Archives of the TeradataForum

Message Posted: Tue, 02 May 2006 @ 09:53:49 GMT

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

<-- Anonymously Posted: Tuesday, May 02, 2006 05:09 -->

Andrea

Thanks for your suggestion, however your solution will not yield the desired result as, if all the col3 values for a particular col1 values were '2', then count (distinct).. will be '1' and therefore col4 would be incorrectly computed to be a '1'. The following solution, using Ordered OLAP function gives the desired results:

     select a.Col1
        , a.Col2
        , a.Col3
        ,  (CASE (MAX(a.Col3) OVER (PARTITION BY a.Col1))
           WHEN 0 THEN 8
           WHEN 1 THEN (CASE(MIN(a.Col3) OVER (PARTITION BY a.Col1)  )
                    WHEN 1 THEN 1
                    ELSE 8 END)
           WHEN 2 THEN (CASE(MIN(a.Col3) OVER (PARTITION BY a.Col1)  )
                    WHEN 2 THEN 2
                    ELSE 8 END)
           ELSE 8
          END) Col4
     from TabA a

Best Regards

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