Wouldn't that be simply:

  df = pd.concat(df,other_df).select(...)