Clarification on the insert function of panda DataFrames-CodePudding

I have a question, I am trying to get more familiar with python and pandas and I was wondering why this does not work:

list_features_countries = [features_AU, features_CA, features_UK, features_US, features_JP, features_DE, features_SW]

for x in list_features_countries:

    x = x.drop(columns=x.columns[0], axis=1, inplace=True)
    x = x.insert(0,'time_period', Eco_AU['time_period'])

(It gave the error: AttributeError: 'NoneType' object has no attribute 'insert')

But, this does work:

list_features_countries = [features_AU, features_CA, features_UK, features_US, features_JP, features_DE, features_SW]

 for x in list_features_countries
     x = x.drop(columns=x.columns[0], axis=1, inplace=True)

 for x in list_features_countries: 
     x = x.insert(0,'time_period', Eco_AU['time_period'])

So basically I tried to remove the first column of every dataframe in the list and then add a column of another dataframe.

CodePudding user response：

The problem with the first variant is the parameter inplace=True. This means that the existing dataframe (x) is changed in place and the call x.drop(... inplace=True) returns None. So x is assigned None at this point, which leads to the error message in the next step.

The solution would be either to remove the inplace=True parameter so that the x.drop(...) function returns the edited dataframe:

x = x.drop(columns=x.columns[0], axis=1)
x = x.insert(0,'time_period', Eco_AU['time_period'])

or to remove the assignment:

x.drop(columns=x.columns[0], axis=1, inplace=True)
x = x.insert(0,'time_period', Eco_AU['time_period'])