Several months ago I wrote about the issue importing measures from an Analysis Services cube in PowerPivot. The short story is that, when you define the MDX query, all columns are imported as Text. Of course, this is not a good thing for the numeric measures. The workaround is to change Data Type and then refresh data of the table. However, you might have two issues:

  1. You import the table twice. If loading data requires is long, this is twice long
  2. You might have an error in conversion when you change Data Type of the column (because existing strings cannot be converted correctly into numbers – changing Data Type the conversion happens on these strings, while another code path is executed when you refresh data from the data source having a column already defined as numeric).

The best thing to do in order to avoid both issues is:

  • Define your MDX query adding a slicer that will return no rows in your table
  • Change data type of the imported Measures in your PivotTable columns
  • Change the MDX query by removing the slicer

I hope this will be fixed in the future, anyway.