Defaultdict of defaultdict?

python

defaultdict

collections

data-structures

byNikita Barsukov·Dec 18, 2024

To create a defaultdict nested inside another defaultdict, you need to utilize the collections.defaultdict module, with the factory setting set to a lambda function that returns an inner defaultdict of type int. Here's an optimized code snippet for quick use:

from collections import defaultdict

nested_defaultdict = defaultdict(lambda: defaultdict(int))
nested_defaultdict['firstlevel']['secondlevel'] = 1

print(nested_defaultdict)

This quickly gives us a working nested defaultdict where the inner values default to integers for hassle-free multi-dimensional dictionary manipulation.

Advanced usage

Building hierarchical data structures

When dealing with multi-dimensional data such as managing hierarchical datasets (e.g., country->city->population), nested defaultdicts allows easy and dynamic building of data structures:

world_population = defaultdict(lambda: defaultdict(lambda: defaultdict(int)))
world_population['Europe']['UK']['London'] += 1
world_population['Asia']['Japan']['Tokyo'] += 1

Notice how we directly accessed non-existent keys at each level and Python didn't throw a tantrum. Neat, right?

Expanding factory function horizons

Let's go beyond mere ints. Say hello to Counter from the collections module—a handy tool when you are playing detective with the frequency of hashable objects:

from collections import defaultdict, Counter

message_frequency = defaultdict(lambda: defaultdict(Counter))
message_frequency['John Doe']['Thanks'].update('Received')

For those who love the clean syntax, functools.partial function can be a stylish alternative to lambda:

from collections import defaultdict
from functools import partial

nested_defaultdict = defaultdict(partial(defaultdict, int))

Delving deeper with custom functions

Working with a maze of defaults? You need defaultdicts of variable depth. Below function could be your trusted torchlight:

def nested_defaultdict(depth, default_type):
    if depth < 1:
        return default_type()
    return defaultdict(lambda: nested_defaultdict(depth - 1, default_type))

nested_dd = nested_defaultdict(3, int)  # 3 layers deep, Like Inception but less confusing.

Flexible defaults

Default doesn't mean boring. defaultdicts can hold lists too. If you are keeping track of to-do lists or inventories, this flexibility comes in handy:

shopping_list = defaultdict(lambda: defaultdict(list))
shopping_list['Groceries']['Fruits'].append('Bananas')  # Don't forget the bananas! 🍌

Useful tips and potential pitfalls

Defaultdict nuances

Working with nested defaultdicts is like walking in an enchanted forest where trees sprout up wherever you step. Accessing any key automatically creates it—an attribute to beware of.

Pythonic options to solve the problem

While defaultdict is handy, good old subclassing can give you more explicit control over behavior. And all this while being just as Pythonic! Also, some external libraries like pandas can simplify data manipulation and you won’t have to sling defaultdicts around.

Legacy of readable code

Readability counts, especially when dealing with multi-level dictionaries. Make sure your structure doesn't turn into a matplotlib of dots and arrows. Remember to refactor generously to keep the codebase clean and maintainable.

explain-codes / Python / Defaultdict of defaultdict?

Linked

'too many values to unpack', iterating over a dict. key=>string, value=>list



Getting key with maximum value in dictionary?



Get key by value in dictionary



Should I use 'has_key()' or 'in' on Python dicts?



Converting a Pandas GroupBy output from Series to DataFrame

