[Python] Why does replacing NaN values with None have different results in Windows vs Unix?

I have a dataframe which had NaN values and I'm replacing the NaN values with None using the below definition with NumPy:

df.replace({np.nan: None})

The result of this value in Windows is NULL, whereas the same code provides the result as a 'nan' string in a Unix environment.

I have even used multiple combinations: df.replace({np.nan: None}).replace({np.NaN: None}).replace({np.NAN: None})

I also tried replacing the NaN string using df.replace('nan': None), this provides me string value, instead of NULL/a 'nan' string, which should not be the case.

Continue reading...