Question

I am trying to create a csv file using python that is truly Excel-compatible (I am using Excel 2007, if that makes any difference).

Here is the sort of thing I’m trying to do

import csv

data = [ ['9-1', '9-2', '9-3'] ]

fd = open('test.csv', 'wb')
try:
    writer = csv.writer(fd, dialect='excel', quotechar='"', quoting=csv.QUOTE_ALL)
    writer.writerows(data)

finally:
    fd.close()

This creates a csv file that contains:

"9-1","9-2","9-3"

When I load it in Excel, I get the columns:

09-Jan  09-Feb  09-Mar

[Grrr, thanks Excel. Seriously, what is the point of quotes?]

Now, I have read elsewhere that in order to get Excel to treat strings as literal they must be prefixed by an equals sign ="Like This". So what I really want to produce is:

="9-1",="9-2",="9-3"

Is there any way to do this with the csv module in python? I’m guessing it will involve creating a custom dialect, but I can’t see any variable that would give me the desired result.

Answer 1

Ok, thanks for all your help, I’ve come up with a rough solution, so I thought I’d share it here.

It’s not the prettiest of solutions but it works for my purposes. A warning though: It will break if you have any quotes in your strings!

import csv
import re

data = [ ['9-1', '9-2', '9-3'] ]


def quoteCsvData(data):
    for x, row in enumerate(data):
        for y, item in enumerate(row):
             if data[x][y] != '':
                 data[x][y] = r'="' + data[x][y] + r'"'
    return data

def unquoteCsvData(data):
    for x, row in enumerate(data):
        for y, item in enumerate(row):
            if data[x][y] != '':
                m = re.match(r'="([^"]*)"',data[x][y])
                if m:
                    data[x][y] =  m.group(1)
    return data


fd = open('test2.csv', 'wb')
data = quoteCsvData(data)
try:
    writer = csv.writer(fd, delimiter=',',
                            quotechar='|', 
                            quoting=csv.QUOTE_NONE, 
                            escapechar='\\')
    writer.writerows(data)

finally:
    fd.close()

fd = open('test2.csv', 'rb')
try:
    reader = csv.reader(fd, delimiter=',', 
                            quotechar='|', 
                            quoting=csv.QUOTE_NONE, 
                            escapechar='\\')
    data = []
    for row in reader:
        data.append(row)

finally:
    fd.close()

data = unquoteCsvData(data)

print data

Update: removed spamming of empty cells with lots of =""