Skip to content

Fix join condition lost after pull up sublink to join #1961

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
leborchuk merged 1 commit into from
Sep 9, 2026
+219 −0
Merged

Fix join condition lost after pull up sublink to join #1961

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
7 changes: 7 additions & 0 deletions src/backend/cdb/cdbsubselect.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -451,6 +451,13 @@ SubqueryToJoinWalker(Node *node, ConvertSubqueryToJoinContext *context)
*/
context->safeToConvert = false;
}
else
{
/*
* For other expressions, we should keep them in original place.
*/
context->innerQual = make_and_qual(context->innerQual, node);
}

return;
}
Expand Down
85 changes: 85 additions & 0 deletions src/test/regress/expected/subselect.out
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2098,3 +2098,88 @@ select (select max((select t.i))) from t;
(1 row)

drop table t;
-- Fix join condition expression lost as pull up sublink to join.
create table tl1(a int, b int, c int, d int) distributed by (a);
create table tl2(a int, b int, c int, d int) distributed by (a);
create table tl3(a int, b int, c int, d int) distributed by (a);
create table tl4(a int, b int, c int, d int) distributed by (a);
insert into tl1 values (-1, 3, 1, 0);
insert into tl2 values (2, 1, 1, 0);
insert into tl2 values (3, 1, 1, 0);
insert into tl2 values (1, 1, 1, 0);
insert into tl3 values (9, 9, 1, 9);
insert into tl4 values (-1, -1, -1, -1);
explain(costs off, verbose on)
select * from tl1
where
tl1.b = (
select
max(tl2.a)
from
tl2 join tl4
on tl4.d = tl2.d
where
tl2.b = tl1.c
);
QUERY PLAN
---------------------------------------------------------------------------------------------------------
Gather Motion 3:1 (slice1; segments: 3)
Output: tl1.a, tl1.b, tl1.c, tl1.d
-> Hash Join
Output: tl1.a, tl1.b, tl1.c, tl1.d
Inner Unique: true
Hash Cond: ((tl1.b = "Expr_SUBQUERY".csq_c1) AND (tl1.c = "Expr_SUBQUERY".csq_c0))
-> Seq Scan on public.tl1
Output: tl1.a, tl1.b, tl1.c, tl1.d
-> Hash
Output: "Expr_SUBQUERY".csq_c1, "Expr_SUBQUERY".csq_c0
-> Broadcast Motion 3:3 (slice2; segments: 3)
Output: "Expr_SUBQUERY".csq_c1, "Expr_SUBQUERY".csq_c0
-> Subquery Scan on "Expr_SUBQUERY"
Output: "Expr_SUBQUERY".csq_c1, "Expr_SUBQUERY".csq_c0
-> Finalize GroupAggregate
Output: tl2.b, max(tl2.a)
Group Key: tl2.b
-> Sort
Output: tl2.b, (PARTIAL max(tl2.a))
Sort Key: tl2.b
-> Redistribute Motion 3:3 (slice3; segments: 3)
Output: tl2.b, (PARTIAL max(tl2.a))
Hash Key: tl2.b
-> Streaming Partial HashAggregate
Output: tl2.b, PARTIAL max(tl2.a)
Group Key: tl2.b
-> Hash Join
Output: tl2.b, tl2.a
Hash Cond: (tl4.d = tl2.d)
-> Broadcast Motion 3:3 (slice4; segments: 3)
Output: tl4.d
-> Seq Scan on public.tl4
Output: tl4.d
-> Hash
Output: tl2.b, tl2.a, tl2.d
-> Seq Scan on public.tl2
Output: tl2.b, tl2.a, tl2.d
Settings: gp_cte_sharing = 'off', optimizer = 'off'
Optimizer: Postgres query optimizer
(39 rows)

select * from tl1
where
tl1.b = (
select
max(tl2.a)
from
tl2 join tl4
on tl4.d = tl2.d
where
tl2.b = tl1.c
);
a | b | c | d
---+---+---+---
(0 rows)

drop table tl1;
drop table tl2;
drop table tl3;
drop table tl4;
84 changes: 84 additions & 0 deletions src/test/regress/expected/subselect_optimizer.out
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2185,3 +2185,87 @@ select (select max((select t.i))) from t;
(1 row)

drop table t;
-- Fix join condition expression lost as pull up sublink to join.
create table tl1(a int, b int, c int, d int) distributed by (a);
create table tl2(a int, b int, c int, d int) distributed by (a);
create table tl3(a int, b int, c int, d int) distributed by (a);
create table tl4(a int, b int, c int, d int) distributed by (a);
insert into tl1 values (-1, 3, 1, 0);
insert into tl2 values (2, 1, 1, 0);
insert into tl2 values (3, 1, 1, 0);
insert into tl2 values (1, 1, 1, 0);
insert into tl3 values (9, 9, 1, 9);
insert into tl4 values (-1, -1, -1, -1);
explain(costs off, verbose on)
select * from tl1
where
tl1.b = (
select
max(tl2.a)
from
tl2 join tl4
on tl4.d = tl2.d
where
tl2.b = tl1.c
);
QUERY PLAN
------------------------------------------------------------------------------------------------
Gather Motion 3:1 (slice1; segments: 3)
Output: tl1.a, tl1.b, tl1.c, tl1.d
-> Hash Join
Output: tl1.a, tl1.b, tl1.c, tl1.d
Hash Cond: ((tl1.b = (max(tl2.a))) AND (tl1.c = tl2.b))
-> Redistribute Motion 3:3 (slice2; segments: 3)
Output: tl1.a, tl1.b, tl1.c, tl1.d
Hash Key: tl1.c
-> Seq Scan on public.tl1
Output: tl1.a, tl1.b, tl1.c, tl1.d
-> Hash
Output: (max(tl2.a)), tl2.b
-> GroupAggregate
Output: max(tl2.a), tl2.b
Group Key: tl2.b
-> Sort
Output: tl2.a, tl2.b
Sort Key: tl2.b
-> Redistribute Motion 3:3 (slice3; segments: 3)
Output: tl2.a, tl2.b
Hash Key: tl2.b
-> Hash Join
Output: tl2.a, tl2.b
Hash Cond: (tl2.d = tl4.d)
-> Redistribute Motion 3:3 (slice4; segments: 3)
Output: tl2.a, tl2.b, tl2.d
Hash Key: tl2.d
-> Seq Scan on public.tl2
Output: tl2.a, tl2.b, tl2.d
-> Hash
Output: tl4.d
-> Redistribute Motion 3:3 (slice5; segments: 3)
Output: tl4.d
Hash Key: tl4.d
-> Seq Scan on public.tl4
Output: tl4.d
Settings: gp_cte_sharing = 'off', optimizer = 'on'
Optimizer: GPORCA
(38 rows)

select * from tl1
where
tl1.b = (
select
max(tl2.a)
from
tl2 join tl4
on tl4.d = tl2.d
where
tl2.b = tl1.c
);
a | b | c | d
---+---+---+---
(0 rows)

drop table tl1;
drop table tl2;
drop table tl3;
drop table tl4;
43 changes: 43 additions & 0 deletions src/test/regress/sql/subselect.sql
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1035,3 +1035,46 @@ select (select max((select t.i))) from t;
select (select max((select t.i))) from t;

drop table t;

-- Fix join condition expression lost as pull up sublink to join.
create table tl1(a int, b int, c int, d int) distributed by (a);
create table tl2(a int, b int, c int, d int) distributed by (a);
create table tl3(a int, b int, c int, d int) distributed by (a);
create table tl4(a int, b int, c int, d int) distributed by (a);

insert into tl1 values (-1, 3, 1, 0);
insert into tl2 values (2, 1, 1, 0);
insert into tl2 values (3, 1, 1, 0);
insert into tl2 values (1, 1, 1, 0);
insert into tl3 values (9, 9, 1, 9);
insert into tl4 values (-1, -1, -1, -1);

explain(costs off, verbose on)
select * from tl1
where
tl1.b = (
select
max(tl2.a)
from
tl2 join tl4
on tl4.d = tl2.d
where
tl2.b = tl1.c
);

select * from tl1
where
tl1.b = (
select
max(tl2.a)
from
tl2 join tl4
on tl4.d = tl2.d
where
tl2.b = tl1.c
);

drop table tl1;
drop table tl2;
drop table tl3;
drop table tl4;
Loading