useMemoWithComparator

It's React.useMemo except that it takes a custom comparator-function so that you have more control over when it re-computes its value.

Quick start

Caution: this hook circumvents some of the guardrails that React put in place. Use it only if you're sure about what you're doing.

const ShoppingCart = props => {
  const {
    items, // e.g. ['apples', 'bananas', 'carrots']
  } = props;

  // first two args are the same as `React.useMemo`
  // last arg is a comparison function
  // if the comparator returns `false`, the memo will be re-computed
  // if it returns `true`, the memo will not re-compute
  const totalPrice = useMemoWithComparator(
    () => {
      computeTotalPrice(items);
    },
    [items],
    // custom comparison function
    (prevDeps, currentDeps) => {
      const [prevItems] = prevDeps;
      const [currentItems] = currentDeps;


      return (
        prevItems.length === currentItems.length &&
        currentItems.every(item => prevItems.includes(item));
      );
    }
  );
};

Why?

React.useMemo is a performance optimization that can remember the result of a previous computation so that you don't need to re-compute on each render. You pass it an array of values that are used in the computation (the dependency array), and it will only re-run the computation if it thinks that one of those values has changed. Simple enough.

However, we have no control over the logic that it uses in order to decide if the values in the dependency array have materially changed or not. React.useMemo uses referential equality to determine whether a previous value is the same as the current value. Referential-equality comparisons are super fast, but will fail to recognize when two values are practically the same but not strictly the same.

e.g.

// triple-equals demonstrates referential equality

// even though these objects are practically the same,
// they are not strictly the same object
// so these return `false`
{ a: 'apple' } === { a: 'apple' } // false
[1, 2, 3] === [1, 2, 3] // false

In order for React to recognize that these values are practically the same, React would have to perform a deep equality check which involves looping over every single value/property (including nested properties) of the object/array. This could cause serious performance issues if we were dealing with really big, gnarly, complex objects and that's why React chose referential equality instead.

useMemoWithComparator removes these guardrails and lets you compare your values however you want

For example, you might know that your values are not very complicated and that performing a deep-equality check wouldn't cause a perf issue. Or you might know that every one of your values has an id prop and that if two values have the same id, they can be treated as equal.

Maybe you're publishing a component to NPM and you don't want your users to have memoize the values that they're sending to you.

See also

A similar concept for React.useEffect - react-use-effect-with-comparator