Problem filtering data without associated standard deviations

[bmanubay]

When trying to filter data by limiting to properties where an associated standard deviation is present (which is a separate column in the pandas dataframe) I get the following error:

TypeError: ufunc 'isnan' not supported for the input types, and the inputs could not be safely coerced to any supported types according to the casting rule ''safe''

Not really sure what this means, but it is an issue worth looking into (I think) for the purposes of the OpenFF project.

[kyleabeauchamp]

It likely means you're calling numpy.isnan() on a column that contains string values, rather than numerical values. See example below:

import pandas as pd
import numpy as np

x = pd.DataFrame([dict(A=1, B=2), dict(A="3", B=4)])

np.isnan(x.A)
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-5-62fa1d22f6fa> in <module>()
----> 1 np.isnan(x.A)

TypeError: ufunc 'isnan' not supported for the input types, and the inputs could not be safely coerced to any supported types according to the casting rule ''safe''


np.isnan(x.B)
Out[6]: 
0    False
1    False
Name: B, dtype: bool

[bmanubay]

That was indeed the case Kyle. It's odd, but some of the columns with no values fill as the string 'nan' rather than the actual data type 'NaN'. Thanks for the input!