Received: from localhost (maia-4.hub.org [200.46.204.183]) by postgresql.org (Postfix) with ESMTP id 009BC9FB4EF for ; Mon, 18 Jun 2007 05:49:36 -0300 (ADT) Received: from postgresql.org ([200.46.204.71]) by localhost (mx1.hub.org [200.46.204.183]) (amavisd-maia, port 10024) with ESMTP id 25859-03 for ; Mon, 18 Jun 2007 05:49:33 -0300 (ADT) X-Greylist: from auto-whitelisted by SQLgrey-1.7.5 Received: from oxford.xeocode.com (unknown [87.127.95.194]) by postgresql.org (Postfix) with ESMTP id D77879FB3FE for ; Mon, 18 Jun 2007 05:49:32 -0300 (ADT) Received: from localhost ([127.0.0.1] helo=oxford.xeocode.com) by oxford.xeocode.com with esmtp (Exim 4.67) (envelope-from ) id 1I0CvQ-0001dR-T8; Mon, 18 Jun 2007 09:49:28 +0100 From: Gregory Stark To: "Heikki Linnakangas" Cc: "Tom Lane" , "Jaime Casanova" , "PostgreSQL-patches" Subject: Re: Maintaining cluster order on insert In-Reply-To: <46761DB7.7070505@enterprisedb.com> (Heikki Linnakangas's message of "Mon, 18 Jun 2007 06:52:55 +0100") Organization: EnterpriseDB References: <200705152223.l4FMNJa17969@momjian.us> <464A33AB.1050401@enterprisedb.com> <20070516003810.GI20707@nasby.net> <464B15EF.4000808@enterprisedb.com> <464D96D6.1030608@enterprisedb.com> <464F410C.8090900@enterprisedb.com> <10328.1182113758@sss.pgh.pa.us> <46761DB7.7070505@enterprisedb.com> X-Draft-From: ("nnimap+mail01.enterprisedb.com:INBOX.patches" 1744) Date: Mon, 18 Jun 2007 09:49:28 +0100 Message-ID: <87wsy1hckn.fsf@oxford.xeocode.com> User-Agent: Gnus/5.110006 (No Gnus v0.6) Emacs/21.4 (gnu/linux) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: Maia Mailguard 1.0.1 X-Archive-Number: 200706/177 X-Sequence-Number: 2793 "Heikki Linnakangas" writes: > The reason for switching to the new API instead of the amsuggestblock API is > CPU overhead. It avoids constructing the IndexTuple twice and descending the > tree twice. I wonder if it's possible to finesse this. Have the suggestblock function remember the last block it suggested and somehow recognize when it's given the same tuple to index. Keeping the block pinned would still be the sticky point though. -- Gregory Stark EnterpriseDB http://www.enterprisedb.com