问Python :胁迫到Unicode:需要字符串或缓冲区，找到文件

EN

import csv, sys, re, random, os, time, io, StringIO

datfile = sys.argv[1] 

outfileName = sys.argv[2]

outfile = open(outfileName, "w")

count = 0

removal_list = set()

tmp = list()

i=0

re_pattern = "\d+" 

with open(datfile, 'r') as fp:

    reader1 = csv.reader(fp)
    for row in reader1:
        if re.match(re_pattern, row[0]):
             for cols in row:  
                    removal_list.add(tuple(cols)) #as tuple is hashable

print "::row>>>>>>",row

print "::removal_list>>>>>>>>",removal_list

convert = list(removal_list)

print "<<<<<<<<<<<<<<<<>>>>>>>>>>>>>>>>"

print convert

f = open(datfile, 'r')

reader2 = csv.reader(f)

print ""

print "Removal List Starts"

print removal_list

print "Removal List Ends\n"

new_a_buf =  io.BytesIO() # StringIO.StringIO() : both 'io' & StringIO' work

writer = csv.writer(new_a_buf)

rr =""

j  = 0

for row in reader2:

    if row not in convert:   # removal_list: not used as list not hashable
        
          writer.writerow(row)  #outfile.write(new_a_buf)
        
         '''
 #below code using char array isn't used as it doesn't copy structure of csv file
    
    for cols in row:  #at indentation level of "if row not in  convert", stated above

          if cols not in convert:   # removal_list: not used as list not hashable

              for j in range(0,len(cols)):

                   rr+=cols[j]  #at indentation level of "if cols not in convert:"
        
         outfile.write(rr)  # at the indentation level of 'if'

         print "<<<<<<<<<<<<<<<<", rr

 
f = open(outfile, 'r')

reader2 = csv.reader(f)


       '''

new_a_buf.seek(0)

reader2 = csv.reader(new_a_buf)

for row in reader2:

      print row