In SQL 2005 Analysis server, the relationship between a dimension and the
facttable is defined in the dimension usage tab of the cube. So, what is the
purpose of defining the relationship link in the datasource view? Cause I had
to dimension tab linked to the fact table but when I process the cube. I was
not getting the results I expected. But, later found out from other newsgroup
that I needed to define the relationship in the dimension usage tab. Once I
set it up, it worked. So, what is the point of defining the relationship link
in the view if it did not tell the cube how to relate it.

It's basically just a layer over the database. Some cube builders may not
have the flexibility in designing the database as they'd want, so the DSV
allows them to add referential integrity, named columns, set PK's, ect.

I prefer to do everything like that in the database and not in the DSV.
However I have a lot of influence in what goes into the RDBMS. Not everyone
is as fortunate...




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

They serve different purposes. The DSV defines relationship between tabbes
in the relational database so that cube know how to issue SQL to retrive the
data through certain joins. The dimension usage tab defines the relationship
between dimensions and facts by grains, and yes it could be same as in DSV,
but it doesn't have to be since you could define different grains on
different facts for the same dimension.

"Roger" wrote: