Most of the tutorials that I've read on arrays in JavaScript (including w3schools and devguru) suggest that you can initialize an array with a certain length by passing an integer to the Array constructor using the var test = new Array(4);
syntax.
After using this syntax liberally in my js files, I ran one of the files through jsLint, and it freaked out:
Error: Problem at line 1 character 22: Expected ')' and instead saw '4'. var test = new Array(4); Problem at line 1 character 23: Expected ';' and instead saw ')'. var test = new Array(4); Problem at line 1 character 23: Expected an identifier and instead saw ')'.
After reading through jsLint's explanation of its behavior, it looks like jsLint doesn't really like the new Array()
syntax, and instead prefers []
when declaring arrays.
So I have a couple questions:
First, why? Am I running any risk by using the new Array()
syntax instead? Are there browser incompatibilities that I should be aware of?
And second, if I switch to the square bracket syntax, is there any way to declare an array and set its length all on one line, or do I have to do something like this:
var test = [];
test.length = 4;
new Array()
in general, but it's okay with specifying size. I think it all comes down to code consistency through the whole context.
Array(5) gives you an array with length 5 but no values, hence you can't iterate over it.
Array.apply(null, Array(5)).map(function () {}) gives you an array with length 5 and undefined as values, now it can be iterated over.
Array.apply(null, Array(5)).map(function (x, i) { return i; }) gives you an array with length 5 and values 0,1,2,3,4.
Array(5).forEach(alert) does nothing, Array.apply(null, Array(5)).forEach(alert) gives you 5 alerts
ES6 gives us Array.from so now you can also use Array.from(Array(5)).forEach(alert)
If you want to initialize with a certain value, these are good to knows... Array.from('abcde'), Array.from('x'.repeat(5)) or Array.from({length: 5}, (v, i) => i) // gives [0, 1, 2, 3, 4]
Why do you want to initialize the length? Theoretically there is no need for this. It can even result in confusing behavior, because all tests that use the length to find out whether an array is empty or not will report that the array is not empty. Some tests show that setting the initial length of large arrays can be more efficient if the array is filled afterwards, but the performance gain (if any) seem to differ from browser to browser. jsLint does not like new Array() because the constructer is ambiguous. new Array(4); creates an empty array of length 4. But new Array('4'); creates an array containing the value '4'.
Regarding your comment: In JS you don't need to initialize the length of the array. It grows dynamically. You can just store the length in some variable, e.g.
var data = [];
var length = 5; // user defined length
for(var i = 0; i < length; i++) {
data.push(createSomeObject());
}
for
loop that iterates over the length of the array and fills it. I guess there would be other ways to do this in JavaScript, so the real answer to "Why do I want to do this?" is "Because of old habits that formed while programming in other languages." :)
Why do you want to initialize the length? Theoretically there is no need for this. And all tests that use the length to find out whether an array is empty or not will fail.Um, performance perhaps? It's faster to set a pre-existing element of an array than it is to add it on the fly.
new Array(10)
does not create an array with 10 undefined elements. It simply creates an empty array with a length of 10. See this answer for the nasty truth: stackoverflow.com/questions/18947892/…
With ES2015 .fill()
you can now simply do:
// `n` is the size you want to initialize your array
// `0` is what the array will be filled with (can be any other value)
Array(n).fill(0)
Which is a lot more concise than Array.apply(0, new Array(n)).map(i => value)
It is possible to drop the 0
in .fill()
and run without arguments, which will fill the array with undefined
. (However, this will fail in Typescript)
When Array is called as a function rather than as a constructor, it also creates and initializes a new Array object. Thus the function call Array(…) is equivalent to the object creation expression new Array(…) with the same arguments
Array(n).fill()
[...Array(6)].map(x => 0);
// [0, 0, 0, 0, 0, 0]
OR
Array(6).fill(0);
// [0, 0, 0, 0, 0, 0]
Note: you can't loop empty slots i.e. Array(4).forEach(() => …)
OR
( typescript safe )
Array(6).fill(null).map((_, i) => i);
// [0, 1, 2, 3, 4, 5]
OR
Classic method using a function ( works in any browser )
function NewArray(size) {
var x = [];
for (var i = 0; i < size; ++i) {
x[i] = i;
}
return x;
}
var a = NewArray(10);
// [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
Creating nested arrays
When creating a 2D array with the fill
intuitively should create new instances. But what actually going to happen is the same array will be stored as a reference.
var a = Array(3).fill([6]);
// [ [6], [6], [6] ]
a[0].push(9);
// [ [6, 9], [6, 9], [6, 9] ]
Solution
var a = [...Array(3)].map(x => []);
a[0].push(4, 2);
// [ [4, 2], [], [] ]
So a 3x2 Array will look something like this:
[...Array(3)].map(x => Array(2).fill(0));
// [ [0, 0], [0, 0], [0, 0] ]
N-dimensional array
function NArray(...dimensions) {
var index = 0;
function NArrayRec(dims) {
var first = dims[0], next = dims.slice().splice(1);
if(dims.length > 1)
return Array(dims[0]).fill(null).map((x, i) => NArrayRec(next ));
return Array(dims[0]).fill(null).map((x, i) => (index++));
}
return NArrayRec(dimensions);
}
var arr = NArray(3, 2, 4);
// [ [ [ 0, 1, 2, 3 ] , [ 4, 5, 6, 7] ],
// [ [ 8, 9, 10, 11] , [ 12, 13, 14, 15] ],
// [ [ 16, 17, 18, 19] , [ 20, 21, 22, 23] ] ]
Initialize a chessboard
var Chessboard = [...Array(8)].map((x, j) => {
return Array(8).fill(null).map((y, i) => {
return `${String.fromCharCode(65 + i)}${8 - j}`;
});
});
// [ [A8, B8, C8, D8, E8, F8, G8, H8],
// [A7, B7, C7, D7, E7, F7, G7, H7],
// [A6, B6, C6, D6, E6, F6, G6, H6],
// [A5, B5, C5, D5, E5, F5, G5, H5],
// [A4, B4, C4, D4, E4, F4, G4, H4],
// [A3, B3, C3, D3, E3, F3, G3, H3],
// [A2, B2, C2, D2, E2, F2, G2, H2],
// [A1, B1, C1, D1, E1, F1, G1, H1] ]
Math filled values
handy little method overload when working with math
function NewArray( size , method, linear )
{
method = method || ( i => i );
linear = linear || false;
var x = [];
for( var i = 0; i < size; ++i )
x[ i ] = method( linear ? i / (size-1) : i );
return x;
}
NewArray( 4 );
// [ 0, 1, 2, 3 ]
NewArray( 4, Math.sin );
// [ 0, 0.841, 0.909, 0.141 ]
NewArray( 4, Math.sin, true );
// [ 0, 0.327, 0.618, 0.841 ]
var pow2 = ( x ) => x * x;
NewArray( 4, pow2 );
// [ 0, 1, 4, 9 ]
NewArray( 4, pow2, true );
// [ 0, 0.111, 0.444, 1 ]
a[0].push(9);
// [ [6, 9], [6], [6] ]
should be a[0].push(9);
// [ [6, 9], [6, 9], [6, 9] ]
because each nested array is stored as a reference, so mutating one array affects the rest. You may have just mistyped it though I think :)
[...Array(6)].map(x => 0);
is an excellent solution and was very helpful for me! With a slight modification it can be used to create an array that counts integers (just like your typescript safe example): [...Array(6)].map((x, i) => i);
results in [0, 1, 2, 3, 4, 5]
map((_, i) => i)
Array(Infinity).fill(0)
create your universe! Array(Infinity).fill(Infinity)
you're God!
The shortest:
let arr = [...Array(10)]; console.log(arr);
[
without using a ;
since it will attempt to dereference the line above. So a safe way to use this line predictably in any part of the code would be: ;[...Array(1000)]//.whateverOpsNeeded()
[...Array(5)].map((item, index) => ({ index }))
try this as well: [...Array(5)]; console.log('hello');
const a = 'a'
and the next line [...Array(5)].//irrelevent
. What do you think the first line would resolve to? It would be const a = 'a'[...Array(5)]
which would result in: Uncaught SyntaxError: Unexpected token ...
const a = 'a'
will be a Lint error in that case. Anyway, it really has nothing to do with this answer.
ES6 introduces Array.from
which lets you create an Array
from any "array-like" or iterables objects:
Array.from({length: 10}, (x, i) => i);
// [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
In this case {length: 10}
represents the minimal definition of an "array-like" object: an empty object with just a length
property defined.
Array.from
allows for a second argument to map over the resulting array.
Sparse arrays are here! 🥳 [2021]
In modern JS engines, sparse arrays are fully supported. You can use []
or new Array(len)
in any way you like, even with random access. Dictionary mode seems to be a thing of the past.
In current Chrome (and I guess any V8 environment), Arrays can have a length of up to 2^32-1 and allocation is sparse (meaning empty chunks don't use up any memory):
https://i.stack.imgur.com/aEmtE.png
https://i.stack.imgur.com/db75k.png
However, there is a catch
On the one hand, for
loops work as intended, however, Array
's builtin higher order functions (such as map
, filter
, find
, some
etc.) ignore unassigned elements. They require fill
(or some other method of population) first:
const a = new Array(10);
const b = new Array(10).fill(0);
a.forEach(x => console.log(x)); // does nothing
b.forEach(x => console.log(x)); // works as intended
Old Version
(I removed most of the old version.) The gist was that creating a large array using new Array(largeNumber)
or random accessing an array in places that have not yet been allocated would tumble it into "dictionary mode". Meaning you are using an array with indexes, but under the hood it would use a dictionary to store the values, thus messing with performance, and also with iteration behavior. Luckily that is a thing of the past.
let
, var
, or const
). I just want to clear things up for the next person who copies StackOverflow code and realizes it doesn't work 😉
.fill
without an argument and then you can iterate over the array: new Array(10).fill().forEach(i => console.log(i))
This will initialize the length property to 4:
var x = [,,,,];
var size = 42; var myArray = eval("["+",".repeat(size)+"]");
? (Not that serious ;)
I'm surprised there hasn't been a functional solution suggested that allows you to set the length in one line. The following is based on UnderscoreJS:
var test = _.map(_.range(4), function () { return undefined; });
console.log(test.length);
For reasons mentioned above, I'd avoid doing this unless I wanted to initialize the array to a specific value. It's interesting to note there are other libraries that implement range including Lo-dash and Lazy, which may have different performance characteristics.
Here is another solution
var arr = Array.apply( null, { length: 4 } );
arr; // [undefined, undefined, undefined, undefined] (in Chrome)
arr.length; // 4
The first argument of apply()
is a this object binding, which we don't care about here, so we set it to null
.
Array.apply(..)
is calling the Array(..)
function and spreading out the { length: 3 }
object value as its arguments.
new Array(n)
for initializing an array like this new Array(n).map(function(notUsed,index){...})
, but with this approach, as @zangw mentioned, you can do it. +1 from me.
Array.apply(null, { length: childrenLength }).map((notUsed, i) => (<div key={i} />)
Please people don't give up your old habits just yet. There is a large difference in speed between allocating memory once then working with the entries in that array (as of old), and allocating it many times as an array grows (which is inevitably what the system does under the hood with other suggested methods).
None of this matters of course, until you want to do something cool with larger arrays. Then it does.
Seeing as there still seems to be no option in JS at the moment to set the initial capacity of an array, I use the following...
var newArrayWithSize = function(size) {
this.standard = this.standard||[];
for (var add = size-this.standard.length; add>0; add--) {
this.standard.push(undefined);// or whatever
}
return this.standard.slice(0,size);
}
There are tradeoffs involved:
This method takes as long as the others for the first call to the function, but very little time for later calls (unless asking for a bigger array).
The standard array does permanently reserve as much space as the largest array you have asked for.
But if it fits with what you're doing there can be a payoff. Informal timing puts
for (var n=10000;n>0;n--) {var b = newArrayWithSize(10000);b[0]=0;}
at pretty speedy (about 50ms for the 10000 given that with n=1000000 it took about 5 seconds), and
for (var n=10000;n>0;n--) {
var b = [];for (var add=10000;add>0;add--) {
b.push(undefined);
}
}
at well over a minute (about 90 sec for the 10000 on the same chrome console, or about 2000 times slower). That won't just be the allocation, but also the 10000 pushes, for loop, etc..
(this was probably better as a comment, but got too long)
So, after reading this I was curious if pre-allocating was actually faster, because in theory it should be. However, this blog gave some tips advising against it http://www.html5rocks.com/en/tutorials/speed/v8/.
So still being unsure, I put it to the test. And as it turns out it seems to in fact be slower.
var time = Date.now();
var temp = [];
for(var i=0;i<100000;i++){
temp[i]=i;
}
console.log(Date.now()-time);
var time = Date.now();
var temp2 = new Array(100000);
for(var i=0;i<100000;i++){
temp2[i] = i;
}
console.log(Date.now()-time);
This code yields the following after a few casual runs:
$ node main.js
9
16
$ node main.js
8
14
$ node main.js
7
20
$ node main.js
9
14
$ node main.js
9
19
var arr=[];
arr[5]=0;
alert("length="+arr.length); // gives 6
[5: 0]
this is a sparse array and probably not what is wanted.
js const foo = []; foo[10] = undefined;
The value of foo
is not [undefined, undefined undefined, ...]
but [empty * 9, undefined]
. If you tried to run any of the sequencial array methods (forEach
, map
, reduce
), then you would realize that it doesn't actually iterate through the empty items. They are just dead space.
Assuming that Array's length is constant. In Javascript, This is what we do:
const intialArray = new Array(specify the value);
The array constructor has an ambiguous syntax, and JSLint just hurts your feelings after all.
Also, your example code is broken, the second var
statement will raise a SyntaxError
. You're setting the property length
of the array test
, so there's no need for another var
.
As far as your options go, array.length
is the only "clean" one. Question is, why do you need to set the size in the first place? Try to refactor your code to get rid of that dependency.
var test
. That was some sloppy copy-and pasting on my part.
In addition to the answers of others, another clever way is to use Float32Array
to create an array and iterate on it.
For this purpose, create an instance from Float32Array
with your desired length like this:
new Float32Array(5)
This code returns an array-like that you can convert it to an array with Array.from()
:
Array.from(new Float32Array(5)) // [0, 0, 0, 0, 0]
You can also use fill()
to change the value of items:
Array.from(new Float32Array(5).fill(2)) // [2, 2, 2, 2, 2]
And of course you can iterate on it:
Array.from(new Float32Array(5)).map(item => /* ... */ )
In most answers it is recommended to fill
the array because otherwise "you can't iterate over it", but this is not true. You can iterate an empty array, just not with forEach
. While loops, for of loops and for i loops work fine.
const count = Array(5);
Does not work.
console.log('---for each loop:---');
count.forEach((empty, index) => {
console.log(`counting ${index}`);
});
These work:
console.log('---for of loop:---');
for (let [index, empty] of count.entries()) {
console.log(`counting for of loop ${index}`);
}
console.log('---for i loop:---');
for (let i = 0, il = count.length; i < il; ++i) {
console.log(`counting for i loop ${i}`);
}
console.log('---while loop:---');
let index = 0;
while (index < count.length) {
console.log(`counting while loop ${index}`);
index++;
}
Check this fiddle with the above examples.
Also angulars *ngFor
works fine with an empty array:
<li *ngFor="let empty of count; let i = index" [ngClass]="
<span>Counting with *ngFor {{i}}</span>
</li>
You can set the array length by using array.length = youValue
So it would be
var myArray = [];
myArray.length = yourValue;
The reason you shouldn't use new Array
is demonstrated by this code:
var Array = function () {};
var x = new Array(4);
alert(x.length); // undefined...
Some other code could mess with the Array variable. I know it's a bit far fetched that anyone would write such code, but still...
Also, as Felix King said, the interface is a little inconsistent, and could lead to some very difficult-to-track-down bugs.
If you wanted an array with length = x, filled with undefined (as new Array(x)
would do), you could do this:
var x = 4;
var myArray = [];
myArray[x - 1] = undefined;
alert(myArray.length); // 4
alert
and undefined
, because some other code could mess with them. The second example is less readable than new Array(4)
, and doesn't give you the same result: jsfiddle.net/73fKd
Success story sharing
Array
is given multiple arguments, it iterates over thearguments
object and explicitly applies each value to the new array. When you callArray.apply
with an array or an object with a length propertyArray
is going to use the length to explicitly set each value of the new array. This is whyArray(5)
gives an array of 5 elisions, whileArray.apply(null, Array(5))
gives an array of 5 undefined's. For more information, see this answer.Array.apply(null, Array(5))
can also be written asArray.apply(null, {length: 5})
. There really is not much difference, but the latter is unambiguously clear that the intent is to create an array oflength 5
, and not an array containing5