12

I've tried many solutions to add a header to my csv file, but nothing's working properly. Here they are :

  1. I used the writerow method, but my data are overwriting the first row.

  2. I used the DictWriter method, but I don't know how to fill it correctly. Here is my code:

    csv = csv.DictWriter(open(directory +'/csv.csv', 'wt'), fieldnames = ["stuff1", "stuff2", "stuff3"], delimiter = ';')
csv.writeheader(["stuff1", "stuff2", "stuff3"])

I got a "2 arguments instead of one" error and I really don't know why.

Any advice?

Martijn Pieters
  • 1,048,767
  • 296
  • 4,058
  • 3,343
Crolle
  • 822
  • 3
  • 10
  • 20

2 Answers2

25

All you need to do is call DictWriter.writeheader() without arguments:

with open(os.path.join(directory, 'csv.csv'), 'wb') as csvfile:
    writer = csv.DictWriter(csvfile, fieldnames = ["stuff1", "stuff2", "stuff3"], delimiter = ';')
    writer.writeheader()

You already told DictWriter() what your headers are.

Martijn Pieters
  • 1,048,767
  • 296
  • 4,058
  • 3,343
  • 1
    Umm - is `t` a valid `open` mode ? – Jon Clements Apr 09 '13 at 16:22
  • @JonClements: I didn't pay attention to the mode; just copied from the OP. Corrected. – Martijn Pieters Apr 09 '13 at 16:23
  • 2
    TypeError: 'fieldnames' is an invalid keyword argument for this function, I'm using Python 2.6 , PyQt4.7. Can you help? – Aleksandar Mar 11 '14 at 14:41
  • 1
    @Aleksandar: `fieldnames` is a keyword argument for the `csv.DictWriter()` and `csv.DictReader()` classes *only*. Make sure you don't confuse this with `csv.writer()` or `csv.reader()`. Python 2.6 supports it just fine, if you use it correctly. – Martijn Pieters Mar 11 '14 at 14:57
  • @MartijnPieters ok, thank you, I've tried it like in your answer and get `AttributeError: DictWriter instance has no attribute 'writeheader'` – Aleksandar Mar 11 '14 at 16:08
  • 5
    @Aleksandar: Right, *that* Python 2.6 doesn't support. The method was added in Python 2.7. You can simulate it with `writer.writerow(dict(zip(writer.fieldnames, writer.fieldnames)))`. – Martijn Pieters Mar 11 '14 at 16:10
2

I encountered a similar problem when writing the CSV file. I had to read the csv file and modify some of the fields in it. To write the header in the CSV file, i used the following code:

reader = csv.DictReader(open(infile))
headers = reader.fieldnames

with open('ChartData.csv', 'wb') as outcsv:
    writer1 = csv.writer(outcsv)
    writer1.writerow(headers)

and when you write the data rows, you can use a DictWriter in the following way

writer = csv.DictWriter(open("ChartData.csv", 'a' ), headers)

In the above code "a" stands for appending.

In conclusion - > use a to append data to csv after you have written your header to the same file

Rudresh Ajgaonkar
  • 779
  • 1
  • 10
  • 27
  • 1
    Why create a separate `csv.writer()` first, then open the *same file* for appending to with a `csv.DictWriter()`? Just use `writer = csv.DictWriter(outcsv, fieldnames=headers)`, then `writer.writeheader()`. – Martijn Pieters Aug 26 '16 at 17:11