Performance Optimizations in Redux' mapStateToProps


I already talked about issues with Redux’ connect/mapStateToProps in my post last week, let’s address the performance issues in this post. In Redux, you connect a component to the store via connect by providing a mapStateToProps function. The problem is that this function gets called on every state update, even if the connected component is currently invisible, or the part of the state tree that got updated has no relevance for this component. For example, you’re doing Navigation in React Native and push routes onto a route stack. You will end up with many mounted components that are currently not visible, but their mapStateToProps will still get called every time the user triggers a state change. This is fine as long as your selectors in mapStateToProps only do easy computations on the state. If the result is however not easily derivable from the state, doing the computations all over again, every time, slows down your performance, and is simply just not necessary.

Caching the result

The solution seems pretty simple: You store the active route somewhere in your state tree, and in mapStateToProps you simply check if the active route is equal to this component, to prevent unnecessary calculations when the mounted component is not rendered, and in the selector you check if the relevant part of the state tree changed, to prevent unnecessary recomputations. This is easier said than done, as in mapStateToProps(state, ownProps) you must return the new injected properties, while only having access to the state and the properties passed down from the parent, but not to the old injected properties. So you have to memoize -cache the result of- mapStateToProps (or the selector). There is a nice library for this task, reselect, and I encourage you to read through their example and see if it can be applied to your setting. To mine it couldn’t, because their selectors only allow access to the state, whereas I need them to also take dynamic arguments. So let’s create our own simpler reselect from scratch as it’s a good learning experience and helps you understand how exactly the memoization works.

Memoizing the selector

We store the arguments and the result from the last call and if the new arguments match the last arguments, we simply return the last result. Otherwise, we recompute.

// reducer.js
const createMemoizedSelector = () => {
  let lastArgs
  let lastResult
  return (state, dynamicArg) => {
    if (lastArgs) {
      let [lastRelevantState, lastDynamicArg] = lastArgs
      if (getRelevantState(state, dynamicArg) === lastRelevantState
          && dynamicArg === lastDynamicArg) return lastResult
    // first call or relevant part changed, recompute it
    lastResult = expensiveComputation(state, dynamicArg)
    lastArgs = [getRelevantState(state, dynamicArg), dynamicArg]
    return lastResult

const selector = createMemoizedSelector()
export selector

Memoizing mapStateToProps

Here, we only call the selector if the component is actually rendered to save on performance. When the component becomes the active route, this will be reflected in a state change and mapStateToProps will be called, and only then we do the computation.

// Component.js
import { selector } from 'reducer'
import { connect } from 'react-redux'

class MyComponent extends Component {

var lastInjectedProps
const mapStateToProps = (state, ownProps) => {
  // if we are not getting rendered, don't do computations
  if(getActiveRoute(state) === 'MyComponent') {
    lastInjectedProps = {
      foo: selector(state,
  return lastInjectedProps

Hi, I'm Christoph Michel 👋

I'm a , , and .

I mostly do and help small to mid-sized teams from all over the world 🌎 on both short and long-term projects on an independent contractor basis.

I strive for efficiency ⏱️ and therefore track many aspects of my life.