Thread: AW: indexes and big tables
> netacc=> EXPLAIN (SELECT SUM(counterfrom) AS from, > SUM(counterto) AS to, > floor((985098900 - date_part('epoch', counterstamp)) / 300) > AS sequence > FROM counters WHERE line='absolonll' AND date_part('epoch', > ) counterstamp > 984978900 GROUP BY sequence, line) ... I would guess the problem is the restriction on counterstamp, because written like that, it probably can't use the index. try something where you avoid the use of the date_part function e.g.:AND counterstamp > '2001-07-26 00:00:00.0' Andreas
> I would guess the problem is the restriction on counterstamp, because > written like that, it probably can't use the index. > > try something where you avoid the use of the date_part function e.g.: > AND counterstamp > '2001-07-26 00:00:00.0' I will try it, but it use the index when there is a few amount of rows. When I insert a lot of rows like me (in milions), index isn't used. I don't know the number of rows which makes border between using and don't using index and I can discover it if you want. Going to try your suggestions ... Best regards, Robert -- _ |-| __ Robert Vojta <vojta-at-ipex.cz> -= Oo.oO =- |=| [Ll] IPEX, s.r.o. "^" ====`o