# coding=utf-8
# 分批讀取共有358086行內(nèi)容的txt文件，每取1000條輸出到一個文件當(dāng)中

txt_path = "E:/torrenthandle.txt"
base_path="E:/torrent_distribution/"


def distribution( ):
 f = open(txt_path,"r") 
 lines = f.readlines()
 f2=open(base_path+"1.txt","w")
 content=""
 for i in range( 1,len(lines) ):
  if ( i%1000!=0 ):
   content+=lines[i-1]
  else:
   content+=lines[i-1]
   f2.write(content.strip('\n'))
   block_path=base_path+str(i)+".txt"
   f2=open(block_path,"w")
   content=""
 #最后的掃尾工作
 content+=lines[i] 
 f2.write(content.strip('\n')) 
 f2.close()
 f.close()

distribution( )

# coding: utf-8

import os
import shutil

sourcepath = "E:\\sample"
distribution_path = "E:\\sample\\distribution\\" 

if __name__ =='__main__':
 rs = unicode(sourcepath , "utf8")
 count = 1
 savepath = unicode(distribution_path+"1", "utf-8")
 if not os.path.exists(savepath):
  os.makedirs(savepath)
 for rt,dirs,files in os.walk(rs):
  for fname in files:
   if ( count%1000!=0 ):
    shutil.copy(rt + os.sep + fname,savepath) 
    #os.remove(rt + os.sep + fname)
   else:
    shutil.copy(rt + os.sep + fname,savepath) 
    #os.remove(rt + os.sep + fname)
    savepath = unicode(distribution_path+str(count), "utf-8")
    if not os.path.exists(savepath):
     os.makedirs(savepath)
   count+=1