Fix for BUG#16710.

The bug was due to a missed case in the detection of whether an index
can be used for loose scan. More precisely, the range optimizer chose
to use loose index scan for queries for which the condition(s) over
an index key part could not be pushed to the index together with the
loose scan.

As a result, loose index scan was selecting the first row in the
index with a given GROUP BY prefix, and was applying the WHERE
clause after that, while it should have inspected all rows with
the given prefix, and apply the WHERE clause to all of them.

The fix detects and skips such cases.


mysql-test/r/group_min_max.result:
  Added test for BUG#16710.
mysql-test/t/group_min_max.test:
  Added test for BUG#16710.
sql/item.cc:
  Added new method [Item | Item_field]::find_item_in_field_list_processor.
sql/item.h:
  Added new method [Item | Item_field]::find_item_in_field_list_processor.
sql/opt_range.cc:
  Handle the case when there is no MIN/MAX aggregate function, and a
  keypart of the index being considered, that is after the GROUP BY
  prefix, is used in the WHERE clause and the condition where it is
  used cannot be pushed to the index.
  If this is the case, we rule out this index.

mysql-test/r/group_min_max.result

@@ -1954,6 +1954,15 @@ id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
 explain select sum(ord(a1)) from t1 where (a1 > 'a') group by a1,a2,b;
 id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
 1	SIMPLE	t1	index	idx_t1_0,idx_t1_1,idx_t1_2	idx_t1_1	163	NULL	128	Using where; Using index
+explain select distinct(a1) from t1 where ord(a2) = 98;
+id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
+1	SIMPLE	t1	index	NULL	idx_t1_1	163	NULL	128	Using where; Using index
+select distinct(a1) from t1 where ord(a2) = 98;
+a1
+a
+b
+c
+d
 explain select a1 from t1 where a2 = 'b' group by a1;
 id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
 1	SIMPLE	t1	range	NULL	idx_t1_1	130	NULL	5	Using where; Using index for group-by

mysql-test/t/group_min_max.test

@@ -641,6 +641,14 @@ explain select a1,a2,count(a2) from t1 group by a1,a2,b;
 explain select a1,a2,count(a2) from t1 where (a1 > 'a') group by a1,a2,b;
 explain select sum(ord(a1)) from t1 where (a1 > 'a') group by a1,a2,b;
 
+
+#
+# Bug #16710: select distinct doesn't return all it should
+#
+
+explain select distinct(a1) from t1 where ord(a2) = 98;
+select distinct(a1) from t1 where ord(a2) = 98;
+
 #
 # BUG#11044: DISTINCT or GROUP BY queries with equality predicates instead of MIN/MAX.
 #

sql/item.cc

@@ -499,7 +499,7 @@ bool Item_ident::remove_dependence_processor(byte * arg)
     arguments in a condition the method must return false.
 
   RETURN
-    false to force the evaluation of collect_item_field_processor
+    FALSE to force the evaluation of collect_item_field_processor
           for the subsequent items.
 */
 
@@ -520,6 +520,38 @@ bool Item_field::collect_item_field_processor(byte *arg)
 }
 
 
+/*
+  Check if an Item_field references some field from a list of fields.
+
+  SYNOPSIS
+    Item_field::find_item_in_field_list_processor
+    arg   Field being compared, arg must be of type Field
+
+  DESCRIPTION
+    Check whether the Item_field represented by 'this' references any
+    of the fields in the keyparts passed via 'arg'. Used with the
+    method Item::walk() to test whether any keypart in a sequence of
+    keyparts is referenced in an expression.
+
+  RETURN
+    TRUE  if 'this' references the field 'arg'
+    FALE  otherwise
+*/
+bool Item_field::find_item_in_field_list_processor(byte *arg)
+{
+  KEY_PART_INFO *first_non_group_part= *((KEY_PART_INFO **) arg);
+  KEY_PART_INFO *last_part= *(((KEY_PART_INFO **) arg) + 1);
+  KEY_PART_INFO *cur_part;
+
+  for (cur_part= first_non_group_part; cur_part != last_part; cur_part++)
+  {
+    if (field->eq(cur_part->field))
+      return TRUE;
+  }
+  return FALSE;
+}
+
+
 bool Item::check_cols(uint c)
 {
   if (c != 1)

sql/item.h

@@ -699,6 +699,7 @@ class Item {
   virtual bool remove_fixed(byte * arg) { fixed= 0; return 0; }
   virtual bool cleanup_processor(byte *arg);
   virtual bool collect_item_field_processor(byte * arg) { return 0; }
+  virtual bool find_item_in_field_list_processor(byte *arg) { return 0; }
   virtual bool change_context_processor(byte *context) { return 0; }
   virtual bool reset_query_id_processor(byte *query_id) { return 0; }
 
@@ -1146,6 +1147,7 @@ class Item_field :public Item_ident
   bool is_null() { return field->is_null(); }
   Item *get_tmp_table_item(THD *thd);
   bool collect_item_field_processor(byte * arg);
+  bool find_item_in_field_list_processor(byte *arg);
   bool reset_query_id_processor(byte *arg)
   {
     field->query_id= *((query_id_t *) arg);

sql/opt_range.cc

@@ -6876,6 +6876,7 @@ cost_group_min_max(TABLE* table, KEY *index_info, uint used_key_parts,
                    bool have_min, bool have_max,
                    double *read_cost, ha_rows *records);
 
+
 /*
   Test if this access method is applicable to a GROUP query with MIN/MAX
   functions, and if so, construct a new TRP object.
@@ -7282,11 +7283,36 @@ get_best_group_min_max(PARAM *param, SEL_TREE *tree)
       }
       else if (min_max_arg_part &&
                (min_max_arg_part - first_non_group_part > 0))
+      {
         /*
           There is a gap but no range tree, thus no predicates at all for the
           non-group keyparts.
         */
         goto next_index;
+      }
+      else if (first_non_group_part && join->conds)
+      {
+        /*
+          If there is no MIN/MAX function in the query, but some index
+          key part is referenced in the WHERE clause, then this index
+          cannot be used because the WHERE condition over the keypart's
+          field cannot be 'pushed' to the index (because there is no
+          range 'tree'), and the WHERE clause must be evaluated before
+          GROUP BY/DISTINCT.
+        */
+        /*
+          Store the first and last keyparts that need to be analyzed
+          into one array that can be passed as parameter.
+        */
+        KEY_PART_INFO *key_part_range[2];
+        key_part_range[0]= first_non_group_part;
+        key_part_range[1]= last_part;
+
+        /* Check if cur_part is referenced in the WHERE clause. */
+        if (join->conds->walk(&Item::find_item_in_field_list_processor,
+                              (byte*) key_part_range))
+          goto next_index;
+      }
     }
 
     /*