Mike Mascari <mascarm(at)mascari(dot)com> writes:
> When I query the view with a simple filter, I get:
> explain analyze select * from p_areas where deactive is null;
The problem seems to be here:
> -> Seq Scan on _areas a (cost=0.00..2.48 rows=1 width=163) (actual time=0.037..0.804 rows=48 loops=1)
> Filter: (deactive IS NULL)
Why is it so completely off about the selectivity of the IS NULL clause?
Are you sure you ANALYZEd this table recently?
regards, tom lane
In response to
pgsql-performance by date
|Next:||From: Dave Page||Date: 2004-11-23 20:12:21|
|Subject: Re: scalability issues on win32|
|Previous:||From: Cott Lang||Date: 2004-11-23 17:20:07|
|Subject: Re: Some quick Opteron 32-bit/64-bit results|