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 88AA85E46CC for ; Fri, 27 Aug 2004 00:39:54 -0300 (ADT) 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 26895-08 for ; Fri, 27 Aug 2004 03:39:58 +0000 (GMT) Received: from stark.xeocode.com (gsstark.mtl.istop.com [66.11.160.162]) by svr1.postgresql.org (Postfix) with ESMTP id 936585E3639 for ; Fri, 27 Aug 2004 00:39:50 -0300 (ADT) Received: from localhost ([127.0.0.1] helo=stark.xeocode.com) by stark.xeocode.com with smtp (Exim 3.36 #1 (Debian)) id 1C0XaZ-0000B7-00; Thu, 26 Aug 2004 23:39:43 -0400 To: Bruce Momjian Cc: "J. Andrew Rogers" , Magnus Hagander , pgsql-performance@postgresql.org, mischa.sandberg@telus.net Subject: Re: Equivalent praxis to CLUSTERED INDEX? References: <200408270145.i7R1jop04185@candle.pha.pa.us> In-Reply-To: <200408270145.i7R1jop04185@candle.pha.pa.us> From: Greg Stark Organization: The Emacs Conspiracy; member since 1992 Date: 26 Aug 2004 23:39:42 -0400 Message-ID: <87pt5dnta9.fsf@stark.xeocode.com> Lines: 28 User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.3 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at hub.org X-Spam-Status: No, hits=0.0 tagged_above=0.0 required=5.0 tests= X-Spam-Level: X-Archive-Number: 200408/382 X-Sequence-Number: 8018 Bruce Momjian writes: > Updated TODO item: > > o Automatically maintain clustering on a table > > This would require some background daemon to maintain clustering > during periods of low usage. It might also require tables to be only > paritally filled for easier reorganization. It also might require > creating a merged heap/index data file so an index lookup would > automatically access the heap data too. Fwiw, I would say the first "would" is also a "might". None of the previous discussions here presumed a maintenance daemon. The discussions before talked about a mechanism to try to place new tuples as close as possible to the proper index position. I would also suggest making some distinction between a cluster system similar to what we have now but improved to maintain the clustering continuously, and an actual index-organized-table where the tuples are actually only stored in a btree structure. They're two different approaches to similar problems. But they might both be useful to have, and have markedly different implementation details. -- greg