处理复杂的CSV文件

Question

I have a CSV file like:

I want to convert to a dictionary using Python, and the dictionary should look like:

[{'Type': ['Date'], 'Value': ['dd/mm/yyyy']}, 
 {'Type': ['Gender'], 'Value': ['Male', 'Female']},
 {'Type': ['Title'], 'Value': ['Mr.', 'Mrs.', 'Ms.']}]

Currently I've tried the below:

import csv
import collections

with open('test2.csv', 'rU') as fin:
   reader = csv.DictReader(fin)

   data = []
   for row in reader:
       data.append(row)
   print data

And the output is:

[{'Type': 'Date', 'Value': 'dd/mm/yyyy'}, 
 {'Type': 'Gender', 'Value': 'Male'}, 
 {'Type': '', 'Value': 'Female'}, 
 {'Type': 'Title', 'Value': 'Mr.'}, 
 {'Type': '', 'Value': 'Mrs.'}, 
 {'Type': '', 'Value': 'Ms.'}]

Answer 1

Try this:

import csv

data=[]
with open(fn, 'rU') as fin:
    reader=csv.reader(fin, dialect='excel')
    header=next(reader)
    for row in reader:
        di={k:[v] for k,v in zip(header, row)}
        if di[header[0]]==['']:
            data[-1][header[1]].extend(di[header[1]]) 
        else:
            data.append(di) 

>>> data
[{'Type': ['Date'], 'Value': ['dd/mm/yyyy']}, {'Type': ['Gender'], 'Value': ['Male', 'Female']}, {'Type': ['Title'], 'Value': ['Mr.', 'Mrs', 'Ms']}]

Answer 2

sadly you can't read that in using csv.DictReader as this is very non standard format of an csv file

you will probably have to read it and parse it manually.

i assume you always expect two columns, and if the type is empty then you use the type from previous line.

as an alternative it may be worth changing the format and making the values in column A mandatory (if it's something you control), which solves some of your problems but not all, you would still have to agregate the results from csv reader.

Answer 3

import csv
from pprint import pprint

with open('test.csv','r') as test_file:
    reader = csv.reader(test_file, delimiter=',')

    output = []
    last_key = None

    for row in reader:
        if row[0]:
            last_key = row[0]
            output.append({row[0]:[row[1]]})
        else:
            output[-1][last_key].append(row[1])

pprint(output)

---

>>> 
[{'Type': ['Value']},
 {'Date': ['dd/mm/yy']},
 {'Gender': ['Male', 'Female']},
 {'Title': ['Mr.', 'Mrs.', 'Ms.']}]

Answer 4

If you know that your csv is going to be two columns and you know that it will always be nicely grouped in the way you've shown, then it might be easiest just to build up your dictionary by hand. The trick it, when there isn't a value in the 1st column, you want to use the previously known value.

from collections import defaultdict
import csv

last_key = None
data = defaultdict(list)
with open('test2.csv', 'rU') as fin:
    csv_reader = csv.reader(fin, delimiter=',')
    for row in csv_reader:
        key, value = row[0], row[1]
        if key:
            data[key].append(value)
            last_key = key
        else:
            data[last_key].append(value)