Is there no constants in Python? Otherwise why can you do this??

import numpy as np
np.nan = 0
print(np.nan == np.nan, np.isnan(np.nan))  # True False

The operator can return different type depending on input and you have to check the return type. And I guess someone thought it is a great idea designing it?

df = pd.DataFrame([{'Name': 'Alice'},
                   {'Name': 'Jack'},
                   {'Name': 'Helen'}],
                   index=['group1','group1','group2'])
print(type(df.loc['group1'])) # pandas.core.frame.DataFrame
print(type(df.loc['group2'])) # pandas.core.series.Series

print(type(df.loc[df.index == 'group2'])) # workaround to always get DataFrame

Another way to treat missing values (the most illogical as to me)

print(np.nan == np.nan, np.nan != np.nan) # False True