Remove Optional Argument from _fill_perfdata

[michaelmckinsey1]

Always fill numerical columns with np.nan to make functionality clearer.

[ilumsden]

Looks good.

One minor comment is that it's probably better to use NaN in the comments, but that's so minor that it's fine if you don't do that.

Also, whenever it's no longer experimental in pandas, we should probably switch to using pandas.NA. It's a pandas-specific type for representing missing data that works for all pandas data types. So, NA can be interpreted as Int64Dtype, StringDtype, Float64Dtype, BooleanDtype, and even the newer ArrowDtype.

@pearce8 this is ready for your review

[michaelmckinsey1]

Looks good.

One minor comment is that it's probably better to use NaN in the comments, but that's so minor that it's fine if you don't do that.

Also, whenever it's no longer experimental in pandas, we should probably switch to using pandas.NA. It's a pandas-specific type for representing missing data that works for all pandas data types. So, NA can be interpreted as Int64Dtype, StringDtype, Float64Dtype, BooleanDtype, and even the newer ArrowDtype.

@pearce8 this is ready for your review

I reverted the change back to "NaN". They actually were "NaN" before and I had changed it to "nan" to match np.nan. However, since you mentioned it this doesn't make sense, as we should only be using NaN or np.nan, and NaN makes more sense in the context of comments.