Creating Data Cubes from OLAP Data Connectors

I would like to experiment with creating a Dundas Data Cube from my Analysis Services Data Connector (MS SSAS) but I am running into some things that I am finding troublesome.

  1. My Analysis Services database has over 200 dimensions and 200 measures, how can I add all of these at once? Do I seriously have to drag over 400 members one at a time into the ‘Selected x box’ ?

  2. These all default to a way they would look in an mdx query, i.e. ‘[Measures].[Total Sales]’. How can I make them default to how they look when they are selected from the cube, i.e. ‘Total Sales’. I hope that I do not have to rename all of these.

  3. How do I add multiple levels of a dimension/hierarchy? I have some hierarchies that go down several levels but I can only drag one level from the hierarchy into the “selected dimensions” area. It won’t let me drag in a second one.

Thank you.

Hi Ryan,

If you are using Analysis Services already, you typically wouldn’t want to re-warehouse all of the work that you’ve already done in a Dundas BI Data Cube. It’s kind of like… ‘i just finished building a data warehouse, i guess i should put all this data from my data warehouse into another data warehouse.’

Now this being said, if your currently SSAS Cube isn’t providing data in the right format for certain metrics that you need you could use the Dundas BI Data Cube for these cases. These Cubes could provide additional calculations, add performance or even be helpful to join to other data sources. Think of these data cubes more like views rather than a new data source in itself. You would only want to bring in the dimensions/measures you actually need to build a dashboard and not all 200 measures / 200 dimensions that you have.

As for your hierarchies, in SSAS they already have full support for hierarchies and by using your cube directly, you wouldn’t have to re-create these hierarchies in Dundas BI. This direct support for OLAP hierarchies is one of those fairly unique differentiators of Dundas BI in the first place.

My recommendation - use your cube as much as you can directly and build Dundas BI Data Cubes only when you’re finding you need a bit more to solve specific problems.

Does this make sense?