Received: from localhost (maia-1.hub.org [200.46.204.191])
	by postgresql.org (Postfix) with ESMTP id 510A19F9D61
	for <pgsql-sql-postgresql.org@postgresql.org>;
	Mon, 13 Aug 2007 16:40:46 -0300 (ADT)
Received: from postgresql.org ([200.46.204.71])
	by localhost (mx1.hub.org [200.46.204.191]) (amavisd-maia, port 10024)
	with ESMTP id 92274-04 for <pgsql-sql-postgresql.org@postgresql.org>;
	Mon, 13 Aug 2007 16:40:43 -0300 (ADT)
X-Greylist: from auto-whitelisted by SQLgrey-1.7.5
Received: from sss.pgh.pa.us (sss.pgh.pa.us [66.207.139.130])
	by postgresql.org (Postfix) with ESMTP id 2587A9F9D2F
	for <pgsql-sql@postgresql.org>; Mon, 13 Aug 2007 16:40:43 -0300 (ADT)
Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1])
	by sss.pgh.pa.us (8.14.1/8.14.1) with ESMTP id l7DJbe6v020124;
	Mon, 13 Aug 2007 15:37:41 -0400 (EDT)
To: ksimpson@mailchannels.com
cc: pgsql-sql@postgresql.org
Subject: Re: Comparing two slices within one table efficiently 
In-reply-to: 
 <1694202581-1187027999-cardhu_decombobulator_blackberry.rim.net-401169594-@bxe119.bisx.prod.on.blackberry>
References: 
 <1694202581-1187027999-cardhu_decombobulator_blackberry.rim.net-401169594-@bxe119.bisx.prod.on.blackberry>
Comments: In-reply-to "=?utf-8?B?S2VuIFNpbXBzb24=?="
	<ksimpson@mailchannels.com>
	message dated "Mon, 13 Aug 2007 17:58:16 -0000"
Date: Mon, 13 Aug 2007 15:37:40 -0400
Message-ID: <20123.1187033860@sss.pgh.pa.us>
From: Tom Lane <tgl@sss.pgh.pa.us>
X-Virus-Scanned: Maia Mailguard 1.0.1
X-Archive-Number: 200708/112
X-Sequence-Number: 29050

"=?utf-8?B?S2VuIFNpbXBzb24=?=" <ksimpson@mailchannels.com> writes:
> select * from t t1 where exists (select 1 from t t2 where
> t2.mta_id=t1.mta_id and t2.domain_id=t1.domain_id and (t2.attribute1
> != t1.attribute1 or t2.attribute2 != t1.attribute2)

> This query takes millenia...

Yeah, because you're effectively forcing the least efficient style of
join --- a nestloop is generally going to suck for a full-table join,
even if you've got indexes.  Try something like this:

select * from
  t t1 join t t2 on (t2.mta_id=t1.mta_id and t2.domain_id=t1.domain_id)
where (t2.attribute1 != t1.attribute1 or t2.attribute2 != t1.attribute2)

Make sure you've got work_mem cranked up to something appropriate.

			regards, tom lane