#Introduction
I recently read through a fairly long piece of machine vision code. It used Python’s defaultdict, which was new to me, so I wrote up how it works.
#What defaultdict does
Definition: defaultdict is a dictionary subclass in Python’s collections module.
Behavior: When you look up a missing key using square brackets, a defaultdict with a default factory creates and stores an initial value.
Typical uses: It is handy for accumulating large datasets without initializing every key by hand and for avoiding KeyError on missing entries.
#When to use defaultdict
#1. Appending a value without first initializing the key
Skipping the setup
A regular dictionary needs an initialization step.
from collections import defaultdict
hobbies = defaultdict(list)
hobbies['James'].append('Piano')
hobbies['Bob'].append('Soccer')
# Output: defaultdict(<class 'list'>, {'James': ['Piano'], 'Bob': ['Soccer']})
print(hobbies)
Looking up a new person’s name creates a list for that person’s hobbies.
#2. Using a default value for missing keys
Supplying zero as the default
scores = defaultdict(int)
print(scores['Tom']) # Output: 0. Tom has no entry, so the default factory supplies zero.
With zero as the default, looking up a missing key returns zero.
#How to use defaultdict
from collections import defaultdict
# Create a defaultdict with int as its default factory.
default_d = defaultdict(int)
# Looking up the missing key 'apple' creates a default value of zero.
print(default_d['apple']) # Output: 0
Built-in types such as list, set, int, and float can serve as the default_factory.