Hello All,
We have cube that have following dimensions:
Customers, Products and Segments
Each customer can have a different segment for each product. Since star or
snowflake schemas do not support this type of relationship we just
implemented Segment dimension as an independant dimension. although ugly,
this solution works as long as non empty keyword is used in mdx queries.
Fact table contains both product, customer and segment id's.
Could somebody explain me if this type of design is supported in yukon (or
an alternative way in AS 2000) through many-to-many dimensions.

great thanks,

Resho

Hi,

in Yukon this is solved, as we are not talking about star or snowflake
schemas but we are talking about relational schema.
You can import a relational schema of your source database and you can even
have more than 1 fact table in 1 cube (better to call Analysis Services
project) and then you can solve this kind of problem by making Perspectives,
which offers you different kind of views on the data.

Very cool stuff and must try!

HTH,
Andrej



"Resho" <Resho (AT) discussions (DOT) microsoft.com> wrote

Thanks Andrej. I think i was not clear enough to explain my problem.
Perspectives are great however I do want to solve sparsity (or outer-join)
problem. Autoexist feature seems to solve this problem however I'm not sure
if there is a good design technique to use it.
Regards,

Resho

"Andrej Hudoklin" wrote: