| Bug #7351 | incomplete result, outer query with NULL <> ALL (empty subquery) | ||
|---|---|---|---|
| Submitted: | 16 Dec 2004 13:15 | Modified: | 20 Jan 2008 21:41 |
| Reporter: | Matthias Leich | Email Updates: | |
| Status: | Duplicate | Impact on me: | |
| Category: | MySQL Server | Severity: | S2 (Serious) |
| Version: | 5.1 | OS: | Any |
| Assigned to: | Assigned Account | CPU Architecture: | Any |
[27 Dec 2004 17:23]
Igor Babaev
ChangeSet
1.2128 04/12/25 19:17:57 igor@rurik.mysql.com +4 -0
subselect.result, subselect.test:
Added test cases for bug #7351.
item_cmpfunc.cc:
Fixed bug #7351: incorrect result for a query with a
subquery returning empty set.
If in the predicate v IN (SELECT a FROM t WHERE cond)
v is null, then the result of the predicate is either
INKNOWN or FALSE. It is FALSE if the subquery returns
an empty set.
item_subselect.cc:
Fixed bug #7351: incorrect result for a query with a
subquery returning empty set.
The problem was due to not a quite legal transformation
for 'IN' subqueries. A subquery containing a predicate
of the form
v IN (SELECT a FROM t WHERE cond)
was transformed into
EXISTS(SELECT a FROM t WHERE cond AND (a=v OR a IS NULL)).
Yet, this transformation is valid only if v is not null.
If v is null, then, in the case when
(SELECT a FROM t WHERE cond) returns an empty set the value
of the predicate is FALSE, otherwise the result of the
predicate is INKNOWN.
The fix resolves this problem by changing the result
of the transformation to
EXISTS(SELECT a FROM t WHERE cond AND (v IS NULL OR (a=v OR a IS NULL)))
in the case when v is nullable.
The new transformation prevents applying the lookup
optimization for IN subqueries. To make it still
applicable we have to introduce guarded access methods.
[9 Feb 2005 18:59]
Oleksandr Byelkin
This bug will not be fixed in 4.1
[20 Jan 2008 21:41]
Sergey Petrunya
Fixed by fixes for BUG324127, BUG#8804. Changing status to duplicate.

Description: test case derived from NIST test script dml079 # TEST:0453 NULL with empty subquery of ALL, SOME, ANY DROP TABLE IF EXISTS t1, t2; CREATE TABLE t1 ( f1 BIGINT ); INSERT INTO t1 SET f1= NULL; INSERT INTO t1 SET f1= 1; CREATE TABLE t2 ( f1 BIGINT ); SELECT f1 FROM t1 WHERE f1 <> ALL ( SELECT f1 FROM t2); f1 1 NIST expects, that the record containing NULL is also selected. MySQL shows this incomplete result in version 4.1 and 5.0 . The behaviour of the MySQl server is not covered by the manual: 13.1.8.4 Subqueries with ALL <cut> Finally, if table t2 is empty, the result is TRUE. You might think the result should be UNKNOWN, but sorry, it's TRUE. So, rather oddly, the following statement is TRUE when table t2 is empty: SELECT * FROM t1 WHERE 1 > ALL (SELECT s1 FROM t2); ... My environment: - Intel PC with Linux(SuSE 9.1) - MySQL compiled from source Version 4.1 ChangeSet@1.2092.1.86, 2004-12-16 Version 5.0 ChangeSet@1.1746, 2004-12-16 How to repeat: Please execute the statements above