X-Original-To: pgsql-performance@postgresql.org Received: from candle.pha.pa.us (momjian.navpoint.com [207.106.42.251]) by postgresql.org (Postfix) with ESMTP id 2803E476368 for ; Thu, 6 Mar 2003 20:51:15 -0500 (EST) Received: (from pgman@localhost) by candle.pha.pa.us (8.11.6/8.10.1) id h271oaW02436; Thu, 6 Mar 2003 20:50:36 -0500 (EST) From: Bruce Momjian Message-Id: <200303070150.h271oaW02436@candle.pha.pa.us> Subject: Re: OIDs as keys In-Reply-To: <1046985774.10528.57.camel@tokyo> To: Neil Conway Date: Thu, 6 Mar 2003 20:50:36 -0500 (EST) Cc: Tom Lane , Christopher Kings-Lynne , daniel alvarez , Richard Huxton , PostgreSQL Performance X-Mailer: ELM [version 2.4ME+ PL99 (25)] MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=US-ASCII X-Archive-Number: 200303/54 X-Sequence-Number: 1352 Neil Conway wrote: > On Thu, 2003-03-06 at 16:13, Bruce Momjian wrote: > > It would be good to somehow SET the use_oids GUC value on restore start, > > and just use SET when the table is different than the default, but then > > there is no mechanism to do that when you restore a single table. > > What if the default value changes? > > IMHO, running a SET per CREATE TABLE isn't too ugly... Not ugly, but a little noisy. However, my idea of having a single SET at the top is never going to work, so I don't have a better idea. The killer for me is that you are never going to know the GUC default when you are loading, so we are _always_ going to have that SET for each table. I suppose we could set the default to off, and set it ON in the dump only when we want OID. If they set GUC to on, they will get oid's from the load, but it will cut down on the cruft and over time, they will only have the SET for cases where they really want an oid. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square, Pennsylvania 19073