1

I am currently using python, to read one CSV file and then just write the last column to a different CSV file.

My code for reading and writing is:

with open('source.csv', 'rb') as csvfile:
    rdr = csv.reader(csvfile, delimiter=',', quotechar='|')
    with open('result.csv', 'wb') as result:
        wtr = csv.writer(result, delimiter=',', quotechar='|' )
        for r in rdr:
            wtr.writerow((r[2]))

The input looks like this:

#SYMBOL_NAME,TIMESTAMP,VALUE
,20140909230500.000000,0.000000
,20140909231000.000000,0.000000
,20140909231500.000000,0.000000
,20140909232000.000000,0.000000
,20140909232500.000000,0.000000

I am not interested in the timestamp and because of the software generating it, in this instance symbol name is an empty column.

I would expect my output to be like this:

VALUE
0.000000
0.000000
0.000000
0.000000
0.000000

However I am actually getting this:

V,A,L,U,E
0,.,0,0,0,0,0,0
0,.,0,0,0,0,0,0
0,.,0,0,0,0,0,0
0,.,0,0,0,0,0,0
0,.,0,0,0,0,0,0

Now I'm pretty sure that's because of delimiter=',' in the csv.writer part. However when I try to set delimiter='' to remove the commas I am told it must be at least one character string.

When I remove the delimiter parameter all together then I get exactly the same output.

Joe Smart
  • 751
  • 3
  • 10
  • 28
  • Your original code was very close. It seems you tried to create a tuple, based on the additional parentheses in your code. If you'd just done this, it would have worked: `wtr.writerow((r[2],))` (although the list solution below is better). See [Returning tuple with a single item from a function](http://stackoverflow.com/q/6682093) for more info. – Duncan Jones Sep 12 '14 at 09:28
  • The additional parentheses were from another question I found when googling how to write only column in a csv. I probably should've looked into the differences between my case and the one in that question, because he was trying to remove a column (therefore listing the ones he wanted to keep in double parenthesis) whereas I want to keep one and remove the rest. I think nico had the same idea as you and his soultion worked. But thanks for the link, it was helpful. – Joe Smart Sep 12 '14 at 09:28

1 Answers1

5

writerow() expects a list as argument. When you give a single string like this ...

wtr.writerow((r[2]))

... it will interpret the strings characters as list items and put them as individual columns into the csv. Do this instead

wtr.writerow([r[2]])
Nico Heidtke
  • 525
  • 4
  • 7