defaultdict
If you're tired of encountering KeyErrors when accessing a non-existent key in a dict, and instead want to have a predefined default value returned when accessing the existent key, defaultdict might be for you.
The problem
Suppose you want to keep track of the scores of people in a game.
And now suppose you directly access the dict using a key that doesn't exist in the dict, you get a KeyError.
scores_dict = {}
# We expect that scores_dict['John'] is a list
scores_dict['John'].append(10)
# KeyError: John is not a key inside scores_dict
Simple workaround
The native solution would be to just declare the key John inside scores_dict beforehand so as to avoid the KeyError.
scores_dict = {}
if 'John' not in scores_dict:
scores_dict['John'] = []
scores_dict['John'].append(10)
print(scores_dict) # { 'John' : [10] }
Using defaultdict
By using defaultdict, we pass a function as an argument, so that we are able to define what the default value will be returned in place of the KeyError.
from collections import defaultdict
# Remember, you must pass a function, not a value
scores_dict = defaultdict(list)
scores_dict['John'].append(10)
# scores_dict is first evaluated, and returns a list before appending the value 10 to that list
print(scores_dict) # { 'John' : [10] }
Arbitrary default values
list is actually a constructor, and calling list() returns the empty list [].
But what if we wan't to set arbitrary values of our own? We use lambda
from collections import defaultdict
# We are passing a function, that returns a value
names_dict = defaultdict(lambda _: 'James')
names_dict[007] += " Bond"
# names_dict[007] is first evaluated, and returns 'James'
names_dict[009] = "Moneypenny"
print(names_dict) # { 007: 'James Bond', 009: "Moneypenny"}
We've just saved ourselves a few lines, thanks to the syntactic sugar of Python. But more importantly, we are able to predictably set a default value for every new key in the dict.