X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (av.hub.org [200.46.204.144]) by postgresql.org (Postfix) with ESMTP id AD4589DCC84 for ; Sun, 18 Dec 2005 06:27:53 -0400 (AST) Received: from postgresql.org ([200.46.204.71]) by localhost (av.hub.org [200.46.204.144]) (amavisd-new, port 10024) with ESMTP id 60230-06 for ; Sun, 18 Dec 2005 06:27:54 -0400 (AST) X-Greylist: from auto-whitelisted by SQLgrey- Received: from smtp.nildram.co.uk (smtp.nildram.co.uk [195.112.4.54]) by postgresql.org (Postfix) with ESMTP id 9495B9DCC88 for ; Sun, 18 Dec 2005 06:27:49 -0400 (AST) Received: from [192.168.0.4] (unknown [84.12.183.55]) by smtp.nildram.co.uk (Postfix) with ESMTP id 3FCCA26B6BF; Sun, 18 Dec 2005 10:27:49 +0000 (GMT) Subject: Re: Should Oracle outperform PostgreSQL on a complex From: Simon Riggs To: Mark Kirkwood Cc: Tom Lane , Bruce Momjian , =?ISO-8859-1?Q?P=E5l?= Stenslet , pgsql-performance@postgresql.org In-Reply-To: <43A4C348.7050703@paradise.net.nz> References: <200512170428.jBH4S2W23715@candle.pha.pa.us> <1134810953.2964.117.camel@localhost.localdomain> <8627.1134843222@sss.pgh.pa.us> <43A4C348.7050703@paradise.net.nz> Content-Type: text/plain Date: Sun, 18 Dec 2005 10:27:50 +0000 Message-Id: <1134901670.2964.140.camel@localhost.localdomain> Mime-Version: 1.0 X-Mailer: Evolution 2.2.3 (2.2.3-2.fc4) Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at hub.org X-Spam-Status: No, score=0.022 required=5 tests=[AWL=0.022] X-Spam-Score: 0.022 X-Spam-Level: X-Archive-Number: 200512/438 X-Sequence-Number: 16259 On Sun, 2005-12-18 at 15:02 +1300, Mark Kirkwood wrote: > Yeah - the quoted method of "make a cartesian product of the dimensions > and then join to the fact all at once" is not actually used (as written) > in many implementations But it is used in some, which is why I mentioned it. I gave two implementations, that is just (1) > - probably for the reasons you are pointing out. > I found these two papers whilst browsing: > > > http://www.cs.brown.edu/courses/cs227/Papers/Indexing/O'NeilGraefe.pdf > http://www.dama.upc.edu/downloads/jaguilar-2005-4.pdf > > > They seem to be describing a more subtle method making use of join > indexes and bitmapped indexes. Which is the option (2) I described. Best Regards, Simon Riggs