Archives of the TeradataForum
Message Posted: Thu, 10 Apr 2003 @ 17:52:31 GMT
<-- Anonymously Posted: Thursday, April 10, 2003 12:54 -->
An mload job which is currently running in my production for over a month or so, is running almost 3 hours a day for loding 25 million rows.
I would like to find out
(1) the ways to determine best/optimum performance for loading this volume of data by mload in my NT environment.
(2) The primary factors (e.g. other loads, queries, hardware bottleneck etc) that make the load to perform poorly. The best performance I've seen for the load was 1:40 minutes, where as most of the days 95% the load finishes with in 2:45 hrs., some worst days had taken more than 3:30 hours).
(3) Does the target table size (number of rows) matter in the performance of MLOAD? (Practically on this table, even after reducing the table size to less than half I haven't seen any faster mlods). Currently I don't collect statistics on this table. Does collecting stats improve the load performance?
(4) Does compressing columns impact MLOAD performance?
(5) Do I need to write a PM/API to closely watch the performance of this load? If yes, Can some one you please provide me with a sample program to start with?
Thanks and apprciate your suggestions
|Copyright 2016 - All Rights Reserved|
|Last Modified: 28 Jun 2020|