Re: *****SPAM***** RE: Manipulate DATE histograms

From: Wolfgang Breitling <breitliw_at_centrexcc.com>
Date: Thu, 14 Dec 2006 14:30:29 -0700
Message-Id: <20061214213003.055CB56C9E3@turing.freelists.org>

You're right. avgclen should be 7. As I said I just took the example from my paper. I changed the density (so some fictitious value) but forgot to change avgclen.
The flags value I think I took from dbms_stats.set_column_stats calls created by export. I suppose you could just let it default. The flag(s) correspond to GLOBAL_STATS and USER_STATS of dba_tab_columns and thus can take on the values 0, 1, 2, and 3. (binary 00, 01, 10, 11).

At 02:11 PM 12/14/2006, Milen Kulev wrote:
>Hi Wolfgang,
>Thank you very much for your reply/Christmas present ;)
>I have noticed that you are always using (..., 2,2) for the last
>two paramemers of DBMS_STATS.SET_COLUMN_STATS
>(avgclen and flags )
>
>Shouldn't in this case avgclen to be set to 7 (for DATE type) ?
>avgclen should play a role somewhere in the CBO calculations, I think ...
>
>And can "flags" get any other value other than 2 (asking just of curiosity) ?
>
>Best Regards. Milen
>

Regards

Wolfgang Breitling
Centrex Consulting Corporation
www.centrexcc.com

This email has been scanned by the MessageLabs Email Security System. For more information please visit http://www.messagelabs.com/email

--
http://www.freelists.org/webpage/oracle-l

Received on Thu Dec 14 2006 - 15:30:29 CST