[MDEV-5203] Different results with record_cond_statistics=on and record_cond_statistics=off with SOME subquery - JIRA

Details

Type: Bug
Status: Closed
Priority: Major
Resolution: Fixed
Affects Version/s: None
Fix Version/s: 10.0.6
Component/s: None
Labels:
None

Description

The following test case produces different result sets for the first and the second SELECT. SELECT is the same, only the value of record_cond_statistics differs.
The same query on 10.0-base or on MariaDB 5.3 or on MySQL 5.6 produces 2 rows, which I assume to be a correct result.

CREATE TABLE t1 (i1 INT) ENGINE=MyISAM;
INSERT INTO t1 VALUES (9),(8);

CREATE TABLE t2 (id2 INT, i2 INT) ENGINE=MyISAM;
INSERT INTO t2 VALUES (1,5),(2,6);

CREATE TABLE t3 (pk3 INT PRIMARY KEY, i3 INT) ENGINE=MyISAM;
INSERT INTO t3 VALUES (1,0),(2,6);

SELECT * FROM t1 AS outer_t1, t2, t3 
WHERE pk3 = id2 AND i2 < SOME ( SELECT i1 FROM t1 WHERE i1 < outer_t1.i1 );

SET optimizer_switch='record_cond_statistics=on';

SELECT * FROM t1 AS outer_t1, t2, t3 
WHERE pk3 = id2 AND i2 < SOME ( SELECT i1 FROM t1 WHERE i1 < outer_t1.i1 );

Results:

SELECT * FROM t1 AS outer_t1, t2, t3 
WHERE pk3 = id2 AND i2 < SOME ( SELECT i1 FROM t1 WHERE i1 < outer_t1.i1 );
i1	id2	i2	pk3	i3
9	1	5	1	0
9	2	6	2	6
SET optimizer_switch='record_cond_statistics=on';
SELECT * FROM t1 AS outer_t1, t2, t3 
WHERE pk3 = id2 AND i2 < SOME ( SELECT i1 FROM t1 WHERE i1 < outer_t1.i1 );
i1	id2	i2	pk3	i3
9	1	5	1	0
8	1	5	1	0
9	2	6	2	6
8	2	6	2	6

EXPLAIN with record_cond_statistics=off:

id	select_type	table	type	possible_keys	key	key_len	ref	rows	filtered	Extra
1	PRIMARY	outer_t1	ALL	NULL	NULL	NULL	NULL	2	100.00	
1	PRIMARY	t2	ALL	NULL	NULL	NULL	NULL	2	100.00	Using where; Subqueries: 2; Using join buffer (flat, BNL join)
1	PRIMARY	t3	eq_ref	PRIMARY	PRIMARY	4	test.t2.id2	1	100.00	
2	DEPENDENT SUBQUERY	t1	ALL	NULL	NULL	NULL	NULL	2	100.00	Using where
Warnings:
Note	1276	Field or reference 'test.outer_t1.i1' of SELECT #2 was resolved in SELECT #1
Note	1003	select `test`.`outer_t1`.`i1` AS `i1`,`test`.`t2`.`id2` AS `id2`,`test`.`t2`.`i2` AS `i2`,`test`.`t3`.`pk3` AS `pk3`,`test`.`t3`.`i3` AS `i3` from `test`.`t1` `outer_t1` join `test`.`t2` join `test`.`t3` where ((`test`.`t3`.`pk3` = `test`.`t2`.`id2`) and <nop>(<expr_cache><`test`.`t2`.`i2`,`test`.`outer_t1`.`i1`>(<in_optimizer>(`test`.`t2`.`i2`,<exists>(select `test`.`t1`.`i1` from `test`.`t1` where ((`test`.`t1`.`i1` < `test`.`outer_t1`.`i1`) and trigcond(((<cache>(`test`.`t2`.`i2`) < `test`.`t1`.`i1`) or isnull(`test`.`t1`.`i1`)))) having trigcond(<is_not_null_test>(`test`.`t1`.`i1`)))))))

EXPLAIN with record_cond_statistics=on:

id	select_type	table	type	possible_keys	key	key_len	ref	rows	filtered	Extra
1	PRIMARY	outer_t1	ALL	NULL	NULL	NULL	NULL	2	100.00	
1	PRIMARY	t2	ALL	NULL	NULL	NULL	NULL	2	100.00	Using where; Subqueries: 2; Using join buffer (flat, BNL join)
1	PRIMARY	t3	eq_ref	PRIMARY	PRIMARY	4	test.t2.id2	1	100.00	Using where
2	DEPENDENT SUBQUERY	t1	ALL	NULL	NULL	NULL	NULL	2	100.00	Using where
Warnings:
Note	1276	Field or reference 'test.outer_t1.i1' of SELECT #2 was resolved in SELECT #1
Note	1003	select `test`.`outer_t1`.`i1` AS `i1`,`test`.`t2`.`id2` AS `id2`,`test`.`t2`.`i2` AS `i2`,`test`.`t3`.`pk3` AS `pk3`,`test`.`t3`.`i3` AS `i3` from `test`.`t1` `outer_t1` join `test`.`t2` join `test`.`t3` where ((`test`.`t3`.`pk3` = `test`.`t2`.`id2`) and <nop>(<expr_cache><`test`.`t2`.`i2`,`test`.`outer_t1`.`i1`>(<in_optimizer>(`test`.`t2`.`i2`,<exists>(select `test`.`t1`.`i1` from `test`.`t1` where ((`test`.`t1`.`i1` < `test`.`outer_t1`.`i1`) and trigcond(((<cache>(`test`.`t2`.`i2`) < `test`.`t1`.`i1`) or isnull(`test`.`t1`.`i1`)))) having trigcond(<is_not_null_test>(`test`.`t1`.`i1`)))))))

Gliffy Diagrams

Attachments

Issue Links

relates to

MDEV-83 Cost-based choice for the pushdown of subqueries to joined tables

Stalled

Activity

Hide

Permalink

Timour Katchaounov added a comment - 2013-11-06 10:41

Already fixed by the following commit:

revno: 3752
fixes bug: https://mariadb.atlassian.net/browse/MDEV-5178
committer: timour@askmonty.org <timour@askmonty.org>
branch nick: 10.0-md83-4170-static-dyn-push
timestamp: Thu 2013-10-31 16:19:45 +0200
message:

Addressed review comments:
moved the calculation of the extra cost of subquery predicates during join optimization
into a separate method JOIN::static_pushdown_cost()
fixed small errors in static_pushdown_cost()
added a call to this method into optimize_straight_join, so that SELECT STRAIGHT_JOIN
is optimized in the same way as regular joins
added test cases for SELECT STRAIGHT_JOIN
Fixed bug ~~MDEV-5178~~ Valgrind warnings (Conditional jump or move depends on uninitialised value) with static_cond_pushdown=on, FROM subquery

Show

Timour Katchaounov added a comment - 2013-11-06 10:41 Already fixed by the following commit: revno: 3752 fixes bug: https://mariadb.atlassian.net/browse/MDEV-5178 committer: timour@askmonty.org <timour@askmonty.org> branch nick: 10.0-md83-4170-static-dyn-push timestamp: Thu 2013-10-31 16:19:45 +0200 message: Addressed review comments: moved the calculation of the extra cost of subquery predicates during join optimization into a separate method JOIN::static_pushdown_cost() fixed small errors in static_pushdown_cost() added a call to this method into optimize_straight_join, so that SELECT STRAIGHT_JOIN is optimized in the same way as regular joins added test cases for SELECT STRAIGHT_JOIN Fixed bug MDEV-5178 Valgrind warnings (Conditional jump or move depends on uninitialised value) with static_cond_pushdown=on, FROM subquery

Different results with record_cond_statistics=on and record_cond_statistics=off with SOME subquery

Details

Description

Gliffy Diagrams

Attachments

Issue Links

Activity

People

Dates

Agile