Bug #30596 GROUP BY optimization gives wrong result order

The optimization that uses a unique index to remove GROUP BY, did not 
ensure that the index was actually used, thus violating the ORDER BY
that is impled by GROUP BY.
Fixed by replacing GROUP BY with ORDER BY if the GROUP BY clause contains
a unique index. In case GROUP BY ... ORDER BY null is used, GROUP BY is
simply removed.

.bzrignore

@@ -1345,3 +1345,8 @@ zlib/*.vcproj
 debian/control
 debian/defs.mk
 include/abi_check
+support-files/mysqld_multi.server
+tests/bug25714
+cscope.in.out
+cscope.out
+cscope.po.out

mysql-test/r/distinct.result

@@ -526,10 +526,10 @@ id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
 1	SIMPLE	t1	index	NULL	PRIMARY	4	NULL	3	Using index
 EXPLAIN SELECT a,b FROM t1 GROUP BY a,b;
 id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
-1	SIMPLE	t1	ALL	NULL	NULL	NULL	NULL	3	
+1	SIMPLE	t1	ALL	NULL	NULL	NULL	NULL	3	Using filesort
 EXPLAIN SELECT DISTINCT a,b FROM t1 GROUP BY a,b;
 id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
-1	SIMPLE	t1	ALL	NULL	NULL	NULL	NULL	3	
+1	SIMPLE	t1	ALL	NULL	NULL	NULL	NULL	3	Using filesort
 CREATE TABLE t2(a INT, b INT NOT NULL, c INT NOT NULL, d INT, 
 PRIMARY KEY (a,b));
 INSERT INTO t2 VALUES (1,1,1,50), (1,2,3,40), (2,1,3,4);
@@ -554,7 +554,7 @@ id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
 CREATE UNIQUE INDEX c_b_unq ON t2 (c,b);
 EXPLAIN SELECT DISTINCT a,b,d FROM t2 GROUP BY c,b,d;
 id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
-1	SIMPLE	t2	ALL	NULL	NULL	NULL	NULL	3	
+1	SIMPLE	t2	ALL	NULL	NULL	NULL	NULL	3	Using filesort
 DROP TABLE t1,t2;
 create table t1 (id int, dsc varchar(50));
 insert into t1 values (1, "line number one"), (2, "line number two"), (3, "line number three");

mysql-test/r/group_by.result

@@ -1064,3 +1064,52 @@ select t1.f1,t.* from t1, t1 t group by 1;
 ERROR 42000: 'test.t.f1' isn't in GROUP BY
 drop table t1;
 SET SQL_MODE = '';
+CREATE TABLE t1(
+a INT, 
+b INT NOT NULL, 
+c INT NOT NULL, 
+d INT, 
+UNIQUE KEY (c,b)
+);
+INSERT INTO t1 VALUES (1,1,1,50), (1,2,3,40), (2,1,3,4);
+EXPLAIN SELECT c,b,d FROM t1 GROUP BY c,b,d;
+id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
+1	SIMPLE	t1	ALL	NULL	NULL	NULL	NULL	3	Using filesort
+SELECT c,b,d FROM t1 GROUP BY c,b,d;
+c	b	d
+1	1	50
+3	1	4
+3	2	40
+EXPLAIN SELECT c,b,d FROM t1 GROUP BY c,b,d ORDER BY NULL;
+id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
+1	SIMPLE	t1	ALL	NULL	NULL	NULL	NULL	3	
+SELECT c,b,d FROM t1 GROUP BY c,b,d ORDER BY NULL;
+c	b	d
+1	1	50
+3	2	40
+3	1	4
+EXPLAIN SELECT c,b,d FROM t1 ORDER BY c,b,d;
+id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
+1	SIMPLE	t1	ALL	NULL	NULL	NULL	NULL	3	Using filesort
+SELECT c,b,d FROM t1 ORDER BY c,b,d;
+c	b	d
+1	1	50
+3	1	4
+3	2	40
+EXPLAIN SELECT c,b,d FROM t1 GROUP BY c,b;
+id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
+1	SIMPLE	t1	ALL	NULL	NULL	NULL	NULL	3	Using filesort
+SELECT c,b,d FROM t1 GROUP BY c,b;
+c	b	d
+1	1	50
+3	1	4
+3	2	40
+EXPLAIN SELECT c,b   FROM t1 GROUP BY c,b;
+id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
+1	SIMPLE	t1	index	NULL	c	8	NULL	3	Using index
+SELECT c,b   FROM t1 GROUP BY c,b;
+c	b
+1	1
+3	1
+3	2
+DROP TABLE t1;

mysql-test/r/innodb_mysql.result

@@ -1047,4 +1047,53 @@ t1	CREATE TABLE `t1` (
   KEY `a` (`a`(255))
 ) ENGINE=InnoDB DEFAULT CHARSET=utf8
 drop table t1;
+CREATE TABLE t1(
+a INT, 
+b INT NOT NULL, 
+c INT NOT NULL, 
+d INT, 
+UNIQUE KEY (c,b)
+) engine=innodb;
+INSERT INTO t1 VALUES (1,1,1,50), (1,2,3,40), (2,1,3,4);
+EXPLAIN SELECT c,b,d FROM t1 GROUP BY c,b,d;
+id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
+1	SIMPLE	t1	ALL	NULL	NULL	NULL	NULL	3	Using filesort
+SELECT c,b,d FROM t1 GROUP BY c,b,d;
+c	b	d
+1	1	50
+3	1	4
+3	2	40
+EXPLAIN SELECT c,b,d FROM t1 GROUP BY c,b,d ORDER BY NULL;
+id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
+1	SIMPLE	t1	ALL	NULL	NULL	NULL	NULL	3	
+SELECT c,b,d FROM t1 GROUP BY c,b,d ORDER BY NULL;
+c	b	d
+1	1	50
+3	1	4
+3	2	40
+EXPLAIN SELECT c,b,d FROM t1 ORDER BY c,b,d;
+id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
+1	SIMPLE	t1	ALL	NULL	NULL	NULL	NULL	3	Using filesort
+SELECT c,b,d FROM t1 ORDER BY c,b,d;
+c	b	d
+1	1	50
+3	1	4
+3	2	40
+EXPLAIN SELECT c,b,d FROM t1 GROUP BY c,b;
+id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
+1	SIMPLE	t1	index	NULL	c	8	NULL	3	
+SELECT c,b,d FROM t1 GROUP BY c,b;
+c	b	d
+1	1	50
+3	1	4
+3	2	40
+EXPLAIN SELECT c,b   FROM t1 GROUP BY c,b;
+id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
+1	SIMPLE	t1	index	NULL	c	8	NULL	3	Using index
+SELECT c,b   FROM t1 GROUP BY c,b;
+c	b
+1	1
+3	1
+3	2
+DROP TABLE t1;
 End of 5.0 tests

mysql-test/t/group_by.test

@@ -788,3 +788,30 @@ select * from t1 group by f1, f2;
 select t1.f1,t.* from t1, t1 t group by 1;
 drop table t1;
 SET SQL_MODE = '';
+
+#
+# Bug#30596: GROUP BY optimization gives wrong result order
+#
+CREATE TABLE t1(
+  a INT, 
+  b INT NOT NULL, 
+  c INT NOT NULL, 
+  d INT, 
+  UNIQUE KEY (c,b)
+);
+
+INSERT INTO t1 VALUES (1,1,1,50), (1,2,3,40), (2,1,3,4);
+
+EXPLAIN SELECT c,b,d FROM t1 GROUP BY c,b,d;
+SELECT c,b,d FROM t1 GROUP BY c,b,d;
+EXPLAIN SELECT c,b,d FROM t1 GROUP BY c,b,d ORDER BY NULL;
+SELECT c,b,d FROM t1 GROUP BY c,b,d ORDER BY NULL;
+EXPLAIN SELECT c,b,d FROM t1 ORDER BY c,b,d;
+SELECT c,b,d FROM t1 ORDER BY c,b,d;
+
+EXPLAIN SELECT c,b,d FROM t1 GROUP BY c,b;
+SELECT c,b,d FROM t1 GROUP BY c,b;
+EXPLAIN SELECT c,b   FROM t1 GROUP BY c,b;
+SELECT c,b   FROM t1 GROUP BY c,b;
+
+DROP TABLE t1;

mysql-test/t/innodb_mysql.test

@@ -882,4 +882,31 @@ alter table t1 add index(a(1024));
 show create table t1;
 drop table t1;
 
+#
+# Bug#30596: GROUP BY optimization gives wrong result order
+#  
+CREATE TABLE t1(
+  a INT, 
+  b INT NOT NULL, 
+  c INT NOT NULL, 
+  d INT, 
+  UNIQUE KEY (c,b)
+) engine=innodb;
+
+INSERT INTO t1 VALUES (1,1,1,50), (1,2,3,40), (2,1,3,4);
+
+EXPLAIN SELECT c,b,d FROM t1 GROUP BY c,b,d;
+SELECT c,b,d FROM t1 GROUP BY c,b,d;
+EXPLAIN SELECT c,b,d FROM t1 GROUP BY c,b,d ORDER BY NULL;
+SELECT c,b,d FROM t1 GROUP BY c,b,d ORDER BY NULL;
+EXPLAIN SELECT c,b,d FROM t1 ORDER BY c,b,d;
+SELECT c,b,d FROM t1 ORDER BY c,b,d;
+
+EXPLAIN SELECT c,b,d FROM t1 GROUP BY c,b;
+SELECT c,b,d FROM t1 GROUP BY c,b;
+EXPLAIN SELECT c,b   FROM t1 GROUP BY c,b;
+SELECT c,b   FROM t1 GROUP BY c,b;
+
+DROP TABLE t1;
+
 --echo End of 5.0 tests

sql/sql_select.cc

@@ -1030,6 +1030,14 @@ JOIN::optimize()
                                  find_field_in_order_list,
                                  (void *) group_list))
     {
+      /*
+        We have found that grouping can be removed since groups correspond to
+        only one row anyway, but we still have to guarantee correct result
+        order. The line below effectively rewrites the query from GROUP BY
+        <fields> to ORDER BY <fields>. One exception is if skip_sort_order is
+        set (see above), then we can simply skip GROUP BY.
+       */
+      order= skip_sort_order ? 0 : group_list;
       group_list= 0;
       group= 0;
     }