Okay, now I'm annoyed ... I *swear* I ran VACUUM ANALYZE a couple of times
since populating those tables *sigh* now its bringing up indices I'm
expecting :(
On Wed, 6 Feb 2002, Tom Lane wrote:
> "Marc G. Fournier" <scrappy@hub.org> writes:
> > iwantu=# \d poc_uid
> > Index "poc_uid"
> > Column | Type
> > --------+--------
> > uid | bigint
> > btree
>
> > iwantu=# explain select count(1) from orientation_c poc where uid = 1;
> > NOTICE: QUERY PLAN:
>
> > Aggregate (cost=2264.97..2264.97 rows=1 width=0)
> > -> Seq Scan on orientation_c poc (cost=0.00..2264.96 rows=1 width=0)
>
> > EXPLAIN
>
> You're forgetting ye olde constant-casting problem. You need something
> like
>
> select count(1) from orientation_c poc where uid = 1::bigint;
>
> to use an index on a bigint column.
>
> Not sure about the other thing; have you VACUUM ANALYZEd (or at least
> ANALYZEd) since filling the tables? It looks like the system thinks
> the tables are much smaller than they really are.
>
> regards, tom lane
>