Re: Making Vars outer-join aware - Mailing list pgsql-hackers

From Richard Guo
Subject Re: Making Vars outer-join aware
Date
Msg-id CAMbWs4-K8_yqARnrmRB5=xAWTJGgTpimQinvaa-k6c8f6Hideg@mail.gmail.com
Whole thread Raw
In response to Re: Making Vars outer-join aware  (Justin Pryzby <pryzby@telsasoft.com>)
Responses Re: Making Vars outer-join aware
Re: Making Vars outer-join aware
List pgsql-hackers

On Mon, Feb 13, 2023 at 7:58 AM Justin Pryzby <pryzby@telsasoft.com> wrote:
The patch broke this query:

select from pg_inherits inner join information_schema.element_types
right join (select from pg_constraint as sample_2) on true
on false, lateral (select scope_catalog, inhdetachpending from pg_publication_namespace limit 3);
ERROR:  could not devise a query plan for the given query
 
Thanks for the report!  I've looked at it a little bit and traced down
to function have_unsafe_outer_join_ref().  The comment there says

 * In practice, this test never finds a problem ...
 * ...
 * It still seems worth checking
 * as a backstop, but we don't go to a lot of trouble: just reject if the
 * unsatisfied part includes any outer-join relids at all.

This seems not correct as showed by the counterexample.  ISTM that we
need to do the check honestly as what the other comment says

 * If the parameterization is only partly satisfied by the outer rel,
 * the unsatisfied part can't include any outer-join relids that could
 * null rels of the satisfied part.

The NOT_USED part of code is doing this check.  But I think we need a
little tweak.  We should check the nullable side of related outer joins
against the satisfied part, rather than inner_paramrels.  Maybe
something like attached.

However, this test seems to cost some cycles after the change.  So I
wonder if it's worthwhile to perform it, considering that join order
restrictions should be able to guarantee there is no problem here.

BTW, here is a simplified query that can trigger this issue on HEAD.

select * from t1 inner join t2 left join (select null as c from t3 left join t4 on true) as sub on true on true, lateral (select c, t1.a from t5 offset 0 ) ss;

Thanks
Richard
Attachment

pgsql-hackers by date:

Previous
From: "Drouvot, Bertrand"
Date:
Subject: Re: Reconcile stats in find_tabstat_entry() and get rid of PgStat_BackendFunctionEntry
Next
From: Michael Paquier
Date:
Subject: Re: recovery modules