Re: postgres 9 query performance

["Ross J. Reedstrom" <reedstrm(at)rice(dot)edu>]

On Sun, Jan 30, 2011 at 05:18:15PM -0500, Tom Lane wrote:
> Andres Freund <andres(at)anarazel(dot)de> writes:
> > What happens if you change the
> >     left join event.origin on event.id = origin.eventid
> > into
> >     join event.origin on event.id = origin.eventid
> > ?
> 
> > The EXISTS() requires that origin is not null anyway. (Not sure why the 
> > planner doesn't recognize that though).
> 
> Sloppy thinking in reduce_outer_joins() is why.  Fixed now:
> http://git.postgresql.org/gitweb?p=postgresql.git;a=commitdiff;h=1df57f63f3f60c684aa8918910ac410e9c780713
> 
> 			regards, tom lane

This is one of the reasons I love open source in general, and PostgreSQL
in particular: Tom has the bandwidth to notice these kinds of
workarounds being discussed on support lists, and turn them immediately
into improvements in the planner. Partly because (I assume, based on
the commit message) Andres's parenthetical comment red-flagged it for
him, since he knew he could trust Andres's opinion that there was
probably a planner improvement hiding here. Amazing!

Ross
-- 
Ross Reedstrom, Ph.D.                                 reedstrm(at)rice(dot)edu
Systems Engineer & Admin, Research Scientist        phone: 713-348-6166
Connexions                  http://cnx.org            fax: 713-348-3665
Rice University MS-375, Houston, TX 77005
GPG Key fingerprint = F023 82C8 9B0E 2CC6 0D8E  F888 D3AE 810E 88F0 BEDE