In the code I actually use the alias 'REQQTR1' etc but I didn't know rollup was a keyword. Michael, I am going to try and impliment your shorter solution tomorrow. Is there a way to get the job done with a single SELECT for my 13 or so sum fields? Or is the method here the way. I anticipate this table will have between 300,000 to 600,000 records on upper end equipment. So is the multiple SELECT a bad hit?

UPDATE foo
SET myField = (SELECT SUM(myField) FROM foo WHERE ikey LIKE '1%' AND ikey != '100')
SET myField1 = (SELECT SUM(myField1) FROM foo WHERE ikey LIKE '1%' AND ikey != '100')
SET myField2 = (SELECT SUM(myField2) FROM foo WHERE ikey LIKE '1%' AND ikey != '100')
-- ETC 
WHERE iKey='100'

My thanks again for your help,
Joe

>It doesn't work because Rollup is an alias for a column and the results were returned to the client. To get the behavior using your example, you'd probably do something like this:
>
>

>-- Roll up the values
>DECLARE @rollup int
>select
>    @rollup = sum(MyField)
>from foo
>    where left(IKey,1) = '1'
>
> -- insert the rolled up values
> update foo
>     set
>        MyField = @RollUp
>where IKey = '100'
>

>
>Now, a better solution would be to stay in the set based operations. So, you could also write:
>
>

>UPDATE foo
>SET myField = (SELECT SUM(myField) FROM foo WHERE ikey LIKE '1%' AND ikey != '100')
>WHERE iKey='100'
>

>
>-Mike
>
>>Why does this not Work? I am certain I have oversimplified the mighty T-SQL language. I have some records that get a rollup value based
[SNIP]