Hi!
Hm, the title is not that meaningful, hm? Sorry.
The Problem:

I have a fact table with the following columns (in principle):

o f_key
o class
o count

f_key is a foreign key, linking to another table with further dimension-information
class is a one-level-dimension
count is a measure; it means: in subcube f_key, there are count hits in the value-class class. The aggregator of this measure is "sum"

Now I want to create a "virtual" measure that shows the ratio of the classes of one f_key.

e.g.:
A part of the fact table is:
Code:
f_key | class | count
------|-------|------
1     |  ...  |  ...
2     |   1   |  4
2     |   2   |  10
2     |   3   |  6
3     |  ...  |  ...
The result for the virtual ratio-measure for the tuples (f_key, class) should be
Code:
(2, 1)  =>  0,2
(2, 2)  =>  0,5
(2, 3)  =>  0,3
That means: "For subcube f_key=2, there are 20% hits in valueclass 1, 50% hits in valueclass 2 and 30% hits in valueclass 3."
Hope, you understood my problem....

Has anyone an idea for me?!?
Thanks for any help!