Summing NamedSets Help

[RichardDBA]

Hi,
I have created several named sets.
I need to find the percentage of sales of one named set over the sum of all sales for all named sets.
Looking like this in pseudo code:
Jan1, Sales, NamedSet1
-----------------------
Jan1, Sales, NamedSet1+NamedSet2+NamedSet3


Here is some mdx that works EXCEPT for the SumOfSets:
with
member Measures.SumOfSets as '([Measures].[Sales], {NamedSet1+NamedSet2+NamedSet3})'

select
{[Measures].[Sales], Measures.SumOfSets}
on 0,
CROSSJOIN ( {[TimeYMD].[All TimeYMD].[2009].[January]:[TimeYMD].[All TimeYMD].[2009].[March]}, [MediaChannels].[Vendors])
on 1
from (
select [NamedSet1] on 0 from rebills

Thanks all!
Richard

[RichardDBA]

Hi,
I found this solution:

with
member measures.NewSet as 'AGGREGATE({[ChartAffiliate],[ChartOrganic]}, [Measures].[Sales])'

SELECT NON EMPTY { [Measures].[Sales] , measures.NewSet }
ON COLUMNS,


which gave me the numerator and the denominator respectively.
But if anyone has a better way please let me know.
Richard

[Island1]

I prefer your ultimate solution to that which you originally sought ..... Mainly because it actually uses the dimensional structure (dims across measures, etc.) versus using Named Sets - which, in my experience, are best leveraged as framers of dims for presentation purposes, groupings that may go beyond the dimensional structure, and the like.

Accumulation is best done with measures / calculated members, and then the dims (even within Named Sets) can simply be juxtaposed to give the correct values / aggregations.

We appreciate your sharing your solution, though ... that way, others with similar needs / questions might benefit! Let us know if we can offer further assistance or insight.

Good Luck,

Bill