Re: Index clustering factor

On Jan 9, 2008 1:54 PM, Orysia Husak <Orysia.Husak@xxxxxxxxxxxxx> wrote:

>
>
> When I generate an explain plan manually, the appropriate index is used.
> When I view the sql running through OEM, I see the execution plan is a
> full tablescan. Since the clustering factor for this index is very high,
> my assumption is that the index isn't being selected due to the high
> clustering factor.
>

Cluster  factor should not come into play if only 1 or two blocks need to be
read.

It seems as if you may have been unlucky, as in the SQL in the cache was
parsed
with values that required an FTS due to data skew.

One of the optimizer experts is sure to chime in here soon.


-- 
Jared Still
Certifiable Oracle DBA and Part Time Perl Evangelist

Other related posts: