FM and Indexes

[kazanec]

How does FM determine what indexes to use when generating SQL?

I'm creating a FM package that will ultimately feed into Transformer. When I run a test within FM, it takes a long time for the results to be returned. I ran the FM-generated SQL in Query Analyzer with the Include Actual Execution Plan option activated. I see that 95% of my cost is spent doing a table scan, instead of using an index.

Is there a way for me to specify indexes in FM, or am I going about this wrong? 

[kazanec]

What I ended up doing is using Native SQL for each of my imported Query Subjects, which allowed me to specify indexes using "with (index())". End result is cube build times went from 17 minutes to less then 2.