What use does the JavaScript forEach method have (that map can't do)?

Question

The only difference I see in map and foreach is that map is returning an array and forEach is not. However, I don't even understand the last line of the forEach method "func.call(scope, this[i], i, this);". For example, isn't "this" and "scope" referring to same object and isn't this[i] and i referring to the current value in the loop?

I noticed on another post someone said "Use forEach when you want to do something on the basis of each element of the list. You might be adding things to the page, for example. Essentially, it's great for when you want "side effects". I don't know what is meant by side effects.

Array.prototype.map = function(fnc) {
    var a = new Array(this.length);
    for (var i = 0; i < this.length; i++) {
        a[i] = fnc(this[i]);
    }
    return a;
}

Array.prototype.forEach = function(func, scope) {
    scope = scope || this;
    for (var i = 0, l = this.length; i < l; i++) {
        func.call(scope, this[i], i, this);
    }
}

Finally, are there any real uses for these methods in JavaScript (since we aren't updating a database) other than to manipulate numbers like the following?

alert([1,2,3,4].map(function(x){ return x + 1})); // This is the only example I ever see of map in JavaScript.

Solution 1

The essential difference between map and forEach in your example is that forEach operates on the original array elements, whereas map explicitly returns a new array as a result.

With forEach you are taking some action with -- and optionally changing -- each element in the original array. The forEach method runs the function you provide for each element, but returns nothing (undefined). On the other hand, map walks through the array, applies a function to each element, and emits the result as a new array.

The "side effect" with forEach is that the original array is being changed. "No side effect" with map means that, in idiomatic usage, the original array elements are not changed; the new array is a one-to-one mapping of each element in the original array -- the mapping transform being your provided function.

The fact that there's no database involved does not mean that you won't have to operate on data structures, which, after all, is one of the essences of programming in any language. As for your last question, your array can contain not only numbers, but objects, strings, functions, etc.

Solution 2

The main difference between the two methods is conceptual and stylistic: You use forEach when you want to do something to or with each element of an array (doing "with" is what the post you cite meant by "side-effects", I think), whereas you use map when you want to copy and transform each element of an array (without changing the original).

Because both map and forEach call a function on each item in an array, and that function is user-defined, there is almost nothing you can do with one and not with the other. It's possible, though ugly, to use map to modify an array in-place and/or do something with array elements:

var a = [{ val: 1 }, { val: 2 }, { val: 3 }];
a.map(function(el) {
    el.val++; // modify element in-place
    alert(el.val); // do something with each element
});
// a now contains [{ val: 2 }, { val: 3 }, { val: 4 }]

but much cleaner and more obvious as to your intent to use forEach:

var a = [{ val: 1 }, { val: 2 }, { val: 3 }];
a.forEach(function(el) {
    el.val++;
    alert(el.val);
});

Especially if, as is usually the case in the real world, el is a usefully human-readable variable:

cats.forEach(function(cat) {
    cat.meow(); // nicer than cats[x].meow()
});

In the same way, you can easily use forEach to make a new array:

var a = [1,2,3],
    b = [];
a.forEach(function(el) {
    b.push(el+1);
});
// b is now [2,3,4], a is unchanged

but it's cleaner to use map:

var a = [1,2,3],
    b = a.map(function(el) {
        return el+1;
    });

Note as well that, because map makes a new array, it likely incurs at least some performance/memory hit when all you need is iteration, particularly for large arrays - see http://jsperf.com/map-foreach

As for why you'd want to use these functions, they're helpful any time you need to do array manipulation in JavaScript, which (even if we're just talking about JavaScript in a browser environment) is pretty often, almost any time you're accessing an array that you're not writing down by hand in your code. You might be dealing with an array of DOM elements on the page, or data pulled from an Ajax request, or data entered in a form by the user. One common example I run into is pulling data from an external API, where you might want to use map to transform the data into the format you want and then use forEach to iterate over your new array in order to display it to your user.

Solution 3

The voted answer (from Ken Redler) is misleading.

A side effect in computer science means that a property of a function/method alters a global state [Wikipedia]. In some narrow sense, this may also include reading from a global state, rather than from arguments. In imperative or OO programming, side effects appear most of the time. And you are probably making use of it without realizing.

The significant difference between forEach and map is that map allocates memory and stores the returning value, while forEach throws it away. See the ECMA specification for more information.

As for the reason why people say forEach is used when you want a side effect is that the return value of forEach is always undefined. If it has no side effect (does not change global state), then the function is just wasting CPU time. An optimizing compiler will eliminate this code block and replace the it with the final value (undefined).

By the way, it should be noted that JavaScript has no restriction on side effects. You can still modify the original array inside map.

var a = [1,2,3]; //original
var b = a.map( function(x,i){a[i] = 2*x; return x+1} );
console.log("modified=%j\nnew array=%j",a,b);
// output:
// modified=[2,4,6]
// new array=[2,3,4]

Solution 4

This is a beautiful question with an unexpected answer.

The following is based on the official description of Array.prototype.map().

There is nothing that forEach() can do that map() cannot. That is, map() is a strict super-set of forEach().

Although map() is usually used to create a new array, it may also be used to change the current array. The following example illustrates this:

var a = [0, 1, 2, 3, 4], mapped = null;
mapped = a.map(function (x) { a[x] = x*x*x; return x*x; });
console.log(mapped); // logs [0, 1, 4, 9, 16]  As expected, these are squares.
console.log(a); // logs [0, 1, 8, 27, 64] These are cubes of the original array!!

In the above example, a was conveniently set such that a[i] === i for i < a.length. Even so, it demonstrates the power of map(), and in particular its ability to change the array on which it is called.

Note1:
The official description implies that map() may even change length the array on which it is called! However, I cannot see (a good) reason to do this.

Note 2:
While map() map is a super-set of forEach(), forEach() should still be used where one desires the change a given array. This makes your intentions clear.

Solution 5

You can use map as though it were forEach.

It will do more than it has to, however.

scope can be an arbitrary object; it's by no means necessarily this.

As for whether there are real uses for map and forEach, as well to ask if there are real uses for for or while loops.

View more solutions

Comments

[JohnMerlino]

But why is both "scope" and "this" being called here: func.call(scope, this[i], i, this); Isn't scope a parameter that is equal to the current object, which is "this"?

[wombleton]

No, it can be equal to the current object. The object itself is passed as the third parameter to the array. scope = scope || this means "if scope is falsy (undefined, null, false, etc) set scope to this instead and carry on".

[JohnMerlino]

Can you link me to an example when it's not equal to this?

[wombleton]

developer.mozilla.org/en/Core_JavaScript_1.5_Reference/… has one under "Printing the contents of an array with an object method"

[Antti Haapala -- Слава Україні]

Actually there is 1 thing that forEach can do that map can't do - not return an array.

[pdoherty926]

You can also use the third argument to the mapping function to mutate the target array, instead of the scoped variable: mapped = a.map(function (x, i, arr) { arr[i] = x * x * x; return x * x; });.

[Travis Webb]

note from the future: this answer is actually nonsense; .map can do everything .forEach can do (including modifying elements), it just returns a new list constructed from the iterator function.

[Ken Redler]

Responding from the past: I respectfully disagree. .map() creates a new array, and does not alter the original. You could indeed modify the array that is itself being mapped, but that would be at the very least non-idiomatic, if not nonsensical. .forEach(), while similar, applies its function to each element, but always returns undefined. Notwithstanding all the above, the OP is also asking about his own specific functions added to the array prototype; there was no ES5 back here in the past.

[poke]

There’s still nothing that forEach does that you can’t do with map. You could simply not care about the return value from map and you’d have a forEach. The difference is that it’s super ineffective to use map for tasks where you don’t want to create a new array based on results. So for those, you use forEach.

[Ken Redler]

@poke: ..and likewise, you could do whatever you need with a plain old for loop. I don't disagree that there's some difference in "effectiveness", but also I don't think anyone's arguing that one has some magic the other can't somehow achieve. Although there is, in fact, one thing that map cannot do: not return an array. There's value in having JS live up to expectations from other languages as to the behavior of functional favorites like map, reduce, filter, etc. E.g., you could implement reduceRight using similar building blocks, but why not just use reduceRight?

[conny]

@pdoherty926 true, but so can a.forEach(function(x, i, arr) { ..., which, again, is a conceptually more correct way when you there is an intention to mutate each original item as opposed to map values from one array to another

[Sumukh Barve]

@conny: Agreed. Also, please see this answer, to a similar question.

[maschwenk]

I think this is misleading because unless your map closure returns an element, you can easily modify the original array if you're not careful.

[Ken Redler]

@maschwenk, how do you mean? If the function passed to map returns nothing (i.e. undefined), you'd end up emitting a new array of undefined elements. The original array would remain unchanged unless you really went out of your way to change it.

[maschwenk]

look at the first example in @nrabinowitz answer

[Ken Redler]

@mashwenk, I think this is just a matter of your viewing the modification of original array elements as being easy to do if you're not careful, and my viewing it as departing from the idiom enough to be something you'd have to go out of your way to do. Likewise, @ nrabinowitz just considers it "ugly". We're all in agreement it's possible; "easy" vs "awkward" vs "ugly" are matters of opinion and style. As a side note, it's amusing how much voting activity, commentary, and even occasional vitriol (I don't mean your comment) this question generates.

[Chinoto Vokro]

The output of a=[{val:1},{val:2},{val:3},{val:4}]; a.map((b)=>{b.val=b.val**2}); JSON.stringify(a); is [{"val":1},{"val":4},{"val":9},{"val":16}]. I'm not sure if you're wrong or merely being unclear, but .map() does operate on the original array elements, just as .forEach() does, so I'm not sure why you specifically make that point with .forEach().

[chovy]

i see people using .map() all the time to modify inline elements. I was under the impression that was the main benefit of using .map over .forEach

[leewz]

@chovy I believe that you should choose based on whether you want to make an array of results. .map can modify elements, yes, but it would create an array you don't need. You should also consider how choosing one or the other lets the reader guess whether or not you have side-effects

[Ken Redler]

@ChinotoVokro, your example is explicitly mutating the original array by assigning b.val = b.val**2 inside the returned object. That's the precise thing we're saying is indeed possible but confusing and non-idiomatic. Normally you simply return the new value, not also assign it to the original array: a=[{val:1},{val:2},{val:3},{val:4}]; a.map((b)=>{b.val**2}); JSON.stringify(a);

[Chinoto Vokro]

@KenRedler I was confused by my own comment until I thought to look at your edit history. When I originally commented on this, there was no mention of "idiomatic usage", just that .forEach() operates on the original array elements, which in my mind implied that .map() somehow doesn't, so I was showing that it can. You are certainly correct that, idiomatically, map shouldn't change the original array.

[ruffin]

I think the take-home is that you can't return a new array from a call to forEach and chain operations dare he say "[LINQ-style"?], which means map's behavior lends itself to functional logic in a way forEach simply can't. If you want to be anti-pattern with map, you can, of course, but... why? Why not use the right tool for your job? That's precisely why we have forEach & for. Self-comment your code, please. Someday, someone else might have to maintain it. ;^)

[AndrewL64]

The discussion on this thread is gold (and I'm not even being sarcastic) lol

[Ken Redler]

@AndrewL64 I stopped answering questions on SO ages ago; this is the only one that continually attracts new comments. There have even been a bunch of comments that have had to be removed because they violated community rules for polite discourse. People literally insulting each other. It's quite enjoyable.

[Ken Redler]

This question and its answers and comments are always fun to circle back on every couple of years. The way map is implemented in JS, in terms of memory allocation, is another nice angle.