Re: How to know the stats is corrupted
Date: Tue, 30 Apr 2013 14:08:54 +0100
Message-ID: <517FC266.1020304_at_dunbar-it.co.uk>
On 30/04/13 13:56, Powell, Mark wrote:
> K R, I do not think "corrupt" is the right term. If the statistics in question were actually corrupt I would expect that an error to be thrown and you did not post any error messages. Perhaps you mean the optimizer statistics are stale or otherwise inaccurate?
The cynic in me thinks that the OP was posting the question because "corrupted stats" sounds like a get out clause for something like "we don't know"! I could be wrong.
I'm thinking that maybe the OP's DBAs don't gather stats regularly or correctly, and rather than admit this, they come back with an excuse that sounds plausible - until someone asks "just how did you determine that the stats were corrupted?" and "and given this fact, what do you intend to do to prevent this from happening again?".
I'm in agreement with Mark, I too think that really corrupted stats would probably give an error message - which one, I have no idea, but I imagine that there will be one, somewhere - maybe even an ORA-00600, who knows.
I'm not fond of fobbing off my users or developers with vague reasons, not without evidence anyway, and the evidence usually removes the vagueness!
Just my thoughts, for whatever they are worth!
Cheers,
Norm.
-- Norman Dunbar Dunbar IT Consultants Ltd Registered address: Thorpe House 61 Richardshaw Lane Pudsey West Yorkshire United Kingdom LS28 7EL Company Number: 05132767 -- http://www.freelists.org/webpage/oracle-lReceived on Tue Apr 30 2013 - 15:08:54 CEST