Try analyze table
ANALYZE TABLE table_name COMPUTE STATISTICS FOR ALL INDEXES
ANALYZE TABLE table_name COMPUTE STATISTICS
Use count(0)
Użytkownik "Martin Pein" <martin.pein_at_fernuni-hagen.de> napisał w wiadomości
news:aalhi6$4ud$1_at_beech.fernuni-hagen.de...
> Hi there,
>
> I've got a generated table here (in an Oracla 8 db) with about 400 columns
> and about 4000 sets. About 95% of the fields are empty. One of the columns
> is a unique and indexed and not-null'd ID.
>
> Accessing the data is kind-of slow: a "select count(*)" takes about 5
> seconds.
> It also takes about 5 seconds for a "select id from .. where id=17" to
> return with the result (17 :-). The same is basically true for the other
> fields in the row.
>
> All the other tables with fewer columns are much quicker.
> From you experience, is that likely to be an architectural or a technical
> problem?
>
> Thanks,
>
> Martin
>
>
>
>
>
>
Received on Tue Apr 30 2002 - 03:42:06 CDT
