Re: 10g optimizer and dynamic sampling

On 2 Jun 2006 12:17:11 -0700, p.santos000_at_gmail.com wrote:

>

>
>Well that's a good point, but here is the difficult thing about this
>system. Our customers
>upload data anytime they want at any point of the day. So in the
>morning COL1='X' could
>return 100 rows, but the same query in the evenning could yield a
>million rows.
>
>Most of the queries executed on our system take anywhere from 5 minutes
>to 30-45 minutes, so a couple of extra seconds is for the sampling is
>insignificant.
>
>-peter
>

Hi Peter,

You may also start reconsider your design. Partitioning could be a way to keep data distribution stable within each partition.

• Kenneth