|
Archives of the TeradataForumMessage Posted: Thu, 18 Oct 2007 @ 10:16:59 GMT
Since your work table does not have unique index to force the logical key, you can take a couple approaches: 1) Use the Row_Number() window aggregate and partition by the UPI and qualify where Row_Number equals 1. 2) Put a timestamp on your work table and use the MIN(<timestamp>) window aggregate partitioned by the UPI and Qualify where the timestamp on the record equals the MIN(<timestamp>). 3) Multiload your work table and leverage the error table functionality it provides in lieu of Teradata 12's logging table options on bulk SQL operations. Number 1 would probably be the best solution since it doesn't require you to modify your tables or drastically alter your ETL process. Hope this helps.
| ||||||||||||||||||||||||||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||
Copyright 2016 - All Rights Reserved | ||||||||||||||||||||||||||||||||||||||||||||||||
Last Modified: 15 Jun 2023 | ||||||||||||||||||||||||||||||||||||||||||||||||