#df.nlargest(How many largest Values, 'Col_name') >>> df.nlargest(3, 'a') a b c 3 11 c 3 1 10 b 2 2 8 d NaN