* Igor V. Rafienko <igorr@ifi.uio.no> [001013 05:09] wrote:
>
>
> Hi,
>
>
> I've got a slight optimization problem with postgres and I was hoping
> someone could give me a clue as to what could be tweaked.
>
> I have a couple of tables which contain little data (around 500,000 tuples
> each), and most operations take insanely long time to complete. The
> primary keys in both tables are ints (int8, iirc). When I perform a delete
> (with a where clause on a part of a primary key), an strace shows that
> postgres reads the entire table sequentially (lseek() and read()). Since
> each table is around 200MB, things take time.
Postgresql fails to use the index on several of our tables, an
'EXPLAIN <query>' would probably output a lot of lines about
doing a 'sequential scan'.
The only solution that I've been able to come across is to issue
a 'set enable_seqscan=off;' SQL statement on most of my queries
to force postgresql to use an index.
hope this helps,
-Alfred
