Re: SQL group totals

[CRPence <CRPbottle@xxxxxxxxx>]

  Add a row with a NULL value for ORDER# to see the difference.  See 
the nuances of COUNT here, for row versus value count and the NULL value.
http://publib.boulder.ibm.com/infocenter/iseries/v5r4/topic/db2/rbafzmstcolfunc.htm

Regards, Chuck

rob@xxxxxxxxx wrote:
> I do not think your "count" implementation works.  COUNT seems to
> ignore what's in the parenthesis.
>
> select co, cust, PO#, order#, pick#
> from qtemp.craig 
> 1....+....2....+....3....+....4....+....5....+....6....+.
>   CO            CUST   PO#         ORDER#           PICK#
>  100             101   a1               1               1
>  100             101   a1               1               2
>  100             101   a1               1               3
>  100             101   a1               2               4
>  100             101   a1               2               5
>  End of data  ******** 
>
> select co, cust, PO#, count(order#) AS NbrOrders, count(pick#) as
> nbrPicks 
> from qtemp.craig 
> group by co, cust, po# 
> 1....+....2....+....3....+....4....+....5....+....6....+..
>   CO            CUST   PO#      NBRORDERS        NBRPICKS 
>  100             101   a1               5               5 
>  End of data  ******** 
>
> Perhaps with some complicated mess involving DISTINCT or some such 
> thing...
>
> Rob Berendt