Paul Drake wrote:
>>I haven't found that things get any easier in 9i (9.2.0.5.0).
>>I'm still seeing a 40x difference between using a GTT
>>and a real table, with the GTT on the wrong end of the
>>discrepancy (see my other post here).
>>
>>The SET_TABLE_STATS solution seems to work, however.
>
>
> Jim,
>
> what is your setting for the value of
>
> v$parameter.name='optimizer_dynamic_sampling'
>
> and optimizer_features_enable ?
Someone else suggested reading up on dynamic_sampling, which
I did. It solved the problem, and the specific solution is
described in another thread. Thanks for your reply.
Received on Thu Sep 23 2004 - 12:49:09 CDT
