Archives of the TeradataForum

Message Posted: Thu, 06 Apr 2006 @ 11:47:03 GMT

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

Your problem is the choice of primary index on the table...

SET table doesn't allow duplicate rows. (identical in each column)

As you specified sfra, the PI is not unique on column a, so duplicate on index column are accepted, while complete duplicate rows will be insert one time and the other discarded automatically without errors.

Remember that SET table with a not unique primary index always do a duplicate row check, with a lot of rows it will degrade performance...

MULTISET table, allow duplicate rows, but if you declare an unique primary index, you will take the error 2801.

If you want allow duplicates, you must specify a NOT unique primary index....

Sorry for my bad english, I hope you have understand what I mean....

Andrea Cappelli
NCR Italia
Teradata Professional Services
Teradata Certified Implementation Specialist V2R5

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