Re: 7.4 dramatically slower than 7.3?
От | Stephan Szabo |
---|---|
Тема | Re: 7.4 dramatically slower than 7.3? |
Дата | |
Msg-id | 20040406080800.I36214@megazone.bigpanda.com обсуждение исходный текст |
Ответ на | Re: 7.4 dramatically slower than 7.3? (Nabil Sayegh <postgresql@e-trolley.de>) |
Список | pgsql-novice |
On Tue, 6 Apr 2004, Nabil Sayegh wrote: > Stephan Szabo wrote: > > On Tue, 6 Apr 2004, Nabil Sayegh wrote: > > > > > >>When I take the query of my first post and EXPLAIN ANALYZE it with 7.3 I get the following output: > >>---------------------------------------------------------------------------------------------------- > >> Hash Join (cost=85.66..213.16 rows=1 width=240) (actual time=34.01..34.47 rows=1 loops=1) > > > > > >>Now I tested it with 7.4 to see if gets faster, but guess what? > >>It's about 40 times slower(!): > >> > >>---------------------------------------------------------------------------------------------------- > >> Merge Left Join (cost=142.22..142.28 rows=2 width=0) (actual time=18.681..18.684 rows=1 loops=1) > > > > > > Hmm, this actual time number seems about 1/2 the above unless I'm missing > > something (34 vs 19). Is it possible you missed some lines or something > > while posting? > > I posted only the first few lines (it's repeating). Like I said, if that's the first line up there, it's saying that 7.4 is twice as fast as 7.3 (at 19 ms vs 34 ms). We don't know where the other 39.5x the cost is going since it isn't into the plan AFAICS.
В списке pgsql-novice по дате отправления: