Re: Density statistic calculation in case of histogram
From: Greg Rahn <greg_at_structureddata.org>
Date: Mon, 6 Apr 2009 07:42:41 -0700
Message-ID: <a9c093440904060742r3148343ehda10e8e6419e1d43_at_mail.gmail.com>
On Mon, Apr 6, 2009 at 6:56 AM, Neeraj Bhatia <neeraj.dba_at_gmail.com> wrote:
> 2) Without histograms, density should be equal to 1/number of distinct
> values. But in my case, it is null. I am not using method_opt while
> collecting statistics. The global default setting is FOR ALL INDEXED COLUMNS
> SIZE 1.
I would strongly suggest revisiting that METHOD_OPT choice. It is probably one of the worst ones to choose IMO. Here is why: http://structureddata.org/2008/10/14/dbms_stats-method_opt-and-for-all-indexed-columns/
Date: Mon, 6 Apr 2009 07:42:41 -0700
Message-ID: <a9c093440904060742r3148343ehda10e8e6419e1d43_at_mail.gmail.com>
On Mon, Apr 6, 2009 at 6:56 AM, Neeraj Bhatia <neeraj.dba_at_gmail.com> wrote:
> 2) Without histograms, density should be equal to 1/number of distinct
> values. But in my case, it is null. I am not using method_opt while
> collecting statistics. The global default setting is FOR ALL INDEXED COLUMNS
> SIZE 1.
I would strongly suggest revisiting that METHOD_OPT choice. It is probably one of the worst ones to choose IMO. Here is why: http://structureddata.org/2008/10/14/dbms_stats-method_opt-and-for-all-indexed-columns/
-- Regards, Greg Rahn http://structureddata.org -- http://www.freelists.org/webpage/oracle-lReceived on Mon Apr 06 2009 - 09:42:41 CDT