X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [64.117.224.130]) by svr1.postgresql.org (Postfix) with ESMTP id 71923D1B4E3 for ; Mon, 13 Oct 2003 06:16:30 +0000 (GMT) Received: from svr1.postgresql.org ([64.117.224.193]) by localhost (neptune.hub.org [64.117.224.130]) (amavisd-new, port 10024) with ESMTP id 30483-05 for ; Mon, 13 Oct 2003 03:15:41 -0300 (ADT) Received: from boats_exch01.boats.com (h209-17-182-106.gtconnect.net [209.17.182.106]) by svr1.postgresql.org (Postfix) with ESMTP id 40762D1B502 for ; Mon, 13 Oct 2003 03:15:40 -0300 (ADT) Received: by BOATS_EXCH01 with Internet Mail Service (5.5.2653.19) id <4S92KHVS>; Sun, 12 Oct 2003 23:13:37 -0700 Received: from GRIFFITHS2 ([192.168.12.192]) by boats_exch01.boats.com with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2653.13) id 4S92KHVQ; Sun, 12 Oct 2003 23:13:29 -0700 From: David Griffiths To: pgsql-performance@postgresql.org Message-ID: <098b01c39152$921632e0$6501a8c0@griffiths2> References: <052001c39030$1ccc8020$6501a8c0@griffiths2> <05a401c3903f$a8524850$6501a8c0@griffiths2> <20031011153251.F57866@megazone.bigpanda.com> <080d01c3912a$1356dea0$6501a8c0@griffiths2> <20031012184536.N5307@megazone.bigpanda.com> Subject: Re: Another weird one with an UPDATE Date: Sun, 12 Oct 2003 23:23:33 -0700 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 5.50.4922.1500 X-MIMEOLE: Produced By Microsoft MimeOLE V5.50.4925.2800 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200310/410 X-Sequence-Number: 4158 It's a slight improvement, but that could be other things as well. I'd read that how you tune Postgres will determine how the optimizer works on a query (sequential scan vs index scan). I am going to post all I've done with tuning tommorow, and see if I've done anything dumb. I've found some contradictory advice, and I'm still a bit hazy on how/why Postgres trusts the OS to do caching. I'll post it all tommorow. ---------------------------------------------------------------------------- ---------------------------------------------------------------------------- ---------------- Merge Join (cost=11819.21..15258.55 rows=12007 width=752) (actual time=4107.64..5587.81 rows=20880 loops=1) Merge Cond: ("outer".commercial_entity_id = "inner".commercial_entity_id) -> Index Scan using comm_serv_comm_ent_id_i on commercial_service cs (cost=0.00..3015.53 rows=88038 width=12) (actual time=0.05..487.23 rows=88038 loops=1) -> Sort (cost=11819.21..11846.08 rows=10752 width=740) (actual time=3509.07..3955.15 rows=25098 loops=1) Sort Key: ce.commercial_entity_id -> Merge Join (cost=0.00..9065.23 rows=10752 width=740) (actual time=0.18..2762.13 rows=7990 loops=1) Merge Cond: ("outer".user_account_id = "inner".user_account_id) -> Index Scan using user_account_pkey on user_account (cost=0.00..8010.39 rows=72483 width=716) (actual time=0.05..2220.86 rows=72483 loops=1) -> Index Scan using comm_ent_usr_acc_id_i on commercial_entity ce (cost=0.00..4787.69 rows=78834 width=24) (actual time=0.02..55.64 rows=7991 loops=1) Total runtime: 226239.77 msec (10 rows) David ----- Original Message ----- From: "Stephan Szabo" To: "David Griffiths" Cc: Sent: Sunday, October 12, 2003 6:48 PM Subject: Re: [PERFORM] Another weird one with an UPDATE > On Sun, 12 Oct 2003, David Griffiths wrote: > > > [snip] > > > > > I think you want something like: > > > UPDATE user_account SET last_name = 'abc' > > > WHERE EXISTS (SELECT 1 FROM commercial_entity ce, commercial_service cs > > > WHERE user_account.user_account_id = ce.user_account_id AND > > > ce.commercial_entity_id = cs.commercial_entity_id); > > > > Unfort, this is still taking a long time. > > Hmm, does > UPDATE user_account SET last_name='abc' > FROM commercial_entity ce, commercial_service cs > WHERE user_account.user_account_id = ce.user_account_id AND > ce.commercial_entity_id=cs.commercial_entity_id; > give the right results... That might end up being faster.