TopolearnThese containers turn ten lines of bookkeeping into one and are fast, tested building blocks.
The collections module provides specialised containers that come up in real code daily.
Counter counts hashable items: Counter(words) gives a dict of item to count, most_common(n) returns the top n, and counters support +, - and update(). Missing keys read as 0 instead of raising.
defaultdict(factory) calls factory() to create a value the first time a missing key is accessed. defaultdict(list) makes grouping trivial: groups[key].append(item). Only d[key] triggers the factory; d.get(key) and key in d don't insert anything.
deque ("deck") is a double-ended queue with O(1) append, appendleft, pop and popleft — use it for queues and BFS instead of list.pop(0). With maxlen=n it keeps only the last n items, perfect for "recent history" buffers.
namedtuple("Point", "x y") creates a tuple subclass with named fields: lightweight, immutable, still unpackable and indexable. The typed equivalent, typing.NamedTuple, uses class syntax with annotations and is usually nicer to read.
Rounding out the module, ChainMap layers several dicts (e.g. CLI args over environment over defaults), and OrderedDict survives mainly for its move_to_end method, since plain dicts are ordered.
Neighbouring modules worth knowing: itertools for iterator plumbing, functools (lru_cache/cache, partial, reduce), heapq for priority queues, bisect for sorted lists, datetime and zoneinfo for time, json, csv, re and logging.
Example
from collections import Counter, defaultdict, deque
from typing import NamedTuple
words = "to be or not to be that is the question".split()
counts = Counter(words)
print(counts.most_common(2)) # [('to', 2), ('be', 2)]
print(counts["missing"]) # 0
by_length = defaultdict(list)
for w in words:
by_length[len(w)].append(w)
print(by_length[2]) # ['to', 'be', 'or', 'to', 'be', 'is']
recent = deque(["home", "docs", "faq", "blog"], maxlen=3)
print(list(recent)) # ['docs', 'faq', 'blog']
print(recent.popleft()) # docs
class Point(NamedTuple):
x: float
y: float = 0.0
p = Point(3, 4)
x, y = p
print(p, p.x, p[1]) # Point(x=3, y=4) 3 4Remember
Countercounts,defaultdictgroups,dequequeuesdeque(maxlen=n)keeps a rolling windowtyping.NamedTuplegives readable immutable records
Watch out for
- Using
list.pop(0)as a queue, which is O(n) per pop.