>>> df.sort_values(by='col1', ascending=False) col1 col2 col3 col4 4 D 7 2 e 5 C 4 3 F 2 B 9 9 c 0 A 2 0 a 1 A 1 1 B 3 NaN 8 4 D