Is there a better, pythonic way to do this?

Question

This is my first python program -

Requirement: Read a file consisting of {adId UserId} in each line. For each adId, print the number of unique userIds.

Here is my code, put together from reading the python docs. Could you give me feedback on how I can write this in more python-ish way?

CODE :

import csv

adDict = {}
reader = csv.reader(open("some.csv"), delimiter=' ')
for row in reader:
    adId = row[0]
    userId = row[1]
    if ( adId in adDict ):
        adDict[adId].add(userId)
    else:
        adDict[adId] = set(userId)

for key, value in adDict.items():
    print (key, ',' , len(value))

Thanks.

Answer 1

Congratulations, your code is very nice. There are a few little tricks you could use to make it shorter/simpler.

There is a nifty object type called defaultdict which is provided by the collections module. Instead of having to check if adDict has an adId key, you can set up a defaultdict which acts like a regular dict, except that it automatically provides you with an empty set() when there is no key. So you can change

if ( adId in adDict ):
    adDict[adId].add(userId)
else:
    adDict[adId] = set(userId)

to simply

adDict[adId].add(userId)

Also, instead of

for row in reader:
    adId = row[0]
    userId = row[1]

you could shorten that to

for adId,userId in reader:

Edit: As Parker kindly points out in the comments,

for key, value in adDict.iteritems():

is the most efficient way to iterate over a dict, if you are going to use both the key and value in the loop. In Python3, you can use

for key, value in adDict.items():

since items() returns an iterator.

#!/usr/bin/env python
import csv
from collections import defaultdict

adDict = defaultdict(set)
reader = csv.reader(open("some.csv"), delimiter=' ')
for adId,userId in reader:
    adDict[adId].add(userId)
for key,value in adDict.iteritems():
    print (key, ',' , len(value))