Re: Removing unneeded self joins

Hi, Alexander!

On Tue, Apr 30, 2024 at 9:00 AM Alexander Lakhin <exclusion(at)gmail(dot)com> wrote:
> 23.10.2023 12:47, Alexander Korotkov wrote:
> > I think this patch makes substantial improvement to query planning.
> > It has received plenty of reviews. The code is currently in quite
> > good shape. I didn't manage to find the cases when this optimization
> > causes significant overhead to planning time. Even if such cases will
> > be spotted there is a GUC option to disable this feature. So, I'll
> > push this if there are no objections.
>
> I've discovered another failure, introduced by d3d55ce57.
> Please try the following:
> CREATE TABLE t (a int unique, b float);
> SELECT * FROM t NATURAL JOIN LATERAL
> (SELECT * FROM t t2 TABLESAMPLE SYSTEM (t.b)) t2;
>
> With asserts enabled, it triggers
> TRAP: failed Assert("!bms_is_member(rti, lateral_relids)"), File: "initsplan.c", Line: 697, PID: 3074054
> ExceptionalCondition at assert.c:52:13
> create_lateral_join_info at initsplan.c:700:8
> query_planner at planmain.c:257:2
> grouping_planner at planner.c:1523:17
> subquery_planner at planner.c:1098:2
> standard_planner at planner.c:415:9
> planner at planner.c:282:12
> pg_plan_query at postgres.c:904:9
> pg_plan_queries at postgres.c:996:11
> exec_simple_query at postgres.c:1193:19
> PostgresMain at postgres.c:4684:27
>
> With no asserts, I get:
> ERROR: failed to construct the join relation
>
> Please take a look at this.

I'm looking into this, thank you!

------
Regards,
Alexander Korotkov