From: michael.spoden on
Hi,

is there a possibility to use the dbms_stats package on schema level
to just see which tables would be affected by the gather_auto option?

I want first see which tables would be analyzed depending on the
output I want to start the dbms_stats package for this schema, or
maybe just for some less tables of this output.

Regards

Michael

From: fitzjarrell on
Comments embedded.
On Jun 23, 7:17 am, michael.spo...(a)web.de wrote:
> Hi,
>
> is there a possibility to use the dbms_stats package on schema level
> to just see which tables would be affected by the gather_auto option?
>

Not without actually generating those statistics.

> I want first see which tables would be analyzed depending on the
> output I want to start the dbms_stats package for this schema, or
> maybe just for some less tables of this output.

Again, not without actually generating those statistics.
GATHER_SCHEMA_STATS isn't a preprocessor to output which objects might
be processed, it is a procedure to actually generate the statistics
and report which objects were considered STALE or empty.

>
> Regards
>
> Michael


David Fitzjarrell
From: michael.spoden on
> Again, not without actually generating those statistics.
> GATHER_SCHEMA_STATS isn't a preprocessor to output which objects might
> be processed, it is a procedure to actually generate the statistics
> and report which objects were considered STALE or empty.

David,

any other ideas or possibilites?
From: fitzjarrell on
On Jun 23, 8:12 am, michael.spo...(a)web.de wrote:
> > Again, not without actually generating those statistics.
> > GATHER_SCHEMA_STATS isn't a preprocessor to output which objects might
> > be processed, it is a procedure to actually generate the statistics
> > and report which objects were considered STALE or empty.
>
> David,
>
> any other ideas or possibilites?

You might try querying the DBA_TAB_MODIFICATIONS view, which lists the
tables which have been modified since the last statistics run.


David Fitzjarrell
From: joel garry on
On Jun 23, 6:12 am, michael.spo...(a)web.de wrote:
> > Again, not without actually generating those statistics.
> > GATHER_SCHEMA_STATS isn't a preprocessor to output which objects might
> > be processed, it is a procedure to actually generate the statistics
> > and report which objects were considered STALE or empty.
>
> David,
>
> any other ideas or possibilites?

Use RMAN to duplicate the db and run it there.

Also check out the "Managing Optimizer Statistics" chapter in the
performance manual. You can save and restore statistics. If you are
worried about it, you can lock statistics. Also note the older
statistics are saved automatically by default, for a while.

jg
--
@home.com is bogus.
Biggest loser ever finds new career as motivational speaker:
http://www.catalyst.burtongroup.com/NA08/nickleeson.pdf