When to use Iterator instead of normal looping constructs - Javascript

Question

I have been reading articles regarding Iterators and Iterables in ES6. I have seen one example like below.

let someArr = [1, 2, 3, 4]; let iteratorVar = someArr[Symbol.iterator](); console.log(iteratorVar.next().value); // If we log for 5 times, it will print all array elements.

But my doubt is that we can simply use for, while and etc loops to get the each element from array like below

for (let i = 0; i < someArr.length; i++) { console.log(someArr[i]); }

So, what is the advantage of using Iterables and Iterators over normal looping constructs like for, while etc...

I didn't get any clarification in the articles I read. Thank you for your help.

Ever use for..of or ...values? Those are iterator driven. Things like Array.prototype.map are not but are very useful. I would turn it around and ask why you would ever use an index driven for loop? — Aluan Haddad
– Aluan Haddad, Commented Apr 24, 2018 at 2:20
@AluanHaddad I have used for..of and ...values also. My doubt is why we need to write let iteratorVar = someArr[Symbol.iterator](); to loop instead of simply using available looping constructs — Sivakumar Tadisetti
– Sivakumar Tadisetti, Commented Apr 24, 2018 at 2:24
You usually don't need to. It is occasionally useful but should be the exception not the rule — Aluan Haddad
– Aluan Haddad, Commented Apr 24, 2018 at 2:27
@JSSA in most cases, it's not useful, no. But when you do need it, it can be quite a valuable tool. Since anything can be an iterable if it implements this, then you can set up, say, an infinite source of information and just call the next value from the iterator as long as you want. You could have a pRNG underneath or maybe consuming some sort of potentially infinite feed and processing it, say tweets or something. Another use might be to process something finite but that might be expensive in one go, like reading a file a chunk at a time. — VLAZ
– VLAZ, Commented Apr 24, 2018 at 2:44

Aluan Haddad · Accepted Answer · 2018-04-25 14:30:47Z

A number of facilities in the language are defined in terms of iterators.

This includes syntax such as the for..of loop as well as the spread (...values) of arrays, maps, and sets.

That said you almost never have cause to directly manipulate an iterator itself.

However, there are situations where being able to do so is essential.

For example, imagine we want to write a zip function that combines the corresponding elements of several iterables.

Since iterables can be infinite, we can't create an array and return it. On the one hand we can't allocate an infinite amount of memory and on the other we can't limit our input to finite sequences.

In this case, directly manipulating the iterator gives us exactly what we need

function* zip(iterables = []) { if (iterables.length < 2) { throw RangeError('at least 2 iterables must be provided'); } const iterators = iterables.map(iterable => iterable[Symbol.iterator]); while (true) { const iterations = iterators.map(iterator => iterator.next()); if (iterations.some(iteration => iteration.done)) { return; } yield iterations.map(iteration => iteration.value); } }

That is zip takes an array of 2 or more iteratables and consumes them element wise. It yields their corresponding elements together and ceases when any of them are done.

The above example is fairly complex but has the advantage of not being contrived.

These sorts of scenarios are important but relatively rare making direct use of iterators the exception rather than the rule.

Nice write up! That is why I say an unified, predictable API is important in the language. Especially for people who want to build up a much complex applications.

Collectives™ on Stack Overflow

When to use Iterator instead of normal looping constructs - Javascript

1 Answer 1

1 Comment

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Related