Archives of the TeradataForum

Message Posted: Wed, 30 Jun 2004 @ 13:21:52 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

<-- Anonymously Posted: Wednesday, June 30, 2004 09:17 -->

What do you specifically want to know?

I don't think you can get cumulative CPU, I/O, etc., during the course of a query other than via PM/API. That is what PMON uses.. it doesn't query any table for that info, but rather gets the info from system buffers that are refreshed/updated at specified intervals. That requires writing a program, not mere SQL.

You can get the final performance info after the query is finished, from dbc.acctg/ampusage. That data is given on amp by amp basis so you can get a view of parallel efficiency of the query.

Also AFAIK you can't determine via SQL what step a request is currently running.

With R5's DBQL feature you can join these tables to ACCTG/AMPUSAGE to match the query's SQL with the performance info.

	<Prev	Next>		<<First	<Prev	Next>	Last>>