From: Tom Lane <tgl@sss.pgh.pa.us>
To: Josh Berkus <josh@agliodbs.com>
cc: pgsql-performance@postgresql.org
Subject: Re: Poor plan when joining against a union containing a join
In-reply-to: <5137CDFC.1060902@agliodbs.com>
References: 
 <CALc3eMXjdWDMRr+48R2LH9fimKHD4JG9aWHDfqmmG=SQPTBVeA@mail.gmail.com>
	<5137CDFC.1060902@agliodbs.com>
Comments: In-reply-to Josh Berkus <josh@agliodbs.com>
	message dated "Wed, 06 Mar 2013 15:15:08 -0800"
Date: Thu, 07 Mar 2013 00:52:35 -0500
Message-ID: <21700.1362635555@sss.pgh.pa.us>
Precedence: bulk
Sender: pgsql-performance-owner@postgresql.org

Josh Berkus <josh@agliodbs.com> writes:
> On 03/06/2013 06:54 AM, David Leverton wrote:
>> I'm encountering very poor query plans when joining against a union,

> Actually, in case #4, Postgres *is* pushing down the join qual into the
> segments of the Union.

Yeah, but not further.  I believe the core issue here (as of 9.2) is
that we're not willing to generate parameterized paths for subquery
relations.  We could do that without a huge amount of new code,
I think, but the scary thing is how much time it might take to generate
(and then discard most of the) plans for assorted parameterizations of
complicated subqueries.

			regards, tom lane


-- 
Sent via pgsql-performance mailing list (pgsql-performance@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-performance