python csv help

Question

Sometimes I need to parse string that is CSV, but I am having trouble whit quoted comas. As this code demonstrated. I am using python 2.4

import csv
for row in csv.reader(['one",f",two,three']):
    print row

i get 4 elements ['one"', 'f"', 'two', 'three'] but I would like to get this ['one", f"', 'two', 'three'] or 3 elements even if I try to use quotechar = '"' option (this is according to documentation default) still the same, how can I ignore coma in quotes?

Edit: Thank you all for answers obviously I mistaken my input for CSV, et the end I parsed strig for key values (NAME,DESCR...)

This is input

NAME: "2801 chassis", DESCR: "2801 chassis, Hw Serial#: xxxxxxx, Hw Revision: 6.0",PID: CISCO2801 , VID: V03 , SN: xxxxxxxxx

Answer 1

You can get the csv module to tell you, just feed your desired output into the writer

In [1]: import sys,csv

In [2]: csv.writer(sys.stdout).writerow(['one", f"', 'two', 'three'])  
"one"", f""",two,three

In [3]: csv.reader(['"one"", f""",two,three']).next()  
Out[3]: ['one", f"', 'two', 'three']

Answer 2

Actually the result you get is correct—your CSV syntax is wrong.

If you want to quote commas or other characters in a CSV value, you have to use quotes surrounding the whole value, not parts of it. If a value does not start with the quote character, Python's CSV implementation does not assume the value is quoted.

So, instead of using

one",f",two,three

you should be using

"one,f",two,three

Answer 3

Your input string is not really CSV. Instead your input contains the column name in each row. If your input looks like this:

NAME: "2801 chassis", DESCR: "2801 chassis, Hw Serial#: xxxxxxx, Hw Revision: 6.0",PID: CISCO2801 , VID: V03 , SN: xxxxxxxxx
NAME: "2802 wroomer", DESCR: "2802 wroomer, Hw Serial#: xxxxxxx, Hw Revision: 6.0",PID: CISCO2801 , VID: V03 , SN: xxxxxxxxx
NAME: "2803 foobars", DESCR: "2803 foobars, Hw Serial#: xxxxxxx, Hw Revision: 6.0",PID: CISCO2801 , VID: V03 , SN: xxxxxxxxx

The simplest you can do is probably to filter out the column names first, in the whole file. That would then give you a CSV file you can parse. But that assumes each line has the same columns in the same order.

However, if the data is not that consistent, you might want to parse it based on the names. Perhaps it looks like this:

NAME: "2801 chassis", PID: CISCO2801 , VID: V03 , SN: xxxxxxxxx, DESCR: "2801 chassis, Hw Serial#: xxxxxxx, Hw Revision: 6.0"
NAME: "2802 wroomer", DESCR: "2802 wroomer, Hw Serial#: xxxxxxx, Hw Revision: 6.0",PID: CISCO2801 , VID: V03 , SN: xxxxxxxxx
NAME: "2803 foobars",  VID: V03 ,PID: CISCO2801 ,SN: xxxxxxxxx

Or something. In that case I'd parse each line by looking for the first ':', split out the column head from that, then parse the value (including looking for quotes), and then continue with the rest of the line. Something like this (completely untested code):

def parseline(line):
    result = {}
    while ':' in line:
        column, rest = line.split(':',1)
        column = column.strip()
        rest = rest.strip()
        if rest[0] in ('"', '"'): # It's quoted.
            quotechar = rest[0]
            end = rest.find(quotechar, 1) # Find the end of the quote
            value = rest[1:end]
            end = rest.find(',', end) # Find the next comma
        else: #Not quoted, just find the next comma:
            end = rest.find(',', 1) # Find the end of the value
            value = rest[0:end]
        result[column] = value
        line = rest[end+1:]
        line.strip()
    return result