I think the problem that you are seeing is that if you create a calculated
member based on the two measures (Amount and Price) that the data doesn't
roll up correctly in Analysis Services. Here's how you solve that problem.

Use a view in which you have a column that multiplies out amount and price
and use that view as the fact table. You can then create a regular measure
based on that column and the data will roll up correctly. You can also use
your existing fact table and just specify the following for the source
column of the measure

"tablename"."amount" * "tablename"."price". There's actually an example of
this in one of the Foodmart 2000 cubes.

Sean


--
Sean Boon
SQL Server BI Product Unit

Disclaimer : This posting is provided "AS IS" with no warranties, and
confers no rights.

"michael" <ekleinm (AT) gmx (DOT) de> wrote

Hi Sean,

thank you very much for your reply. I forgot to tell you,
that it would be
a planning projekt, so I work with the write back table
and it dont work as
you describe.

Description:
amount is the measure to plan
price is the fix measure (the roll up must be a average)

amount * price would be calculated

The roll up of amount works correktly.
The roll up of price dont work correktly, I would need the
average of the
price Or
the roll up of the calculated (amount * price) must work,
so I can calculate
the price for the aggregated levels by the sum(amount *
price from the
leafes) / sum(amount)

I tested a "before insert" trigger to calculate the amount
* price before
inserting the records in the write back table, but this
didn't work because
there is in every row only the delta of one measure, the
other are 0.

Do you have any idea to solve these problem.

Thank you very much,
Michael