Movatterモバイル変換

[0]ホーム
URL:

Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

 Sign up
Appearance settings

Commit798c017

Browse files
committed
remove_rel_from_query() must clean up PlaceHolderVar.phrels fields.
While we got away with this sloppiness before, it's not okay nowthatfee7b77 caused build_joinrel_tlist() to make use of phrels.Per report from Robins Tharakan.Richard Guo (some cosmetic tweaks by me)Discussion:https://postgr.es/m/CAMbWs4_ngw9sKxpTE8hqk=-ooVX_CQP3DarA4HzkRMz_JKpTrA@mail.gmail.com
1 parentb7e84c6 commit798c017

File tree

3 files changed

+33
-7
lines changed

3 files changed

+33
-7
lines changed

‎src/backend/optimizer/plan/analyzejoins.c

Lines changed: 10 additions & 7 deletions
Original file line numberDiff line numberDiff line change
@@ -316,12 +316,8 @@ join_is_removable(PlannerInfo *root, SpecialJoinInfo *sjinfo)
316316
* Remove the target relid from the planner's data structures, having
317317
* determined that there is no need to include it in the query.
318318
*
319-
* We are not terribly thorough here. We must make sure that the rel is
320-
* no longer treated as a baserel, and that attributes of other baserels
321-
* are no longer marked as being needed at joins involving this rel.
322-
* Also, join quals involving the rel have to be removed from the joininfo
323-
* lists, but only if they belong to the outer join identified by ojrelid
324-
* and joinrelids.
319+
* We are not terribly thorough here. We only bother to update parts of
320+
* the planner's data structures that will actually be consulted later.
325321
*/
326322
staticvoid
327323
remove_rel_from_query(PlannerInfo*root,intrelid,intojrelid,
@@ -429,11 +425,18 @@ remove_rel_from_query(PlannerInfo *root, int relid, int ojrelid,
429425
}
430426
else
431427
{
428+
PlaceHolderVar*phv=phinfo->ph_var;
429+
432430
phinfo->ph_eval_at=bms_del_member(phinfo->ph_eval_at,relid);
433431
phinfo->ph_eval_at=bms_del_member(phinfo->ph_eval_at,ojrelid);
434-
Assert(!bms_is_empty(phinfo->ph_eval_at));
432+
Assert(!bms_is_empty(phinfo->ph_eval_at));/* checked previously */
435433
phinfo->ph_needed=bms_del_member(phinfo->ph_needed,relid);
436434
phinfo->ph_needed=bms_del_member(phinfo->ph_needed,ojrelid);
435+
/* ph_needed might or might not become empty */
436+
phv->phrels=bms_del_member(phv->phrels,relid);
437+
phv->phrels=bms_del_member(phv->phrels,ojrelid);
438+
Assert(!bms_is_empty(phv->phrels));
439+
Assert(phv->phnullingrels==NULL);/* no need to adjust */
437440
}
438441
}
439442

‎src/test/regress/expected/join.out

Lines changed: 16 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -5052,6 +5052,22 @@ from int4_tbl as t1
50525052
One-Time Filter: false
50535053
(5 rows)
50545054

5055+
-- per further discussion of bug #17781
5056+
explain (costs off)
5057+
select ss1.x
5058+
from (select f1/2 as x from int4_tbl i4 left join a on a.id = i4.f1) ss1
5059+
right join int8_tbl i8 on true
5060+
where current_user is not null; -- this is to add a Result node
5061+
QUERY PLAN
5062+
-----------------------------------------------
5063+
Result
5064+
One-Time Filter: (CURRENT_USER IS NOT NULL)
5065+
-> Nested Loop Left Join
5066+
-> Seq Scan on int8_tbl i8
5067+
-> Materialize
5068+
-> Seq Scan on int4_tbl i4
5069+
(6 rows)
5070+
50555071
-- check that join removal works for a left join when joining a subquery
50565072
-- that is guaranteed to be unique by its GROUP BY clause
50575073
explain (costs off)

‎src/test/regress/sql/join.sql

Lines changed: 7 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -1813,6 +1813,13 @@ from int4_tbl as t1
18131813
ont5.q1=t7.q2)
18141814
on false;
18151815

1816+
-- per further discussion of bug #17781
1817+
explain (costs off)
1818+
selectss1.x
1819+
from (select f1/2as xfrom int4_tbl i4left join aona.id=i4.f1) ss1
1820+
right join int8_tbl i8on true
1821+
wherecurrent_useris not null;-- this is to add a Result node
1822+
18161823
-- check that join removal works for a left join when joining a subquery
18171824
-- that is guaranteed to be unique by its GROUP BY clause
18181825
explain (costs off)

0 commit comments

Comments
 (0)
[8]ページ先頭
©2009-2025 Movatter.jp