Thank you all, You were right on the analyze. Insert statement with an aggregated subquery had a problem on an empty table.
I had to change the queries to do a simple insert then analyze on the table followed by an update with an aggregated sub query. That goes thru very fast. -mridula From: Віталій Тимчишин [mailto:tiv...@gmail.com] Sent: Wednesday, February 03, 2010 8:11 AM To: Mridula Mahadevan Cc: pgsql-performance@postgresql.org Subject: Re: [PERFORM] Queries within a function 2010/2/2 Mridula Mahadevan <mmahade...@stratify.com<mailto:mmahade...@stratify.com>> Hi, I am running a bunch of queries within a function, creating some temp tables and populating them. When the data exceeds say, 100k the queries start getting really slow and timeout (30 min). when these are run outside of a transaction(in auto commit mode), they run in a few seconds. Any ideas on what may be going on and any postgresql.conf parameters etc that might help? Thanks Have you tried to analyze temp tables after you've populated them? Because AFAIK it won't do it automatically for tables created, filled and then used in same transaction.
