Python：csv.Dictreader 列上的额外逗号

Question

I have this read function where it reads a csv file using csv.DictReader. The file.csv is separated by commas and it fully reads. However, this part of my file has a column that contains multiple commas. My question is, how can I make sure that comma is counted as part of a column? I cannot alter my csv file to meet the criteria.

Text File :

ID,Name,University,Street,ZipCode,Country
12,Jon Snow,U of Winterfell,Winterfell #45,60434,Westeros
13,Steve Rogers,NYU,108, Chelsea St.,23333,United States
20,Peter Parker,Yale,34, Tribeca,32444,United States
34,Tyrion Lannister,U of Casterly Rock,Kings Landing #89, 43543,Westeros

The desired output is this:

{'ID': '12', 'Name': 'Jon Snow', 'University': 'U of Winterfell', 'Street': 'Winterfell #45', 'ZipCode': '60434', 'Country': 'Westeros'}
{'ID': '13', 'Name': 'Steve Rogers', 'University': 'NYU', 'Street': '108, Chelsea St.', 'ZipCode': '23333', 'Country': 'United States'}
{'ID': '20', 'Name': 'Peter Parker', 'University': 'Yale', 'Street': '34, Tribeca', 'ZipCode': '32444', 'Country': 'United States'}
{'ID': '34', 'Name': 'Tyrion Lannister', 'University': 'U of Casterly Rock', 'Street': 'Kings Landing #89', 'ZipCode': '43543', 'Country': 'Westeros'}

As you can tell the 'Street' has at least two commas due to the numbers:

13,Steve Rogers,NYU, 108, Chelsea St. ,23333,United States

20,Peter Parker,Yale, 34, Tribeca ,32444,United States

Note: Most of the columns being read splits by a str,str BUT under the 'Street' column it is followed by a str, str (there is an extra space after the comma). I hope this makes sense.

The options I tried looking out is using re.split, but I don't know how to implement it on my read file. I was thinking re.split(r'(?!\\s),(?!\\s)',x[:-1]) ? How can I make sure the format from my file will count as part of any column? I can't use pandas.

My current output looks like this right now:

{'ID': '12', 'Name': 'Jon Snow', 'University': 'U of Winterfell', 'Street': 'Winterfell #45', 'ZipCode': '60434', 'Country': 'Westeros'}
{'ID': '13', 'Name': 'Steve Rogers', 'University': 'NYU', 'Street': '108', 'ZipCode': 'Chelsea St.', 'Country': '23333', None: ['United States']}
{'ID': '20', 'Name': 'Peter Parker', 'University': 'Yale', 'Street': '34', 'ZipCode': 'Tribeca', 'Country': '32444', None: ['United States']}
{'ID': '34', 'Name': 'Tyrion Lannister', 'University': 'U of Casterly Rock', 'Street': 'Kings Landing #89', 'ZipCode': '43543', 'Country': 'Westeros'}

This is my read function:

import csv

list = []
with open('file.csv', mode='r') as csv_file:
  csv_reader = csv.DictReader(csv_file, delimiter=",", skipinitialspace=True)

  for col in csv_reader:
    list.append(dict(col))
    print(dict(col))

Answer 1

You can't use csv if the file isn't valid CSV format.

You need to call re.split() on ordinary lines, not on dictionaries.

list = []
with open('file.csv', mode='r') as csv_file:
    keys = csv_file.readline().strip().split(',') # Read header line
    for line in csv_file:
        line = line.strip()
        row = re.split(r'(?!\s),(?!\s)',line)
        list.append(dict(zip(keys, row)))

Answer 2

The actual solution for the problem is modifying the script that generates the csv file.

If you have a chance to modify that output you can do 2 things

• Use a delimiter other than a comma such as | symbol or ; whatever you believe it doesn't exist in the string.
• Or enclose all columns with " so you'll be able to split them by , which are actual separators.

If you don't have a chance to modify the output.

And if you are sure about that multiple commas are only in the street column; then you should use csv.reader instead of DictReader this way you can get the columns by Indexes that you are already sure. for instance row[0] will be ID row[1] will be Name and row[-1] will be Country row[-2] will be ZipCode so row[2:-2] would give you what you need i guess. Indexes can be arranged but the idea is clear I guess.

Hope that helps.

---

Edit:

import csv

list = []
with open('file.csv', mode='r') as csv_file:
  csv_reader = csv.reader(csv_file, delimiter=",", skipinitialspace=True)
  # pass the header row
  next(csv_reader)
  for row in csv_reader:
  list.append({"ID": row[0],
               "Name": row[1],
               "University": row[2],
               "Street": ' '.join(row[3:-2]),
               "Zipcode": row[-2],
               "Country": row[-1]})
print(list)

-- Here is the output (with pprint)

[{'Country': 'Westeros',
'ID': '12',
'Name': 'Jon Snow',
'Street': 'Winterfell #45',
'University': 'U of Winterfell',
'Zipcode': '60434'},
{'Country': 'United States',
'ID': '13',
'Name': 'Steve Rogers',
'Street': '108 Chelsea St.',
'University': 'NYU',
'Zipcode': '23333'},
 {'Country': 'United States',
'ID': '20',
'Name': 'Peter Parker',
'Street': '34 Tribeca',
'University': 'Yale',
'Zipcode': '32444'},
 {'Country': 'Westeros',
'ID': '34',
'Name': 'Tyrion Lannister',
'Street': 'Kings Landing #89',
'University': 'U of Casterly Rock',
'Zipcode': '43543'}]

-- second edit edited the index on the street. Regards.