Archives of the TeradataForum

Message Posted: Sat, 23 Apr 2006 @ 00:02:07 GMT

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

Well, this might be a made a little more difficult by the fact that the MemID doesn't seem to be unique for the names you want. But if the data is actually like:

     MemID  Name  Value
     1      A     Good
     1      B     Better
     2      A     
     3      B     Wurst (for Dieter's sake)

Then the following select:

     SELECT MemID,
            CASE
               WHEN Value IS NULL
                  THEN t1.Name
                  ELSE t2.Name
             END AS Name,
            COALESCE(t1.Value, t2.Value) AS Value
       FROM TheTable t1
       JOIN TheTable t2 -- Could be an left outer join if there's
                           not always a B for an A
         ON t1.MemID = t2.MemID
        AND t2.Name = 'B'
      WHERE t1.Name = 'A';

Got that? The results would be:

     MemID  Name  Value
     1      A     Good
     2      B     Wurst

Zounds good?

iv

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