CSV.writerow在每个字符之间都有逗号吗？

Question

I am currently calling a python script within my Python script and trying to save the output of my call in a CSV file. At the moment it will work, however, there is a comma between every character so the output is incorrect.

What is causing this?

import csv
import GetAlexRanking #External Method exposed here
import subprocess
import pandas as p
import tai
import numpy as np

loadData = lambda f: np.genfromtxt(open(f,'r'), delimiter=' ')
with open('train.tsv','rb') as tsvin, open('PageRanks.csv', 'wb') as csvout:
    tsvin = list(np.array(p.read_table('train.tsv'))[:,0])
    csvout = csv.writer(csvout)

    for row in tsvin:
        count = 0
        cmd = subprocess.Popen("python GetAlexRanking.py " + row ,
                           stdout=subprocess.PIPE,
                           stderr=subprocess.PIPE,
                           shell=True)
        (output, err) = cmd.communicate()
        exit_code = cmd.wait()
        print exit_code #testing
        print output
        print err
        csvout.writerow(row + "\t" + output) #writing,error here
        count+=1

Edit:

A sample row returned from the function when called in cmd like so "python GetAlexRanking.py www.google.com" is :

www.google.com
AlexaTrafficRank:1
GooglePageRank:9

I would like this to be saved in a tsv as so (spaces added to make formatting more clear, all columns separated by a tab only :))

URL \t AlexaRank \t GoogleRank
www.google.com \t 1 \t 9

Answer 1

You are passing a string to the csv.write, which it then interprets as a list and therefore splits it by each list element, ie character. I've made this mistake so many times...

Try this:

# add coustom code to split the row up into the values, hint user row.split()
csvout.writerow([row, output]) 

Answer 2

It looks as if you want your input to be a list as well as your output. Thus, let your input stay a set of strings and split it into a list at each row.

Your sample row as returned is showing in three lines. Does this mean it is one long string with the column separators? If that is the case split the output and insert the tabs.

  outrow = row # row is already a list
  outrow.append(output.split('\t'))
  csvout.writerow(outrow)

Looking again at your sample, it appears that you want to output two tsv rows, one with the "header" and one with the "rank". Thus (with extra lines for ease of reading)

outlist = output.split('\t')
outname1 = outlist[1][0:outlist[1].index(':')-1]
outname2 = outlist[2][0:outlist[2].index(':')-1]
outrank1 = outlist[1][outlist[1].index(':')+1:]
outrank2 = outlist[2][outlist[2].index(':')+1:]
outrow1 = ['URL', outname1, outname2]
outrow2 = [outlist[0], outrank1, outrank2]

You would then write the two output rows as you seem to have put in your sample output