Learn about the useMemo Hook, its use cases and implementation.
Now we come to one of the slightly more complex Hooks. The
useMemo Hook accepts a function and a list of dependencies, much like the
useEffect Hook. It returns a memoized value (yes, that is a weird word!)
What’s special about this memoized value is that it is only recomputed when there is a change in one of the values in the dependency array.
The dependency array you pass is important. If you don’t pass in a dependency array to the
useMemoHook, its contents will be regenerated on each render, thus undermining its use.
The use of the
useMemo Hook is quite simple. The function
expensiveComputedFunction will only be called when either
Suitable applications for useMemo#
Much like its sibling (of sorts) the
useCallback Hook (which we’ll learn about in the next lesson), the
useMemo Hook is usually found when it comes time to profile your application for performance optimization.
The best employment of this particular Hook is to prevent multiple frequent calls to functions that are computationally expensive, especially when their expected value hasn’t changed or is not expected to have changed. These extra calls can create optimization bottlenecks and are largely unnecessary.
You’ll often find
useMemo dealing with big data items such as graphs, charts, plotting data, and large mathematical equations.
It’s worth a quick mention here that there is a temptation to over-optimize your code, either too liberally or too early. For me, optimization should be kept in mind, but it should be based on as much empirical data as possible.
Try to resist the temptation to pepper your code with
useCallbackat every opportunity as the complexity you add and the readability you lose are often not worth the trade-off for slight gains in perceived performance.
Building the useMemo example#
For this demo we don’t have any extra styles to add, so fire up the
UseMemoExample.jsx file and let’s start by defining some imports and helper functions. We’ll be using a few Hooks here, so we’re importing them upfront.
Next, our helper functions:
countLettersInWord we’re employing a bit of a cheat here, or at least, a forced ‘expensive’ computation. In the real world we’d be looking at a big data calculation or math equation processing function, but here, we have a simple
while expression that increases a counter,
i, around three billion times. Once that’s finished, the function just returns the length of the string
word that we pass in as an argument.
We’ll be memoizing the
countLettersInWord function later on as we pretend it is a computationally expensive function.
Defining the default component export#
Now for the main event, defining the default component export. Let’s start by scaffolding out the component before filling it in: