There are dfs with many thousands of columns, which are advisable to delete in a loop and provided they are completely empty. Wrote the code:

for n in df_balance.columns: if df_balance[n].any() == False: df_balance.drop(columns=df_balance.loc[:, n], inplace=True)

but it gives an error:

KeyError: 'labels [nan nan nan ... nan nan nan] not contained in axis'

Tell me where was wrong.

    1 answer 1

    Solution in one line:

     df_balance.dropna(axis='columns',how='all', inplace=True)