I'm having a hard time defining the attribute relationships and attribute types for a couple of dimensions in one of my cubes. Taking into consideration I come from a Cognos 8 data modeling world, I'm trying to leverage some of the concepts I've used in Cognos framework manager with Analysis Services.

I have the following information for a product dimension: Key, Name, ID, Sub Group Name, Sub Group ID, Sub Group Abbv, Group Name, Group ID, Group Abbv.

I created a product dimension with all of the attributes above, in addition of a Product Hierarchy which contains the Product Name > Sub Group Name > Group Name. I would like to give the users the ability to slice, dice and filter by the ID, Name, or Abbv of any of the attributes above.

My issue is...

What attribute relationship do I define for the IDs and the abbreviations for the Sub Group and Group? Do I leave them with their default relationship to the Product Key? Do I define the relationships within the IDs (Prod ID > Sub ID > Group ID) and Abbv (Prod Abbv > Sub Abbv > Group Abbv) ? Does this even make sense to have in a cube?

I know the attribute relationships are very important and can greatly improve the performance, both processing and querying, if set up correctly. However I don't want to add anything that would

Thanks in advanced for your help!!!