--------
        >>> df_not_necessarily_pandas = pd.DataFrame({'A': [1, 2], 'B': [3, 4]})
        >>> interchange_object = df_not_necessarily_pandas.__dataframe__()
        >>> interchange_object.column_names()
        Index(['A', 'B'], dtype='object')
        >>> df_pandas = (pd.api.interchange.from_dataframe
        ...              (interchange_object.select_columns_by_name(['A'])))
        >>> df_pandas
             A
        0    1
        1    2

        These methods (``column_names``, ``select_columns_by_name``) should work
        for any dataframe library which implements the interchange protocol.
        r