Re: hmmm....
From: joel garry <joel-garry_at_home.com>
Date: Mon, 28 Feb 2011 13:21:46 -0800 (PST)
Message-ID: <26b638df-4f54-47e4-84dc-5d613f868834_at_o39g2000prb.googlegroups.com>
On Feb 28, 12:46 pm, Noons <wizofo..._at_gmail.com> wrote:
> How is it that dbms_stats accepts "estimate_percent" as a percentage
> of the rows to sample, when it obviously doesn't know what 100% is
> until it's done a full sample?
> Something's amiss and I reckon it's a thing called logic...
Date: Mon, 28 Feb 2011 13:21:46 -0800 (PST)
Message-ID: <26b638df-4f54-47e4-84dc-5d613f868834_at_o39g2000prb.googlegroups.com>
On Feb 28, 12:46 pm, Noons <wizofo..._at_gmail.com> wrote:
> How is it that dbms_stats accepts "estimate_percent" as a percentage
> of the rows to sample, when it obviously doesn't know what 100% is
> until it's done a full sample?
> Something's amiss and I reckon it's a thing called logic...
Season as with all MOS, but see Why or When does Dbms_stats.Auto_sample_size sample all rows [ID 343849.1]
"...but then dbms_stats will take an accurate row count by using SELECT COUNT(*) from the table and index"
http://structureddata.org/2007/09/17/oracle-11g-enhancements-to-dbms_stats/ shows an interesting test.
jg
-- _at_home.com is bogus. http://www.signonsandiego.com/news/2011/feb/28/missing-swimmer-was-awaiting-sentencing-mail-fraud/Received on Mon Feb 28 2011 - 15:21:46 CST