How to Use Map, Filter and Reduce in JavaScript

Functional programming has been making quite a splash in the
development world these days. And for good reason: Functional techniques
can help you write more declarative code that is easier to understand
at a glance, refactor, and test.

One of the cornerstones of
functional programming is its special use of lists and list operations.
And those things are exactly what the sound like they are: arrays of
things, and the stuff you do to them. But the functional mindset treats
them a bit differently than you might expect.

This article will take a close look at what I like to call the “big three” list operations: map, filter, and reduce.
Wrapping your head around these three functions is an important step
towards being able to write clean functional code, and opens the doors
to the vastly powerful techniques of functional and reactive

Curious? Let’s dive in.

A Map From List to List

we find ourselves needing to take an array and modify every element in
it in exactly the same way. Typical examples of this are squaring every
element in an array of numbers, retrieving the name from a list of
users, or running a regex against an array of strings.

map is a method built to do exactly that. It’s defined on Array.prototype, so you can call it on any array, and it accepts a callback as its first argument. 

The syntax for map is as shown below.

When you call map on an array, it executes that callback on every element within it, returning a new array with all of the values that the callback returned.

Under the hood, map passes three arguments to your callback:

  1. The current item in the array
  2. The array index of the current item
  3. The entire array you called map on 

Let’s look at some code.

map in Practice

Suppose we have an app that maintains an array of your tasks for the day. Each task is an object, each with a name and duration property:

Let’s say we want to create a new array with just the name of
each task, so we can take a look at everything we’ve gotten done today.
Using a for loop, we’d write something like this:

JavaScript also offers a forEach loop. It functions like a for loop, but manages all the messiness of checking our loop index against the array length for us:

Using map, we can simply write:

Here I included the index and  array
parameters to remind you that they’re there if you need them. Since I
didn’t use them here, though, you could leave them out, and the code
would run just fine.

An even more succinct way of writing map in modern JavaScript is with arrow functions.

Arrow functions are a short form for one line functions that just have a return statement. It doesn’t get much more readable than that.

There are a few important differences between the different approaches:

  1. Using map, you don’t have to manage the state of the for loop yourself.
  2. With map you can operate on the element directly, rather than having to index into the array.
  3. You don’t have to create a new array and push into it. map returns the finished product all in one go, so we can simply assign the return value to a new variable.
  4. You do have to remember to include a return statement in your callback. If you don’t, you’ll get a new array filled with undefined

Turns out, all of the functions we’ll look at today share these characteristics.

fact that we don’t have to manually manage the state of the loop makes
our code simpler and more maintainable. The fact that we can operate
directly on the element instead of having to index into the array makes
things more readable. 

Using a forEach loop solves both of these problems for us. But map still has at least two distinct advantages:

  1. forEach returns undefined, so it doesn’t chain with other array methods. map returns an array, so you can chain it with other array methods.
  2. map returns an array with the finished product, rather than requiring us to mutate an array inside the loop. 

the number of places where you modify state to an absolute minimum is
an important tenet of functional programming. It makes for safer and
more intelligible code.


The callback you pass to map must have an explicit return statement, or map will spit out an array full of undefined. It’s not hard to remember to include a return value, but it’s not hard to forget. 

If you do forget, map
won’t complain. Instead, it’ll quietly hand back an array full
of nothing. Silent errors like that can be surprisingly hard to debug. 

Fortunately, this is the only gotcha with map. But it’s a common enough pitfall that I’m obliged to emphasize: Always make sure your callback contains a return statement!


Reading implementations is an important part of understanding. So, let’s write our own lightweight map to better understand what’s going on under the hood. If you want to see a production-quality implementation, check out Mozilla’s polyfill at MDN.

This code accepts an array and a callback function as
arguments. It then creates a new array; executes the callback on each
element on the array we passed in; pushes the results into the new
array; and returns the new array. If you run this in your console,
you’ll get the same result as before. 

While we’re using a for
loop under the hood, wrapping it up into a function hides the details
and lets us work with the abstraction instead. 

That makes our code more declarative—it says what to do, not how to do it. You’ll appreciate how much more readable, maintainable, and, erm, debuggable this can make your code.

Filter Out the Noise

The next of our array operations is filter. It does exactly what it sounds like: It takes an array, and filters out unwanted elements.

The sytnax for filter is:

Just like map, filter passes your callback three arguments:

  1. The current item 
  2. The current index
  3. The array you called filter on

Consider the following example which filters out any string which is less than 8 characters.

The expected result will be:

revisit our task example. Instead of pulling out the names of each
task, let’s say I want to get a list of just the tasks that took me two
hours or more to get done.

Using forEach, we’d write:

With filter we can simply write:

Just like map, filter lets us:

  • avoid mutating an array inside a forEach or for loop
  • assign its result directly to a new variable, rather than push into an array we defined elsewhere


The callback you pass to map has to include a return statement if you want it to function properly. With filter, you also have to include a return statement (unless you’re using arrow functions), and you must make sure it returns a boolean value.

If you forget your return statement, your callback will return undefined, which filter will unhelpfully coerce to false. Instead of throwing an error, it will silently return an empty array! 

f you go the other route, and return something that’s isn’t explicitly true or false, then filter will try to figure out what you meant by applying JavaScript’s coercion rules. More often than not, this is a bug. And, just like forgetting your return statement, it’ll be a silent one. 

Always make sure your callbacks include an explicit return statement. And always make sure your callbacks in filter return true or false. Your sanity will thank you.


Once again, the best way to understand a piece of code is… well, to write it. Let’s roll our own lightweight filter. The good folks at Mozilla have an industrial-strength polyfill for you to read, too.

Reduce Method

The syntax for  reduce array method in JavaScript is:

map creates a new array by transforming every element in an array, individually. filter creates a new array by removing elements that don’t belong. reduce, on the other hand, takes all of the elements in an array, and reduces them into a single value.

Just like map and filterreduce is defined on Array.prototype
and so available on any array, and you pass a callback as its first
argument. But it also takes a second argument: the value to start
combining all your array elements into. 

reduce passes your callback four arguments:

  1. The current value
  2. The previous value 
  3. The current index
  4. The array you called reduce on

Notice that the callback gets a previous value on each iteration. On the first iteration, there is no previous value. This is why you have the option to pass reduce an initial value: It acts as the “previous value” for the first iteration, when there otherwise wouldn’t be one.

Finally, bear in mind that reduce returns a single value, not an array containing a single item. This is more important than it might seem, and I’ll come back to it in the examples.

reduce in Practice

Let’s say you want to find the sum of a list of numbers. Using a loop, it would look like this:

While this isn’t a bad use case for forEachreduce still has the advantage of allowing us to avoid mutation. With reduce, we would write:

First, we call reduce on our list of numbers. We
pass it a callback, which accepts the previous value and current value
as arguments, and returns the result of adding them together.  Since we
passed 0 as a second argument to reduce, it’ll use that as the value of previous on the first iteration.

With arrow functions, we would write it like this:

If we take it step by step, it looks like this:

Iteration  Previous Current Total
1 0 1 1
2 1 2 3
3 3 3 6
4 6 4 10
5 10 5 15

If you’re not a fan of tables, run this snippet in the console:

To recap: reduce iterates over all
the elements of an array, combining them however you specify in your
callback. On every iteration, your callback has access to the previous value, which is the total-so-far, or accumulated value; the current value; the current index; and the entire array, if you need them.

Let’s turn back to our tasks example. We’ve gotten a list of task names from map, and a filtered list of tasks that took a long time with… well, filter

What if we wanted to know the total amount of time we spent working today?

Using a forEach loop, you’d write:

With reduce, that becomes:

That’s almost all there is to it. Almost, because JavaScript provides us with one more little-known method, called reduceRight. In the examples above, reduce started at the first item in the array, iterating from left to right:

reduceRight does the same thing, but in the opposite direction:

I use reduce every day, but I’ve never needed reduceRight. I reckon you probably won’t, either. But in the event you ever do, now you know it’s there.


The three big gotchas with reduce are:

  1. Forgetting to return
  2. Forgetting an initial value
  3. Expecting an array when reduce returns a single value

the first two are easy to avoid. Deciding what your initial value
should be depends on what you’re doing, but you’ll get the hang of it

The last one might seem a bit strange. If reduce only ever returns a single value, why would you expect an array?

There are a few good reasons for that. First, reduce always returns a single value, not always a single number.
If you reduce an array of arrays, for instance, it will return a single
array. If you’re in the habit or reducing arrays, it would be fair to
expect that an array containing a single item wouldn’t be a special

Second, if reduce did return an array with a single value, it would naturally play nice with map and filter, and other functions on arrays that you’re likely to be using with it. 


Time for our last look under the hood. As usual, Mozilla has a bulletproof polyfill for reduce if you want to check it out.

Two things to note, here:

  1. This time, I used the name accumulator instead of previous. This is what you’ll usually see in the wild.
  2. I assign accumulator an initial value, if a user provides one, and default to 0, if not. This is how the real reduce behaves, as well.

Putting It Together: Map, Filter, Reduce, and Chainability

At this point, you might not be that impressed. 

Fair enough: map, filter, and reduce, on their own, aren’t awfully interesting. 

After all, their true power lies in their chainability. 

Let’s say I want to do the following:

  1. Collect two days’ worth of tasks.
  2. Convert the task durations to hours, instead of minutes.
  3. Filter out everything that took two hours or more.
  4. Sum it all up.
  5. Multiply the result by a per-hour rate for billing.
  6. Output a formatted dollar amount.

First, let’s define our tasks for Monday and Tuesday:

And now, our lovely-looking transformation:

If you’ve made it this far, this should be pretty straightforward. There are two bits of weirdness to explain, though. 

First, on line 10, I have to write:

Two things to explain here:

  1. The plus signs in front of accumulator and current coerce their values to numbers. If you don’t do this, the return value will be the rather useless string, "12510075100".
  2. If don’t wrap that sum in brackets, reduce will spit out a single value, not an array. That would end up throwing a TypeError, because you can only use map on an array! 

The second bit that might make you a bit uncomfortable is the last reduce, namely:

That call to map returns an array containing a single value. Here, we call reduce to pull out that value.

Finally, let’s see how our friend the forEach loop would get it done:

Tolerable, but noisy.

Conclusion and Next Steps

In this tutorial, you’ve learned how mapfilter, and reduce
work; how to use them; and roughly how they’re implemented. You’ve seen
that they all allow you to avoid mutating state, which using for and forEach loops requires, and you should now have a good idea of how to chain them all together. 

now, I’m sure you’re eager for practice and further reading. For a
masterclass in functional programming in JavaScript, check out our
online course.

Questions, comments, or confusions? Leave them in the comment section.

Leave a comment

Your email address will not be published.