10.01.2012 11:15, ma_golyo wrote:

> Currently the non index read count is over 30K, and this is just a simle
> (few database object) test case.
> Index reads means less record read -> less I/O -> faster execution.

Sure, but if these non-indexed reads occupy e.g. less than 10% of the
overall execution time, it makes a lot of sense to ignore them.

> Query
> ------------------------------------------------
> execute procedure dbschema_vc
>
> Query Time
> ------------------------------------------------
> Prepare : 47,00 ms
> Execute : 11 766,00 ms

30K sequential reads cannot last for 12 seconds, so I believe the real
problem is elsewhere.


Dmitry