changed @code{LOAD DATA} to use less memory with tables and @code{BLOB} -columns. +@code{LOAD DATA LOCAL INFILE} didn't work in the Unix version because of +a missing file. @item -Sorting on a function which made a division / 0 produced a wrong set -in some cases. +Fixed problems with @code{VARCHAR}/@code{BLOB} on very short rows (< 4 bytes); +error 127 could occur when deleting rows. @item -Fixed @code{SELECT} problem with @code{LEFT()} when using the czech character -set. +Updating @code{BLOB/TEXT} through formulas didn't work for short (< 256 char) +strings. @item -Fixed problem in @code{isamchk}; it couldn't repair a packed table in a very -unusual case. +When you did a @code{GRANT} on a new host, @code{mysqld} could die on the first +connect from this host. @item -@code{SELECT} statements with @code{&} or @code{|} (bit functions) failed on -columns with @code{NULL} values. +Fixed bug when one used @code{ORDER BY} on column name that was the same +name as an alias. @item -When comparing a field = field, where one of the fields was a part key, -only the length of the part key was compared. +Added @code{BENCHMARK(loop_count,expression)} function to time expressions. @end itemize -@node News-3.21.29, News-3.21.28, News-3.21.30, News-3.21.x -@appendixsubsec Changes in release 3.21.29 +@node News-3.22.14, News-3.22.13, News-3.22.15, News-3.22.x +@appendixsubsec Changes in release 3.22.14 + @itemize @bullet @item -@code{LOCK TABLES} + @code{DELETE from tbl_name} never removed locks properly. -@item -Fixed problem when grouping on an @code{OR} function. +Allow empty arguments to @code{mysqld} to make it easier to start from shell +scripts. @item -Fixed permission problem with @code{umask()} and creating new databases. +Setting a @code{TIMESTAMP} column to @code{NULL} didn't record the timestamp +value in the update log. @item -Fixed permission problem on result file with @code{SELECT ... INTO OUTFILE ...} +Fixed lock handler bug when one did +@code{INSERT INTO TABLE ... SELECT ... GROUP BY}. @item -Fixed problem in range optimizer (core dump) for a very complex query. +Added a patch for @code{localtime_r()} on Windows so that it will not crash +anymore if your date is > 2039, but instead will return a time of all zero. @item -Fixed problem when using @code{MIN(integer)} or @code{MAX(integer)} in -@code{GROUP BY}. +Names for +user-defined functions are no longer case sensitive. @item -Fixed bug on Alpha when using integer keys. (Other keys worked on Alpha). +Added escape of @code{^Z} (ASCII 26) to @code{\Z} as @code{^Z} doesn't +work with pipes on Windows. @item -Fixed bug in @code{WEEK("XXXX-xx-01")}. +@code{mysql_fix_privileges} adds a new column to the @code{mysql.func} to +support aggregate UDF functions in future @strong{MySQL} releases. @end itemize -@node News-3.21.28, News-3.21.27, News-3.21.29, News-3.21.x -@appendixsubsec Changes in release 3.21.28 +@node News-3.22.13, News-3.22.12, News-3.22.14, News-3.22.x +@appendixsubsec Changes in release 3.22.13 + @itemize @bullet @item -Fixed socket permission (clients couldn't connect to Unix socket on Linux). +Saving @code{NOW()}, @code{CURDATE()} or @code{CURTIME()} directly in a +column didn't work. @item -Fixed bug in record caches; for some queries, you could get -@code{Error from table handler: #} on some operating systems. +@code{SELECT COUNT(*) ... LEFT JOIN ...} didn't work with no @code{WHERE} part. +@item +Updated @file{config.guess} to allow @strong{MySQL} to configure on +UnixWare 7.0.x. +@item +Changed the implementation of @code{pthread_cond()} on the Windows version. +@code{get_lock()} now correctly times out on Windows! @end itemize -@node News-3.21.27, News-3.21.26, News-3.21.28, News-3.21.x -@appendixsubsec Changes in release 3.21.27 +@node News-3.22.12, News-3.22.11, News-3.22.13, News-3.22.x +@appendixsubsec Changes in release 3.22.12 + @itemize @bullet @item -Added user level lock functions @code{GET_LOCK(string,timeout)}, -@code{RELEASE_LOCK(string)}. +Fixed problem when using @code{DATE_ADD()} and @code{DATE_SUB()} in a +@code{WHERE} clause. @item -Added @code{opened_tables} to @code{show status}. +You can now set the password for a user with the @code{GRANT ... TO user +IDENTIFIED BY 'password'} syntax. @item -Changed connect timeout to 3 seconds to make it somewhat harder -for crackers to kill @code{mysqld} through telnet + TCP/IP. +Fixed bug in @code{GRANT} checking with @code{SELECT} on many tables. @item -Fixed bug in range optimizer when using -@code{WHERE key_part_1 >= something AND key_part_2 <= something_else}. +Added missing file @code{mysql_fix_privilege_tables} to the RPM +distribution. This is not run by default because it relies on the client +package. @item -Changed @code{configure} for detection of FreeBSD 3.0 9803xx and above +Added option @code{SQL_SMALL_RESULT} to @code{SELECT} to force use of +fast temporary tables when you know that the result set will be small. @item -@code{WHERE} with string_column_key = constant_string didn't always find -all rows if the column had many values differing only with characters of -the same sort value (like e and @'e). +Allow use of negative real numbers without a decimal point. @item -Strings keys looked up with 'ref' were not compared in case-sensitive fashion. +Day number is now adjusted to maximum days in month if the resulting month +after @code{DATE_ADD}/@code{DATE_SUB()} doesn't have enough days. @item -Added @code{umask()} to make log files non-readable for normal users. +Fix that @code{GRANT} compares columns in case-insensitive fashion. @item -Ignore users with old (8-byte) password on startup if not using -@code{--old-protocol} option to @code{mysqld}. +Fixed a bug in @file{sql_list.h} that made @code{ALTER TABLE} dump core in +some contexts. @item -@code{SELECT} which matched all key fields returned the values in the -case of the matched values, not of the found values. (Minor problem.) +The hostname in @code{user@@hostname} can now include @samp{.} and @samp{-} +without quotes in the context of the @code{GRANT}, @code{REVOKE} and +@code{SET PASSWORD FOR ...} statements. +@item +Fix for @code{isamchk} for tables which need big temporary files. @end itemize -@node News-3.21.26, News-3.21.25, News-3.21.27, News-3.21.x -@appendixsubsec Changes in release 3.21.26 +@node News-3.22.11, News-3.22.10, News-3.22.12, News-3.22.x +@appendixsubsec Changes in release 3.22.11 + @itemize @bullet @item -@code{FROM_DAYS(0)} now returns "0000-00-00". +@strong{IMPORTANT}: You must run the @code{mysql_fix_privilege_tables} script +when you upgrade to this version! This is needed because of the new +@code{GRANT} system. If you don't do this, you will get @code{Access +denied} when you try to use @code{ALTER TABLE}, @code{CREATE INDEX} or +@code{DROP INDEX}. @item -In @code{DATE_FORMAT()}, PM and AM were swapped for hours 00 and 12. +@code{GRANT} to allow/deny users table and column access. @item -Extended the default maximum key size to 256. +Changed @code{USER()} to return @code{user@@host} @item -Fixed bug when using @code{BLOB}/@code{TEXT} in @code{GROUP BY} with many -tables. +Changed the syntax for how to set @code{PASSWORD} for another user. @item -An @code{ENUM} field that is not declared @code{NOT NULL} has @code{NULL} as -the default value. -(Previously, the default value was the first enumeration value.) +New command @code{FLUSH STATUS} that sets most status variables to zero. @item -Fixed bug in the join optimizer code when using many part keys -on the same key: @code{INDEX (Organization,Surname(35),Initials(35))}. +New status variables: @code{aborted_threads}, @code{aborted_connects}. @item -Added some tests to the table order optimizer to get some cases with -@code{SELECT ... FROM many_tables} much faster. +New option variable: @code{connection_timeout}. @item -Added a retry loop around @code{accept()} to possibly fix some problems on some -Linux machines. +Added support for Thai sorting (by Pruet Boonma +@email{pruet@@ds90.intanon.nectec.or.th}). +@item +Slovak and japanese error messages. +@item +Configuration and portability fixes. +@item +Added option @code{SET SQL_WARNINGS=1} to get a warning count also for simple +inserts. +@item +@strong{MySQL} now uses @code{SIGTERM} instead of @code{SIGQUIT} with +shutdown to work better on FreeBSD. +@item +Added option @code{\G} (print vertically) to @code{mysql}. +@item +@code{SELECT HIGH_PRIORITY ...} killed @code{mysqld}. +@item +@code{IS NULL} on a @code{AUTO_INCREMENT} column in a @code{LEFT JOIN} didn't +work as expected. +@item +New function @code{MAKE_SET()}. @end itemize -@node News-3.21.25, News-3.21.24, News-3.21.26, News-3.21.x -@appendixsubsec Changes in release 3.21.25 +@node News-3.22.10, News-3.22.9, News-3.22.11, News-3.22.x +@appendixsubsec Changes in release 3.22.10 + @itemize @bullet @item -Changed @code{typedef 'string'} to @code{typedef 'my_string'} for better -portability. +@code{mysql_install_db} no longer starts the @strong{MySQL} server! You +should start @code{mysqld} with @code{safe_mysqld} after installing it! The +@strong{MySQL} RPM will, however, start the server as before. @item -You can now kill threads that are waiting on a disk-full condition. +Added @code{--bootstrap} option to @code{mysqld} and recoded +@code{mysql_install_db} to use it. This will make it easier to install +@strong{MySQL} with RPMs. @item -Fixed some problems with UDF functions. +Changed @code{+}, @code{-} (sign and minus), @code{*}, @code{/}, @code{%}, +@code{ABS()} and @code{MOD()} to be @code{BIGINT} aware (64-bit safe). @item -Added long options to @code{isamchk}. Try @code{isamchk --help}. +Fixed a bug in @code{ALTER TABLE} that caused @code{mysqld} to crash. @item -Fixed a bug when using 8 bytes long (alpha); @code{filesort()} didn't work. -Affects @code{DISTINCT}, @code{ORDER BY} and @code{GROUP BY} on 64-bit -processors. +@strong{MySQL} now always reports the conflicting key values when a +duplicate key entry occurs. (Before this was only reported for @code{INSERT}). +@item +New syntax: @code{INSERT INTO tbl_name SET col_name=value, col_name=value, ...} +@item +Most errors in the @file{.err} log are now prefixed with a time stamp. +@item +Added option @code{MYSQL_INIT_COMMAND} to @code{mysql_options()} to make +a query on connect or reconnect. +@item +Added option @code{MYSQL_READ_DEFAULT_FILE} and +@code{MYSQL_READ_DEFAULT_GROUP} to @code{mysql_options()} to read the +following parameters from the @strong{MySQL} option files: @code{port}, +@code{socket}, @code{compress}, @code{password}, @code{pipe}, @code{timeout}, +@code{user}, @code{init-command}, @code{host} and @code{database}. +@item +Added @code{maybe_null} to the UDF structure. +@item +Added option @code{IGNORE} to @code{INSERT} statements with many rows. +@item +Fixed some problems with sorting of the koi8 character sets; Users of koi8 +@strong{MUST} run @code{isamchk -rq} on each table that has an index on +a @code{CHAR} or @code{VARCHAR} column. +@item +New script @code{mysql_setpermission}, by Luuk de Boer, allows one +to easily create new users with permissions for specific databases. +@item +Allow use of hexadecimal strings (0x...) when specifying a constant string +(like in the column separators with @code{LOAD DATA INFILE}). +@item +Ported to OS/2 (thanks to Antony T. Curtis @email{antony.curtis@@olcs.net}). +@item +Added more variables to @code{SHOW STATUS} and changed format of output to +be like @code{SHOW VARIABLES}. +@item +Added @code{extended-status} command to @code{mysqladmin} which will show the +new status variables. @end itemize -@node News-3.21.24, News-3.21.23, News-3.21.25, News-3.21.x -@appendixsubsec Changes in release 3.21.24 +@node News-3.22.9, News-3.22.8, News-3.22.10, News-3.22.x +@appendixsubsec Changes in release 3.22.9 + @itemize @bullet @item -Dynamic loadable functions. Based on source from Alexis Mikhailov. +@code{SET SQL_LOG_UPDATE=0} caused a lockup of the server. @item -You couldn't delete from a table if no one had done a @code{SELECT} on the -table. +New SQL command: @code{FLUSH [ TABLES | HOSTS | LOGS | PRIVILEGES ] [, ...]} @item -Fixed problem with range optimizer with many @code{OR} operators on key parts -inside each other. +New SQL command: @code{KILL} @code{thread_id}. @item -Recoded @code{MIN()} and @code{MAX()} to work properly with strings and -@code{HAVING}. +Added casts and changed include files to make @strong{MySQL} easier to +compile on AIX and DEC OSF1 4.x @item -Changed default umask value for new files from @code{0664} to @code{0660}. +Fixed conversion problem when using @code{ALTER TABLE} from a @code{INT} +to a short @code{CHAR()} column. @item -Fixed problem with @code{LEFT JOIN} and constant expressions in the @code{ON} -part. +Added @code{SELECT HIGH_PRIORITY}; This will get a lock for the +@code{SELECT} even if there is a thread waiting for another +@code{SELECT} to get a @code{WRITE LOCK}. @item -Added Italian error messages from @email{brenno@@dewinter.com}. +Moved wild_compare to string class to be able to use @code{LIKE} on +@code{BLOB}/@code{TEXT} columns with @code{\0}. @item -@code{configure} now works better on OSF1 (tested on 4.0D). +Added @code{ESCAPE} option to @code{LIKE}. @item -Added hooks to allow @code{LIKE} optimization with international character -support. +Added a lot more output to @code{mysqladmin debug}. @item -Upgraded @code{DBI} to 0.93. +You can now start @code{mysqld} on Windows with the @code{--flush} option. +This will flush all tables to disk after each update. This makes things +much safer on NT/Win98 but also @strong{MUCH} slower. @end itemize -@node News-3.21.23, News-3.21.22, News-3.21.24, News-3.21.x -@appendixsubsec Changes in release 3.21.23 +@node News-3.22.8, News-3.22.7, News-3.22.9, News-3.22.x +@appendixsubsec Changes in release 3.22.8 + @itemize @bullet @item -The following symbols are now reserved words: -@code{TIME}, @code{DATE}, @code{TIMESTAMP}, @code{TEXT}, @code{BIT}, -@code{ENUM}, @code{NO}, @code{ACTION}, @code{CHECK}, @code{YEAR}, -@code{MONTH}, @code{DAY}, @code{HOUR}, @code{MINUTE}, @code{SECOND}, -@code{STATUS}, @code{VARIABLES}. -@item -Setting a @code{TIMESTAMP} to @code{NULL} in @code{LOAD DATA INFILE ...} didn't -set the current time for the @code{TIMESTAMP}. +Czech character sets should now work much better. You must also install +@uref{http://www.mysql.com/Downloads/Patches/czech-3.22.8-patch}. +This patch should also be installed if you are using a character set with +uses @code{my_strcoll()}! The patch should always be safe to install (for any system), +but as this patch changes ISAM internals it's not yet in the default +distribution. @item -Fix @code{BETWEEN} to recognize binary strings. Now @code{BETWEEN} is -case sensitive. +@code{DATE_ADD()} and @code{DATE_SUB()} didn't work with group functions. @item -Added @code{--skip-thread-priority} option to @code{mysqld}, for systems -where @code{mysqld}'s thread scheduling doesn't work properly (BSDI 3.1). +@code{mysql} will now also try to reconnect on @code{USE DATABASE} commands. @item -Added ODBC functions @code{DAYNAME()} and @code{MONTHNAME()}. +Fix problem with @code{ORDER BY} and @code{LEFT JOIN} and @code{const} tables. @item -Added function @code{TIME_FORMAT()}. This works like @code{DATE_FORMAT()}, -but takes a time string (@code{'HH:MM:DD'}) as argument. +Fixed problem with @code{ORDER BY} if the first @code{ORDER BY} column +was a key and the rest of the @code{ORDER BY} columns wasn't part of the key. @item -Fixed unlikely(?) key optimizer bug when using @code{OR}s of key parts -inside @code{AND}s. +Fixed a big problem with @code{OPTIMIZE TABLE}. @item -Added command @code{variables} to @code{mysqladmin}. +@strong{MySQL} clients on NT will now by default first try to connect with +named pipes and after this with TCP/IP. @item -A lot of small changes to the binary releases. +Fixed a problem with @code{DROP TABLE} and @code{mysqladmin shutdown} on Windows +(a fatal bug from 3.22.6). @item -Fixed a bug in the new protocol from @strong{MySQL} Version 3.21.20. +Fixed problems with @code{TIME columns} and negative strings. @item -Changed @code{ALTER TABLE} to work with Windows (Windows can't rename -open files). Also fixed a couple of small bugs in the Windows version. +Added an extra thread signal loop on shutdown to avoid some error messages +from the client. @item -All standard @strong{MySQL} clients are now ported to @strong{MySQL}-Windows. +@strong{MySQL} now uses the next available number as extension for the update +log file. @item -@strong{MySQL} can now be started as a service on NT. +Added patches for UNIXWARE 7. @end itemize -@node News-3.21.22, News-3.21.21a, News-3.21.23, News-3.21.x -@appendixsubsec Changes in release 3.21.22 +@node News-3.22.7, News-3.22.6, News-3.22.8, News-3.22.x +@appendixsubsec Changes in release 3.22.7 + @itemize @bullet @item -Starting with this version, all @strong{MySQL} distributions will be -configured, compiled and tested with @code{crash-me} and the benchmarks on -the following platforms: SunOS 5.6 sun4u, SunOS 5.5.1 sun4u, SunOS 4.14 sun4c, -SunOS 5.6 i86pc, Irix 6.3 mips5k, HP-UX 10.20 hppa, AIX 4.2.1 ppc, -OSF1 V4.0 alpha, FreeBSD 2.2.2 i86pc and BSDI 3.1 i386. -@item -Fix @code{COUNT(*)} problems when the @code{WHERE} clause didn't match any -records. (Bug from 3.21.17.) -@item -Removed that @code{NULL = NULL} is true. Now you must use @code{IS NULL} -or @code{IS NOT NULL} to test whether or not a value is @code{NULL}. -(This is according to ANSI SQL but may break -old applications that are ported from @code{mSQL}.) -You can get the old behavior by compiling with @code{-DmSQL_COMPLIANT}. -@item -Fixed bug that core dumped when using many @code{LEFT OUTER JOIN} clauses. +Added @code{LIMIT} clause for the @code{DELETE} statement. @item -Fixed bug in @code{ORDER BY} on string formula with possible @code{NULL} values. +You can now use the @code{/*! ... */} syntax to hide @strong{MySQL}-specific +keywords when you write portable code. @strong{MySQL} will parse the code +inside the comments as if the surrounding @code{/*!} and @code{*/} comment +characters didn't exist. @item -Fixed problem in range optimizer when using <= on sub index. +@code{OPTIMIZE TABLE tbl_name} can now be used to reclaim disk space +after many deletes. Currently, this uses @code{ALTER TABLE} to +regenerate the table, but in the future it will use an integrated +@code{isamchk} for more speed. @item -Added functions @code{DAYOFYEAR()}, @code{DAYOFMONTH()}, @code{MONTH()}, -@code{YEAR()}, @code{WEEK()}, @code{QUARTER()}, @code{HOUR()}, @code{MINUTE()}, -@code{SECOND()} and @code{FIND_IN_SET()}. +Upgraded @code{libtool} to get the configure more portable. @item -Added command @code{SHOW VARIABLES}. +Fixed slow @code{UPDATE} and @code{DELETE} operations when using +@code{DATETIME} or @code{DATE} keys. @item -Added support of ``long constant strings'' from ANSI SQL: -@example -mysql> SELECT 'first ' 'second'; -> 'first second' -@end example +Changed optimizer to make it better at deciding when to do a full join +and when using keys. @item -Upgraded mSQL-Mysql-modules to 1.1825. +You can now use @code{mysqladmin proc} to display information about your own +threads. Only users with the @strong{Process_priv} privilege can get +information about all threads. @item -Upgraded @code{mysqlaccess} to 2.02. +Added handling of formats @code{YYMMDD}, @code{YYYYMMDD}, +@code{YYMMDDHHMMSS} for numbers when using @code{DATETIME} and +@code{TIMESTAMP} types. (Formerly these formats only worked with strings.) @item -Fixed problem with Russian character set and @code{LIKE}. +Added connect option @code{CLIENT_IGNORE_SPACE} to allow use of spaces +after function names and before @samp{(} (Powerbuilder requires this). +This will make all function names reserved words. @item -Ported to OpenBSD 2.1. +Added the @code{--log-long-format} option to @code{mysqld} to enable timestamps +and INSERT_ID's in the update log. @item -New Dutch error messages. -@end itemize - -@node News-3.21.21a, News-3.21.21, News-3.21.22, News-3.21.x -@appendixsubsec Changes in release 3.21.21a -@itemize @bullet +Added @code{--where} option to @code{mysqldump} (patch by Jim Faucette). @item -Configure changes for some operating systems. +The lexical analyzer now uses ``perfect hashing'' for faster parsing of SQL +statements. @end itemize -@node News-3.21.21, News-3.21.20, News-3.21.21a, News-3.21.x -@appendixsubsec Changes in release 3.21.21 -@itemize @bullet -@item -Fixed optimizer bug when using -@code{WHERE data_field = date_field2 AND date_field2 = constant}. -@item -Added command @code{SHOW STATUS}. -@item -Removed @file{manual.ps} from the source distribution to make it smaller. -@end itemize +@node News-3.22.6, News-3.22.5, News-3.22.7, News-3.22.x +@appendixsubsec Changes in release 3.22.6 -@node News-3.21.20, News-3.21.19, News-3.21.21, News-3.21.x -@appendixsubsec Changes in release 3.21.20 @itemize @bullet @item -Changed the maximum table name and column name lengths from 32 to 64. -@item -Aliases can now be of ``any'' length. +Faster @code{mysqldump}. @item -Fixed @code{mysqladmin stat} to return the right number of queries. +For the @code{LOAD DATA INFILE} statement, you can now use the new @code{LOCAL} +keyword to read the file from the client. @code{mysqlimport} will +automatically use @code{LOCAL} when importing with the TCP/IP protocol. @item -Changed protocol (downward compatible) to mark if a column has the -@code{AUTO_INCREMENT} attribute or is a @code{TIMESTAMP}. This is needed for -the new Java driver. +Fixed small optimize problem when updating keys. @item -Added Hebrew sorting order by Zeev Suraski. +Changed makefiles to support shared libraries. @item -Solaris 2.6: Fixed @code{configure} bugs and increased maximum table size -from 2G to 4G. +@strong{MySQL}-NT can now use named pipes, which means that you can now use +@strong{MySQL}-NT without having to install TCP/IP. @end itemize -@node News-3.21.19, News-3.21.18, News-3.21.20, News-3.21.x -@appendixsubsec Changes in release 3.21.19 +@node News-3.22.5, News-3.22.4, News-3.22.6, News-3.22.x +@appendixsubsec Changes in release 3.22.5 + @itemize @bullet @item -Upgraded @code{DBD} to 1823. This version implements @code{mysql_use_result} in -@code{DBD-Mysql}. +All table lock handing is changed to avoid some very subtle +deadlocks when using @code{DROP TABLE}, @code{ALTER TABLE}, @code{DELETE FROM +TABLE} and @code{mysqladmin flush-tables} under heavy usage. +Changed locking code to get better handling of locks of different types. @item -Benchmarks updated for empress (by Luuk). +Updated @code{DBI} to 1.00 and @code{DBD} to 1.2.0. @item -Fixed a case of slow range searching. +Added a check that the error message file contains error messages suitable for +the current version of @code{mysqld}. (To avoid errors if you accidentally +try to use an old error message file.) @item -Configure fixes (@file{Docs} directory). +All count structures in the client (@code{affected_rows()}, +@code{insert_id()}, ...) are now of type @code{BIGINT} to allow 64-bit values +to be used. +This required a minor change in the @strong{MySQL} protocol which should affect +only old clients when using tables with @code{AUTO_INCREMENT} values > 16M. @item -Added function @code{REVERSE()} (by Zeev Suraski). -@end itemize - -@node News-3.21.18, News-3.21.17, News-3.21.19, News-3.21.x -@appendixsubsec Changes in release 3.21.18 -@itemize @bullet +The return type of @code{mysql_fetch_lengths()} has changed from @code{uint *} +to @code{ulong *}. This may give a warning for old clients but should work +on most machines. @item -Issue error message if client C functions are called in wrong order. +Change @code{mysys} and @code{dbug} libraries to allocate all thread variables +in one struct. This makes it easier to make a threaded @file{libmysql.dll} +library. @item -Added automatic reconnect to the @file{libmysql.c} library. If a write command -fails, an automatic reconnect is done. +Use the result from @code{gethostname()} (instead of @code{uname()}) when +constructing @file{.pid} file names. @item -Small sort sets no longer use temporary files. +New better compressed server/client protocol. @item -Upgraded @code{DBI} to 0.91. +@code{COUNT()}, @code{STD()} and @code{AVG()} are extended to handle more than +4G rows. @item -Fixed a couple of problems with @code{LEFT OUTER JOIN}. +You can now store values in the range @code{-838:59:59} <= x <= +@code{838:59:59} in a @code{TIME} column. @item -Added @code{CROSS JOIN} syntax. @code{CROSS} is now a reserved word. +@strong{WARNING: INCOMPATIBLE CHANGE!!} +If you set a @code{TIME} column to too short a value, @strong{MySQL} now +assumes the value is given as: @code{[[[D ]HH:]MM:]SS} instead of +@code{HH[:MM[:SS]]}. @item -Recoded @code{yacc}/@code{bison} stack allocation to be even safer and to allow -@strong{MySQL} to handle even bigger expressions. +@code{TIME_TO_SEC()} and @code{SEC_TO_TIME()} can now handle negative times +and hours up to 32767. @item -Fixed a couple of problems with the update log. +Added new option +@code{SET OPTION SQL_LOG_UPDATE=@{0|1@}} to allow users with +the @strong{process} privilege to bypass the update log. +(Modified patch from Sergey A Mukhin @email{violet@@rosnet.net}.) @item -@code{ORDER BY} was slow when used with key ranges. -@end itemize - -@node News-3.21.17, News-3.21.16, News-3.21.18, News-3.21.x -@appendixsubsec Changes in release 3.21.17 -@itemize @bullet +Fixed fatal bug in @code{LPAD()}. @item -Changed documentation string of @code{--with-unix-socket-path} to avoid -confusion. +Initialize line buffer in @file{mysql.cc} to make @code{BLOB} reading from +pipes safer. @item -Added ODBC and ANSI SQL style @code{LEFT OUTER JOIN}. +Added @code{-O max_connect_errors=#} option to @code{mysqld}. +Connect errors are now reset for each correct connection. @item -The following are new reserved words: @code{LEFT}, @code{NATURAL}, -@code{USING}. +Increased the default value of @code{max_allowed_packet} to @code{1M} in +@code{mysqld}. @item -The client library now uses the value of the environment variable -@code{MYSQL_HOST} as the default host if it's defined. +Added @code{--low-priority-updates} option to @code{mysqld}, to give +table-modifying operations (@code{INSERT}, @code{REPLACE}, @code{UPDATE}, +@code{DELETE}) lower priority than retrievals. You can now use +@code{@{INSERT | REPLACE | UPDATE | DELETE@} LOW_PRIORITY ...} You can +also use @code{SET OPTION SQL_LOW_PRIORITY_UPDATES=@{0|1@}} to change +the priority for one thread. One side effect is that @code{LOW_PRIORITY} +is now a reserved word. :( @item -@code{SELECT col_name, SUM(expr)} now returns @code{NULL} for @code{col_name} -when there are matching rows. +Add support for @code{INSERT INTO table ... VALUES(...),(...),(...)}, +to allow inserting multiple rows with a single statement. @item -Fixed problem with comparing binary strings and @code{BLOB}s with ASCII -characters over 127. +@code{INSERT INTO tbl_name} is now also cached when used with @code{LOCK TABLES}. +(Previously only @code{INSERT ... SELECT} and @code{LOAD DATA INFILE} were +cached.) @item -Fixed lock problem: when freeing a read lock on a table with multiple -read locks, a thread waiting for a write lock would have been given the lock. -This shouldn't affect data integrity, but could possibly make @code{mysqld} -restart if one thread was reading data that another thread modified. +Allow @code{GROUP BY} functions with @code{HAVING}: +@example +mysql> SELECT col FROM table GROUP BY col HAVING COUNT(*)>0; +@end example @item -@code{LIMIT offset,count} didn't work in @code{INSERT ... SELECT}. +@code{mysqld} will now ignore trailing @samp{;} characters in queries. This +is to make it easier to migrate from some other SQL servers that require the +trailing @samp{;}. @item -Optimized key block caching. This will be quicker than the old algorithm when -using bigger key caches. -@end itemize - -@node News-3.21.16, News-3.21.15, News-3.21.17, News-3.21.x -@appendixsubsec Changes in release 3.21.16 - -@itemize @bullet +Fix for corrupted fixed-format output generated by @code{SELECT INTO OUTFILE}. @item -Added ODBC 2.0 & 3.0 functions @code{POWER()}, @code{SPACE()}, -@code{COT()}, @code{DEGREES()}, @code{RADIANS()}, @code{ROUND(2 arg)} -and @code{TRUNCATE()}. +@strong{WARNING: INCOMPATIBLE CHANGE!!} +Added Oracle @code{GREATEST()} and @code{LEAST()} functions. You must now use +these instead of the @code{MAX()} and @code{MIN()} functions to get the +largest/smallest value from a list of values. These can now handle @code{REAL}, +@code{BIGINT} and string (@code{CHAR} or @code{VARCHAR}) values. @item -@strong{WARNING: INCOMPATIBLE CHANGE!!} @code{LOCATE()} parameters were -swapped according to ODBC standard. Fixed. +@strong{WARNING: INCOMPATIBLE CHANGE!!} +@code{DAYOFWEEK()} had offset 0 for Sunday. Changed the offset to 1. @item -Added function @code{TIME_TO_SEC()}. +Give an error for queries that mix @code{GROUP BY} columns and fields when +there is no @code{GROUP BY} specification. @item -In some cases, default values were not used for @code{NOT NULL} fields. +Added @code{--vertical} option to @code{mysql}, for printing results in +vertical mode. @item -Timestamp wasn't always updated properly in @code{UPDATE SET ...} statements. +Index-only optimization; some queries are now resolved using +only indexes. Until @strong{MySQL} 4.0, this works only for numeric columns. +@xref{MySQL indexes, , @strong{MySQL} indexes}. @item -Allow empty strings as default values for @code{BLOB} and @code{TEXT}, to -be compatible with @code{mysqldump}. +Lots of new benchmarks. +@item +A new C API chapter and lots of other improvements in the manual. @end itemize -@node News-3.21.15, News-3.21.14b, News-3.21.16, News-3.21.x -@appendixsubsec Changes in release 3.21.15 +@node News-3.22.4, News-3.22.3, News-3.22.5, News-3.22.x +@appendixsubsec Changes in release 3.22.4 @itemize @bullet @item -@strong{WARNING: INCOMPATIBLE CHANGE!!} @code{mysqlperl} is now from -Msql-Mysql-modules. This means that @code{connect()} now takes -@code{host}, @code{database}, @code{user}, @code{password} arguments! The old -version took @code{host}, @code{database}, @code{password}, @code{user}. -@item -Allow @code{DATE '1997-01-01'}, @code{TIME '12:10:10'} and -@code{TIMESTAMP '1997-01-01 12:10:10'} formats required by ANSI SQL. -@strong{WARNING: INCOMPATIBLE CHANGE!!} This has the unfortunate -side-effect that you no longer can have columns named @code{DATE}, @code{TIME} -or @code{TIMESTAMP}. :( Old columns can still be accessed through -@code{tablename.columnname}!) -@item -Changed Makefiles to hopefully work better with BSD systems. Also, -@file{manual.dvi} is now included in the distribution to avoid having stupid -@code{make} programs trying to rebuild it. -@item -@code{readline} library upgraded to version 2.1. +Added @code{--tmpdir} option to @code{mysqld}, for specifying the location +of the temporary file directory. @item -A new sortorder german-1. That is a normal ISO-Latin1 with a german sort -order. +@strong{MySQL} now automatically changes a query from an ODBC client: +@example +SELECT ... FROM table WHERE auto_increment_column IS NULL +@end example +to: +@example +SELECT ... FROM table WHERE auto_increment_column == LAST_INSERT_ID() +@end example +This allows some ODBC programs (Delphi, Access) to retrieve the newly +inserted row to fetch the @code{AUTO_INCREMENT} id. @item -Perl @code{DBI}/@code{DBD} is now included in the distribution. @code{DBI} -is now the recommended way to connect to @strong{MySQL} from Perl. +@code{DROP TABLE} now waits for all users to free a table before deleting it. @item -New portable benchmark suite with @code{DBD}, with test results from -@code{mSQL} 2.0.3, @strong{MySQL}, PostgreSQL 6.2.1 and Solid server 2.2. +Fixed small memory leak in the new connect protocol. @item -@code{crash-me} is now included with the benchmarks; This is a Perl program -designed to find as many limits as possible in a SQL server. Tested with -@code{mSQL}, PostgreSQL, Solid and @strong{MySQL}. +New functions @code{BIN()}, @code{OCT()}, @code{HEX()} and @code{CONV()} for +converting between different number bases. @item -Fixed bug in range-optimizer that crashed @strong{MySQL} on some queries. +Added function @code{SUBSTRING()} with 2 arguments. @item -Table and column name completion for @code{mysql} command line tool, by Zeev -Suraski and Andi Gutmans. +If you created a table with a record length smaller than 5, you couldn't +delete rows from the table. @item -Added new command @code{REPLACE} that works like @code{INSERT} but -replaces conflicting records with the new record. @code{REPLACE INTO -TABLE ... SELECT ...} works also. +Added optimization to remove const reference tables from @code{ORDER BY} and +@code{GROUP BY}. @item -Added new commands @code{CREATE DATABASE db_name} and @code{DROP -DATABASE db_name}. +@code{mysqld} now automatically disables system locking on Linux and Windows, +and for systems that use MIT-pthreads. You can force the use of locking +with the @code{--enable-locking} option. @item -Added @code{RENAME} option to @code{ALTER TABLE}: @code{ALTER TABLE name -RENAME TO new_name}. +Added @code{--console} option to @code{mysqld}, to force a console window +(for error messages) when using Windows. @item -@code{make_binary_distribution} now includes @file{libgcc.a} in -@file{libmysqlclient.a}. This should make linking work for people who don't -have @code{gcc}. +Fixed table locks for Windows. @item -Changed @code{net_write()} to @code{my_net_write()} because of a name -conflict with Sybase. +Allow @samp{$} in identifiers. @item -@cindex ODBC compatibility -@cindex Compatibility, with ODBC -New function @code{DAYOFWEEK()} compatible with ODBC. +Changed name of user-specific configuration file from @file{my.cnf} to +@file{.my.cnf} (Unix only). @item -Stack checking and @code{bison} memory overrun checking to make @strong{MySQL} -safer with weird queries. +Added @code{DATE_ADD()} and @code{DATE_SUB()} functions. @end itemize -@node News-3.21.14b, News-3.21.14a, News-3.21.15, News-3.21.x -@appendixsubsec Changes in release 3.21.14b +@node News-3.22.3, News-3.22.2, News-3.22.4, News-3.22.x +@appendixsubsec Changes in release 3.22.3 @itemize @bullet @item -Fixed a couple of small @code{configure} problems on some platforms. +Fixed a lock problem (bug in @strong{MySQL} Version 3.22.1) when closing temporary tables. +@item +Added missing @code{mysql_ping()} to the client library. +@item +Added @code{--compress} option to all @strong{MySQL} clients. +@item +Changed @code{byte} to @code{char} in @file{mysql.h} and @file{mysql_com.h}. @end itemize -@node News-3.21.14a, News-3.21.13, News-3.21.14b, News-3.21.x -@appendixsubsec Changes in release 3.21.14a +@node News-3.22.2, News-3.22.1, News-3.22.3, News-3.22.x +@appendixsubsec Changes in release 3.22.2 @itemize @bullet @item -Ported to SCO Openserver 5.0.4 with FSU Pthreads. -@item -HP-UX 10.20 should work. -@item -Added new function @code{DATE_FORMAT()}. +Searching on multiple constant keys that matched more than 30% of the rows +didn't always use the best possible key. @item -Added @code{NOT IN}. +New functions @code{<<}, @code{>>}, @code{RPAD()} and @code{LPAD()}. @item -Added automatic removal of 'ODBC function conversions': @code{@{fn now() @}} +You can now save default options (like passwords) in a configuration file +(@file{my.cnf}). @item -Handle ODBC 2.50.3 option flags. +Lots of small changes to get @code{ORDER BY} to work when no records are found +when using fields that are not in @code{GROUP BY} (@strong{MySQL} extension). @item -Fixed comparison of @code{DATE} and @code{TIME} values with @code{NULL}. +Added @code{--chroot} option to @code{mysqld}, to start @code{mysqld} in +a chroot environment (by Nikki Chumakov @email{nikkic@@cityline.ru}). @item -Changed language name from germany to german to be consistent with -the other language names. +Trailing spaces are now ignored when comparing case-sensitive strings; +this should fix some problems with ODBC and flag 512! @item -Fixed sorting problem on functions returning a @code{FLOAT}. Previously, the -values were converted to @code{INT}s before sorting. +Fixed a core-dump bug in the range optimizer. @item -Fixed slow sorting when sorting on key field when using -@code{key_column=constant}. +Added @code{--one-thread} option to @code{mysqld}, for debugging with +LinuxThreads (or @code{glibc}). (This replaces the @code{-T32} flag) @item -Sorting on calculated @code{DOUBLE} values sorted on integer results instead. +Added @code{DROP TABLE IF EXISTS} to prevent an error from occurring if the +table doesn't exist. @item -@code{mysql} no longer needs a database argument. +@code{IF} and @code{EXISTS} are now reserved words (they would have to +be sooner or later). @item -Changed the place where @code{HAVING} should be. According to ANSI, it should -be after @code{GROUP BY} but before @code{ORDER BY}. @strong{MySQL} Version 3.20 -incorrectly had it last. +Added lots of new options to @code{mysqldump}. @item -Added Sybase command @code{USE DATABASE} to start using another database. +Server error messages are now in @file{mysqld_error.h}. @item -Added automatic adjusting of number of connections and table cache size -if the maximum number of files that can be opened is less than needed. -This should fix that @code{mysqld} doesn't crash even if you haven't done a -@code{ulimit -n 256} before starting @code{mysqld}. +The server/client protocol now supports compression. @item -Added lots of limit checks to make it safer when running with too little -memory or when doing weird queries. +All bug fixes from @strong{MySQL} Version 3.21.32. @end itemize -@node News-3.21.13, News-3.21.12, News-3.21.14a, News-3.21.x -@appendixsubsec Changes in release 3.21.13 +@node News-3.22.1, News-3.22.0, News-3.22.2, News-3.22.x +@appendixsubsec Changes in release 3.22.1 @itemize @bullet @item -Added retry of interrupted reads and clearing of @code{errno}. -This makes Linux systems much safer! +Added new C API function @code{mysql_ping()}. @item -Fixed locking bug when using many aliases on the same table in the same -@code{SELECT}. +Added new API functions @code{mysql_init()} and @code{mysql_options()}. +You now MUST call @code{mysql_init()} before you call +@code{mysql_real_connect()}. +You don't have to call @code{mysql_init()} if you only use +@code{mysql_connect()}. @item -Fixed bug with @code{LIKE} on number key. +Added @code{mysql_options(...,MYSQL_OPT_CONNECT_TIMEOUT,...)} so you can set a +timeout for connecting to a server. @item -New error message so you can check whether the connection was lost while -the command was running or whether the connection was down from the start. +Added @code{--timeout} option to @code{mysqladmin}, as a test of +@code{mysql_options()}. @item -Added @code{--table} option to @code{mysql} to print in table format. -Moved time and row information after query result. -Added automatic reconnect of lost connections. +Added @code{AFTER column} and @code{FIRST} options to +@code{ALTER TABLE ... ADD columns}. +This makes it possible to add a new column at some specific location +within a row in an existing table. +@item +@code{WEEK()} now takes an optional argument to allow handling of weeks when +the week starts on Monday (some European countries). By default, +@code{WEEK()} assumes the week starts on Sunday. +@item +@code{TIME} columns weren't stored properly (bug in @strong{MySQL} Version 3.22.0). @item -Added @code{!=} as a synonym for @code{<>}. +@code{UPDATE} now returns information about how many rows were +matched and updated, and how many ``warnings'' occurred when doing the update. @item -Added function @code{VERSION()} to make easier logs. +Fixed incorrect result from @code{FORMAT(-100,2)}. @item -New multi-user test @file{tests/fork_test.pl} to put some strain on the -thread library. +@code{ENUM} and @code{SET} columns were compared in binary (case-sensitive) +fashion; changed to be case insensitive. @end itemize -@node News-3.21.12, News-3.21.11, News-3.21.13, News-3.21.x -@appendixsubsec Changes in release 3.21.12 +@node News-3.22.0, , News-3.22.1, News-3.22.x +@appendixsubsec Changes in release 3.22.0 @itemize @bullet @item -Fixed @code{ftruncate()} call in MIT-pthreads. This made @code{isamchk} -destroy the @file{.ISM} files on (Free)BSD 2.x systems. -@item -Fixed broken @code{__P_} patch in MIT-pthreads. -@item -Many memory overrun checks. All string functions now return @code{NULL} -if the returned string should be longer than @code{max_allowed_packet} bytes. -@item -Changed the name of the @code{INTERVAL} type to @code{ENUM}, because -@code{INTERVAL} is used in ANSI SQL. -@item -In some cases, doing a @code{JOIN} + @code{GROUP} + @code{INTO OUTFILE}, -the result wasn't grouped. +New (backward compatible) connect protocol that allows you to +specify the database to use when connecting, to get much faster +connections to a specific database. + +The @code{mysql_real_connect()} call is changed to: + +@example +mysql_real_connect(MYSQL *mysql, const char *host, const char *user, + const char *passwd, const char *db, uint port, + const char *unix_socket, uint client_flag) +@end example @item -@code{LIKE} with @code{'_'} as last character didn't work. Fixed. +Each connection is handled by its own thread, rather than by the +master @code{accept()} thread. This fixes permanently the telnet bug +that was a topic on the mail list some time ago. @item -Added extended ANSI SQL @code{TRIM()} function. +All TCP/IP connections are now checked with backward resolution of +the hostname to get better security. @code{mysqld} now has a local hostname +resolver cache so connections should actually be faster than before, +even with this feature. @item -Added @code{CURTIME()}. +A site automatically will be blocked from future connections if someone +repeatedly connects with an ``improper header'' (like when one uses telnet). @item -Added @code{ENCRYPT()} function by Zeev Suraski. +You can now refer to tables in different databases with references of the form +@code{tbl_name@@db_name} or @code{db_name.tbl_name}. This makes it possible to +give a user read access to some tables and write access to others simply by +keeping them in different databases! @item -Fixed better @code{FOREIGN KEY} syntax skipping. New reserved words: -@code{MATCH}, @code{FULL}, @code{PARTIAL}. +Added @code{--user} option to @code{mysqld}, to allow it to run +as another Unix user (if it is started as the Unix @code{root} user). @item -@code{mysqld} now allows IP number and hostname to the @code{--bind-address} -option. +Added caching of users and access rights (for faster access rights checking) @item -Added @code{SET OPTION CHARACTER SET cp1251_koi8} to enable conversions of -data to/from cp1251_koi8. +Normal users (not anonymous ones) can change their password with +@code{mysqladmin password 'new_password'}. This uses encrypted passwords +that are not logged in the normal @strong{MySQL} log! @item -Lots of changes for Win95 port. In theory, this version should now be -easily portable to Win95. +All important string functions are now coded in assembler for x86 Linux +machines. This gives a speedup of 10% in many cases. @item -Changed the @code{CREATE COLUMN} syntax of @code{NOT NULL} columns to be after -the @code{DEFAULT} value, as specified in the ANSI SQL standard. This will -make @code{mysqldump} with @code{NOT NULL} and default values incompatible with -@strong{MySQL} Version 3.20. +For tables that have many columns, the column names are now hashed for +much faster column name lookup (this will speed up some benchmark +tests a lot!) @item -Added many function name aliases so the functions can be used with -ODBC or ANSI SQL92 syntax. +Some benchmarks are changed to get better individual timing. +(Some loops were so short that a specific test took < 2 seconds. The +loops have been changed to take about 20 seconds to make it easier +to compare different databases. A test that took 1-2 seconds before +now takes 11-24 seconds, which is much better) @item -Fixed syntax of @code{ALTER TABLE tbl_name ALTER COLUMN col_name SET DEFAULT -NULL}. +Re-arranged @code{SELECT} code to handle some very specific queries +involving group functions (like @code{COUNT(*)}) without a @code{GROUP BY} but +with @code{HAVING}. The following now works: +@example +mysql> SELECT count(*) as C FROM table HAVING C > 1; +@end example @item -Added @code{CHAR} and @code{BIT} as synonyms for @code{CHAR(1)}. +Changed the protocol for field functions to be faster and avoid some +calls to @code{malloc()}. @item -Fixed core dump when updating as a user who has only @strong{select} privilege. +Added @code{-T32} option to @code{mysqld}, for running all queries under the +main thread. This makes it possible to debug @code{mysqld} under Linux with +@code{gdb}! @item -@code{INSERT ... SELECT ... GROUP BY} didn't work in some cases. An -@code{Invalid use of group function} error occurred. +Added optimization of @code{not_null_column IS NULL} (needed for some Access +queries). @item -When using @code{LIMIT}, @code{SELECT} now always uses keys instead of record -scan. This will give better performance on @code{SELECT} and a @code{WHERE} -that matches many rows. +Allow @code{STRAIGHT_JOIN} to be used between two tables to force the optimizer +to join them in a specific order. @item -Added Russian error messages. -@end itemize - -@node News-3.21.11, News-3.21.10, News-3.21.12, News-3.21.x -@appendixsubsec Changes in release 3.21.11 - -@itemize @bullet +String functions now return @code{VARCHAR} rather than @code{CHAR} and +the column type is now @code{VARCHAR} for fields saved as @code{VARCHAR}. +This should make the @strong{MyODBC} driver better, but may break some old +@strong{MySQL} clients that don't handle @code{FIELD_TYPE_VARCHAR} the same +way as @code{FIELD_TYPE_CHAR}. @item -Configure changes. +@code{CREATE INDEX} and @code{DROP INDEX} are now implemented through +@code{ALTER TABLE}. +@code{CREATE TABLE} is still the recommended (fast) way to create indexes. @item -@strong{MySQL} now works with the new thread library on BSD/OS 3.0. +Added @code{--set-variable} option @code{wait_timeout} to @code{mysqld}. @item -Added new group functions @code{BIT_OR()} and @code{BIT_AND()}. +Added time column to @code{mysqladmin processlist} to show how long a query +has taken or how long a thread has slept. @item -Added compatibility functions @code{CHECK} and @code{REFERENCES}. -@code{CHECK} is now a reserved word. +Added lots of new variables to @code{show variables} and some new to +@code{show status}. @item -Added @code{ALL} option to @code{GRANT} for better compatibility. (@code{GRANT} -is still a dummy function.) +Added new type @code{YEAR}. @code{YEAR} is stored in 1 byte with allowable +values of 0, and 1901 to 2155. @item -Added partly-translated dutch messages. +Added new @code{DATE} type that is stored in 3 bytes rather than 4 bytes. +All new tables are created with the new date type if you don't use the +@code{--old-protocol} option to @code{mysqld}. @item -Fixed bug in @code{ORDER BY} and @code{GROUP BY} with @code{NULL} columns. +Fixed bug in record caches; for some queries, you could get +@code{Error from table handler: #} on some operating systems. @item -Added function @code{last_insert_id()} to retrieve last @code{AUTO_INCREMENT} -value. This is intended for clients to ODBC that can't use the -@code{mysql_insert_id()} API function, but can be used by any client. +Added @code{--enable-assembler} option to @code{configure}, for x86 machines +(tested on Linux + @code{gcc}). This will enable assembler functions for the +most important string functions for more speed! +@end itemize + +@node News-3.21.x, News-3.20.x, News-3.22.x, News +@appendixsec Changes in release 3.21.x + +Version 3.21 is quite old now, and should be avoided if possible. This +information is kept here for historical purposes only. + +@menu +* News-3.21.33:: Changes in release 3.21.33 +* News-3.21.32:: Changes in release 3.21.32 +* News-3.21.31:: Changes in release 3.21.31 +* News-3.21.30:: Changes in release 3.21.30 +* News-3.21.29:: Changes in release 3.21.29 +* News-3.21.28:: Changes in release 3.21.28 +* News-3.21.27:: Changes in release 3.21.27 +* News-3.21.26:: Changes in release 3.21.26 +* News-3.21.25:: Changes in release 3.21.25 +* News-3.21.24:: Changes in release 3.21.24 +* News-3.21.23:: Changes in release 3.21.23 +* News-3.21.22:: Changes in release 3.21.22 +* News-3.21.21a:: Changes in release 3.21.21a +* News-3.21.21:: Changes in release 3.21.21 +* News-3.21.20:: Changes in release 3.21.20 +* News-3.21.19:: Changes in release 3.21.19 +* News-3.21.18:: Changes in release 3.21.18 +* News-3.21.17:: Changes in release 3.21.17 +* News-3.21.16:: Changes in release 3.21.16 +* News-3.21.15:: Changes in release 3.21.15 +* News-3.21.14b:: Changes in release 3.21.14b +* News-3.21.14a:: Changes in release 3.21.14a +* News-3.21.13:: Changes in release 3.21.13 +* News-3.21.12:: Changes in release 3.21.12 +* News-3.21.11:: Changes in release 3.21.11 +* News-3.21.10:: Changes in release 3.21.10 +* News-3.21.9:: Changes in release 3.21.9 +* News-3.21.8:: Changes in release 3.21.8 +* News-3.21.7:: Changes in release 3.21.7 +* News-3.21.6:: Changes in release 3.21.6 +* News-3.21.5:: Changes in release 3.21.5 +* News-3.21.4:: Changes in release 3.21.4 +* News-3.21.3:: Changes in release 3.21.3 +* News-3.21.2:: Changes in release 3.21.2 +* News-3.21.0:: Changes in release 3.21.0 +@end menu + +@node News-3.21.33, News-3.21.32, News-3.21.x, News-3.21.x +@appendixsubsec Changes in release 3.21.33 + +@itemize @bullet @item -Added @code{--flush-logs} option to @code{mysqladmin}. +Fixed problem when sending @code{SIGHUP} to @code{mysqld}; +@code{mysqld} core dumped when starting from boot on some systems. @item -Added command @code{STATUS} to @code{mysql}. +Fixed problem with losing a little memory for some connections. @item -Fixed problem with @code{ORDER BY}/@code{GROUP BY} because of bug in @code{gcc}. +@code{DELETE FROM tbl_name} without a @code{WHERE} condition is now done the +long way when you use @code{LOCK TABLES} or if the table is in use, to +avoid race conditions. @item -Fixed problem with @code{INSERT ... SELECT ... GROUP BY}. +@code{INSERT INTO TABLE (timestamp_column) VALUES (NULL);} didn't set timestamp. @end itemize -@node News-3.21.10, News-3.21.9, News-3.21.11, News-3.21.x -@appendixsubsec Changes in release 3.21.10 +@node News-3.21.32, News-3.21.31, News-3.21.33, News-3.21.x +@appendixsubsec Changes in release 3.21.32 @itemize @bullet @item -New @code{mysqlaccess}. -@item -@code{CREATE} now supports all ODBC types and the @code{mSQL} @code{TEXT} type. -All ODBC 2.5 functions are also supported (added @code{REPEAT}). This provides -better portability. -@item -Added text types @code{TINYTEXT}, @code{TEXT}, @code{MEDIUMTEXT} and -@code{LONGTEXT}. These are actually @code{BLOB}types, but all searching is -done in case-insensitive fashion. +Fixed some possible race conditions when doing many reopen/close on the same +tables under heavy load! This can happen if you execute @code{mysqladmin +refresh} often. This could in some very rare cases corrupt the header of the +index file and cause error 126 or 138. + @item -All old @code{BLOB} fields are now @code{TEXT} fields. This only -changes that all searching on strings is done in case-sensitive fashion. -You must do an @code{ALTER TABLE} and change the field type to @code{BLOB} -if you want to have tests done in case-sensitive fashion. +Fixed fatal bug in @code{refresh()} when running with the +@code{--skip-locking} option. There was a ``very small'' time gap after +a @code{mysqladmin refresh} when a table could be corrupted if one +thread updated a table while another thread did @code{mysqladmin +refresh} and another thread started a new update ont the same table +before the first thread had finished. A refresh (or +@code{--flush-tables}) will now not return until all used tables are +closed! @item -Fixed some @code{configure} issues. +@code{SELECT DISTINCT} with a @code{WHERE} clause that didn't match any rows +returned a row in some contexts (bug only in 3.21.31). @item -Made the locking code a bit safer. Fixed very unlikely -deadlock situation. +@code{GROUP BY} + @code{ORDER BY} returned one empty row when no rows where +found. @item -Fixed a couple of bugs in the range optimizer. Now the new range benchmark -@code{test-select} works. +Fixed a bug in the range optimizer that wrote +@code{Use_count: Wrong count for ...} in the error log file. @end itemize -@node News-3.21.9, News-3.21.8, News-3.21.10, News-3.21.x -@appendixsubsec Changes in release 3.21.9 - +@node News-3.21.31, News-3.21.30, News-3.21.32, News-3.21.x +@appendixsubsec Changes in release 3.21.31 @itemize @bullet @item -Added @code{--enable-unix-socket=pathname} option to @code{configure}. +Fixed a sign extension problem for the @code{TINYINT} type on Irix. @item -Fixed a couple of portability problems with include files. +Fixed problem with @code{LEFT("constant_string",function)}. @item -Fixed bug in range calculation that could return empty -set when searching on multiple key with only one entry (very rare). +Fixed problem with @code{FIND_IN_SET()}. @item -Most things ported to FSU Pthreads, which should allow @strong{MySQL} to -run on SCO. @xref{SCO}. +@code{LEFT JOIN} core dumped if the second table is used with a constant +@code{WHERE/ON} expression that uniquely identifies one record. +@item +Fixed problems with @code{DATE_FORMAT()} and incorrect dates. +@code{DATE_FORMAT()} now ignores @code{'%'} to make it possible to extend +it more easily in the future. @end itemize -@node News-3.21.8, News-3.21.7, News-3.21.9, News-3.21.x -@appendixsubsec Changes in release 3.21.8 - +@node News-3.21.30, News-3.21.29, News-3.21.31, News-3.21.x +@appendixsubsec Changes in release 3.21.30 @itemize @bullet @item -Works now in Solaris 2.6. -@item -Added handling of calculation of @code{SUM()} functions. -For example, you can now use @code{SUM(column)/COUNT(column)}. +@code{mysql} now returns an exit code > 0 if the query returned an error. @item -Added handling of trigometric functions: -@code{PI()}, @code{ACOS()}, @code{ASIN()}, @code{ATAN()}, @code{COS()}, -@code{SIN()} and @code{TAN()}. +Saving of command line history to file in @code{mysql} client. +By Tommy Larsen @email{tommy@@mix.hive.no}. @item -New languages: norwegian, norwegian-ny and portuguese. +Fixed problem with empty lines that were ignored in @file{mysql.cc}. @item -Fixed parameter bug in @code{net_print()} in @file{procedure.cc}. +Save the pid of the signal handler thread in the pid file instead +of the pid of the main thread. @item -Fixed a couple of memory leaks. +Added patch by @email{tommy@@valley.ne.jp} to support Japanese characters +SJIS and UJIS. @item -Now allow also the old @code{SELECT ... INTO OUTFILE} syntax. +Changed @code{safe_mysqld} to redirect startup messages to +@code{'hostname'.err} instead +of @code{'hostname'.log} to reclaim file space on @code{mysqladmin refresh}. @item -Fixed bug with @code{GROUP BY} and @code{SELECT} on key with many values. +@code{ENUM} always had the first entry as default value. @item -@code{mysql_fetch_lengths()} sometimes returned incorrect lengths when you used -@code{mysql_use_result()}. This affected at least some cases of -@code{mysqldump --quick}. +@code{ALTER TABLE} wrote two entries to the update log. @item -Fixed bug in optimization of @code{WHERE const op field}. +@code{sql_acc()} now closes the @code{mysql} grant tables after a reload to +save table space and memory. @item -Fixed problem when sorting on @code{NULL} fields. +Changed @code{LOAD DATA} to use less memory with tables and @code{BLOB} +columns. @item -Fixed a couple of 64-bit (Alpha) problems. +Sorting on a function which made a division / 0 produced a wrong set +in some cases. @item -Added @code{--pid-file=#} option to @code{mysqld}. +Fixed @code{SELECT} problem with @code{LEFT()} when using the czech character +set. @item -Added date formatting to @code{FROM_UNIXTIME()}, originally by Zeev Suraski. +Fixed problem in @code{isamchk}; it couldn't repair a packed table in a very +unusual case. @item -Fixed bug in @code{BETWEEN} in range optimizer (Did only test = of the first -argument). +@code{SELECT} statements with @code{&} or @code{|} (bit functions) failed on +columns with @code{NULL} values. @item -Added machine-dependent files for MIT-pthreads i386-SCO. There is probably -more to do to get this to work on SCO 3.5. +When comparing a field = field, where one of the fields was a part key, +only the length of the part key was compared. @end itemize -@node News-3.21.7, News-3.21.6, News-3.21.8, News-3.21.x -@appendixsubsec Changes in release 3.21.7 - +@node News-3.21.29, News-3.21.28, News-3.21.30, News-3.21.x +@appendixsubsec Changes in release 3.21.29 @itemize @bullet @item -Changed @file{Makefile.am} to take advantage of Automake 1.2. +@code{LOCK TABLES} + @code{DELETE from tbl_name} never removed locks properly. @item -Added the beginnings of a benchmark suite. +Fixed problem when grouping on an @code{OR} function. @item -Added more secure password handling. +Fixed permission problem with @code{umask()} and creating new databases. @item -Added new client function @code{mysql_errno()}, to get the error number of -the error message. This makes error checking in the client much easier. -This makes the new server incompatible with the 3.20.x server when running -without @code{--old-protocol}. The client code is backward compatible. -More information can be found in the @file{README} file! +Fixed permission problem on result file with @code{SELECT ... INTO OUTFILE ...} @item -Fixed some problems when using very long, illegal names. -@end itemize - -@node News-3.21.6, News-3.21.5, News-3.21.7, News-3.21.x -@appendixsubsec Changes in release 3.21.6 - -@itemize @bullet +Fixed problem in range optimizer (core dump) for a very complex query. @item -Fixed more portability issues (incorrect @code{sigwait} and @code{sigset} -defines). +Fixed problem when using @code{MIN(integer)} or @code{MAX(integer)} in +@code{GROUP BY}. @item -@code{configure} should now be able to detect the last argument to -@code{accept()}. +Fixed bug on Alpha when using integer keys. (Other keys worked on Alpha). +@item +Fixed bug in @code{WEEK("XXXX-xx-01")}. @end itemize -@node News-3.21.5, News-3.21.4, News-3.21.6, News-3.21.x -@appendixsubsec Changes in release 3.21.5 - +@node News-3.21.28, News-3.21.27, News-3.21.29, News-3.21.x +@appendixsubsec Changes in release 3.21.28 @itemize @bullet @item -Should now work with FreeBSD 3.0 if used with -@file{FreeBSD-3.0-libc_r-1.0.diff}, -which can be found at @uref{http://www.mysql.com/Downloads/Patches/}. -@item -Added new option @code{-O tmp_table_size=#} to @code{mysqld}. -@item -New function @code{FROM_UNIXTIME(timestamp)} which returns a date string in -'YYYY-MM-DD HH:MM:DD' format. -@item -New function @code{SEC_TO_TIME(seconds)} which returns a string in -'HH:MM:SS' format. +Fixed socket permission (clients couldn't connect to Unix socket on Linux). @item -New function @code{SUBSTRING_INDEX()}, originally by Zeev Suraski. +Fixed bug in record caches; for some queries, you could get +@code{Error from table handler: #} on some operating systems. @end itemize -@node News-3.21.4, News-3.21.3, News-3.21.5, News-3.21.x -@appendixsubsec Changes in release 3.21.4 - +@node News-3.21.27, News-3.21.26, News-3.21.28, News-3.21.x +@appendixsubsec Changes in release 3.21.27 @itemize @bullet @item -Should now configure and compile on OSF1 4.0 with the DEC compiler. +Added user level lock functions @code{GET_LOCK(string,timeout)}, +@code{RELEASE_LOCK(string)}. @item -Configuration and compilation on BSD/OS 3.0 works, but due to some bugs in -BSD/OS 3.0, @code{mysqld} doesn't work on it yet. +Added @code{opened_tables} to @code{show status}. @item -Configuration and compilation on FreeBSD 3.0 works, but I couldn't get -@code{pthread_create} to work. -@end itemize - -@node News-3.21.3, News-3.21.2, News-3.21.4, News-3.21.x -@appendixsubsec Changes in release 3.21.3 - -@itemize @bullet +Changed connect timeout to 3 seconds to make it somewhat harder +for crackers to kill @code{mysqld} through telnet + TCP/IP. @item -Added reverse check lookup of hostnames to get better security. +Fixed bug in range optimizer when using +@code{WHERE key_part_1 >= something AND key_part_2 <= something_else}. @item -Fixed some possible buffer overflows if filenames that are too long are used. +Changed @code{configure} for detection of FreeBSD 3.0 9803xx and above @item -@code{mysqld} doesn't accept hostnames that start with digits followed by a -@code{'.'}, because the hostname may look like an IP number. +@code{WHERE} with string_column_key = constant_string didn't always find +all rows if the column had many values differing only with characters of +the same sort value (like e and @'e). @item -Added @code{--skip-networking} option to @code{mysqld}, to only allow socket -connections. (This will not work with MIT-pthreads!) +Strings keys looked up with 'ref' were not compared in case-sensitive fashion. @item -Added check of too long table names for alias. +Added @code{umask()} to make log files non-readable for normal users. @item -Added check if database name is okay. +Ignore users with old (8-byte) password on startup if not using +@code{--old-protocol} option to @code{mysqld}. @item -Added check if too long table names. +@code{SELECT} which matched all key fields returned the values in the +case of the matched values, not of the found values. (Minor problem.) +@end itemize + +@node News-3.21.26, News-3.21.25, News-3.21.27, News-3.21.x +@appendixsubsec Changes in release 3.21.26 +@itemize @bullet @item -Removed incorrect @code{free()} that killed the server on -@code{CREATE DATABASE} or @code{DROP DATABASE}. +@code{FROM_DAYS(0)} now returns "0000-00-00". @item -Changed some @code{mysqld} @code{-O} options to better names. +In @code{DATE_FORMAT()}, PM and AM were swapped for hours 00 and 12. @item -Added @code{-O join_cache_size=#} option to @code{mysqld}. +Extended the default maximum key size to 256. @item -Added @code{-O max_join_size=#} option to @code{mysqld}, to be able to set a -limit how big queries (in this case big = slow) one should be able to handle -without specifying @code{SET OPTION SQL_BIG_SELECTS=1}. A # = is about 10 -examined records. The default is ``unlimited''. +Fixed bug when using @code{BLOB}/@code{TEXT} in @code{GROUP BY} with many +tables. @item -When comparing a @code{TIME}, @code{DATE}, @code{DATETIME} or @code{TIMESTAMP} -column to a constant, the constant is converted to a time value before -performing the comparison. -This will make it easier to get ODBC (particularly Access97) to work with -the above types. It should also make dates easier to use and the comparisons -should be quicker than before. +An @code{ENUM} field that is not declared @code{NOT NULL} has @code{NULL} as +the default value. +(Previously, the default value was the first enumeration value.) @item -Applied patch from Jochen Wiedmann that allows @code{query()} in -@code{mysqlperl} to take a query with @code{\0} in it. +Fixed bug in the join optimizer code when using many part keys +on the same key: @code{INDEX (Organization,Surname(35),Initials(35))}. @item -Storing a timestamp with a 2-digit year (@code{YYMMDD}) didn't work. +Added some tests to the table order optimizer to get some cases with +@code{SELECT ... FROM many_tables} much faster. @item -Fix that timestamp wasn't automatically updated if set in an @code{UPDATE} -clause. +Added a retry loop around @code{accept()} to possibly fix some problems on some +Linux machines. +@end itemize + +@node News-3.21.25, News-3.21.24, News-3.21.26, News-3.21.x +@appendixsubsec Changes in release 3.21.25 +@itemize @bullet @item -Now the automatic timestamp field is the FIRST timestamp field. +Changed @code{typedef 'string'} to @code{typedef 'my_string'} for better +portability. @item -@code{SELECT * INTO OUTFILE}, which didn't correctly if the outfile already -existed. +You can now kill threads that are waiting on a disk-full condition. @item -@code{mysql} now shows the thread ID when starting or doing a reconnect. +Fixed some problems with UDF functions. @item -Changed the default sort buffer size from 2M to 1M. +Added long options to @code{isamchk}. Try @code{isamchk --help}. +@item +Fixed a bug when using 8 bytes long (alpha); @code{filesort()} didn't work. +Affects @code{DISTINCT}, @code{ORDER BY} and @code{GROUP BY} on 64-bit +processors. @end itemize -@node News-3.21.2, News-3.21.0, News-3.21.3, News-3.21.x -@appendixsubsec Changes in release 3.21.2 - +@node News-3.21.24, News-3.21.23, News-3.21.25, News-3.21.x +@appendixsubsec Changes in release 3.21.24 @itemize @bullet @item -The range optimizer is coded, but only 85% tested. It can be enabled with -@code{--new}, but it crashes core a lot yet... +Dynamic loadable functions. Based on source from Alexis Mikhailov. @item -More portable. Should compile on AIX and alpha-digital. -At least the @code{isam} library should be relatively 64-bit clean. +You couldn't delete from a table if no one had done a @code{SELECT} on the +table. @item -New @code{isamchk} which can detect and fix more problems. +Fixed problem with range optimizer with many @code{OR} operators on key parts +inside each other. @item -New options for @code{isamlog}. +Recoded @code{MIN()} and @code{MAX()} to work properly with strings and +@code{HAVING}. @item -Using new version of Automake. +Changed default umask value for new files from @code{0664} to @code{0660}. @item -Many small portability changes (from the AIX and alpha-digital port) -Better checking of pthread(s) library. +Fixed problem with @code{LEFT JOIN} and constant expressions in the @code{ON} +part. @item -czech error messages by @email{snajdr@@pvt.net}. +Added Italian error messages from @email{brenno@@dewinter.com}. +@item +@code{configure} now works better on OSF1 (tested on 4.0D). +@item +Added hooks to allow @code{LIKE} optimization with international character +support. @item -Decreased size of some buffers to get fewer problems on systems with little -memory. Also added more checks to handle ``out of memory'' problems. +Upgraded @code{DBI} to 0.93. +@end itemize + +@node News-3.21.23, News-3.21.22, News-3.21.24, News-3.21.x +@appendixsubsec Changes in release 3.21.23 +@itemize @bullet @item -@code{mysqladmin}: you can now do @code{mysqladmin kill 5,6,7,8} to kill -multiple threads. +The following symbols are now reserved words: +@code{TIME}, @code{DATE}, @code{TIMESTAMP}, @code{TEXT}, @code{BIT}, +@code{ENUM}, @code{NO}, @code{ACTION}, @code{CHECK}, @code{YEAR}, +@code{MONTH}, @code{DAY}, @code{HOUR}, @code{MINUTE}, @code{SECOND}, +@code{STATUS}, @code{VARIABLES}. @item -When the maximum connection limit is reached, one extra connection by a user -with the @strong{PROCESS_ACL} privilege is granted. +Setting a @code{TIMESTAMP} to @code{NULL} in @code{LOAD DATA INFILE ...} didn't +set the current time for the @code{TIMESTAMP}. @item -Added @code{-O backlog=#} option to @code{mysqld}. +Fix @code{BETWEEN} to recognize binary strings. Now @code{BETWEEN} is +case sensitive. @item -Increased maximum packet size from 512K to 1024K for client. +Added @code{--skip-thread-priority} option to @code{mysqld}, for systems +where @code{mysqld}'s thread scheduling doesn't work properly (BSDI 3.1). @item -Almost all of the function code is now tested in the internal test suite. +Added ODBC functions @code{DAYNAME()} and @code{MONTHNAME()}. @item -@code{ALTER TABLE} now returns warnings from field conversions. +Added function @code{TIME_FORMAT()}. This works like @code{DATE_FORMAT()}, +but takes a time string (@code{'HH:MM:DD'}) as argument. @item -Port changed to 3306 (got it reserved from ISI). +Fixed unlikely(?) key optimizer bug when using @code{OR}s of key parts +inside @code{AND}s. @item -Added a fix for Visual FoxBase so that any schema name from a table -specification is automatically removed. +Added command @code{variables} to @code{mysqladmin}. @item -New function @code{ASCII()}. +A lot of small changes to the binary releases. @item -Removed function @code{BETWEEN(a,b,c)}. Use the standard ANSI -syntax instead: @code{expr BETWEEN expr AND expr}. +Fixed a bug in the new protocol from @strong{MySQL} Version 3.21.20. @item -@strong{MySQL} no longer has to use an extra temporary table when sorting -on functions or @code{SUM()} functions. +Changed @code{ALTER TABLE} to work with Windows (Windows can't rename +open files). Also fixed a couple of small bugs in the Windows version. @item -Fixed bug that you couldn't use @code{tbl_name.field_name} in @code{UPDATE}. +All standard @strong{MySQL} clients are now ported to @strong{MySQL}-Windows. @item -Fixed @code{SELECT DISTINCT} when using 'hidden group'. For example: -@example -mysql> SELECT DISTINCT MOD(some_field,10) FROM test - GROUP BY some_field; -@end example -Note: @code{some_field} is normally in the @code{SELECT} part. ANSI SQL should -require it. +@strong{MySQL} can now be started as a service on NT. @end itemize -@node News-3.21.0, , News-3.21.2, News-3.21.x -@appendixsubsec Changes in release 3.21.0 - +@node News-3.21.22, News-3.21.21a, News-3.21.23, News-3.21.x +@appendixsubsec Changes in release 3.21.22 @itemize @bullet @item -New reserved words used: @code{INTERVAL}, @code{EXPLAIN}, @code{READ}, -@code{WRITE}, @code{BINARY}. +Starting with this version, all @strong{MySQL} distributions will be +configured, compiled and tested with @code{crash-me} and the benchmarks on +the following platforms: SunOS 5.6 sun4u, SunOS 5.5.1 sun4u, SunOS 4.14 sun4c, +SunOS 5.6 i86pc, Irix 6.3 mips5k, HP-UX 10.20 hppa, AIX 4.2.1 ppc, +OSF1 V4.0 alpha, FreeBSD 2.2.2 i86pc and BSDI 3.1 i386. @item -Added ODBC function @code{CHAR(num,...)}. +Fix @code{COUNT(*)} problems when the @code{WHERE} clause didn't match any +records. (Bug from 3.21.17.) @item -New operator @code{IN}. This uses a binary search to find a match. +Removed that @code{NULL = NULL} is true. Now you must use @code{IS NULL} +or @code{IS NOT NULL} to test whether or not a value is @code{NULL}. +(This is according to ANSI SQL but may break +old applications that are ported from @code{mSQL}.) +You can get the old behavior by compiling with @code{-DmSQL_COMPLIANT}. @item -New command @code{LOCK TABLES tbl_name [AS alias] @{READ|WRITE@} ...} +Fixed bug that core dumped when using many @code{LEFT OUTER JOIN} clauses. @item -Added @code{--log-update} option to @code{mysqld}, to get a log suitable for -incremental updates. +Fixed bug in @code{ORDER BY} on string formula with possible @code{NULL} values. @item -New command @code{EXPLAIN SELECT ...} to get information about how the -optimizer will do the join. +Fixed problem in range optimizer when using <= on sub index. @item -For easier client code, the client should no longer use -@code{FIELD_TYPE_TINY_BLOB}, @code{FIELD_TYPE_MEDIUM_BLOB}, -@code{FIELD_TYPE_LONG_BLOB} or @code{FIELD_TYPE_VAR_STRING} (as -previously returned by @code{mysql_list_fields}). You should instead only use -@code{FIELD_TYPE_BLOB} or @code{FIELD_TYPE_STRING}. If you want exact -types, you should use the command @code{SHOW FIELDS}. +Added functions @code{DAYOFYEAR()}, @code{DAYOFMONTH()}, @code{MONTH()}, +@code{YEAR()}, @code{WEEK()}, @code{QUARTER()}, @code{HOUR()}, @code{MINUTE()}, +@code{SECOND()} and @code{FIND_IN_SET()}. @item -Added varbinary syntax: @code{0x######} which can be used as a string -(default) or a number. +Added command @code{SHOW VARIABLES}. @item -@code{FIELD_TYPE_CHAR} is renamed to @code{FIELD_TYPE_TINY}. +Added support of ``long constant strings'' from ANSI SQL: +@example +mysql> SELECT 'first ' 'second'; -> 'first second' +@end example @item -Changed all fields to C++ classes. +Upgraded mSQL-Mysql-modules to 1.1825. @item -Removed FORM struct. +Upgraded @code{mysqlaccess} to 2.02. @item -Fields with @code{DEFAULT} values no longer need to be @code{NOT NULL}. +Fixed problem with Russian character set and @code{LIKE}. @item -New field types: -@table @code -@item ENUM -A string which can take only a couple of defined values. The value is -stored as a 1-3 byte number that is mapped automatically to a string. -This is sorted according to string positions! -@item SET -A string which may have one or many string values separated with ','. -The string is stored as a 1-, 2-, 3-, 4- or 8-byte number where each bit stands -for a specific set member. This is sorted according to the unsigned value -of the stored packed number. -@end table +Ported to OpenBSD 2.1. @item -Now all function calculation is done with @code{double} or @code{long long}. -This will provide the full 64-bit range with bit functions and fix some -conversions that previously could result in precision losses. One should -avoid using @code{unsigned long long} columns with full 64-bit range -(numbers bigger than 9223372036854775807) because calculations are done -with @code{signed long long}. +New Dutch error messages. +@end itemize + +@node News-3.21.21a, News-3.21.21, News-3.21.22, News-3.21.x +@appendixsubsec Changes in release 3.21.21a +@itemize @bullet @item -@code{ORDER BY} will now put @code{NULL} field values first. @code{GROUP BY} -will also work with @code{NULL} values. +Configure changes for some operating systems. +@end itemize + +@node News-3.21.21, News-3.21.20, News-3.21.21a, News-3.21.x +@appendixsubsec Changes in release 3.21.21 +@itemize @bullet @item -Full @code{WHERE} with expressions. +Fixed optimizer bug when using +@code{WHERE data_field = date_field2 AND date_field2 = constant}. @item -New range optimizer that can resolve ranges when some keypart prefix is -constant. Example: -@example -mysql> SELECT * FROM tbl_name - WHERE key_part_1="customer" - AND key_part_2>=10 AND key_part_2<=10; -@end example +Added command @code{SHOW STATUS}. +@item +Removed @file{manual.ps} from the source distribution to make it smaller. @end itemize -@node News-3.20.x, News-3.19.x, News-3.21.x, News -@appendixsec Changes in release 3.20.x - -Version 3.20 is quite old now, and should be avoided if possible. This -information is kept here for historical purposes only. - -Changes from 3.20.18 to 3.20.32b are not documented here because the -3.21 release branched here. And the relevant changes are also -documented as changes to the 3.21 version. - -@menu -* News-3.20.18:: Changes in release 3.20.18 -* News-3.20.17:: Changes in release 3.20.17 -* News-3.20.16:: Changes in release 3.20.16 -* News-3.20.15:: Changes in release 3.20.15 -* News-3.20.14:: Changes in release 3.20.14 -* News-3.20.13:: Changes in release 3.20.13 -* News-3.20.11:: Changes in release 3.20.11 -* News-3.20.10:: Changes in release 3.20.10 -* News-3.20.9:: Changes in release 3.20.9 -* News-3.20.8:: Changes in release 3.20.8 -* News-3.20.7:: Changes in release 3.20.7 -* News-3.20.6:: Changes in release 3.20.6 -* News-3.20.3:: Changes in release 3.20.3 -* News-3.20.0:: Changes in releases 3.20.0 -@end menu - -@node News-3.20.18, News-3.20.17, News-3.20.x, News-3.20.x -@appendixsubsec Changes in release 3.20.18 - +@node News-3.21.20, News-3.21.19, News-3.21.21, News-3.21.x +@appendixsubsec Changes in release 3.21.20 @itemize @bullet @item -Added @code{-p#} (remove @code{#} directories from path) to @code{isamlog}. -All files are written with a relative path from the database directory -Now @code{mysqld} shouldn't crash on shutdown when using the -@code{--log-isam} option. -@item -New @code{mysqlperl} version. It is now compatible with @code{msqlperl-0.63}. +Changed the maximum table name and column name lengths from 32 to 64. @item -New @code{DBD} module available at @uref{http://www.mysql.com/Downloads/Contrib/} -site. +Aliases can now be of ``any'' length. @item -Added group function @code{STD()} (standard deviation). +Fixed @code{mysqladmin stat} to return the right number of queries. @item -The @code{mysqld} server is now compiled by default without debugging -information. This will make the daemon smaller and faster. +Changed protocol (downward compatible) to mark if a column has the +@code{AUTO_INCREMENT} attribute or is a @code{TIMESTAMP}. This is needed for +the new Java driver. @item -Now one usually only has to specify the @code{--basedir} option to -@code{mysqld}. All other paths are relative in a normal installation. +Added Hebrew sorting order by Zeev Suraski. @item -@code{BLOB} columns sometimes contained garbage when used with a @code{SELECT} -on more than one table and @code{ORDER BY}. +Solaris 2.6: Fixed @code{configure} bugs and increased maximum table size +from 2G to 4G. +@end itemize + +@node News-3.21.19, News-3.21.18, News-3.21.20, News-3.21.x +@appendixsubsec Changes in release 3.21.19 +@itemize @bullet @item -Fixed that calculations that are not in @code{GROUP BY} work as expected -(ANSI SQL extension). -Example: -@example -mysql> SELECT id,id+1 FROM table GROUP BY id; -@end example +Upgraded @code{DBD} to 1823. This version implements @code{mysql_use_result} in +@code{DBD-Mysql}. @item -The test of using @code{MYSQL_PWD} was reversed. Now @code{MYSQL_PWD} is -enabled as default in the default release. +Benchmarks updated for empress (by Luuk). @item -Fixed conversion bug which caused @code{mysqld} to core dump with -Arithmetic error on Sparc-386. +Fixed a case of slow range searching. @item -Added @code{--unbuffered} option to @code{mysql}, for new @code{mysqlaccess}. +Configure fixes (@file{Docs} directory). @item -When using overlapping (unnecessary) keys and join over many tables, -the optimizer could get confused and return 0 records. +Added function @code{REVERSE()} (by Zeev Suraski). @end itemize -@node News-3.20.17, News-3.20.16, News-3.20.18, News-3.20.x -@appendixsubsec Changes in release 3.20.17 - +@node News-3.21.18, News-3.21.17, News-3.21.19, News-3.21.x +@appendixsubsec Changes in release 3.21.18 @itemize @bullet @item -You can now use @code{BLOB} columns and the functions @code{IS NULL} and -@code{IS NOT NULL} in the @code{WHERE} clause. -@item -All communication packets and row buffers are now allocated dynamically -on demand. The default value of @code{max_allowed_packet} is now 64K for -the server and 512K for the client. This is mainly used to catch -incorrect packets that could trash all memory. The server limit may be -changed when it is started. +Issue error message if client C functions are called in wrong order. @item -Changed stack usage to use less memory. +Added automatic reconnect to the @file{libmysql.c} library. If a write command +fails, an automatic reconnect is done. @item -Changed @code{safe_mysqld} to check for running daemon. +Small sort sets no longer use temporary files. @item -The @code{ELT()} function is renamed to @code{FIELD()}. The new -@code{ELT()} function returns a value based on an index: @code{FIELD()} -is the inverse of @code{ELT()} Example: @code{ELT(2,"A","B","C")} returns -@code{"B"}. @code{FIELD("B","A","B","C")} returns @code{2}. +Upgraded @code{DBI} to 0.91. @item -@code{COUNT(field)}, where @code{field} could have a @code{NULL} value, now -works. +Fixed a couple of problems with @code{LEFT OUTER JOIN}. @item -A couple of bugs fixed in @code{SELECT ... GROUP BY}. +Added @code{CROSS JOIN} syntax. @code{CROSS} is now a reserved word. @item -Fixed memory overrun bug in @code{WHERE} with many unoptimizable brace levels. +Recoded @code{yacc}/@code{bison} stack allocation to be even safer and to allow +@strong{MySQL} to handle even bigger expressions. @item -Fixed some small bugs in the grant code. +Fixed a couple of problems with the update log. @item -If hostname isn't found by @code{get_hostname}, only the IP is checked. -Previously, you got @code{Access denied}. +@code{ORDER BY} was slow when used with key ranges. +@end itemize + +@node News-3.21.17, News-3.21.16, News-3.21.18, News-3.21.x +@appendixsubsec Changes in release 3.21.17 +@itemize @bullet @item -Inserts of timestamps with values didn't always work. +Changed documentation string of @code{--with-unix-socket-path} to avoid +confusion. @item -@code{INSERT INTO ... SELECT ... WHERE} could give the error -@code{Duplicated field}. +Added ODBC and ANSI SQL style @code{LEFT OUTER JOIN}. @item -Added some tests to @code{safe_mysqld} to make it ``safer''. +The following are new reserved words: @code{LEFT}, @code{NATURAL}, +@code{USING}. @item -@code{LIKE} was case sensitive in some places and case insensitive in others. -Now @code{LIKE} is always case insensitive. +The client library now uses the value of the environment variable +@code{MYSQL_HOST} as the default host if it's defined. @item -@file{mysql.cc}: Allow @code{'#'} anywhere on the line. +@code{SELECT col_name, SUM(expr)} now returns @code{NULL} for @code{col_name} +when there are matching rows. @item -New command @code{SET OPTION SQL_SELECT_LIMIT=#}. See the FAQ for more details. +Fixed problem with comparing binary strings and @code{BLOB}s with ASCII +characters over 127. @item -New version of the @code{mysqlaccess} script. +Fixed lock problem: when freeing a read lock on a table with multiple +read locks, a thread waiting for a write lock would have been given the lock. +This shouldn't affect data integrity, but could possibly make @code{mysqld} +restart if one thread was reading data that another thread modified. @item -Change @code{FROM_DAYS()} and @code{WEEKDAY()} to also take a full -@code{TIMESTAMP} or @code{DATETIME} as argument. Before they only took a -number of type @code{YYYYMMDD} or @code{YYMMDD}. +@code{LIMIT offset,count} didn't work in @code{INSERT ... SELECT}. @item -Added new function @code{UNIX_TIMESTAMP(timestamp_column)}. +Optimized key block caching. This will be quicker than the old algorithm when +using bigger key caches. @end itemize -@node News-3.20.16, News-3.20.15, News-3.20.17, News-3.20.x -@appendixsubsec Changes in release 3.20.16 +@node News-3.21.16, News-3.21.15, News-3.21.17, News-3.21.x +@appendixsubsec Changes in release 3.21.16 @itemize @bullet @item -More changes in MIT-pthreads to get them safer. Fixed also some link -bugs at least in SunOS. -@item -Changed @code{mysqld} to work around a bug in MIT-pthreads. This makes multiple -small @code{SELECT} operations 20 times faster. Now @code{lock_test.pl} should -work. -@item -Added @code{mysql_FetchHash(handle)} to @code{mysqlperl}. +Added ODBC 2.0 & 3.0 functions @code{POWER()}, @code{SPACE()}, +@code{COT()}, @code{DEGREES()}, @code{RADIANS()}, @code{ROUND(2 arg)} +and @code{TRUNCATE()}. @item -The @code{mysqlbug} script is now distributed built to allow for reporting -bugs that appear during the build with it. +@strong{WARNING: INCOMPATIBLE CHANGE!!} @code{LOCATE()} parameters were +swapped according to ODBC standard. Fixed. @item -Changed @file{libmysql.c} to prefer @code{getpwuid()} instead of -@code{cuserid()}. +Added function @code{TIME_TO_SEC()}. @item -Fixed bug in @code{SELECT} optimizer when using many tables with the same -column used as key to different tables. +In some cases, default values were not used for @code{NOT NULL} fields. @item -Added new latin2 and Russian KOI8 character tables. +Timestamp wasn't always updated properly in @code{UPDATE SET ...} statements. @item -Added support for a dummy @code{GRANT} command to satisfy Powerbuilder. +Allow empty strings as default values for @code{BLOB} and @code{TEXT}, to +be compatible with @code{mysqldump}. @end itemize -@node News-3.20.15, News-3.20.14, News-3.20.16, News-3.20.x -@appendixsubsec Changes in release 3.20.15 +@node News-3.21.15, News-3.21.14b, News-3.21.16, News-3.21.x +@appendixsubsec Changes in release 3.21.15 @itemize @bullet @item -Fixed fatal bug @code{packets out of order} when using MIT-pthreads. +@strong{WARNING: INCOMPATIBLE CHANGE!!} @code{mysqlperl} is now from +Msql-Mysql-modules. This means that @code{connect()} now takes +@code{host}, @code{database}, @code{user}, @code{password} arguments! The old +version took @code{host}, @code{database}, @code{password}, @code{user}. @item -Removed possible loop when a thread waits for command from client -and @code{fcntl()} fails. Thanks to Mike Bretz for finding this bug. +Allow @code{DATE '1997-01-01'}, @code{TIME '12:10:10'} and +@code{TIMESTAMP '1997-01-01 12:10:10'} formats required by ANSI SQL. +@strong{WARNING: INCOMPATIBLE CHANGE!!} This has the unfortunate +side-effect that you no longer can have columns named @code{DATE}, @code{TIME} +or @code{TIMESTAMP}. :( Old columns can still be accessed through +@code{tablename.columnname}!) @item -Changed alarm loop in @file{mysqld.cc} because shutdown didn't always -succeed in Linux. +Changed Makefiles to hopefully work better with BSD systems. Also, +@file{manual.dvi} is now included in the distribution to avoid having stupid +@code{make} programs trying to rebuild it. @item -Removed use of @code{termbits} from @file{mysql.cc}. This conflicted with -@code{glibc} 2.0. +@code{readline} library upgraded to version 2.1. @item -Fixed some syntax errors for at least BSD and Linux. +A new sortorder german-1. That is a normal ISO-Latin1 with a german sort +order. @item -Fixed bug when doing a @code{SELECT} as superuser without a database. +Perl @code{DBI}/@code{DBD} is now included in the distribution. @code{DBI} +is now the recommended way to connect to @strong{MySQL} from Perl. @item -Fixed bug when doing @code{SELECT} with group calculation to outfile. +New portable benchmark suite with @code{DBD}, with test results from +@code{mSQL} 2.0.3, @strong{MySQL}, PostgreSQL 6.2.1 and Solid server 2.2. +@item +@code{crash-me} is now included with the benchmarks; This is a Perl program +designed to find as many limits as possible in a SQL server. Tested with +@code{mSQL}, PostgreSQL, Solid and @strong{MySQL}. +@item +Fixed bug in range-optimizer that crashed @strong{MySQL} on some queries. +@item +Table and column name completion for @code{mysql} command line tool, by Zeev +Suraski and Andi Gutmans. +@item +Added new command @code{REPLACE} that works like @code{INSERT} but +replaces conflicting records with the new record. @code{REPLACE INTO +TABLE ... SELECT ...} works also. +@item +Added new commands @code{CREATE DATABASE db_name} and @code{DROP +DATABASE db_name}. +@item +Added @code{RENAME} option to @code{ALTER TABLE}: @code{ALTER TABLE name +RENAME TO new_name}. +@item +@code{make_binary_distribution} now includes @file{libgcc.a} in +@file{libmysqlclient.a}. This should make linking work for people who don't +have @code{gcc}. +@item +Changed @code{net_write()} to @code{my_net_write()} because of a name +conflict with Sybase. +@item +@cindex ODBC compatibility +@cindex Compatibility, with ODBC +New function @code{DAYOFWEEK()} compatible with ODBC. +@item +Stack checking and @code{bison} memory overrun checking to make @strong{MySQL} +safer with weird queries. @end itemize -@node News-3.20.14, News-3.20.13, News-3.20.15, News-3.20.x -@appendixsubsec Changes in release 3.20.14 +@node News-3.21.14b, News-3.21.14a, News-3.21.15, News-3.21.x +@appendixsubsec Changes in release 3.21.14b @itemize @bullet @item -If one gives @code{-p} or @code{--password} option to @code{mysql} without -an argument, the user is solicited for the password from the tty. +Fixed a couple of small @code{configure} problems on some platforms. +@end itemize + +@node News-3.21.14a, News-3.21.13, News-3.21.14b, News-3.21.x +@appendixsubsec Changes in release 3.21.14a + +@itemize @bullet @item -Added default password from @code{MYSQL_PWD} (by Elmar Haneke). +Ported to SCO Openserver 5.0.4 with FSU Pthreads. @item -Added command @code{kill} to @code{mysqladmin} to kill a specific -@strong{MySQL} thread. +HP-UX 10.20 should work. @item -Sometimes when doing a reconnect on a down connection this succeeded -first on second try. +Added new function @code{DATE_FORMAT()}. @item -Fixed adding an @code{AUTO_INCREMENT} key with @code{ALTER_TABLE}. +Added @code{NOT IN}. @item -@code{AVG()} gave too small value on some @code{SELECT}s with -@code{GROUP BY} and @code{ORDER BY}. +Added automatic removal of 'ODBC function conversions': @code{@{fn now() @}} @item -Added new @code{DATETIME} type (by Giovanni Maruzzelli -@email{maruzz@@matrice.it}). +Handle ODBC 2.50.3 option flags. @item -Fixed that define @code{DONT_USE_DEFAULT_FIELDS} works. +Fixed comparison of @code{DATE} and @code{TIME} values with @code{NULL}. @item -Changed to use a thread to handle alarms instead of signals on Solaris to -avoid race conditions. +Changed language name from germany to german to be consistent with +the other language names. @item -Fixed default length of signed numbers. (George Harvey -@email{georgeh@@pinacl.co.uk}.) +Fixed sorting problem on functions returning a @code{FLOAT}. Previously, the +values were converted to @code{INT}s before sorting. @item -Allow anything for @code{CREATE INDEX}. +Fixed slow sorting when sorting on key field when using +@code{key_column=constant}. @item -Add prezeros when packing numbers to @code{DATE}, @code{TIME} and -@code{TIMESTAMP}. +Sorting on calculated @code{DOUBLE} values sorted on integer results instead. @item -Fixed a bug in @code{OR} of multiple tables (gave empty set). +@code{mysql} no longer needs a database argument. @item -Added many patches to MIT-pthreads. This fixes at least one lookup bug. +Changed the place where @code{HAVING} should be. According to ANSI, it should +be after @code{GROUP BY} but before @code{ORDER BY}. @strong{MySQL} Version 3.20 +incorrectly had it last. +@item +Added Sybase command @code{USE DATABASE} to start using another database. +@item +Added automatic adjusting of number of connections and table cache size +if the maximum number of files that can be opened is less than needed. +This should fix that @code{mysqld} doesn't crash even if you haven't done a +@code{ulimit -n 256} before starting @code{mysqld}. +@item +Added lots of limit checks to make it safer when running with too little +memory or when doing weird queries. @end itemize -@node News-3.20.13, News-3.20.11, News-3.20.14, News-3.20.x -@appendixsubsec Changes in release 3.20.13 +@node News-3.21.13, News-3.21.12, News-3.21.14a, News-3.21.x +@appendixsubsec Changes in release 3.21.13 @itemize @bullet @item -Added ANSI SQL94 @code{DATE} and @code{TIME} types. -@item -Fixed bug in @code{SELECT} with @code{AND}-@code{OR} levels. -@item -Added support for Slovenian characters. The @file{Contrib} directory contains -source and instructions for adding other character sets. -@item -Fixed bug with @code{LIMIT} and @code{ORDER BY}. -@item -Allow @code{ORDER BY} and @code{GROUP BY} on items that aren't in the -@code{SELECT} list. -(Thanks to Wim Bonis @email{bonis@@kiss.de}, for pointing this out.) +Added retry of interrupted reads and clearing of @code{errno}. +This makes Linux systems much safer! @item -Allow setting of timestamp values in @code{INSERT}. +Fixed locking bug when using many aliases on the same table in the same +@code{SELECT}. @item -Fixed bug with @code{SELECT ... WHERE ... = NULL}. +Fixed bug with @code{LIKE} on number key. @item -Added changes for @code{glibc} 2.0. To get @code{glibc} to work, you should -add the @file{gibc-2.0-sigwait-patch} before compiling @code{glibc}. +New error message so you can check whether the connection was lost while +the command was running or whether the connection was down from the start. @item -Fixed bug in @code{ALTER TABLE} when changing a @code{NOT NULL} field to -allow @code{NULL} values. +Added @code{--table} option to @code{mysql} to print in table format. +Moved time and row information after query result. +Added automatic reconnect of lost connections. @item -Added some ANSI92 synonyms as field types to @code{CREATE TABLE}. -@code{CREATE TABLE} now allows @code{FLOAT(4)} and @code{FLOAT(8)} to mean -@code{FLOAT} and @code{DOUBLE}. +Added @code{!=} as a synonym for @code{<>}. @item -New utility program @code{mysqlaccess} by @email{Yves.Carlier@@rug.ac.be}. -This program shows the access rights for a specific user and the grant -rows that determine this grant. +Added function @code{VERSION()} to make easier logs. @item -Added @code{WHERE const op field} (by @email{bonis@@kiss.de}). +New multi-user test @file{tests/fork_test.pl} to put some strain on the +thread library. @end itemize -@node News-3.20.11, News-3.20.10, News-3.20.13, News-3.20.x -@appendixsubsec Changes in release 3.20.11 +@node News-3.21.12, News-3.21.11, News-3.21.13, News-3.21.x +@appendixsubsec Changes in release 3.21.12 @itemize @bullet @item -When using @code{SELECT ... INTO OUTFILE}, all temporary tables are ISAM -instead of HEAP to allow big dumps. -@item -Changed date functions to be string functions. This fixed some ``funny'' -side effects when sorting on dates. -@item -Extended @code{ALTER TABLE} according to SQL92. +Fixed @code{ftruncate()} call in MIT-pthreads. This made @code{isamchk} +destroy the @file{.ISM} files on (Free)BSD 2.x systems. @item -Some minor compatibility changes. +Fixed broken @code{__P_} patch in MIT-pthreads. @item -Added @code{--port} and @code{--socket} options to all utility programs and -@code{mysqld}. +Many memory overrun checks. All string functions now return @code{NULL} +if the returned string should be longer than @code{max_allowed_packet} bytes. @item -Fixed MIT-pthreads @code{readdir_r()}. Now @code{mysqladmin create database} -and @code{mysqladmin drop database} should work. +Changed the name of the @code{INTERVAL} type to @code{ENUM}, because +@code{INTERVAL} is used in ANSI SQL. @item -Changed MIT-pthreads to use our @code{tempnam()}. This should fix the ``sort -aborted'' bug. +In some cases, doing a @code{JOIN} + @code{GROUP} + @code{INTO OUTFILE}, +the result wasn't grouped. @item -Added sync of records count in @code{sql_update}. This fixed slow updates -on first connection. (Thanks to Vaclav Bittner for the test.) -@end itemize - -@node News-3.20.10, News-3.20.9, News-3.20.11, News-3.20.x -@appendixsubsec Changes in release 3.20.10 - -@itemize @bullet +@code{LIKE} with @code{'_'} as last character didn't work. Fixed. @item -New insert type: @code{INSERT INTO ... SELECT ...} +Added extended ANSI SQL @code{TRIM()} function. @item -@code{MEDIUMBLOB} fixed. +Added @code{CURTIME()}. @item -Fixed bug in @code{ALTER TABLE} and @code{BLOB}s. +Added @code{ENCRYPT()} function by Zeev Suraski. @item -@code{SELECT ... INTO OUTFILE} now creates the file in the current -database directory. +Fixed better @code{FOREIGN KEY} syntax skipping. New reserved words: +@code{MATCH}, @code{FULL}, @code{PARTIAL}. @item -@code{DROP TABLE} now can take a list of tables. +@code{mysqld} now allows IP number and hostname to the @code{--bind-address} +option. @item -Oracle synonym @code{DESCRIBE} (@code{DESC}). +Added @code{SET OPTION CHARACTER SET cp1251_koi8} to enable conversions of +data to/from cp1251_koi8. @item -Changes to @code{make_binary_distribution}. +Lots of changes for Win95 port. In theory, this version should now be +easily portable to Win95. @item -Added some comments to installation instructions about @code{configure}'s -C++ link test. +Changed the @code{CREATE COLUMN} syntax of @code{NOT NULL} columns to be after +the @code{DEFAULT} value, as specified in the ANSI SQL standard. This will +make @code{mysqldump} with @code{NOT NULL} and default values incompatible with +@strong{MySQL} Version 3.20. @item -Added @code{--without-perl} option to @code{configure}. +Added many function name aliases so the functions can be used with +ODBC or ANSI SQL92 syntax. @item -Lots of small portability changes. -@end itemize - -@node News-3.20.9, News-3.20.8, News-3.20.10, News-3.20.x -@appendixsubsec Changes in release 3.20.9 - -@itemize @bullet +Fixed syntax of @code{ALTER TABLE tbl_name ALTER COLUMN col_name SET DEFAULT +NULL}. @item -@code{ALTER TABLE} didn't copy null bit. As a result, fields that were allowed -to have @code{NULL} values were always @code{NULL}. +Added @code{CHAR} and @code{BIT} as synonyms for @code{CHAR(1)}. @item -@code{CREATE} didn't take numbers as @code{DEFAULT}. +Fixed core dump when updating as a user who has only @strong{select} privilege. @item -Some compatibility changes for SunOS. +@code{INSERT ... SELECT ... GROUP BY} didn't work in some cases. An +@code{Invalid use of group function} error occurred. @item -Removed @file{config.cache} from old distribution. -@end itemize - -@node News-3.20.8, News-3.20.7, News-3.20.9, News-3.20.x -@appendixsubsec Changes in release 3.20.8 - -@itemize @bullet +When using @code{LIMIT}, @code{SELECT} now always uses keys instead of record +scan. This will give better performance on @code{SELECT} and a @code{WHERE} +that matches many rows. @item -Fixed bug with @code{ALTER TABLE} and multi-part keys. +Added Russian error messages. @end itemize -@node News-3.20.7, News-3.20.6, News-3.20.8, News-3.20.x -@appendixsubsec Changes in release 3.20.7 +@node News-3.21.11, News-3.21.10, News-3.21.12, News-3.21.x +@appendixsubsec Changes in release 3.21.11 @itemize @bullet @item -New commands: @code{ALTER TABLE}, @code{SELECT ... INTO OUTFILE} and -@code{LOAD DATA INFILE}. +Configure changes. @item -New function: @code{NOW()}. +@strong{MySQL} now works with the new thread library on BSD/OS 3.0. @item -Added new field @strong{file_priv} to @code{mysql/user} table. +Added new group functions @code{BIT_OR()} and @code{BIT_AND()}. @item -New script @code{add_file_priv} which adds the new field @strong{file_priv} -to the @code{user} table. This script must be executed if you want to -use the new @code{SELECT ... INTO} and @code{LOAD DATA INFILE ...} commands -with a version of @strong{MySQL} earlier than 3.20.7. +Added compatibility functions @code{CHECK} and @code{REFERENCES}. +@code{CHECK} is now a reserved word. @item -Fixed bug in locking code, which made @code{lock_test.pl} test fail. +Added @code{ALL} option to @code{GRANT} for better compatibility. (@code{GRANT} +is still a dummy function.) @item -New files @file{NEW} and @file{BUGS}. +Added partly-translated dutch messages. @item -Changed @file{select_test.c} and @file{insert_test.c} to include -@file{config.h}. +Fixed bug in @code{ORDER BY} and @code{GROUP BY} with @code{NULL} columns. @item -Added command @code{status} to @code{mysqladmin} for short logging. +Added function @code{last_insert_id()} to retrieve last @code{AUTO_INCREMENT} +value. This is intended for clients to ODBC that can't use the +@code{mysql_insert_id()} API function, but can be used by any client. @item -Increased maximum number of keys to 16 and maximum number of key parts to 15. +Added @code{--flush-logs} option to @code{mysqladmin}. @item -Use of sub keys. A key may now be a prefix of a string field. +Added command @code{STATUS} to @code{mysql}. @item -Added @code{-k} option to @code{mysqlshow}, to get key information for a table. +Fixed problem with @code{ORDER BY}/@code{GROUP BY} because of bug in @code{gcc}. @item -Added long options to @code{mysqldump}. +Fixed problem with @code{INSERT ... SELECT ... GROUP BY}. @end itemize -@node News-3.20.6, News-3.20.3, News-3.20.7, News-3.20.x -@appendixsubsec Changes in release 3.20.6 +@node News-3.21.10, News-3.21.9, News-3.21.11, News-3.21.x +@appendixsubsec Changes in release 3.21.10 @itemize @bullet @item -Portable to more systems because of MIT-pthreads, which will -be used automatically if @code{configure} cannot find a @code{-lpthreads} -library. +New @code{mysqlaccess}. @item -Added GNU-style long options to almost all programs. Test with -@code{@kbd{program} --help}. +@code{CREATE} now supports all ODBC types and the @code{mSQL} @code{TEXT} type. +All ODBC 2.5 functions are also supported (added @code{REPEAT}). This provides +better portability. @item -Some shared library support for Linux. +Added text types @code{TINYTEXT}, @code{TEXT}, @code{MEDIUMTEXT} and +@code{LONGTEXT}. These are actually @code{BLOB}types, but all searching is +done in case-insensitive fashion. @item -The FAQ is now in @file{.texi} format and is available in @file{.html}, -@file{.txt} and @file{.ps} formats. +All old @code{BLOB} fields are now @code{TEXT} fields. This only +changes that all searching on strings is done in case-sensitive fashion. +You must do an @code{ALTER TABLE} and change the field type to @code{BLOB} +if you want to have tests done in case-sensitive fashion. @item -Added new SQL function @code{RAND([init])}. +Fixed some @code{configure} issues. @item -Changed @code{sql_lex} to handle @code{\0} unquoted, but the client can't send -the query through the C API, because it takes a str pointer. -You must use @code{mysql_real_query()} to send the query. +Made the locking code a bit safer. Fixed very unlikely +deadlock situation. @item -Added API function @code{mysql_get_client_info()}. +Fixed a couple of bugs in the range optimizer. Now the new range benchmark +@code{test-select} works. +@end itemize + +@node News-3.21.9, News-3.21.8, News-3.21.10, News-3.21.x +@appendixsubsec Changes in release 3.21.9 + +@itemize @bullet @item -@code{mysqld} now uses the @code{N_MAX_KEY_LENGTH} from @file{nisam.h} as -the maximum allowed key length. +Added @code{--enable-unix-socket=pathname} option to @code{configure}. @item -The following now works: -@example -mysql> SELECT filter_nr,filter_nr FROM filter ORDER BY filter_nr; -@end example -Previously, this resulted in the error: -@code{Column: 'filter_nr' in order clause is ambiguous}. +Fixed a couple of portability problems with include files. @item -@code{mysql} now outputs @code{'\0'}, @code{'\t'}, @code{'\n'} and @code{'\\'} -when encountering ASCII 0, tab, newline or @code{'\'} while writing -tab-separated output. -This is to allow printing of binary data in a portable format. -To get the old behavior, use @code{-r} (or @code{--raw}). +Fixed bug in range calculation that could return empty +set when searching on multiple key with only one entry (very rare). @item -Added german error messages (60 of 80 error messages translated). +Most things ported to FSU Pthreads, which should allow @strong{MySQL} to +run on SCO. @xref{SCO}. +@end itemize + +@node News-3.21.8, News-3.21.7, News-3.21.9, News-3.21.x +@appendixsubsec Changes in release 3.21.8 + +@itemize @bullet @item -Added new API function @code{mysql_fetch_lengths(MYSQL_RES *)}, which -returns an array of column lengths (of type @code{uint}). +Works now in Solaris 2.6. @item -Fixed bug with @code{IS NULL} in @code{WHERE} clause. +Added handling of calculation of @code{SUM()} functions. +For example, you can now use @code{SUM(column)/COUNT(column)}. @item -Changed the optimizer a little to get better results when searching on a key -part. +Added handling of trigometric functions: +@code{PI()}, @code{ACOS()}, @code{ASIN()}, @code{ATAN()}, @code{COS()}, +@code{SIN()} and @code{TAN()}. @item -Added @code{SELECT} option @code{STRAIGHT_JOIN} to tell the optimizer that -it should join tables in the given order. +New languages: norwegian, norwegian-ny and portuguese. @item -Added support for comments starting with @code{'--'} in @file{mysql.cc} -(Postgres syntax). +Fixed parameter bug in @code{net_print()} in @file{procedure.cc}. @item -You can have @code{SELECT} expressions and table columns in a @code{SELECT} -which are not used in the group part. This makes it efficient to implement -lookups. The column that is used should be a constant for each group because -the value is calculated only once for the first row that is found for a group. -@example -mysql> SELECT id,lookup.text,sum(*) FROM test,lookup - WHERE test.id=lookup.id GROUP BY id; -@end example +Fixed a couple of memory leaks. @item -Fixed bug in @code{SUM(function)} (could cause a core dump). +Now allow also the old @code{SELECT ... INTO OUTFILE} syntax. @item -Changed @code{AUTO_INCREMENT} placement in the SQL query: -@example -INSERT into table (auto_field) values (0); -@end example -inserted 0, but it should insert an @code{AUTO_INCREMENT} value. +Fixed bug with @code{GROUP BY} and @code{SELECT} on key with many values. @item -@file{mysqlshow.c}: Added number of records in table. Had to change the -client code a little to fix this. +@code{mysql_fetch_lengths()} sometimes returned incorrect lengths when you used +@code{mysql_use_result()}. This affected at least some cases of +@code{mysqldump --quick}. @item -@code{mysql} now allows doubled @code{''} or @code{""} within strings for -embedded @code{'} or @code{"}. +Fixed bug in optimization of @code{WHERE const op field}. @item -New math functions: -@code{EXP()}, @code{LOG()}, @code{SQRT()}, @code{ROUND()}, @code{CEILING()}. +Fixed problem when sorting on @code{NULL} fields. +@item +Fixed a couple of 64-bit (Alpha) problems. +@item +Added @code{--pid-file=#} option to @code{mysqld}. +@item +Added date formatting to @code{FROM_UNIXTIME()}, originally by Zeev Suraski. +@item +Fixed bug in @code{BETWEEN} in range optimizer (Did only test = of the first +argument). +@item +Added machine-dependent files for MIT-pthreads i386-SCO. There is probably +more to do to get this to work on SCO 3.5. @end itemize -@node News-3.20.3, News-3.20.0, News-3.20.6, News-3.20.x -@appendixsubsec Changes in release 3.20.3 +@node News-3.21.7, News-3.21.6, News-3.21.8, News-3.21.x +@appendixsubsec Changes in release 3.21.7 @itemize @bullet @item -The @code{configure} source now compiles a thread-free client library -@code{-lmysqlclient}. This is the only library that needs to be linked -with client applications. When using the binary releases, you must -link with @code{-lmysql -lmysys -ldbug -lstrings} as before. -@item -New @code{readline} library from @code{bash-2.0}. +Changed @file{Makefile.am} to take advantage of Automake 1.2. @item -LOTS of small changes to @code{configure} and makefiles (and related source). +Added the beginnings of a benchmark suite. @item -It should now be possible to compile in another directory using -@code{VPATH}. Tested with GNU Make 3.75. +Added more secure password handling. @item -@code{safe_mysqld} and @code{mysql.server} changed to be more compatible -between the source and the binary releases. +Added new client function @code{mysql_errno()}, to get the error number of +the error message. This makes error checking in the client much easier. +This makes the new server incompatible with the 3.20.x server when running +without @code{--old-protocol}. The client code is backward compatible. +More information can be found in the @file{README} file! @item -@code{LIMIT} now takes one or two numeric arguments. -If one argument is given, it indicates the maximum number of rows in -a result. If two arguments are given, the first argument indicates the offset -of the first row to return, the second is the maximum number of rows. -With this it's easy to do a poor man's next page/previous page WWW -application. +Fixed some problems when using very long, illegal names. +@end itemize + +@node News-3.21.6, News-3.21.5, News-3.21.7, News-3.21.x +@appendixsubsec Changes in release 3.21.6 + +@itemize @bullet @item -Changed name of SQL function @code{FIELDS()} to @code{ELT()}. -Changed SQL function @code{INTERVALL()} to @code{INTERVAL()}. +Fixed more portability issues (incorrect @code{sigwait} and @code{sigset} +defines). @item -Made @code{SHOW COLUMNS} a synonym for @code{SHOW FIELDS}. -Added compatibility syntax @code{FRIEND KEY} to @code{CREATE TABLE}. In -@strong{MySQL}, this creates a non-unique key on the given columns. +@code{configure} should now be able to detect the last argument to +@code{accept()}. +@end itemize + +@node News-3.21.5, News-3.21.4, News-3.21.6, News-3.21.x +@appendixsubsec Changes in release 3.21.5 + +@itemize @bullet @item -Added @code{CREATE INDEX} and @code{DROP INDEX} as compatibility functions. -In @strong{MySQL}, @code{CREATE INDEX} only checks if the index exists and -issues an error if it doesn't exist. @code{DROP INDEX} always succeeds. +Should now work with FreeBSD 3.0 if used with +@file{FreeBSD-3.0-libc_r-1.0.diff}, +which can be found at @uref{http://www.mysql.com/Downloads/Patches/}. @item -@file{mysqladmin.c}: added client version to version information. +Added new option @code{-O tmp_table_size=#} to @code{mysqld}. @item -Fixed core dump bug in @code{sql_acl} (core on new connection). +New function @code{FROM_UNIXTIME(timestamp)} which returns a date string in +'YYYY-MM-DD HH:MM:DD' format. @item -Removed @code{host}, @code{user} and @code{db} tables from database @code{test} -in the distribution. +New function @code{SEC_TO_TIME(seconds)} which returns a string in +'HH:MM:SS' format. @item -@code{FIELD_TYPE_CHAR} can now be signed (-128 - 127) or unsigned (0 - 255) -Previously, it was always unsigned. +New function @code{SUBSTRING_INDEX()}, originally by Zeev Suraski. +@end itemize + +@node News-3.21.4, News-3.21.3, News-3.21.5, News-3.21.x +@appendixsubsec Changes in release 3.21.4 + +@itemize @bullet @item -Bug fixes in @code{CONCAT()} and @code{WEEKDAY()}. +Should now configure and compile on OSF1 4.0 with the DEC compiler. @item -Changed a lot of source to get @code{mysqld} to be compiled with SunPro -compiler. +Configuration and compilation on BSD/OS 3.0 works, but due to some bugs in +BSD/OS 3.0, @code{mysqld} doesn't work on it yet. @item -SQL functions must now have a @code{'('} immediately after the function name -(no intervening space). -For example, @code{'user('} is regarded as beginning a function call, and -@code{'user ('} is regarded as an identifier @code{user} followed by a -@code{'('}, not as a function call. +Configuration and compilation on FreeBSD 3.0 works, but I couldn't get +@code{pthread_create} to work. @end itemize -@node News-3.20.0, , News-3.20.3, News-3.20.x -@appendixsubsec Changes in release 3.20.0 +@node News-3.21.3, News-3.21.2, News-3.21.4, News-3.21.x +@appendixsubsec Changes in release 3.21.3 @itemize @bullet @item -The source distribution is done with @code{configure} and Automake. -It will make porting much easier. The @code{readline} library is included -in the distribution. +Added reverse check lookup of hostnames to get better security. @item -Separate client compilation: -the client code should be very easy to compile on systems which -don't have threads. +Fixed some possible buffer overflows if filenames that are too long are used. @item -The old Perl interface code is automatically compiled and installed. -Automatic compiling of @code{DBD} will follow when the new @code{DBD} code -is ported. +@code{mysqld} doesn't accept hostnames that start with digits followed by a +@code{'.'}, because the hostname may look like an IP number. @item -Dynamic language support: @code{mysqld} can now be started with Swedish -or English (default) error messages. +Added @code{--skip-networking} option to @code{mysqld}, to only allow socket +connections. (This will not work with MIT-pthreads!) @item -New functions: @code{INSERT()}, @code{RTRIM()}, @code{LTRIM()} and -@code{FORMAT()}. +Added check of too long table names for alias. @item -@code{mysqldump} now works correctly for all field types (even -@code{AUTO_INCREMENT}). The format for @code{SHOW FIELDS FROM tbl_name} -is changed so the @code{Type} column contains information suitable for -@code{CREATE TABLE}. In previous releases, some @code{CREATE TABLE} -information had to be patched when re-creating tables. +Added check if database name is okay. @item -Some parser bugs from 3.19.5 (@code{BLOB} and @code{TIMESTAMP}) are corrected. -@code{TIMESTAMP} now returns different date information depending on its -create length. +Added check if too long table names. @item -Changed parser to allow a database, table or field name to -start with a number or @code{'_'}. +Removed incorrect @code{free()} that killed the server on +@code{CREATE DATABASE} or @code{DROP DATABASE}. @item -All old C code from Unireg changed to C++ and cleaned up. This makes -the daemon a little smaller and easier to understand. +Changed some @code{mysqld} @code{-O} options to better names. @item -A lot of small bug fixes done. +Added @code{-O join_cache_size=#} option to @code{mysqld}. @item -New @file{INSTALL} files (not final version) and some information regarding porting. -@end itemize - -@node News-3.19.x, , News-3.20.x, News -@appendixsec Changes in release 3.19.x - -Version 3.19 is quite old now, and should be avoided if possible. This -information is kept here for historical purposes only. - -@menu -* News-3.19.5:: Changes in release 3.19.5 -* News-3.19.4:: Changes in release 3.19.4 -* News-3.19.3:: Changes in release 3.19.3 -@end menu - -@node News-3.19.5, News-3.19.4, News-3.19.x, News-3.19.x -@appendixsubsec Changes in release 3.19.5 - -@itemize @bullet +Added @code{-O max_join_size=#} option to @code{mysqld}, to be able to set a +limit how big queries (in this case big = slow) one should be able to handle +without specifying @code{SET OPTION SQL_BIG_SELECTS=1}. A # = is about 10 +examined records. The default is ``unlimited''. @item -Some new functions, some more optimization on joins. +When comparing a @code{TIME}, @code{DATE}, @code{DATETIME} or @code{TIMESTAMP} +column to a constant, the constant is converted to a time value before +performing the comparison. +This will make it easier to get ODBC (particularly Access97) to work with +the above types. It should also make dates easier to use and the comparisons +should be quicker than before. @item -Should now compile clean on Linux (2.0.x). +Applied patch from Jochen Wiedmann that allows @code{query()} in +@code{mysqlperl} to take a query with @code{\0} in it. @item -Added functions @code{DATABASE()}, @code{USER()}, @code{POW()}, -@code{LOG10()} (needed for ODBC). +Storing a timestamp with a 2-digit year (@code{YYMMDD}) didn't work. @item -In a @code{WHERE} with an @code{ORDER BY} on fields from only one table, -the table is now preferred as first table in a multi-join. +Fix that timestamp wasn't automatically updated if set in an @code{UPDATE} +clause. @item -@code{HAVING} and @code{IS NULL} or @code{IS NOT NULL} now works. +Now the automatic timestamp field is the FIRST timestamp field. @item -A group on one column and a sort on a group function (@code{SUM()}, -@code{AVG()}...) didn't work together. Fixed. +@code{SELECT * INTO OUTFILE}, which didn't correctly if the outfile already +existed. @item -@code{mysqldump}: Didn't send password to server. +@code{mysql} now shows the thread ID when starting or doing a reconnect. +@item +Changed the default sort buffer size from 2M to 1M. @end itemize -@node News-3.19.4, News-3.19.3, News-3.19.5, News-3.19.x -@appendixsubsec Changes in release 3.19.4 +@node News-3.21.2, News-3.21.0, News-3.21.3, News-3.21.x +@appendixsubsec Changes in release 3.21.2 @itemize @bullet @item -Fixed horrible locking bug when inserting in one thread and reading -in another thread. +The range optimizer is coded, but only 85% tested. It can be enabled with +@code{--new}, but it crashes core a lot yet... @item -Fixed one-off decimal bug. 1.00 was output as 1.0. +More portable. Should compile on AIX and alpha-digital. +At least the @code{isam} library should be relatively 64-bit clean. @item -Added attribute @code{'Locked'} to process list as info if a query is -locked by another query. +New @code{isamchk} which can detect and fix more problems. @item -Fixed full magic timestamp. Timestamp length may now be 14, 12, 10, 8, 6, 4 -or 2 bytes. +New options for @code{isamlog}. @item -Sort on some numeric functions could sort incorrectly on last number. +Using new version of Automake. @item -@code{IF(arg,syntax_error,syntax_error)} crashed. +Many small portability changes (from the AIX and alpha-digital port) +Better checking of pthread(s) library. @item -Added functions @code{CEILING()}, @code{ROUND()}, @code{EXP()}, @code{LOG()} and @code{SQRT()}. +czech error messages by @email{snajdr@@pvt.net}. @item -Enhanced @code{BETWEEN} to handle strings. -@end itemize - -@node News-3.19.3, , News-3.19.4, News-3.19.x -@appendixsubsec Changes in release 3.19.3 - -@itemize @bullet +Decreased size of some buffers to get fewer problems on systems with little +memory. Also added more checks to handle ``out of memory'' problems. @item -Fixed @code{SELECT} with grouping on @code{BLOB} columns not to return -incorrect @code{BLOB} info. Grouping, sorting and distinct on @code{BLOB} -columns will not yet work as -expected (probably it will group/sort by the first 7 characters in the -@code{BLOB}). Grouping on formulas with a fixed string size (use @code{MID()} -on a @code{BLOB}) should work. +@code{mysqladmin}: you can now do @code{mysqladmin kill 5,6,7,8} to kill +multiple threads. @item -When doing a full join (no direct keys) on multiple tables with @code{BLOB} -fields, the @code{BLOB} was garbage on output. +When the maximum connection limit is reached, one extra connection by a user +with the @strong{PROCESS_ACL} privilege is granted. @item -Fixed @code{DISTINCT} with calculated columns. -@end itemize - -@cindex bugs, known -@cindex errors, known -@cindex design, issues -@cindex known errors -@node Bugs, TODO, News, Top -@appendix Known errors and design deficiencies in MySQL - -The following problems are known and have a very high priority to get -fixed: - -@itemize @bullet +Added @code{-O backlog=#} option to @code{mysqld}. @item -@code{ANALYZE TABLE} on a BDB table may in some case make the table -unusable until one has restarted @code{mysqld}. When this happens you will -see errors like the following in the @strong{MySQL} error file: - -@example -001207 22:07:56 bdb: log_flush: LSN past current end-of-log -@end example - +Increased maximum packet size from 512K to 1024K for client. @item -Don't execute @code{ALTER TABLE} on a @code{BDB} table on which you are -running not completed multi-statement transactions. (The transaction -will probably be ignored). - +Almost all of the function code is now tested in the internal test suite. @item -@code{ANALYZE TABLE}, @code{OPTIMIZE TABLE} and @code{REPAIR TABLE} may -cause problems on tables for which you are using @code{INSERT DELAYED}. - +@code{ALTER TABLE} now returns warnings from field conversions. +@item +Port changed to 3306 (got it reserved from ISI). +@item +Added a fix for Visual FoxBase so that any schema name from a table +specification is automatically removed. +@item +New function @code{ASCII()}. @item -Doing a @code{LOCK TABLE ..} and @code{FLUSH TABLES ..} doesn't -guarantee that there isn't a half-finished transaction in progress on the -table. - +Removed function @code{BETWEEN(a,b,c)}. Use the standard ANSI +syntax instead: @code{expr BETWEEN expr AND expr}. @item -BDB tables are a bit slow to open. If you have many BDB tables in a -database, it will take a long time to use the @code{mysql} client on the -database if you are not using the @code{-A} option or if you are using -@code{rehash}. This is especially notable when you have a big table -cache. - +@strong{MySQL} no longer has to use an extra temporary table when sorting +on functions or @code{SUM()} functions. @item -Th current replication protocol cannot deal with @code{LOAD DATA INFILE} -and line terminator characters of more than 1 character. +Fixed bug that you couldn't use @code{tbl_name.field_name} in @code{UPDATE}. +@item +Fixed @code{SELECT DISTINCT} when using 'hidden group'. For example: +@example +mysql> SELECT DISTINCT MOD(some_field,10) FROM test + GROUP BY some_field; +@end example +Note: @code{some_field} is normally in the @code{SELECT} part. ANSI SQL should +require it. @end itemize -The following problems are known and will be fixed in due time: +@node News-3.21.0, , News-3.21.2, News-3.21.x +@appendixsubsec Changes in release 3.21.0 @itemize @bullet @item -For the moment @code{MATCH} only works with @code{SELECT} statements. +New reserved words used: @code{INTERVAL}, @code{EXPLAIN}, @code{READ}, +@code{WRITE}, @code{BINARY}. @item -When using @code{SET CHARACTER SET}, one can't use translated -characters in database, table and column names. +Added ODBC function @code{CHAR(num,...)}. @item -@code{DELETE FROM merge_table} used without a @code{WHERE} -will only clear the mapping for the table, not delete everything in the -mapped tables +New operator @code{IN}. This uses a binary search to find a match. @item -You cannot build in another directory when using -MIT-pthreads. Because this requires changes to MIT-pthreads, we are not -likely to fix this. +New command @code{LOCK TABLES tbl_name [AS alias] @{READ|WRITE@} ...} @item -@code{BLOB} values can't ``reliably'' be used in @code{GROUP BY} or -@code{ORDER BY} or @code{DISTINCT}. Only the first @code{max_sort_length} -bytes (default 1024) are used when comparing @code{BLOB}bs in these cases. -This can be changed with the @code{-O max_sort_length} option to -@code{mysqld}. A workaround for most cases is to use a substring: -@code{SELECT DISTINCT LEFT(blob,2048) FROM tbl_name}. +Added @code{--log-update} option to @code{mysqld}, to get a log suitable for +incremental updates. @item -Calculation is done with @code{BIGINT} or @code{DOUBLE} (both are -normally 64 bits long). It depends on the function which precision one -gets. The general rule is that bit functions are done with @code{BIGINT} -precision, @code{IF}, and @code{ELT()} with @code{BIGINT} or @code{DOUBLE} -precision and the rest with @code{DOUBLE} precision. One should try to -avoid using bigger unsigned long long values than 63 bits -(9223372036854775807) for anything else than bit fields! +New command @code{EXPLAIN SELECT ...} to get information about how the +optimizer will do the join. @item -All string columns, except @code{BLOB} and @code{TEXT} columns, automatically -have all trailing spaces removed when retrieved. For @code{CHAR} types this -is okay, and may be regarded as a feature according to ANSI SQL92. The bug is -that in @strong{MySQL}, @code{VARCHAR} columns are treated the same way. +For easier client code, the client should no longer use +@code{FIELD_TYPE_TINY_BLOB}, @code{FIELD_TYPE_MEDIUM_BLOB}, +@code{FIELD_TYPE_LONG_BLOB} or @code{FIELD_TYPE_VAR_STRING} (as +previously returned by @code{mysql_list_fields}). You should instead only use +@code{FIELD_TYPE_BLOB} or @code{FIELD_TYPE_STRING}. If you want exact +types, you should use the command @code{SHOW FIELDS}. @item -You can only have up to 255 @code{ENUM} and @code{SET} columns in one table. +Added varbinary syntax: @code{0x######} which can be used as a string +(default) or a number. @item -@code{safe_mysqld} re-directs all messages from @code{mysqld} to the -@code{mysqld} log. One problem with this is that if you execute -@code{mysqladmin refresh} to close and reopen the log, -@code{stdout} and @code{stderr} are still redirected to the old log. -If you use @code{--log} extensively, you should edit @code{safe_mysqld} to -log to @file{'hostname'.err} instead of @file{'hostname'.log} so you can -easily reclaim the space for the old log by deleting the old one and -executing @code{mysqladmin refresh}. +@code{FIELD_TYPE_CHAR} is renamed to @code{FIELD_TYPE_TINY}. @item -In the @code{UPDATE} statement, columns are updated from left to right. -If you refer to an updated column, you will get the updated value instead of the -original value. For example: -@example -mysql> UPDATE tbl_name SET KEY=KEY+1,KEY=KEY+1; -@end example -will update @code{KEY} with @code{2} instead of with @code{1}. +Changed all fields to C++ classes. @item -You can't use temporary tables more than once in the same query. -For example, the following doesn't work. - -@example -select * from temporary_table, temporary_table as t2; -@end example - +Removed FORM struct. @item -@code{RENAME} doesn't work with @code{TEMPORARY} tables. - +Fields with @code{DEFAULT} values no longer need to be @code{NOT NULL}. @item -The optimizer may handle @code{DISTINCT} differently if you are using -'hidden' columns in a join or not. In a join, hidden columns are -counted as part of the result (even if they are not shown) while in -normal queries hidden columns doesn't participate in the @code{DISTINCT} -comparison. We will probably change this in the future to never compare -the hidden columns when executing @code{DISTINCT} - -An example of this is: - +New field types: +@table @code +@item ENUM +A string which can take only a couple of defined values. The value is +stored as a 1-3 byte number that is mapped automatically to a string. +This is sorted according to string positions! +@item SET +A string which may have one or many string values separated with ','. +The string is stored as a 1-, 2-, 3-, 4- or 8-byte number where each bit stands +for a specific set member. This is sorted according to the unsigned value +of the stored packed number. +@end table +@item +Now all function calculation is done with @code{double} or @code{long long}. +This will provide the full 64-bit range with bit functions and fix some +conversions that previously could result in precision losses. One should +avoid using @code{unsigned long long} columns with full 64-bit range +(numbers bigger than 9223372036854775807) because calculations are done +with @code{signed long long}. +@item +@code{ORDER BY} will now put @code{NULL} field values first. @code{GROUP BY} +will also work with @code{NULL} values. +@item +Full @code{WHERE} with expressions. +@item +New range optimizer that can resolve ranges when some keypart prefix is +constant. Example: @example -SELECT DISTINCT mp3id FROM band_downloads WHERE userid = 9 ORDER BY id -DESC; - -and - -SELECT DISTINCT band_downloads.mp3id, FROM band_downloads,band_mp3 -WHERE band_downloads.userid = 9 AND band_mp3.id = band_downloads.mp3id -ORDER BY band_downloads.id DESC; +mysql> SELECT * FROM tbl_name + WHERE key_part_1="customer" + AND key_part_2>=10 AND key_part_2<=10; @end example +@end itemize -In the second case you may in @strong{MySQL} 3.23.x get two identical rows -in the result set (because the hidden 'id' column may differ). +@node News-3.20.x, News-3.19.x, News-3.21.x, News +@appendixsec Changes in release 3.20.x -Note that the this only happens for queries where you don't have the -ORDER BY columns in the result, something that is you are not allowed -to do in ANSI SQL. +Version 3.20 is quite old now, and should be avoided if possible. This +information is kept here for historical purposes only. -@item -Because @strong{MySQL} allows you to work with table types that doesn't -support transactions (and thus can't @code{rollback} data) some things -behaves a little different in @strong{MySQL} than in other SQL servers: -(This is just to ensure that @strong{MySQL} never need to do a rollback -for a SQL command). This may be a little awkward at times as column -Because @strong{MySQL} allows you to work with table types that don't -support transactions, and thus can't @code{rollback} data, some things -behave a little differently in @strong{MySQL} than in other SQL servers. -This is just to ensure that @strong{MySQL} never need to do a rollback -for a SQL command. This may be a little awkward at times as column -values must be checked in the application, but this will actually give -you a nice speed increase as it allows @strong{MySQL} to do some -optimizations that otherwise would be very hard to do. +Changes from 3.20.18 to 3.20.32b are not documented here because the +3.21 release branched here. And the relevant changes are also +documented as changes to the 3.21 version. -If you set a column to an incorrect value, @strong{MySQL} will, instead of -doing a rollback, store the @code{best possible value} in the column: +@menu +* News-3.20.18:: Changes in release 3.20.18 +* News-3.20.17:: Changes in release 3.20.17 +* News-3.20.16:: Changes in release 3.20.16 +* News-3.20.15:: Changes in release 3.20.15 +* News-3.20.14:: Changes in release 3.20.14 +* News-3.20.13:: Changes in release 3.20.13 +* News-3.20.11:: Changes in release 3.20.11 +* News-3.20.10:: Changes in release 3.20.10 +* News-3.20.9:: Changes in release 3.20.9 +* News-3.20.8:: Changes in release 3.20.8 +* News-3.20.7:: Changes in release 3.20.7 +* News-3.20.6:: Changes in release 3.20.6 +* News-3.20.3:: Changes in release 3.20.3 +* News-3.20.0:: Changes in releases 3.20.0 +@end menu + +@node News-3.20.18, News-3.20.17, News-3.20.x, News-3.20.x +@appendixsubsec Changes in release 3.20.18 @itemize @bullet @item -If you try to store a value outside the range in a numerical column, -@strong{MySQL} will instead store the smallest or biggest possible value in -the column. +Added @code{-p#} (remove @code{#} directories from path) to @code{isamlog}. +All files are written with a relative path from the database directory +Now @code{mysqld} shouldn't crash on shutdown when using the +@code{--log-isam} option. @item -If you try to store a string that doesn't start with a number into a -numerical column, @strong{MySQL} will store 0 into it. +New @code{mysqlperl} version. It is now compatible with @code{msqlperl-0.63}. @item -If you try to store @code{NULL} into a column that doesn't take -@code{NULL} values, @strong{MySQL} will store 0 or @code{''} (empty -string) in it instead. (This behavior can, however, be changed with the --DDONT_USE_DEFAULT_FIELDS compile option). +New @code{DBD} module available at @uref{http://www.mysql.com/Downloads/Contrib/} +site. @item -@strong{MySQL} allows you to store some wrong date values into -@code{DATE} and @code{DATETIME} columns. (Like 2000-02-31 or 2000-02-00). -If the date is totally wrong, @strong{MySQL} will store the special -0000-00-00 date value in the column. +Added group function @code{STD()} (standard deviation). @item -If you set an @code{enum} to an unsupported value, it will be set to -the error value 'empty string', with numeric value 0. -@end itemize - +The @code{mysqld} server is now compiled by default without debugging +information. This will make the daemon smaller and faster. @item -If you execute a @code{PROCEDURE} on a query that returns an empty set, -in some cases the @code{PROCEDURE} will not transform the columns. +Now one usually only has to specify the @code{--basedir} option to +@code{mysqld}. All other paths are relative in a normal installation. @item -Creation of a table of type @code{MERGE} doesn't check if the underlying -tables are of compatible types. +@code{BLOB} columns sometimes contained garbage when used with a @code{SELECT} +on more than one table and @code{ORDER BY}. @item -@strong{MySQL} can't yet handle @code{NaN}, @code{-Inf} and @code{Inf} -values in double. Using these will cause problems when trying to export -and import data. We should as an intermediate solution change @code{NaN} to -@code{NULL} (if possible) and @code{-Inf} and @code{Inf} to the -Minimum respective maximum possible @code{double} value. +Fixed that calculations that are not in @code{GROUP BY} work as expected +(ANSI SQL extension). +Example: +@example +mysql> SELECT id,id+1 FROM table GROUP BY id; +@end example @item -@code{LIMIT} on negative numbers are treated as big positive numbers. +The test of using @code{MYSQL_PWD} was reversed. Now @code{MYSQL_PWD} is +enabled as default in the default release. +@item +Fixed conversion bug which caused @code{mysqld} to core dump with +Arithmetic error on Sparc-386. +@item +Added @code{--unbuffered} option to @code{mysql}, for new @code{mysqlaccess}. +@item +When using overlapping (unnecessary) keys and join over many tables, +the optimizer could get confused and return 0 records. @end itemize -The following are known bugs in earlier versions of @strong{MySQL}: +@node News-3.20.17, News-3.20.16, News-3.20.18, News-3.20.x +@appendixsubsec Changes in release 3.20.17 @itemize @bullet @item -You can get a hung thread if you do a @code{DROP TABLE} on a table that is -one among many tables that is locked with @code{LOCK TABLES}. - +You can now use @code{BLOB} columns and the functions @code{IS NULL} and +@code{IS NOT NULL} in the @code{WHERE} clause. @item -In the following case you can get a core dump: -@enumerate +All communication packets and row buffers are now allocated dynamically +on demand. The default value of @code{max_allowed_packet} is now 64K for +the server and 512K for the client. This is mainly used to catch +incorrect packets that could trash all memory. The server limit may be +changed when it is started. @item -Delayed insert handler has pending inserts to a table. +Changed stack usage to use less memory. @item -@code{LOCK table} with @code{WRITE} +Changed @code{safe_mysqld} to check for running daemon. @item -@code{FLUSH TABLES} -@end enumerate - +The @code{ELT()} function is renamed to @code{FIELD()}. The new +@code{ELT()} function returns a value based on an index: @code{FIELD()} +is the inverse of @code{ELT()} Example: @code{ELT(2,"A","B","C")} returns +@code{"B"}. @code{FIELD("B","A","B","C")} returns @code{2}. @item -Before @strong{MySQL} Version 3.23.2 an @code{UPDATE} that updated a key with -a @code{WHERE} on the same key may have failed because the key was used to -search for records and the same row may have been found multiple times: - -@example -UPDATE tbl_name SET KEY=KEY+1 WHERE KEY > 100; -@end example - -A workaround is to use: - -@example -mysql> UPDATE tbl_name SET KEY=KEY+1 WHERE KEY+0 > 100; -@end example - -This will work because @strong{MySQL} will not use index on expressions in -the @code{WHERE} clause. +@code{COUNT(field)}, where @code{field} could have a @code{NULL} value, now +works. @item -Before @strong{MySQL} Version 3.23, all numeric types where treated as fixed-point -fields. That means you had to specify how many decimals a floating-point -field shall have. All results were returned with the correct number of -decimals. -@end itemize - -For platform-specific bugs, see the sections about compiling and porting. - -@cindex ToDo list for MySQL -@node TODO, Porting, Bugs, Top -@appendix MySQL and the future (The TODO) - -@menu -* TODO MySQL 4.0:: Things that should be in Version 4.0 -* TODO future:: Things that must be done in the near future -* TODO sometime:: Things that have to be done sometime -* TODO unplanned:: Some things we don't have any plans to do -@end menu - -This appendix lists the features that we plan to implement in @strong{MySQL}. - -Everything in this list is approximately in the order it will be done. If you -want to affect the priority order, please register a license or support us and -tell us what you want to have done more quickly. @xref{Licensing and Support}. - -The plan is that we in the future will support the full ANSI SQL99 -standard, but with a lot of useful extensions. The challenge is to do -this without sacrifying the speed or compromise the code. - -@node TODO MySQL 4.0, TODO future, TODO, TODO -@appendixsec Things that should be in 4.0 - -We plan to make @strong{MySQL} Version 4.0 a ``quick'' release where we only -add some new stuff to enable others to help us with developing new features -into Version 4.1. The @strong{MySQL} 4.0 version should only take us about -a month to make after which we want to stabilize it and start working on -Version 4.1. Version 4.0 should have the following new features: - -The news section for 4.0 includes a list of the features we have already -implemented in the 4.0 tree. @xref{News-4.0.x}. - -@itemize @bullet +A couple of bugs fixed in @code{SELECT ... GROUP BY}. @item -New table definition file format (@code{.frm} files) This will enable us -to not run out of bits when adding more table options. One will still -be able to use the old @code{.frm} file format with 4.0. All newly created -tables will, however, use the new format. - -The new file format will enable us to add new column types, more options -for keys and @code{FOREIGN KEY} support. +Fixed memory overrun bug in @code{WHERE} with many unoptimizable brace levels. @item -@code{mysqld} as a library. This will have the same interface as the -standard @strong{MySQL} client (with an extra function to just set up -startup parameters) but will be faster (no TCP/IP or socket overhead), -smaller and much easier to use for embedded products. - -One will be able to define at link time if one wants to use the -client/server model or a stand-alone application just by defining which -library to link with. - -The @code{mysqld} will support all standard @strong{MySQL} features and -one can use it in a threaded client to run different queries in each -thread. +Fixed some small bugs in the grant code. @item -Replication should work with @code{RAND()} and user variables @code{@@var}. +If hostname isn't found by @code{get_hostname}, only the IP is checked. +Previously, you got @code{Access denied}. @item -Online backup with very low performance penalty. The online backup will -make it easy to add a new replication slave without taking down the -master. +Inserts of timestamps with values didn't always work. @item -@code{DELETE FROM table_name} will return the number of deleted rows. For -fast execution one should use @code{TRUNCATE table_name}. +@code{INSERT INTO ... SELECT ... WHERE} could give the error +@code{Duplicated field}. @item -Allow @code{DELETE} on @code{MyISAM} tables to use the record cache. -To do this, we need to update the threads record cache when we update -the @code{.MYD} file. +Added some tests to @code{safe_mysqld} to make it ``safer''. @item -Better replication. +@code{LIKE} was case sensitive in some places and case insensitive in others. +Now @code{LIKE} is always case insensitive. @item -More functions for full-text search. -@xref{Fulltext Features to Appear in MySQL 4.0}. +@file{mysql.cc}: Allow @code{'#'} anywhere on the line. @item -Character set casts and syntax for handling multiple character sets. +New command @code{SET OPTION SQL_SELECT_LIMIT=#}. See the FAQ for more details. @item -Allow users to change startup options without taking down the server. +New version of the @code{mysqlaccess} script. @item -Help for all commands from the client. +Change @code{FROM_DAYS()} and @code{WEEKDAY()} to also take a full +@code{TIMESTAMP} or @code{DATETIME} as argument. Before they only took a +number of type @code{YYYYMMDD} or @code{YYMMDD}. @item -Secure connections (with SSL). +Added new function @code{UNIX_TIMESTAMP(timestamp_column)}. +@end itemize + +@node News-3.20.16, News-3.20.15, News-3.20.17, News-3.20.x +@appendixsubsec Changes in release 3.20.16 + +@itemize @bullet @item -Extend the optimizer to be able to optimize some @code{ORDER BY key_name DESC} -queries. +More changes in MIT-pthreads to get them safer. Fixed also some link +bugs at least in SunOS. @item -@code{SHOW COLUMNS FROM table_name} (used by @code{mysql} client to allow -expansions of column names) should not open the table, but only the -definition file. This will require less memory and be much faster. +Changed @code{mysqld} to work around a bug in MIT-pthreads. This makes multiple +small @code{SELECT} operations 20 times faster. Now @code{lock_test.pl} should +work. @item -New key cache +Added @code{mysql_FetchHash(handle)} to @code{mysqlperl}. @item -When using @code{SET CHARACTER SET} we should translate the whole query -at once and not only strings. This will enable users to use the translated -characters in database, table and column names. +The @code{mysqlbug} script is now distributed built to allow for reporting +bugs that appear during the build with it. @item -Add a portable interface over @code{gethostbyaddr_r()} so that we can change -@code{ip_to_hostname()} to not block other threads while doing DNS lookups. +Changed @file{libmysql.c} to prefer @code{getpwuid()} instead of +@code{cuserid()}. @item -Add @code{record_in_range()} method to @code{MERGE} tables to be -able to choose the right index when there is many to choose from. We should -also extend the info interface to get the key distribution for each index, -of @code{analyze} is run on all sub tables. +Fixed bug in @code{SELECT} optimizer when using many tables with the same +column used as key to different tables. @item -@code{SET SQL_DEFAULT_TABLE_TYPE=[MyISAM | INNODB | BDB | HEAP]}. +Added new latin2 and Russian KOI8 character tables. +@item +Added support for a dummy @code{GRANT} command to satisfy Powerbuilder. @end itemize -@node TODO future, TODO sometime, TODO MySQL 4.0, TODO -@appendixsec Things that must be done in the real near future +@node News-3.20.15, News-3.20.14, News-3.20.16, News-3.20.x +@appendixsubsec Changes in release 3.20.15 @itemize @bullet @item -Fail safe replication. +Fixed fatal bug @code{packets out of order} when using MIT-pthreads. @item -Subqueries. -@code{select id from t where grp in (select grp from g where u > 100)} +Removed possible loop when a thread waits for command from client +and @code{fcntl()} fails. Thanks to Mike Bretz for finding this bug. @item -Derived tables. -@example -select a.col1, b.col2 from (select max(col1) as col1 from root_table ) a, -other_table b where a.col1=b.col1 -@end example - -This could be done by automatically creating temporary tables for the -derived tables for the duration of the query. +Changed alarm loop in @file{mysqld.cc} because shutdown didn't always +succeed in Linux. @item -Add @code{PREPARE} of statements and sending of parameters to @code{mysqld}. +Removed use of @code{termbits} from @file{mysql.cc}. This conflicted with +@code{glibc} 2.0. @item -Extend the server/client protocol to support warnings. +Fixed some syntax errors for at least BSD and Linux. @item -Add options to the server/protocol protocol to get progress notes -for long running commands. +Fixed bug when doing a @code{SELECT} as superuser without a database. @item -Add database and real table name (in case of alias) to the MYSQL_FIELD -structure. +Fixed bug when doing @code{SELECT} with group calculation to outfile. +@end itemize + +@node News-3.20.14, News-3.20.13, News-3.20.15, News-3.20.x +@appendixsubsec Changes in release 3.20.14 + +@itemize @bullet @item -Don't allow more than a defined number of threads to run MyISAM recover -at the same time. +If one gives @code{-p} or @code{--password} option to @code{mysql} without +an argument, the user is solicited for the password from the tty. @item -Change @code{INSERT ... SELECT} to optionally use concurrent inserts. +Added default password from @code{MYSQL_PWD} (by Elmar Haneke). @item -Implement @code{RENAME DATABASE}. To make this safe for all table handlers, -it should work as follows: -@itemize @bullet +Added command @code{kill} to @code{mysqladmin} to kill a specific +@strong{MySQL} thread. @item -Create the new database. +Sometimes when doing a reconnect on a down connection this succeeded +first on second try. @item -For every table do a rename of the table to another database, as -we do with the @code{RENAME} command. +Fixed adding an @code{AUTO_INCREMENT} key with @code{ALTER_TABLE}. @item -Drop the old database. -@end itemize +@code{AVG()} gave too small value on some @code{SELECT}s with +@code{GROUP BY} and @code{ORDER BY}. @item -Return the original field types() when doing @code{SELECT MIN(column) -... GROUP BY}. +Added new @code{DATETIME} type (by Giovanni Maruzzelli +@email{maruzz@@matrice.it}). @item -Multiple result sets. +Fixed that define @code{DONT_USE_DEFAULT_FIELDS} works. @item -Change the protocol to allow binary transfer of values. To do this -efficiently, we need to add an API to allow binding of variables. +Changed to use a thread to handle alarms instead of signals on Solaris to +avoid race conditions. @item -Make it possible to specify @code{long_query_time} with a granularity -in microseconds. +Fixed default length of signed numbers. (George Harvey +@email{georgeh@@pinacl.co.uk}.) @item -Add a configurable prompt to the @code{mysql} command line client, with -options like database in use, time and date... +Allow anything for @code{CREATE INDEX}. +@item +Add prezeros when packing numbers to @code{DATE}, @code{TIME} and +@code{TIMESTAMP}. +@item +Fixed a bug in @code{OR} of multiple tables (gave empty set). +@item +Added many patches to MIT-pthreads. This fixes at least one lookup bug. +@end itemize + +@node News-3.20.13, News-3.20.11, News-3.20.14, News-3.20.x +@appendixsubsec Changes in release 3.20.13 + +@itemize @bullet @item -Add range checking to @code{MERGE} tables. +Added ANSI SQL94 @code{DATE} and @code{TIME} types. @item -Link the @code{myisampack} code into the server. +Fixed bug in @code{SELECT} with @code{AND}-@code{OR} levels. @item -Port of @strong{MySQL} to BeOS. +Added support for Slovenian characters. The @file{Contrib} directory contains +source and instructions for adding other character sets. @item -Port of the @strong{MySQL} clients to LynxOS. +Fixed bug with @code{LIMIT} and @code{ORDER BY}. @item -Add a temporary key buffer cache during @code{INSERT/DELETE/UPDATE} so that we -can gracefully recover if the index file gets full. +Allow @code{ORDER BY} and @code{GROUP BY} on items that aren't in the +@code{SELECT} list. +(Thanks to Wim Bonis @email{bonis@@kiss.de}, for pointing this out.) @item -If you perform an @code{ALTER TABLE} on a table that is symlinked to another -disk, create temporary tables on this disk. +Allow setting of timestamp values in @code{INSERT}. @item -Implement a @code{DATE/DATETIME} type that handles time zone information -properly, so that dealing with dates in different time zones is easier. +Fixed bug with @code{SELECT ... WHERE ... = NULL}. @item -FreeBSD and MIT-pthreads; Do sleeping threads take CPU? +Added changes for @code{glibc} 2.0. To get @code{glibc} to work, you should +add the @file{gibc-2.0-sigwait-patch} before compiling @code{glibc}. @item -Check if locked threads take any CPU. +Fixed bug in @code{ALTER TABLE} when changing a @code{NOT NULL} field to +allow @code{NULL} values. @item -Fix configure so that one can compile all libraries (like @code{MyISAM}) -without threads. +Added some ANSI92 synonyms as field types to @code{CREATE TABLE}. +@code{CREATE TABLE} now allows @code{FLOAT(4)} and @code{FLOAT(8)} to mean +@code{FLOAT} and @code{DOUBLE}. @item -Add an option to periodically flush key pages for tables with delayed keys -if they haven't been used in a while. +New utility program @code{mysqlaccess} by @email{Yves.Carlier@@rug.ac.be}. +This program shows the access rights for a specific user and the grant +rows that determine this grant. @item -Allow join on key parts (optimization issue). +Added @code{WHERE const op field} (by @email{bonis@@kiss.de}). +@end itemize + +@node News-3.20.11, News-3.20.10, News-3.20.13, News-3.20.x +@appendixsubsec Changes in release 3.20.11 + +@itemize @bullet @item -@code{INSERT SQL_CONCURRENT} and @code{mysqld --concurrent-insert} to do -a concurrent insert at the end of the file if the file is read-locked. +When using @code{SELECT ... INTO OUTFILE}, all temporary tables are ISAM +instead of HEAP to allow big dumps. @item -Remember @code{FOREIGN} key definitions in the @file{.frm} file. +Changed date functions to be string functions. This fixed some ``funny'' +side effects when sorting on dates. @item -Cascading @code{DELETE} +Extended @code{ALTER TABLE} according to SQL92. @item -Server side cursors. +Some minor compatibility changes. @item -Check if @code{lockd} works with modern Linux kernels; If not, we have -to fix @code{lockd}! To test this, start @code{mysqld} with -@code{--enable-locking} and run the different fork* test suits. They shouldn't -give any errors if @code{lockd} works. +Added @code{--port} and @code{--socket} options to all utility programs and +@code{mysqld}. @item -Allow SQL variables in @code{LIMIT}, like in @code{LIMIT @@a,@@b}. +Fixed MIT-pthreads @code{readdir_r()}. Now @code{mysqladmin create database} +and @code{mysqladmin drop database} should work. @item -Allow update of variables in @code{UPDATE} statements. For example: -@code{UPDATE TABLE foo SET @@a=a+b,a=@@a, b=@@a+c} +Changed MIT-pthreads to use our @code{tempnam()}. This should fix the ``sort +aborted'' bug. @item -Change when user variables are updated so that one can use them with -@code{GROUP BY}, as in the following example: -@code{SELECT id, @@a:=count(*), sum(sum_col)/@@a FROM table_name GROUP BY id}. +Added sync of records count in @code{sql_update}. This fixed slow updates +on first connection. (Thanks to Vaclav Bittner for the test.) +@end itemize + +@node News-3.20.10, News-3.20.9, News-3.20.11, News-3.20.x +@appendixsubsec Changes in release 3.20.10 + +@itemize @bullet @item -Don't add automatic @code{DEFAULT} values to columns. Give an error when using -an @code{INSERT} that doesn't contain a column that doesn't have a -@code{DEFAULT}. +New insert type: @code{INSERT INTO ... SELECT ...} @item -Caching of queries and results. This should be done as a separated -module that examines each query and if this is query is in the cache -the cached result should be returned. When one updates a table one -should remove as few queries as possible from the cache. -This should give a big speed bost on machines with much RAM where -queries are often repeated (like WWW applications). -One idea would be to only cache queries of type: -@code{SELECT CACHED ...} +@code{MEDIUMBLOB} fixed. @item -Fix @file{libmysql.c} to allow two @code{mysql_query()} commands in a row -without reading results or give a nice error message when one does this. +Fixed bug in @code{ALTER TABLE} and @code{BLOB}s. @item -Optimize @code{BIT} type to take 1 bit (now @code{BIT} takes 1 char). +@code{SELECT ... INTO OUTFILE} now creates the file in the current +database directory. @item -Check why MIT-pthreads @code{ctime()} doesn't work on some FreeBSD systems. +@code{DROP TABLE} now can take a list of tables. @item -Add an @code{IMAGE} option to @code{LOAD DATA INFILE} to not update -@code{TIMESTAMP} and @code{AUTO_INCREMENT} fields. +Oracle synonym @code{DESCRIBE} (@code{DESC}). @item -Added @code{LOAD DATE INFILE.. UPDATE} syntax. -@itemize @bullet +Changes to @code{make_binary_distribution}. @item -For tables with primary keys, if the data contains the primary key, -entries matching that primary key are updated from the remainder of the -columns. However, columns MISSING from the incoming data feed are not -touched. +Added some comments to installation instructions about @code{configure}'s +C++ link test. @item -For tables tables with primary keys that are missing some part of the key -in the incoming data stream, or that have no primary key, the feed is -treated as a @code{LOAD DATA INFILE ... REPLACE INTO} is now. -@end itemize +Added @code{--without-perl} option to @code{configure}. @item -Make @code{LOAD DATA INFILE} understand syntax like: -@example -LOAD DATA INFILE 'file_name.txt' INTO TABLE tbl_name -TEXT_FIELDS (text_field1, text_field2, text_field3) -SET table_field1=concatenate(text_field1, text_field2), table_field3=23 -IGNORE text_field3 +Lots of small portability changes. +@end itemize -This can be used to skip over extra columns in the text file, or update columns -based on expressions of the read data... -@end example -@item -@code{LOAD DATA INFILE 'file_name' INTO TABLE 'table_name' ERRORS TO err_table_name} -This would cause any errors and warnings to be logged into the err_table_name -table. That table would have a structure like: +@node News-3.20.9, News-3.20.8, News-3.20.10, News-3.20.x +@appendixsubsec Changes in release 3.20.9 -@example -line_number - line number in data file -error_message - the error/warning message -and maybe -data_line - the line from the data file -@end example +@itemize @bullet @item -Add true @code{VARCHAR} support (There is already support for this in MyISAM). +@code{ALTER TABLE} didn't copy null bit. As a result, fields that were allowed +to have @code{NULL} values were always @code{NULL}. @item -Automatic output from @code{mysql} to netscape. +@code{CREATE} didn't take numbers as @code{DEFAULT}. @item -@code{LOCK DATABASES}. (with various options) +Some compatibility changes for SunOS. @item -Change sort to allocate memory in ``hunks'' to get better memory utilization. +Removed @file{config.cache} from old distribution. +@end itemize + +@node News-3.20.8, News-3.20.7, News-3.20.9, News-3.20.x +@appendixsubsec Changes in release 3.20.8 + +@itemize @bullet @item -@code{DECIMAL} and @code{NUMERIC} types can't read exponential numbers; -@code{Field_decimal::store(const char *from,uint len)} must be recoded -to fix this. +Fixed bug with @code{ALTER TABLE} and multi-part keys. +@end itemize + +@node News-3.20.7, News-3.20.6, News-3.20.8, News-3.20.x +@appendixsubsec Changes in release 3.20.7 + +@itemize @bullet @item -Fix @code{mysql.cc} to do fewer @code{malloc()} calls when hashing field -names. +New commands: @code{ALTER TABLE}, @code{SELECT ... INTO OUTFILE} and +@code{LOAD DATA INFILE}. @item -Functions: -ADD_TO_SET(value,set) and REMOVE_FROM_SET(value,set) +New function: @code{NOW()}. @item -Add use of @code{t1 JOIN t2 ON ...} and @code{t1 JOIN t2 USING ...} -Currently, you can only use this syntax with @code{LEFT JOIN}. +Added new field @strong{file_priv} to @code{mysql/user} table. @item -Add full support for @code{unsigned long long} type. +New script @code{add_file_priv} which adds the new field @strong{file_priv} +to the @code{user} table. This script must be executed if you want to +use the new @code{SELECT ... INTO} and @code{LOAD DATA INFILE ...} commands +with a version of @strong{MySQL} earlier than 3.20.7. @item -Many more variables for @code{show status}. Counts for: -@code{INSERT}/@code{DELETE}/@code{UPDATE} statements. Records reads and -updated. Selects on 1 table and selects with joins. Mean number of -tables in select. Number of @code{ORDER BY} and @code{GROUP BY} queries. +Fixed bug in locking code, which made @code{lock_test.pl} test fail. @item -If you abort @code{mysql} in the middle of a query, you should open -another connection and kill the old running query. -Alternatively, an attempt should be made to detect this in the server. +New files @file{NEW} and @file{BUGS}. @item -Add a handler interface for table information so you can use it as a system -table. This would be a bit slow if you requested information about all tables, -but very flexible. @code{SHOW INFO FROM tbl_name} for basic table information -should be implemented. +Changed @file{select_test.c} and @file{insert_test.c} to include +@file{config.h}. @item -Add support for UNICODE. +Added command @code{status} to @code{mysqladmin} for short logging. @item -@code{NATURAL JOIN} and @code{UNION JOIN} +Increased maximum number of keys to 16 and maximum number of key parts to 15. @item -Allow @code{select a from crash_me left join crash_me2 using (a)}; In this -case a is assumed to come from the crash_me table. +Use of sub keys. A key may now be a prefix of a string field. @item -Fix that @code{ON} and @code{USING} works with the @code{JOIN} and -@code{INNER JOIN} join types. +Added @code{-k} option to @code{mysqlshow}, to get key information for a table. @item -Oracle like @code{CONNECT BY PRIOR ...} to search hierarchy structures. +Added long options to @code{mysqldump}. +@end itemize + +@node News-3.20.6, News-3.20.3, News-3.20.7, News-3.20.x +@appendixsubsec Changes in release 3.20.6 + +@itemize @bullet @item -@code{RENAME DATABASE} +Portable to more systems because of MIT-pthreads, which will +be used automatically if @code{configure} cannot find a @code{-lpthreads} +library. @item -@code{mysqladmin copy database new-database}. -- Requires COPY command to be -added to @code{mysqld} +Added GNU-style long options to almost all programs. Test with +@code{@kbd{program} --help}. @item -Processlist should show number of queries/thread. +Some shared library support for Linux. @item -@code{SHOW HOSTS} for printing information about the hostname cache. +The FAQ is now in @file{.texi} format and is available in @file{.html}, +@file{.txt} and @file{.ps} formats. @item -@code{DELETE} and @code{REPLACE} options to the @code{UPDATE} statement -(this will delete rows when one gets a duplicate key error while updating). +Added new SQL function @code{RAND([init])}. @item -Change the format of @code{DATETIME} to store fractions of seconds. +Changed @code{sql_lex} to handle @code{\0} unquoted, but the client can't send +the query through the C API, because it takes a str pointer. +You must use @code{mysql_real_query()} to send the query. @item -Add all missing ANSI92 and ODBC 3.0 types. +Added API function @code{mysql_get_client_info()}. @item -Change table names from empty strings to @code{NULL} for calculated columns. +@code{mysqld} now uses the @code{N_MAX_KEY_LENGTH} from @file{nisam.h} as +the maximum allowed key length. @item -Don't use 'Item_copy_string' on numerical values to avoid -number->string->number conversion in case of: -@code{SELECT COUNT(*)*(id+0) FROM table_name GROUP BY id} +The following now works: +@example +mysql> SELECT filter_nr,filter_nr FROM filter ORDER BY filter_nr; +@end example +Previously, this resulted in the error: +@code{Column: 'filter_nr' in order clause is ambiguous}. @item -Make it possible to use the new GNU regexp library instead of the current -one (The GNU library should be much faster than the old one). +@code{mysql} now outputs @code{'\0'}, @code{'\t'}, @code{'\n'} and @code{'\\'} +when encountering ASCII 0, tab, newline or @code{'\'} while writing +tab-separated output. +This is to allow printing of binary data in a portable format. +To get the old behavior, use @code{-r} (or @code{--raw}). @item -Change that @code{ALTER TABLE} doesn't abort clients that executes -@code{INSERT DELAYED}. +Added german error messages (60 of 80 error messages translated). @item -Fix that when columns referenced in an @code{UPDATE} clause contains the old -values before the update started. +Added new API function @code{mysql_fetch_lengths(MYSQL_RES *)}, which +returns an array of column lengths (of type @code{uint}). @item -@code{myisamchk}, @code{REPAIR} and @code{OPTIMIZE TABLE} should be able -to handle cases where the data and/or index files are symbolic links. +Fixed bug with @code{IS NULL} in @code{WHERE} clause. @item -Add simulation of @code{pread()}/@code{pwrite()} on Windows to enable -concurrent inserts. +Changed the optimizer a little to get better results when searching on a key +part. @item -A logfile analyzer that could parsed out information about which tables -are hit most often, how often multi-table joins are executed, etc. It -should help users identify areas or table design that could be optimized -to execute much more efficient queries. +Added @code{SELECT} option @code{STRAIGHT_JOIN} to tell the optimizer that +it should join tables in the given order. @item -Add @code{SUM(DISTINCT)} +Added support for comments starting with @code{'--'} in @file{mysql.cc} +(Postgres syntax). @item -Add @code{ANY()},@code{EVERY()} and @code{SOME()} group functions. In -ANSI SQL these only works on boolean columns, but we can extend these to -work on any columns/expressions by applying: value == 0 -> FALSE and -value <> 0 -> TRUE. +You can have @code{SELECT} expressions and table columns in a @code{SELECT} +which are not used in the group part. This makes it efficient to implement +lookups. The column that is used should be a constant for each group because +the value is calculated only once for the first row that is found for a group. +@example +mysql> SELECT id,lookup.text,sum(*) FROM test,lookup + WHERE test.id=lookup.id GROUP BY id; +@end example @item -Fix that the type for @code{MAX(column)} is the same as the column type. +Fixed bug in @code{SUM(function)} (could cause a core dump). +@item +Changed @code{AUTO_INCREMENT} placement in the SQL query: @example -create table t1 (a DATE); -insert into t1 values (now()); -create table t2 select max(a) from t1; -show columns from t2; +INSERT into table (auto_field) values (0); @end example +inserted 0, but it should insert an @code{AUTO_INCREMENT} value. @item -Come up with a nice syntax for a statement that will @code{UPDATE} the row -if it exists and @code{INSERT} a new row if the row didn't exist. -(Like @code{REPLACE} works with @code{INSERT} / @code{DELETE}) +@file{mysqlshow.c}: Added number of records in table. Had to change the +client code a little to fix this. +@item +@code{mysql} now allows doubled @code{''} or @code{""} within strings for +embedded @code{'} or @code{"}. +@item +New math functions: +@code{EXP()}, @code{LOG()}, @code{SQRT()}, @code{ROUND()}, @code{CEILING()}. @end itemize -@node TODO sometime, TODO unplanned, TODO future, TODO -@appendixsec Things that have to be done sometime +@node News-3.20.3, News-3.20.0, News-3.20.6, News-3.20.x +@appendixsubsec Changes in release 3.20.3 @itemize @bullet @item -Implement function: @code{get_changed_tables(timeout,table1,table2,...)} +The @code{configure} source now compiles a thread-free client library +@code{-lmysqlclient}. This is the only library that needs to be linked +with client applications. When using the binary releases, you must +link with @code{-lmysql -lmysys -ldbug -lstrings} as before. @item -Atomic multi-table updates, eg @code{update items,month set -items.price=month.price where items.id=month.id;}; +New @code{readline} library from @code{bash-2.0}. @item -Change reading through tables to use memmap when possible. Now only -compressed tables use memmap. +LOTS of small changes to @code{configure} and makefiles (and related source). @item -Add a new privilege @strong{'Show_priv'} for @code{SHOW} commands. +It should now be possible to compile in another directory using +@code{VPATH}. Tested with GNU Make 3.75. @item -Make the automatic timestamp code nicer. Add timestamps to the update -log with @code{SET TIMESTAMP=#;} +@code{safe_mysqld} and @code{mysql.server} changed to be more compatible +between the source and the binary releases. @item -Use read/write mutex in some places to get more speed. +@code{LIMIT} now takes one or two numeric arguments. +If one argument is given, it indicates the maximum number of rows in +a result. If two arguments are given, the first argument indicates the offset +of the first row to return, the second is the maximum number of rows. +With this it's easy to do a poor man's next page/previous page WWW +application. @item -Full foreign key support. One probably wants to implement a procedural -language first. +Changed name of SQL function @code{FIELDS()} to @code{ELT()}. +Changed SQL function @code{INTERVALL()} to @code{INTERVAL()}. @item -Simple views (first on one table, later on any expression). +Made @code{SHOW COLUMNS} a synonym for @code{SHOW FIELDS}. +Added compatibility syntax @code{FRIEND KEY} to @code{CREATE TABLE}. In +@strong{MySQL}, this creates a non-unique key on the given columns. @item -Automatically close some tables if a table, temporary table or temporary files -gets error 23 (not enough open files). +Added @code{CREATE INDEX} and @code{DROP INDEX} as compatibility functions. +In @strong{MySQL}, @code{CREATE INDEX} only checks if the index exists and +issues an error if it doesn't exist. @code{DROP INDEX} always succeeds. @item -When one finds a field=#, change all occurrences of field to #. Now this -is only done for some simple cases. +@file{mysqladmin.c}: added client version to version information. @item -Change all const expressions with calculated expressions if possible. +Fixed core dump bug in @code{sql_acl} (core on new connection). @item -Optimize key = expression. At the moment only key = field or key = -constant are optimized. +Removed @code{host}, @code{user} and @code{db} tables from database @code{test} +in the distribution. @item -Join some of the copy functions for nicer code. +@code{FIELD_TYPE_CHAR} can now be signed (-128 - 127) or unsigned (0 - 255) +Previously, it was always unsigned. @item -Change @file{sql_yacc.yy} to an inline parser to reduce its size and get -better error messages (5 days). +Bug fixes in @code{CONCAT()} and @code{WEEKDAY()}. @item -Change the parser to use only one rule per different number of arguments -in function. +Changed a lot of source to get @code{mysqld} to be compiled with SunPro +compiler. @item -Use of full calculation names in the order part. (For ACCESS97) +SQL functions must now have a @code{'('} immediately after the function name +(no intervening space). +For example, @code{'user('} is regarded as beginning a function call, and +@code{'user ('} is regarded as an identifier @code{user} followed by a +@code{'('}, not as a function call. +@end itemize + +@node News-3.20.0, , News-3.20.3, News-3.20.x +@appendixsubsec Changes in release 3.20.0 + +@itemize @bullet @item -@code{UNION}, @code{MINUS}, @code{INTERSECT} and @code{FULL OUTER JOIN}. -(Currently only @code{LEFT OUTER JOIN} is supported) +The source distribution is done with @code{configure} and Automake. +It will make porting much easier. The @code{readline} library is included +in the distribution. @item -Allow @code{UNIQUE} on fields that can be @code{NULL}. +Separate client compilation: +the client code should be very easy to compile on systems which +don't have threads. @item -@code{SQL_OPTION MAX_SELECT_TIME=#} to put a time limit on a query. +The old Perl interface code is automatically compiled and installed. +Automatic compiling of @code{DBD} will follow when the new @code{DBD} code +is ported. @item -Make the update log to a database. +Dynamic language support: @code{mysqld} can now be started with Swedish +or English (default) error messages. @item -Negative @code{LIMIT} to retrieve data from the end. +New functions: @code{INSERT()}, @code{RTRIM()}, @code{LTRIM()} and +@code{FORMAT()}. @item -Alarm around client connect/read/write functions. +@code{mysqldump} now works correctly for all field types (even +@code{AUTO_INCREMENT}). The format for @code{SHOW FIELDS FROM tbl_name} +is changed so the @code{Type} column contains information suitable for +@code{CREATE TABLE}. In previous releases, some @code{CREATE TABLE} +information had to be patched when re-creating tables. @item -Please note the changes to @code{safe_mysqld}: according to FSSTND (which -Debian tries to follow) PID files should go into @file{/var/run/<progname>.pid} -and log files into @file{/var/log}. It would be nice if you could put the -"DATADIR" in the first declaration of "pidfile" and "log", so the -placement of these files can be changed with a single statement. +Some parser bugs from 3.19.5 (@code{BLOB} and @code{TIMESTAMP}) are corrected. +@code{TIMESTAMP} now returns different date information depending on its +create length. @item -Allow a client to request logging. +Changed parser to allow a database, table or field name to +start with a number or @code{'_'}. @item -Add use of @code{zlib()} for @code{gzip}-ed files to @code{LOAD DATA INFILE}. +All old C code from Unireg changed to C++ and cleaned up. This makes +the daemon a little smaller and easier to understand. @item -Fix sorting and grouping of @code{BLOB} columns (partly solved now). +A lot of small bug fixes done. @item -Stored procedures. This is currently not regarded to be very -important as stored procedures are not very standardized yet. -Another problem is that true stored procedures make it much harder for -the optimizer and in many cases the result is slower than before -We will, on the other hand, add a simple (atomic) update language that -can be used to write loops and such in the @strong{MySQL} server. +New @file{INSTALL} files (not final version) and some information regarding porting. +@end itemize + +@node News-3.19.x, , News-3.20.x, News +@appendixsec Changes in release 3.19.x + +Version 3.19 is quite old now, and should be avoided if possible. This +information is kept here for historical purposes only. + +@menu +* News-3.19.5:: Changes in release 3.19.5 +* News-3.19.4:: Changes in release 3.19.4 +* News-3.19.3:: Changes in release 3.19.3 +@end menu + +@node News-3.19.5, News-3.19.4, News-3.19.x, News-3.19.x +@appendixsubsec Changes in release 3.19.5 + +@itemize @bullet @item -Change to use semaphores when counting threads. One should first implement -a semaphore library to MIT-pthreads. +Some new functions, some more optimization on joins. @item -Don't assign a new @code{AUTO_INCREMENT} value when one sets a column to 0. -Use @code{NULL} instead. +Should now compile clean on Linux (2.0.x). @item -Add full support for @code{JOIN} with parentheses. +Added functions @code{DATABASE()}, @code{USER()}, @code{POW()}, +@code{LOG10()} (needed for ODBC). @item -As an alternative for one thread / connection manage a pool of threads -to handle the queries. +In a @code{WHERE} with an @code{ORDER BY} on fields from only one table, +the table is now preferred as first table in a multi-join. @item -Allow one to get more than one lock with @code{GET_LOCK}. When doing this, -one must also handle the possible deadlocks this change will introduce. +@code{HAVING} and @code{IS NULL} or @code{IS NOT NULL} now works. +@item +A group on one column and a sort on a group function (@code{SUM()}, +@code{AVG()}...) didn't work together. Fixed. +@item +@code{mysqldump}: Didn't send password to server. @end itemize -Time is given according to amount of work, not real time. +@node News-3.19.4, News-3.19.3, News-3.19.5, News-3.19.x +@appendixsubsec Changes in release 3.19.4 -@node TODO unplanned, , TODO sometime, TODO -@appendixsec Some things we don't have any plans to do +@itemize @bullet +@item +Fixed horrible locking bug when inserting in one thread and reading +in another thread. +@item +Fixed one-off decimal bug. 1.00 was output as 1.0. +@item +Added attribute @code{'Locked'} to process list as info if a query is +locked by another query. +@item +Fixed full magic timestamp. Timestamp length may now be 14, 12, 10, 8, 6, 4 +or 2 bytes. +@item +Sort on some numeric functions could sort incorrectly on last number. +@item +@code{IF(arg,syntax_error,syntax_error)} crashed. +@item +Added functions @code{CEILING()}, @code{ROUND()}, @code{EXP()}, @code{LOG()} and @code{SQRT()}. +@item +Enhanced @code{BETWEEN} to handle strings. +@end itemize + +@node News-3.19.3, , News-3.19.4, News-3.19.x +@appendixsubsec Changes in release 3.19.3 @itemize @bullet @item -Nothing; In the long run we plan to be fully ANSI 92 / ANSI 99 compliant. +Fixed @code{SELECT} with grouping on @code{BLOB} columns not to return +incorrect @code{BLOB} info. Grouping, sorting and distinct on @code{BLOB} +columns will not yet work as +expected (probably it will group/sort by the first 7 characters in the +@code{BLOB}). Grouping on formulas with a fixed string size (use @code{MID()} +on a @code{BLOB}) should work. +@item +When doing a full join (no direct keys) on multiple tables with @code{BLOB} +fields, the @code{BLOB} was garbage on output. +@item +Fixed @code{DISTINCT} with calculated columns. @end itemize + + @cindex porting, to other systems -@node Porting, Regexp, TODO, Top +@node Porting, Regexp, News, Top @appendix Comments on porting to other systems A working Posix thread library is needed for the server. On Solaris 2.5 -- 2.30.9