If there's an actual combination of values which returns null, you can replace it with a blank string '' using:
Coalesce(YourFormula,'')
However, it is not possible to remove the nulls from pivot table intersections where no combination of dimensions exist, so if that's your situation you'll need to either add all of the possible intersections in your data load, or live with the nulls.
Note that certain third-party solutions, such as Vizlib Pivot Table, may also allow you to achieve what you want. In the case of Vizlib's pivot table, you can set your own Null values rather than having the default dash.
