I'm using hte mod_perl version ... with a database of some 12million words
... except with the database under load with a vacuum or some such, I
dont' find that the search is too slow ... would like to somehow get rid
of that LIKE clause, which would probably spee d thing sup some, but going
from search.c to search.mpl mae a world of difference here ...
On Fri, 12 Jan 2001, Thomas T. Thai wrote:
> On Fri, 12 Jan 2001, Tom Lane wrote:
>
> > "Thomas T. Thai" <tom@minnesota.com> writes:
> > > 'select * from url' from psql monitor took 59 seconds.
> >
> > How big is the table? Your EXPLAIN mentions 99256 rows, but I can't
> > tell if that stat is up-to-date or not.
>
> it is 99256. i don't think it's that big of a table is it? typically the
> query under mnogo takes less than a second, at most a couple seconds but
> not 50+ secs.
>
> maybe Hermit has some input as he runs it for postgresql.org's search.
>
> > A select like that is going to be pretty much all data transfer: read
> > the disk blocks, format the data values, send 'em to the frontend.
> > There's hardly anything that Postgres can do to optimize or pessimize
> > it. You might shave a few milliseconds by using a binary cursor (to
> > avoid formatting the integer values into ASCII) but probably not a lot.
> >
> > If you've done a whole lot of UPDATEs/DELETEs on the table since your
> > last VACUUM, then reading empty disk blocks might be costing you some
> > time.
>
> i did vacuum analyze.
>
>
Marc G. Fournier ICQ#7615664 IRC Nick: Scrappy
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org
