Hi All,

I haven't done anything sophisticated with OLAP, so please bear with
me.

I have a database that holds documents and their ratings by various
users. The documents are also categorized in various ways, such as
topic, author, etc. The users are also categorized by things like
functional area, etc.

The ratings all live in a single fact table.

I can build dimensions for all of these things, and I can look at the
ratings and aggregate them for a single dimension. However, if I build
a cube where I have two dimensions (i.e. Document Topic and User
Organization), the cube has no values in it. The same cube with a just
Document Topic shows me ratings.

Any ideas?

Thanks
Tim

To clarify: does a cube built using just the "User
Organization" dimension produce expected results? And are all of these
regular dimensions, with "All" levels?


- Deepak

Deepak Puri
Microsoft MVP - SQL Server

*** Sent via Developersdex http://www.developersdex.com ***

Deepak Puri wrote:
Yes, a cube built with a single dimension produces the expected
results. One of these dimensions is a parent-child dimension, and the
others are all snowflake dimensions.

Whenever I add a second dimension to the cube, all of the cells are
empty.

The only odd thing I can think of about this situation is that
documents can appear in mulitple places, i.e. they can appear in
multiple topics, and some of our users are "cross-over" users, i.e.
they appear in more than one organization. Could this cause an issue?

Thanks
Tim

Not sure what the exact problem is - could there be any unintended
coupling between schemas of the dimensions?

A couple of suggestions to help troubleshooting:

- Define a measure like RowCount, with 'count' aggregation on some
non-null field of the fact table. This will indicate whether any fact
table rows remain after the dimensions are joined in. If this is empty,
then you can examine the SQL query generated to process the cube.

- If the diemsnions are shared, and are joined to the fact table bye the
key field at the lowest level, then the cube schema can be optimized.
When the cube is then processed, you will get error messages for missing
dimension members.


- Deepak

Deepak Puri
Microsoft MVP - SQL Server

*** Sent via Developersdex http://www.developersdex.com ***