Archives of the TeradataForum

Message Posted: Thu, 27 Apr 2006 @ 10:00:37 GMT

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

Naveen:

In TD, PI is different PK, PK is 3NF mechanism to make sure the row uniquess, but in TD, PI is the way for even distributiion over AMPS, In most cases, OLTP PK usually be mapped to USI in TD.

If you are sure there is no duplicate rows in your source db then you can define your SET table, otherwise in MULTISET, alternativly you can Define NUPI instead of UPI, but UPI stands better performance than NUPI.

FYI

Yong

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