How do I handle statistics on volatile tables

My specific problem relates to the Peoplesoft application, if you're familiar with it, and care about such things.

I am running 7.3.4 on AIX, with the cost-based optimizer. Things run pretty well in most situations. But the application has MANY "temp tables", defined to mean tables which are normally empty, but are filled with data during a batch run, used in some way, and truncated at the end.

I update my statistics every Sunday, like a good boy, but I'm certainly lying to Oracle when I do this, because all the "temp tables" have no rows in them at 6 AM Sunday, but may have hundreds or thousands of rows when I'd like a really good execution plan.

How do you handle this situation? Solutions that come to mind are: 1. Manually update the statistics from within the batch processes, after data is loaded into the table. We don't want to "fool with" a packaged application, however. 2. Somehow (your ideas solicited) load dummy data into the temp tables before running my Sunday statistics. This sounds really kludgy. 3. I don't know if this is legal or kosher, but: Manually update the statistics for selected tables with an UPDATE statement, instead of the "ALTER" statement.

If you've faced and solved this problem, or if you have ideas about how to solve it, I invite your input.

-----== Posted via Deja News, The Leader in Internet Discussion ==----- http://www.dejanews.com/rg_mkgrp.xp Create Your Own Free Member Forum Received on Mon Aug 31 1998 - 11:53:27 CDT