Klint Gore <kg(at)kgb(dot)une(dot)edu(dot)au> writes:
> On Sun, 27 Jun 2004 23:29:46 -0400, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>> [yawn...] Cast the constants to bigint. See previous discussions.
> Would there be any way of adding some sort of indicator to the plan as
> to why sequential was chosen?
Not really ... the plan that's presented is the one that looked the
cheapest out of the feasible plans. How are you going to identify a
single reason as to why any other plan was not generated or lost out
on a cost-estimate basis? Humans might be able to do so (note that
the above quote is an off-the-cuff estimate, not something I'd care
to defend rigorously) but I don't think software can do it.
FWIW, the particular problem here should be fixed in 7.5.
regards, tom lane
In response to
pgsql-performance by date
|Next:||From: Richard Huxton||Date: 2004-06-28 09:14:11|
|Subject: Re: Query performance|
|Previous:||From: Klint Gore||Date: 2004-06-28 05:29:57|
|Subject: Re: SQL stupid query plan... terrible performance ! |