asked on

Group By and Function

This is SQL 2000....

I have the SQl below. I replaced that SUM statement with a function. This gets a summary. Since I added the function call, I need
to add "group by" at the end of the SQL. But that will cause it to give lots of of rows..because of that PO_Date. How can I change this OR change
the function to get me the same rows/results as that original SUM?

---***** Function ************----
ALTER FUNCTION [dbo].[fnCalculateCutOffSum]
(@orderDate datetime, @warehouse varchar(2),@ShipMethodAbbrev varchar(10))
RETURNS int
AS
BEGIN
-- *** Used in WHT_OrderSummary
declare @Today int
  -- minus 45 mins..deduct that from @orderDate 
  select @Today = sum(case 
                    when datepart(hour,DateAdd(minute,-45, @orderDate)) <= Datepart(Hour,cutoffHour) 
                        Then 1 else 0
                    end)  --compare hour portion of dates
   from nbCutOff
     where --Priority = @priority 
        warehouse = @warehouse
        and ShipMethodAbbrev = @ShipMethodAbbrev  

return @Today
end

------------ SQL ------------------
ALTER PROCEDURE [dbo].[OrderSummaryTest]
AS 
BEGIN
	SELECT CASE o.Customer_ID
            WHEN 100 THEN 'X'
            WHEN 200 THEN 'Y'
            WHEN 400 THEN 'W'
            WHEN 500 THEN 'P'
            ELSE 'B'
       END AS Company,
       sum(...) AS UPS,
       sum(...) AS AIR,
       sum(..) AS USPS,
       sum(..) AS MAN,
     
    
       dbo.fnCalculateCutOffSum (o.PO_Date,'FN','UPSA') as Today,--added this but then i need to add "group by" for o.PO_Date
       /*sum(CASE 
                WHEN o.PO_Date < dateadd(hour,19,CONVERT(varchar,getdate(),101)) AND c.Abbrev_Name = 'UPSA' THEN 1
                WHEN o.PO_Date < dateadd(hour,14,CONVERT(varchar,getdate(),101)) AND c.Abbrev_Name = 'USPS' THEN 1
                WHEN o.PO_Date < dateadd(hour,15,convert(varchar,getdate(),101)) THEN 1
       ELSE 0
           END) AS Today,*/
      
       count(*) AS Total,
      
FROM   Orders o
.......
 
GROUP BY
       CASE o.Customer_ID
            WHEN 100 THEN 'X'
            WHEN 200 THEN 'Y'
            WHEN 400 THEN 'W'
            WHEN 500 THEN 'P'
            ELSE 'B'
       END 
       
ORDER BY
       1 desc

END

Open in new window

HainKurt

looks like this is correct

/*sum(CASE
WHEN o.PO_Date < dateadd(hour,19,CONVERT(varchar,getdate(),101)) AND c.Abbrev_Name = 'UPSA' THEN 1
WHEN o.PO_Date < dateadd(hour,14,CONVERT(varchar,getdate(),101)) AND c.Abbrev_Name = 'USPS' THEN 1
WHEN o.PO_Date < dateadd(hour,15,convert(varchar,getdate(),101)) THEN 1
ELSE 0
END) AS Today,*/

you cannot use the function here, that function will be called once, but above statement will run for each row in the group... whats the problem with the code above?

ASKER CERTIFIED SOLUTION

HainKurt

membership

This solution is only available to members.

To access this solution, you must be a member of Experts Exchange.

Start Free Trial

Camillia

ASKER

>>but above statement will run for each row in the group...
I want to run function to run for each row...basically, it should be the same results..

Let me try your second suggestion...the thing is...the function is already summing up...so maybe remove the sum from the function?

HainKurt

the function will be called for each row in the group... so, yes, you should return something for each row, not for whole group...

Camillia

ASKER

I will close it but dont quite get it. Will open another one if i dont figure it out.