MDEV-27066: Fixed scientific notation parsing bug
The bug occurs where the float token containing a dot with an 'e' notation was dropped from the request completely. This causes a manner of invalid SQL statements like: select id 1.e, char 10.e(id 2.e), concat 3.e('a'12356.e,'b'1.e,'c'1.1234e)1.e, 12 1.e*2 1.e, 12 1.e/2 1.e, 12 1.e|2 1.e, 12 1.e^2 1.e, 12 1.e%2 1.e, 12 1.e&2 from test; To be parsed correctly as if it was: select id, char(id), concat('a','b','c'), 12*2, 12/2, 12|2, 12^2, 12%2, 12&2 from test.test; This correct parsing occurs when e is followed by any of: ( ) . , | & % * ^ /
Showing
Please register or sign in to comment