The column of the dataframe df is multi-indexed.
>>>df
c1 c2
d1 d2 d1 d2
r1 11 12 21 22
Therefore, with the code list(df.columns.levels[0] a list of ['c1', 'c2'] is created. The following code can run without error. But PyCharm marks levels in the line print(list(df.columns.levels[0])) as a warning Unresolved attribute reference 'levels' for class 'Index'.
I think it could be solved by more specific type hinting in the function signature of print_columns. In the code below, I only use the type df: pd.DataFrame. PyCharm thinks it is a simple dataframe without multi-index.
Could you please show me how to improve the type hint? Or do you have other solutions to remove the warning from PyCharm?
Thanks.
import pandas as pd
def get_df() -> pd.DataFrame:
data = {
('c1', 'd1'): [11],
('c1', 'd2'): [12],
('c2', 'd1'): [21],
('c2', 'd2'): [22],
}
df = pd.DataFrame.from_dict(data)
df.index = ['r1']
return df
def print_columns(df: pd.DataFrame):
print(list(df.columns.levels[0])) # prints ['c1', 'c2']
if __name__ == '__main__':
df_ = get_df()
print_columns(df=df_)
