Re: [GENERAL] Index scan vs. Seq scan on timestamps

2004-12-11 Thread Per Jensen
Den Tue, Dec 07, 2004 at 03:13:04AM - eller der omkring skrev Andrew - Supernews: > On 2004-12-07, Stephen Frost <[EMAIL PROTECTED]> wrote: > > Is there a way to say "just take the value of this function at the start > > of the transaction and then have it be constant" in a query? > > Why not

Re: [GENERAL] Index scan vs. Seq scan on timestamps

2004-12-11 Thread Per Jensen
Den Tue, Dec 07, 2004 at 03:13:04AM - eller der omkring skrev Andrew - Supernews: > On 2004-12-07, Stephen Frost <[EMAIL PROTECTED]> wrote: > > Is there a way to say "just take the value of this function at the start > > of the transaction and then have it be constant" in a query? > > Why not

Re: [GENERAL] Index scan vs. Seq scan on timestamps

2004-12-06 Thread Per Jensen
Andrew - Supernews wrote: On 2004-12-06, Per Jensen <[EMAIL PROTECTED]> wrote: Why does PG not use the index on the time column in the second select, timeofday() has been cast to a timestamp after all. "timestamp" is "timestamp without time zone" (not the most usefu

[GENERAL] Index scan vs. Seq scan on timestamps

2004-12-06 Thread Per Jensen
ne - '30 days'::interval))::timestamp with time zone) AND ("time" <= ((timeofday())::timestamp without time zone)::timestamp with time zone)) (3 rows) Why does PG not use the index on the time column in the second select, timeofday() has been cast to a timestamp after al