`memoize`

Caches a unary function's results, keyed by its argument.

R Function(A) memoize<A, R>(R Function(A) f)

Lecture

memoize(f) wraps f in a cache: the first time it's called with a given argument, it runs f and remembers the result; every later call with an ==-equal argument returns the cached result instantly, without calling f again. Reach for it when f is expensive (heavy computation, a network call) and is likely to be called repeatedly with the same inputs.

FxDart's memoize is unary only and keys the cache by the argument's ==/hashCode. FxTS's version is variadic and keys on the full argument list via a WeakMap-backed cache — Dart has no direct equivalent (no variadic generics, and WeakMap-style weak keys aren't available for arbitrary objects), so multi-argument functions need to be memoized on a single composite key (a record works well) instead.

Because R is unconstrained, f can return a Future — memoizing an async operation caches the Future itself, so a second call returns an already-completed future immediately rather than re-running the work.

Demo 1 · Basics

Demo 2 · Memoizing an async lookup

The second call to fetchUser(1) returns the cached, already-completed Future — no 150ms wait:

Try it yourself

Exercise: wrap this "slow" cubing function with memoize so that calling it twice with 3 only runs the real computation once.

Related: delay & sleep — used above to simulate a slow async call · debounce — rate-limit calls instead of caching them · identity — the simplest possible function to wrap · always — a constant value, no caching needed