On Tuesday 20 October 2009 10:44:13 am Scott Marlowe's cat walking on the keyboard wrote: > Two things. 1: Actually running the query and receiving the results > isn't the same as just running it and throwing them away (what explain > analyze does) and 2: The query may be getting cached in psql if you're > running it more than once, but it may not run often enough on that > data set to get the same caching each time. > You are right, in fact executing: psql -h localhost -U dataflex cogedb -c "SELECT * FROM GMMOVART WHERE DATA >= '01/01/2006' AND DATA <= '31/12/2006' ORDER BY DATA, CONTATORE, RIGA" -o /dev/null produces a log like the following: cogedb LOG: duration: 8841.152 ms statement: SELECT * FROM GMMOVART WHERE DATA >= '01/01/2006' AND DATA <= '31/12/2006' ORDER BY DATA, CONTATORE, RIGA so 8,8 seconds against 7 seconds, now it sounds compatible. But I was always trusting the time of explain analyze, this make me doubt about it. So how is such time (explain analyze) to mind? Luca -- Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
