print df
0 Bytes Client Ip
0 1 1 1000 192.168.10.2
1 0 0 2000 192.168.10.12
2 2 2 500 192.168.10.4
3 3 3 159 192.168.10.56
print df.nlargest(3, 'Client')
0 Bytes Client Ip
1 0 0 2000 192.168.10.12
0 1 1 1000 192.168.10.2
2 2 2 500 192.168.10.4