Re: BUG #18465: Wrong results from SELECT DISTINCT MIN in scalar subquery using HashAggregate

> On Tue, May 14, 2024 at 09:14:34PM +0000, PG Bug reporting form wrote:
>
> postgres=# SELECT ( SELECT DISTINCT MIN( col_int ) FROM B AS
> SUBQUERY1_t1 WHERE SUBQUERY1_t1 .col_int = table1 .col_int ), table1
> .col_int FROM CC AS table1 ;
> min | col_int
> --------+---------
> <null> | <null>
> <null> | 1 <------- *** wrong ***
> <null> | 2
> (3 rows)
>
> postgres=# SELECT ( SELECT DISTINCT MIN( col_int ) FROM B AS
> SUBQUERY1_t1 WHERE SUBQUERY1_t1 .col_int = table1 .col_int ), table1
> .col_int FROM CC AS table1 ;
> min | col_int
> -----+---------
> 1 | 1 <------- *** correct ***
> 1 | 2 <------- *** wrong ***
> 1 | <null> <------- *** wrong ***
> (3 rows)

Thanks, I could reproduce this in v15 and v16. Interestingly enough it
was fixed in d0d44049d1, but looks like due to lack of error reports the
fix was applied only on the main branch. I assume this one makes it
necessary to back-patch the fix.