Not entirely sure if this "consumes" a row, but this will return the number of columns in the csv file.
col_num = len(pd.read_csv(csv_filename)._mgr)