Re: [PERFORM] Help with EXPLAIN ANALYZE runtimes

2005-01-09 Thread Guenzl, Martin
... with or without the Karnak headgear. Martin -Original Message- From: Guenzl, Martin [mailto:[EMAIL PROTECTED] Sent: Sunday, 9 January 2005 3:45 PM To: pgsql-performance@postgresql.org Subject: [PERFORM] Help with EXPLAIN ANALYZE runtimes Hi, I am a recent convert to Postgresql, and am trying

[PERFORM] Help with EXPLAIN ANALYZE runtimes

2005-01-08 Thread Guenzl, Martin
Hi, I am a recent convert to Postgresql, and am trying to tune a very slow query across ten tables all with only a few rows at this stage (20), and was looking for some help to get me out of a dead-end. It runs very slowly both on a hosted Postgresql 7.3.4 server running on FreeBSD UNIX box,

Re: [PERFORM] Help with EXPLAIN ANALYZE runtimes

2005-01-08 Thread Tom Lane
Guenzl, Martin [EMAIL PROTECTED] writes: On the latter, the EXPLAIN ANALYZE returned what I thought was a strange result - here is the excerpt ... Do you think we are psychics who can guess at your problem when you've shown us none of the table definitions, none of the query, and only a small

Re: [PERFORM] Help with EXPLAIN ANALYZE runtimes

2005-01-08 Thread Guenzl, Martin
LOL ... Excuse my ignorance but what's Karnak headear? It's a SELECT statement. There are no foreign-keys, just primary keys and indexes (some clustered). All joins are through integers / big integers (since anything beginning with id_ is either an integer or big integer). The intention of

Re: [PERFORM] Help with EXPLAIN ANALYZE runtimes

2005-01-08 Thread Bruno Wolff III
On Sun, Jan 09, 2005 at 16:45:18 +1100, Guenzl, Martin [EMAIL PROTECTED] wrote: LOL ... Excuse my ignorance but what's Karnak headear? Jonny Carson used to do sketches on the Tonight show where he was Karnak and would give answers to questions in sealed envelopes which would later be read by