Received: from bubbles.electricutopia.net (bubbles.electricutopia.net [63.214.178.76]) by mail.postgresql.org (8.11.1/8.11.1) with ESMTP id f0OJYmq59714 for ; Wed, 24 Jan 2001 14:34:48 -0500 (EST) (envelope-from dave@slickness.org) Received: from localhost (dave@localhost) by bubbles.electricutopia.net (8.9.3/8.9.3) with ESMTP id LAA23493 for ; Wed, 24 Jan 2001 11:31:57 -0800 Date: Wed, 24 Jan 2001 11:31:57 -0800 (PST) From: David Olbersen X-X-Sender: To: Subject: Please don't kill me! Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Archive-Number: 200101/335 X-Sequence-Number: 1008 I have two statements that accomplish the same task and I'm trying to decide which to use. One uses a sub-select, and the other just does a few more joins. I expect that giving the SELECT statement's themseleves won't get me much help, so here is the output of the EXPLAIN query that I ran on both of them. I read the FAQ on EXPLAIN a bit but I'm still confused. So could somebody help me understand why it appears as though the first query will run much faster (?) than the second? --snip!-- Nested Loop (cost=81.80..114.17 rows=33 width=68) InitPlan -> Seq Scan on l_portal_statuses (cost=0.00..22.50 rows=10 width=4) -> Merge Join (cost=81.80..86.63 rows=3 width=52) -> Merge Join (cost=59.13..63.43 rows=33 width=44) -> Sort (cost=22.67..22.67 rows=10 width=28) -> Seq Scan on contacts m (cost=0.00..22.50 rows=10 width=28) -> Sort (cost=36.47..36.47 rows=333 width=16) -> Seq Scan on buildings b (cost=0.00..22.50 rows=333 width=16) -> Sort (cost=22.67..22.67 rows=10 width=8) -> Seq Scan on contracts c (cost=0.00..22.50 rows=10 width=8) -> Index Scan using executives_pkey on executives e (cost=0.00..8.14 rows=10 width=16) -------------------------------------------------------------------------------- Merge Join (cost=174.38..247.30 rows=333 width=76) -> Index Scan using executives_pkey on executives e (cost=0.00..60.00 rows=1000 width=16) -> Sort (cost=174.38..174.38 rows=33 width=60) -> Merge Join (cost=167.58..173.53 rows=33 width=60) -> Merge Join (cost=59.13..63.43 rows=33 width=44) -> Sort (cost=22.67..22.67 rows=10 width=28) -> Seq Scan on contacts m (cost=0.00..22.50 rows=10 width=28) -> Sort (cost=36.47..36.47 rows=333 width=16) -> Seq Scan on buildings b (cost=0.00..22.50 rows=333 width=16) -> Sort (cost=108.44..108.44 rows=100 width=16) -> Merge Join (cost=92.50..105.12 rows=100 width=16) -> Sort (cost=69.83..69.83 rows=1000 width=12) -> Seq Scan on contracts c (cost=0.00..20.00 rows=1000 width=12) -> Sort (cost=22.67..22.67 rows=10 width=4) -> Seq Scan on l_portal_statuses l (cost=0.00..22.50 rows=10 width=4) --snip!-- Hopefully that's not too ugly. TIA -- Dave