CODEWITHSUNDEEP
sql-serverssasolap
cag
cag

Reputation: 512

In SSAS, can you ignore table relationships at the cube level?

We have a situation where, because OLAP is using a single DataSource view for all cubes, some of our aggregations are getting confused.

The problem is the following: - Multiple fact tables join to another table to make a dimension - Another fact table joins to those fact tables to make dimensions - OLAP doesn't know which join to follow and picks the wrong one

Here is how our cubes are defined

Cube A:
[FactTable A]->[Dim A]

Cube B:
[FactTable B]->[Dim A]

Cube C:
[FactTable C]
    |--------->[FactTable A]->[Dim A]
    |--------->[FactTable B]

Because of this the datasource view looks like this:

[FactTable C]
    |--------->[FactTable A]->[Dim A]
    |--------->[FactTable B]->[Dim A]

OLAP doesn't know how to aggregate Dim A. Is there a way to ignore that join for Cube C? Can you define the relations at the cube level alone?

Upvotes: 0

Views: 281

Answers (1)

cag
cag

Reputation: 512

So it turns out I was doing things wrong. You should, in fact, create one datasource view per cube and define the relationships there.

Upvotes: 0

Related Questions