22

I think this is probably something simple, but after an hour of searching, I've had no luck figuring out what I'm doing wrong.

I'm using the following code to read a CSV file - I have no problem reading the file, but when a line contains a field that is double-quoted because it contains the delimiter, the CSV reader ignores the double-quotes and parses the field into 2 separate fields.

Here's the code I'm using:

myReader = csv.reader(open(inPath, 'r'), dialect='excel', delimiter=',', quotechar='"')
for row in myReader:
    print row,
    print len(row)

My input:

hello, this is row 1, foo1
hello, this is row 2, foo2
goodbye, "this, is row 3", foo3

Which gives me:

['hello', ' this is row 1', ' foo1'] 3
['hello', ' this is row 2', ' foo2'] 3
['goodbye', ' "this', ' is row 3"', ' foo3'] 4

What do I need to change so it will recognize the double-quoted field as one field? I'm using python version 2.6.1.

Thanks!

Mahmoud Abdelkader
  • 23,011
  • 5
  • 41
  • 54
jamz
  • 4,991
  • 4
  • 24
  • 19

2 Answers2

31

If you look at the dialect that you're using, you'll notice that the excel dialect is configured as follows:

class excel(Dialect):
    """Describe the usual properties of Excel-generated CSV files."""
    delimiter = ','
    quotechar = '"'
    doublequote = True
    skipinitialspace = False
    lineterminator = '\r\n'
    quoting = QUOTE_MINIMAL

Notice that skipinitialspace is set to False. Just pass that into your reader. Oh and by the way, all the fields you've passed in are already the defaults when using the excel dialect, which is the default dialect parameter passed to csv.reader

So, I would re-write your code like so:

>>> with open(inPath) as fp:
>>>     reader = csv.reader(fp, skipinitialspace=True)
>>>     for row in reader:
>>>         print row,
>>>         print len(row)
['hello', 'this is row 1', 'foo1'] 3
['hello', 'this is row 2', 'foo2'] 3
['goodbye', 'this, is row 3', 'foo3'] 3
Mahmoud Abdelkader
  • 23,011
  • 5
  • 41
  • 54
5

It's because your csv has spaces before the quotes:

one0, one1, one2
two0, two1, two2
tre0, "tr,e1", tre2

vs

one0,one1,one2
two0,two1,two2
tre0,"tr,e1",tre2

You'll need to remove those extra spaces first.

TorelTwiddler
  • 5,996
  • 2
  • 32
  • 39
  • 2
    That's wrong: `csv.reader()` has the `skipinitialspace` option to deal with those blanks. – MERose May 25 '16 at 09:43