Archives of the TeradataForum
Message Posted: Thu, 15 Feb 2007 @ 17:02:56 GMT
<-- Anonymously Posted: Thursday, February 15, 2007 11:53 -->
It is my understanding that a SET table will need to verify the uniqueness of each row during an INSERT job; this can be done using a unique index if one is available, but if it isn't the rows need to be checked on a row by row basis.
If this is true, would there ever be a reason (other than ETL utility restrictions) when on would create a SET table and make the PI non unique, when it can be unique?
CREATE SET TABLE FACT_STR_ITM_MVMNT ,NO FALLBACK , NO BEFORE JOURNAL, NO AFTER JOURNAL, CHECKSUM = DEFAULT ( CHN_ID INTEGER NOT NULL, ITM_ID INTEGER NOT NULL, STR_ID INTEGER NOT NULL, THE_DT DATE FORMAT 'YYYY-MM-DD' NOT NULL, )PRIMARY INDEX XIE1FACT_STR_ITM_MVMNT ( ITM_ID ,STR_ID ,THE_DT ) ;
|Copyright 2016 - All Rights Reserved|
|Last Modified: 28 Jun 2020|