Archives of the TeradataForum
Message Posted: Wed, 05 Jul 2006 @ 14:20:59 GMT
<-- Anonymously Posted: Wednesday, July 05, 2006 10:10 -->
I'm using a volatile table as an appropriate way of improving throughput in a particular set of circumstances.
I have 2 questions:
- The population of this volatile table takes x time for a given primary index set of columns, though with a PI that would improve performance even more, it is taking at least 30x (not strictly controlled conditions, but stable enough). It is known to be in the MERGE step; i.e. populating the volatile table. This discrepancy seems extreme for redistribution based on different hash values??
- Can one look at the volatile table; perhaps even it's size or something else by which the progress of populating can be monitored? I'ma aware that volatile tables are not controlled as fully as normal tables in the data dictionary etc, but surely there must be some way of detecting their presence?
|Copyright 2016 - All Rights Reserved|
|Last Modified: 28 Jun 2020|