Creating a dictionary within a dictionary in Python

Use nested dictionaries by implementing the perl’s autovivification feature.

class AutoVivification(dict):
    """Implementation of perl's autovivification feature."""
    def __getitem__(self, item):
        try:
            return dict.__getitem__(self, item)
        except KeyError:
            value = self[item] = type(self)()
            return value​

​# For your case
names = AutoVivification()

with open('impacts.csv', 'rb') as csvfile:
    namesreaders = csv.reader(csvfile, delimiter=' ', quotechar='|')
    for row in namesreaders:
        person, letter, value = row[0], row[1], row[2]
        if names[person][letter]:
            names[person][letter].append(value)
        else:
            names[person][letter] = [value]

I would solve it like this. I am sure it could be shortened quite a bit if you make it more recursive and use comprehension. But to understand the principle and make it similar to your start (solution not tested for syntactic errors):

names = {}

with open('impacts.csv', 'rb') as csvfile:
    namesreaders = csv.reader(csvfile, delimiter=' ', quotechar='|')
    for row in namesreaders:
       person, letter, value = row[0], row[1], row[2]
       if person not in names:
          names[person] = {}
       if letter not in names[person]:
           names[person][letter] = []
       names[person][letter].append(value)
 print names

Don't let the code confuse you. You defined pretty clearly what you need to do. Now just take it slow and do it.

First, as you said: value of second dictionary is a list. You don't want

names[person] = { letter: value }

Instead you want:

names[person] = { letter: [value] }

Now in your else, you know that person is already in names. Is value? If it is, add to the list, otherwise, create a new list:

else:
    if letter in names[person]:
        names[person][letter].append(value)
    else:
        names[person][letter] = value

Now, you can probably go about this a whole lot cleaner using the setdefault method of a dict.

This would do the same thing:

names = {}

with open('impacts.csv', 'rb') as csvfile:
    namesreaders = csv.reader(csvfile, delimiter=' ', quotechar='|')
for row in namesreaders:
    person, letter, value = row[0], row[1], row[2]

    # Make sure that if person isn't in names, 
    # person is added to names as a dict
    names.setdefault(person, {})
    # Make sure that if letter isn't in the names[person] dict
    # letter is added as an empty list
    names[person].setdefault(letter, [])
    # At this point, names[person][letter] is an existing, possibly empty
    # list.  Simply add the new value
    names[person][letter].append(value)

Quite a well known pythonic technique is to use defaultdict() to specify complex dictionary structures at initialization time.

Disclaimer: I would highly recommend considering class structures to handle this data, as these kinds of nested dictionaries can lead to code smells rather quickly.

import functools
from collections import defaultdict

d_inner = functools.partial(defaultdict, list)
d = defaultdict(d_inner)

# Do something with data here
for name, letter, number in row:
    d[name][letter].append(number)

Sample Output:

defaultdict(<functools.partial object at 0x109de71b0>, {'Bob': defaultdict(<type 'list'>, {'A': [2, 3, 4], 'B': [2, 3]}), 'Jill': defaultdict(<type 'list'>, {'A': [1], 'B': [2]})})