Archives of the TeradataForum

Message Posted: Thu, 19 Aug 2004 @ 18:52:42 GMT

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

Hi TIA,

Below is a sql that would do something close from what you want I think:

     delete from table
     where (PK1, PK2) IN (
     select A.PK1, A.PK2 from table A, table B
             where  A.PK1 = B.PK1
             and A.PK2=B.PK2
             and A.PK3 != B.PK3
             group by A.PK1, A.PK2
     having count(*) > 1)

In the example that you gave it would delete the two first rows, it would also delete a third one with similar PK1 and PK2 whatever the value of PK3 if there was one, if this is what you want..

You can run the select SQL alone first in order to see what are the combination of PK1 And PK2 that will be deleted.

Arnaud

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