You might want to check out this KB article: http://kb.tableausoftware.com/articles/knowledgebase/functional-differences-olap-relational
There are definitely some changes in functionality when connected to cubes. Specifically, in answer to your questions:
- Calculated fields can be created for measures. However you cannot create calculated fields that use dimensions. You can write custom MDX calculated members.
- Parameters can be created and used, but again not in calculations on dimensions. You can also use table calculations.
- Tableau will use the hierarchies defined by the cube. Ad-hoc hierarchies are not available.
All of this is based on the underlying architecture of a cube and the way data can be queried in cubes. When you use a cube you are trading some of the flexibility inherent in other data sources for a more rigid, pre-defined, pre-aggregated structure that may be more performant in some circumstances and possibly more structured and defined.
Personally, I enjoy the freedom and flexibility of relational data with Tableau; but cubes can be nice too
1 of 1 people found this helpful
Joshua's article is the one must read. There is also a design patterns workbook that I found invaluable at Design Patterns: http://kbcdn.tableausoftware.com/workbooks/adventureworks_cube_on_scdemo-dbs.twbx .
Just in case you might have overlooked it, when working with an SSAS cube and you want to create a calculation, you need to reference full member name in MDX syntax i.e [Customer].[Customers].[Country].&[USA] By default, that syntax will not be populated nicely like it does for other types of data sources. I have a few blogs on Tableau & SSAS at Impact Analytix: Business Intelligence, Predictive Analytics & Excel.
Hope that helps.
Thanks Guys! The article explains it all.