Question

I was checking out Peter Norvig’s code on how to write simple spell checkers. At the beginning, he uses this code to insert words into a dictionary.

def train(features):
    model = collections.defaultdict(lambda: 1)
    for f in features:
        model[f] += 1
    return model

What is the difference between a Python dict and the one that was used here? In addition, what is the lambda for? I checked the API documentation here and it says that defaultdict is actually derived from dict but how does one decide which one to use?

Answer 1

The difference is that a defaultdict will “default” a value if that key has not been set yet. If you didn’t use a defaultdict you’d have to check to see if that key exists, and if it doesn’t, set it to what you want.

The lambda is defining a factory for the default value. That function gets called whenever it needs a default value. You could hypothetically have a more complicated default function.

Help on class defaultdict in module collections:

class defaultdict(__builtin__.dict)
 |  defaultdict(default_factory) --> dict with default factory
 |  
 |  The default factory is called without arguments to produce
 |  a new value when a key is not present, in __getitem__ only.
 |  A defaultdict compares equal to a dict with the same items.
 |  

(from help(type(collections.defaultdict())))

{}.setdefault is similar in nature, but takes in a value instead of a factory function. It’s used to set the value if it doesn’t already exist… which is a bit different, though.