Assert a pandas column type with NaN allowed-CodePudding

I'm coding some tests to verify if some columns of a pandas dataframe are integer (or floats for other columns), but with authorized NaN values.

I'm currently trying with assert statements:

import pandas as pd
import pandas.api.types as ptypes

df = pd.read_excel('the_df', na_values='')

try:
    assert(ptypes.is_numeric_dtype(df['a_column']))
except AssertionError:
    return False
else:
    return True

But this method returns False when there are NaN and I can't figure out how to fix this. How can I proceed ?

Thanks.

CodePudding user response：

Check for the dtype, not the entire array:

import pandas as pd
import pandas.api.types as ptypes

df = pd.read_excel('the_df', na_values='')

return ptypes.is_numeric_dtype(df['a_column'].dtype)

(You don't need the try-except-else chain: the above achieves the exact same thing: it returns True or False).