Re: Last time update stats ran

Okay, this one is getting fun - and I hope each of you are smiling through this - I am. I'll cut out the details since we are poking fun and semantics now. Bill Glidden <glidden@informix.com> wrote in article <5unapn$d19@news.informix.com>... > Chuck Ludwigsen (cludwigsen@harrahs.com) wrote: > : Art S. Kagel wrote: > : > Carl Gruber wrote: > : > > Is there an easy way to tell when the last time an "update stats" ran for > : > > a particular table? -- Carl > : > > : > No. -- Art > > : Beg to differ Art. Look at the contents of sysdistrib -- Chuck > > Does update statistics always generate distributions? No. So, Art's answer > is correct. You can't tell when statistics were last updated, but you can > tell the last time distributions were created. -- Bill > Now, since Update statistics is the utility that BUILDS the distributions, then is it not true that the date in sysdistrib IS the date that UPDATE STATISTICS was run at a level that produces distributions. As has been so repeatedly noted, an upstats low could have been run AFTER that date without keeping any history of it. This is true as long as the up stats low did not include "drop distributions". If it did, it would have eliminated the date from sysdistrib and you would be none the wiser as to when it was done. This action usually results in someone working REALLY late at night trying to figure out why performance just shot through the FLOOR. Okay guys, be fair now, have I covered my bases on this one yet??? ;-) Take care all. Chuck.