Functools

Python · Reference cheat sheet

📋 Overview

functools provides higher-order helpers: freeze arguments with partial, memoize with lru_cache, fold with reduce, and preserve metadata with wraps when writing decorators.

🔧 Core concepts

API	Role
`partial(fn, a, *kw)`	Pre-fill args → new callable
`lru_cache(maxsize=128)`	Memoize pure functions
`cache`	Unbounded memoize (3.9+)
`reduce(fn, iterable[, init])`	Fold sequence to one value
`wraps(fn)`	Copy `__name__`/`__doc__` onto wrappers
`singledispatch`	Function overloading by type
`total_ordering`	Fill rich comparisons from a few

💡 Examples

partial:

from functools import partial

def power(base, exp):
    return base**exp

square = partial(power, exp=2)
cube = partial(power, exp=3)
print(square(5), cube(3))  # 25 27

lru_cache:

from functools import lru_cache

@lru_cache(maxsize=256)
def fib(n: int) -> int:
    if n < 2:
        return n
    return fib(n - 1) + fib(n - 2)

print(fib(30), fib.cache_info())
fib.cache_clear()

reduce:

from functools import reduce
import operator

nums = [1, 2, 3, 4]
print(reduce(operator.mul, nums, 1))  # 24
print(reduce(lambda a, b: a + b, nums))  # 10

wraps for decorators:

from functools import wraps

def trace(fn):
    @wraps(fn)
    def wrapper(*args, **kwargs):
        print("call", fn.__name__)
        return fn(*args, **kwargs)
    return wrapper

@trace
def greet(name):
    """Say hi."""
    return f"hi {name}"

print(greet.__name__, greet.__doc__)  # greet, Say hi.

⚠️ Pitfalls

Cache only pure functions: same args → same result; no hidden I/O/mutation.
Args to @lru_cache must be hashable (no lists/dicts).
Unbounded @cache can grow forever — prefer lru_cache(maxsize=...).
Without @wraps, stack traces and docs point at wrapper.
Prefer comprehensions / explicit loops over reduce when clarity matters.

Functools

Functools

📋 Overview

🔧 Core concepts

💡 Examples

⚠️ Pitfalls

🔗 Related

On this page