X-Original-To: pgsql-sql-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.144]) by svr1.postgresql.org (Postfix) with ESMTP id BCE3152872 for ; Thu, 2 Jun 2005 12:34:00 -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 36470-02 for ; Thu, 2 Jun 2005 15:33:59 +0000 (GMT) Received: from mx-00.sil.at (mx-00.sil.at [62.116.68.196]) by svr1.postgresql.org (Postfix) with ESMTP id 8D5D35286E for ; Thu, 2 Jun 2005 12:33:56 -0300 (ADT) Received: (qmail-ldap/ctrl 75713 invoked from network); 2 Jun 2005 15:33:58 -0000 Received: from unknown (HELO io.sonorys.at) ([213.235.243.42]) (envelope-sender ) by mx-00.sil.at (qmail-ldap-1.03) with SMTP for ; 2 Jun 2005 15:33:58 -0000 Received: from neptun.sonorys.at (neptun.sonorys.at [172.24.1.17]) by io.sonorys.at (8.11.7+Sun/8.11.7) with ESMTP id j52FY0G08584 for ; Thu, 2 Jun 2005 17:34:00 +0200 (MEST) Received: by neptun.sonorys.at with Internet Mail Service (5.5.2653.19) id ; Thu, 2 Jun 2005 17:34:00 +0200 Message-ID: From: =?ISO-8859-1?Q?K=D6PFERL_Robert?= To: pgsql-sql@postgresql.org Subject: Splitting a table for performance reasons Date: Thu, 2 Jun 2005 17:33:59 +0200 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Type: text/plain; charset="ISO-8859-1" X-Virus-Scanned: by amavisd-new at hub.org X-Spam-Status: No, hits=0 tagged_above=0 required=5 tests= X-Spam-Level: X-Archive-Number: 200506/26 X-Sequence-Number: 21627 Hi, I've got a table whose records are more less big. There's however jus one Int-column changed frequently. According to postgres' MVCC a whole record gets written, even if just one bit was changed. I think of splitting the table now in two parts, connected via the former PK. so like: a|b|data | othercol | int -> a|b| data|othercol + a|int Will this solve my ongoing performance problem? Or is the overhead for referencial intergrity and system columns bigger than what this gives? BTW: How can I find out how many bytes a record consumes (having just fixed size data)?