X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.144]) by svr1.postgresql.org (Postfix) with ESMTP id 84C553A5941 for ; Thu, 13 Jan 2005 13:45:15 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (av.hub.org [200.46.204.144]) (amavisd-new, port 10024) with ESMTP id 45099-05 for ; Thu, 13 Jan 2005 13:45:06 +0000 (GMT) Received: from flake.decibel.org (flake.decibel.org [66.143.173.58]) by svr1.postgresql.org (Postfix) with ESMTP id CD8623A44A9 for ; Thu, 13 Jan 2005 13:45:09 +0000 (GMT) Received: by flake.decibel.org (Postfix, from userid 1001) id 2DDA91C8FE; Thu, 13 Jan 2005 07:45:09 -0600 (CST) Date: Thu, 13 Jan 2005 07:45:09 -0600 From: "Jim C. Nasby" To: Hasnul Fadhly bin Hasan Cc: Richard Huxton , pgsql-performance@postgresql.org Subject: Re: Performance delay Message-ID: <20050113134509.GG67721@decibel.org> References: <41E63294.9040605@mimos.my> <41E6552C.3020801@archonet.com> <41E65802.5070400@mimos.my> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <41E65802.5070400@mimos.my> X-Operating-System: FreeBSD 4.10-RELEASE-p3 i386 X-Distributed: Join the Effort! http://www.distributed.net User-Agent: Mutt/1.5.6i X-Virus-Scanned: by amavisd-new at hub.org X-Spam-Status: No, hits=0.001 tagged_above=0 required=5 tests=AWL X-Spam-Level: X-Archive-Number: 200501/167 X-Sequence-Number: 9857 On Thu, Jan 13, 2005 at 07:14:10PM +0800, Hasnul Fadhly bin Hasan wrote: > Hi Richard, > > Thanks for the reply.. is that the case? i thought it would comply to > the where condition first.. > and after that it will format the output to what we want.. That is in fact exactly what it's doing. The second query is faster not because of the where clause, but because of the limit clause. The first query builds a list of id, long2ip(srcip), long2ip(dstip) for the timestamp range, then it orders that list and gives you the first 30. The second query builds a list of everything from sometable for the timestamp range, orders it, keeps the first 30, THEN in calculates long2ip based on that list of 30 items. > Hasnul > > Richard Huxton wrote: > > >Hasnul Fadhly bin Hasan wrote: > > > >>Hi, > >> > >>just want to share with all of you a wierd thing that i found when i > >>tested it. > >> > >>i was doing a query that will call a function long2ip to convert > >>bigint to ips. > >> > >>so the query looks something like this. > >> > >>select id, long2ip(srcip), long2ip(dstip) from sometable > >>where timestamp between timestamp '01-10-2005' and timestamp > >>'01-10-2005 23:59' order by id limit 30; > >> > >>for your info, there are about 300k rows for that timeframe. > >> > >>it cost me about 57+ secs to get the list. > >> > >>which is about the same if i query > >>select id, long2ip(srcip), long2ip(dstip) from sometable > >>where timestamp between timestamp '01-10-2005' and timestamp > >>'01-10-2005 23:59' > >> > >>it will cost me about 57+ secs also. > >> > >>Now if i did this > >>select id,long2ip(srcip), long2ip(dstip) from ( > >>* from sometable > >>where timestamp between timestamp '01-10-2005' and timestamp > >>'01-10-2005 23:59' order by id limit 30) as t; > >> > >>it will cost me about 3+ secs > > > > > >The difference will be that in the final case you only make 30 calls > >to long2ip() whereas in the first two you call it 300,000 times and > >then throw away most of them. > >Try running EXPLAIN ANALYSE ... for both - that will show how PG is > >planning the query. > >-- > > Richard Huxton > > Archonet Ltd > > > > > > > ---------------------------(end of broadcast)--------------------------- > TIP 6: Have you searched our list archives? > > http://archives.postgresql.org > -- Jim C. Nasby, Database Consultant decibel@decibel.org Give your computer some brain candy! www.distributed.net Team #1828 Windows: "Where do you want to go today?" Linux: "Where do you want to go tomorrow?" FreeBSD: "Are you guys coming, or what?"