parse_dealer_ids.py 888 B

12345678910111213141516171819202122232425262728293031323334353637
  1. # -*- coding: utf-8 -*-
  2. # !/usr/bin/env python
  3. import os
  4. import pandas as pd
  5. import simplejson as json
  6. total = 0
  7. split = {}
  8. file_type_list = ['csv']
  9. def get_file_list(folder):
  10. filelist = [] #存储要copy的文件全名
  11. for dirpath,dirnames,filenames in os.walk(folder):
  12. for file in filenames:
  13. file_type = file.split('.')[-1]
  14. if(file_type in file_type_list):
  15. file_fullname = os.path.join(dirpath, file) #文件全名
  16. filelist.append(file_fullname)
  17. return filelist
  18. filelist = get_file_list('/var/www/UserServer/script/transaction')
  19. dealerids = set()
  20. for file in filelist:
  21. ttt = pd.read_csv(file)
  22. for i in ttt.index.values:
  23. id_dicts = json.loads(str(ttt.ix[i, 3]))
  24. for id_dict in id_dicts:
  25. dealerids.add(id_dict["$oid"])
  26. for dealerid in dealerids:
  27. print dealerid