My Dataframe look quite like this. I would like to drop column since it contains a error message.

col1, col2, col3
"module 'numpy' has no attribute 'float'......(ABCGG/mordred._graph_matrix.DistanceMatrix(False, False, False))",0,0,0

However, pandas(or python) treat it as a number.

descriptors['col1'].isna()
0    False
Name: col1, dtype: bool

Therefore, dropna won’t work for me and ironically, I can add this error message by one. Then it will become nan. Is there a way to drop this kind of error message(for instances:<mordred.error.Error at 0x77d429a789a0>)?