TopolearnCounting words, events, errors or votes is one of the most common data tasks.
A Counter is a dict subclass that maps items to counts. Build one from any iterable of hashable items:
from collections import Counter
words = Counter(text.split())It behaves like a dict, with conveniences:
- Missing items count as 0 instead of raising
KeyError:words["zebra"]is0. Reading doesn't insert the key. c.most_common(3)returns the three highest(item, count)pairs, sorted by count; ties keep first-seen order. Without an argument it returns all items.c.update(iterable)adds more counts;c[item] += 1works directly too.c.total()(3.10+) sums all counts.
You can also build a Counter from a mapping or keyword arguments: Counter({"a": 2}), Counter(a=2, b=1).
Counting a string counts its characters, a common surprise: Counter("hello") has l: 2. Split first if you want words.
A Counter replaces the classic pattern of if key not in d: d[key] = 0 followed by d[key] += 1, and it's implemented with a C helper, so counting a large iterable is fast.
Example
from collections import Counter
text = "the cat and the hat and the bat"
words = Counter(text.split())
print(words["the"], words["dog"]) # 3 0
print(words.most_common(2)) # [('the', 3), ('and', 2)]
print(words.total()) # 8
words.update(["cat", "cat"])
words["hat"] += 1
print(words["cat"], words["hat"]) # 3 2
print(Counter("hello")) # Counter({'l': 2, 'h': 1, 'e': 1, 'o': 1})
status = Counter(code for code in [200, 404, 200, 500, 200])
print(status.most_common(1)[0]) # (200, 3)Remember
Counter(iterable)tallies items; missing keys count as 0most_common(n)returns the top items with counts- Counting a string counts characters, not words
Watch out for
- Passing a sentence string directly to Counter when you meant to count words.