Re: GEQO and join_collapse_limit correlation - Mailing list pgsql-performance

From Tom Lane
Subject Re: GEQO and join_collapse_limit correlation
Date
Msg-id 12214.1515184156@sss.pgh.pa.us
Whole thread Raw
In response to Re: GEQO and join_collapse_limit correlation  (Juan José Santamaría Flecha <juanjo.santamaria@gmail.com>)
Responses Re: GEQO and join_collapse_limit correlation
List pgsql-performance
=?UTF-8?Q?Juan_Jos=C3=A9_Santamar=C3=ADa_Flecha?= <juanjo.santamaria@gmail.com> writes:
> In an environment with geqo_threshold=1 and join_collapse_limit=1, would
> the planner be GEQO exclusive (and syntactic)?

GEQO's only function, basically, is to search for the join order to use.
If you're constraining the join order completely with
join_collapse_limit=1 then forcing the GEQO path to be taken would just
add pointless overhead.  (If it does anything at all ... I don't remember
the logic exactly but we might be bright enough not to bother with GEQO in
such a situation, regardless of geqo_threshold.)

            regards, tom lane


pgsql-performance by date:

Previous
From: Juan José Santamaría Flecha
Date:
Subject: Re: GEQO and join_collapse_limit correlation
Next
From: Juan José Santamaría Flecha
Date:
Subject: Re: GEQO and join_collapse_limit correlation