If it is the same "IPM Provider Names" that always give the 'inf' or 'nan' result then you can use eliminations to remove the rows.  This will hide the problem but not solve it.  I would recommend you take a look at how this values are being generated in your source dataset.  Generally NAN = not a number and INF = infinity.  Neither value has any place in the result set of an analytics query.  If the values where to calculate to empty/null then Pyramid would exclude them by default.  

It looks like you have quite a few formulations in this grid. the inf and nan values are most likely a result of division by 0.

I would find the formulate where this division takes place, and replace the "/" operator with the divide function:

Divide(numerator, denominator, NULL).

if the entire row has null values, by default, it will not show in the grid.

 

Let us know how it went.