Archives of the TeradataForum
Message Posted: Thu, 18 Jun 2015 @ 19:47:36 GMT
Hi -- U have come across one of the toughest things in our line of work. Staff use the stats diagnostic and apply all rec'd stats en masse. If the outcome is a good one, they move on. U have to take the time to id then apply only the useful stats. When that does not happen, your situation arises.
Another problem (fairly rare) is when a good stat goes bad. The only alert is when U receive a request to fix a query that has run without issue "since anyone can remember ..." In my case this is the last thing I consider investigating. Finding the bad stat in a large, complex qry is often time consuming.
The Optimizer does a great jobs these days, but it is not perfect.
|Copyright 2016 - All Rights Reserved|
|Last Modified: 23 Jun 2019|