1. # -*- coding:utf-8 -*-
    2. # @auth ivan
    3. # @time 2018-07-16 11:45:41
    4. # @goal clean the big file
    6. import os
    7. import pandas as pd
    8. root_list = ['/data/project/Temp']
    9. result = []
    11. # IF NOT os.walk, YOU CAN USE THE os.listdir
    12. print("Start!")
    13. for root in root_list:
    14.     for i, j, k in os.walk(root):
    15.         for t in k:
    16.             p = i + os.path.sep + t
    17.             print(p)
    18.             data = pd.DataFrame([[p, os.path.getsize(p)]],
    19.                                 columns=['File', 'Size'])
    20.             result.append(data)
    21. print("Finis!")
    22. data_all = pd.DataFrame(pd.concat(result))
    23. data_all.sort_values('Size', inplace=True)
    24. print(data_all)