| From: | Stephan Szabo <sszabo(at)megazone(dot)bigpanda(dot)com> |
|---|---|
| To: | Martin Sarsale <martin(at)emepe3(dot)net> |
| Cc: | pgsql-performance(at)postgresql(dot)org |
| Subject: | Re: seqscan instead of index scan |
| Date: | 2004-08-30 20:04:00 |
| Message-ID: | 20040830130016.T98089@megazone.bigpanda.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-performance |
On Mon, 30 Aug 2004, Martin Sarsale wrote:
> On Mon, 2004-08-30 at 15:06, Merlin Moncure wrote:
> > create function is_somethingable (ctype, dtype) returns boolean as
>
> Thanks, but I would prefer a simpler solution.
>
> I would like to know why this uses a seqscan instead of an index scan:
>
> create index t_idx on t((c+d));
> select * from t where c+d > 0;
As a geuss, since 7.4 and earlier have no statistics on the distribution
of c+d it has to guess about how likely that is to be true and is probably
overestimating. 8.0beta might handle this better.
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Greg Stark | 2004-08-30 20:18:39 | Re: Why does a simple query not use an obvious index? |
| Previous Message | Bruno Wolff III | 2004-08-30 19:40:50 | Re: Why does a simple query not use an obvious index? |