This is also referred to as "deep copying", which I've found some articles on. Closest seems to be this one but it's for jQuery - I'm trying to do this without a library.
I've also seen, in two places, that it's possible to do something like:
arr2 = JSON.decode(JSON.encode(arr1));
But that's apparently inefficient. It's also possible to loop and copy each value individually, and recurs through all the arrays. That seems tiring and inefficient as well.
So what's the most efficient, non-library way to copy a JavaScript multi-dimensional array [[a],[b],[c]]? I am completely happy with a "non-IE" method if necessary.
Thanks!
Since it sounds like you're dealing with an Array of Arrays to some unknown level of depth, but you only need to deal with them at one level deep at any given time, then it's going to be simple and fast to use .slice().
var newArray = [];
for (var i = 0; i < currentArray.length; i++)
newArray[i] = currentArray[i].slice();
Or using .map() instead of the for loop:
var newArray = currentArray.map(function(arr) {
return arr.slice();
});
So this iterates the current Array, and builds a new Array of shallow copies of the nested Arrays. Then when you go to the next level of depth, you'd do the same thing.
Of course if there's a mixture of Arrays and other data, you'll want to test what it is before you slice.
I'm not sure how much better JSON.stringify and JSON.parse than encode and decode, but you could try:
JSON.parse(JSON.stringify(array));
Something else I found (although I'd modify it a little):
http://www.xenoveritas.org/blog/xeno/the-correct-way-to-clone-javascript-arrays
function deepCopy(obj) {
if (typeof obj == 'object') {
if (isArray(obj)) {
var l = obj.length;
var r = new Array(l);
for (var i = 0; i < l; i++) {
r[i] = deepCopy(obj[i]);
}
return r;
} else {
var r = {};
r.prototype = obj.prototype;
for (var k in obj) {
r[k] = deepCopy(obj[k]);
}
return r;
}
}
return obj;
}
As you asked for performance, I guess you also would go with a non-generic solution. To copy a multi-dimensional array with a known number of levels, you should go with the easiest solution, some nested for-loops. For your two-dimensional array, it simply would look like this:
var len = arr.length,
copy = new Array(len); // boost in Safari
for (var i=0; i<len; ++i)
copy[i] = arr[i].slice(0);
To extend to higher-dimensional arrays, either use recursion or nested for loops!
The native slice method is more efficient than a custom for loop, yet it does not create deep copies, so we can use it only at the lowest level.
Any recursive algorithm that doesn't visit the same node twice will be about as efficient as you get with javascript (at least in a browser) - in certain situations in other languages you might get away with copying chucks of memory, but javascript obviously doesn't have that ability.
I'd suggest finding someone who's already done it and using their implementation to make sure you get it right - it only needs to be defined once.
Related
I know the classic way of looping through an array arr is:
for(var i=0 ; i<arr.length ; i++) {
// code
}
But someone recently showed me a different way of implementing the condition inside that loop, like this:
for(var i=0 ; arr[i] !== undefined ; i++) {
I think this solution is interesting because this is exactly what you need when you loop through an array: you don't want to get undefineds when you try to access an undefined index.
I realize that if you count the characters it looks longer, and also that you might have some problems with arrays like this: ["Hello", , "World"], but apart from that - is there anything else I'm missing here? Why shouldn't we be using this technique instead?
Why shouldn't we be using this technique instead?
It doesn't work on sparse arrays (as you mentioned)
It doesn't work on arrays that contain undefined values
It's not as easily optimised (assuming the .length stays constant during the loop)
(In the old days, the undefined identifier could be overwritten, you'd need to use typeof)
Of cource, wheter it "works" for you depends on the use case, and sometimes you might want to use it. Most times, you simply don't.
And even if both ways would work in your case, it's better practise to use the standard approach (i<arr.length) as there is lower mental overhead. Everyone recognises that pattern and knows what it does, while with arr[i]!==undefined one would need to think about why the uncommon approach was chosen.
Sometimes arrays have empty values and your way of iteration will fail.
var arr = [];
arr[5] = 5;
for (var i = 0; arr[i] !== undefined; ++i) {
console.log(arr[i]);
}
console.log('done');
If you want to iterate real array values and skip undefined's, i suggest you to filter the array first and do iteration after. So your code will be more understandable. Example:
var arr = [];
arr[5] = 5;
arr.filter(Boolean).forEach(function (e) {
console.log(e);
});
console.log('done');
This question already has answers here:
Closed 10 years ago.
Possible Duplicate:
Functional approach to basic array construction
I am new to js. I just want to know which one is the right approach. Below I have posted my code.
var doubles = (i*20 for (i in it));
var doubles ={i*20 for (i in it)};
Thanks
You should use ordinary for loops when iterating over arrays. for...in is used for iterating over object properties.
So, the correct way is:
for (var i = 0; i < array.length; ++i) {
// do something
}
To iterate over object properties:
for (var prop in obj) {
// do something with prop
}
Assuming it is an array, you can use .map():
var doubles = it.map(function(i){ return i*20; });
Also you might want to have a look at how to write List/Array comprehensions in JavaScript
Assuming you want to use Mozilla's Generator expressions (where it is an existing Iterator), you need to use square brackets:
var twentyfolds = [i*20 for (i in it)];
For future reference, ECMAScript 6 (aka Harmony) will most likely introduce a new sane way of iterating over objects (arrays included):
for(var x of array) {
// do something with x
}
It will also introduce array comprehensions and generator expressions to the core language:
var arr = [1, 2, 3];
console.log([i*20 for (i of arr)])
Both of these options are syntax errors. If 'it' is a list, then you can iterate through it with a for loop or the forEach method. In your case however, it looks like you are really looking for the map method.
var doubles = it.map(function (i) { return i * 20; });
Are there any performance difference between
var a = [10,20,30,40];// Assume we have thousands of values here
// Approach 1
var i, len = a.length;
for(i=0;i<len;i++){
alert(i);
alert(a[i]);
}
// Approach 2
for( i in a ){
alert(i);
alert(a[i]);
}
Use for (var i = 0, len = a.length; i < len; i++) because it's way faster and it's the correct way or iterating the items in an array.
First: It's not correct to iterate arrays with for (i in a) because that iteration will include enumerable properties in addition to array elements. If any methods or properties have been added to the array, they will be part of the iteration when using for (i in a) which is never what you want when trying to traverse the elements of the array.
Second: The correct option is a lot faster (9-20x faster). See this jsPerf test which shows the for (var i = 0; i < len; i++) option to be about 9x faster in Chrome and even more of a speed difference in Firefox: http://jsperf.com/for-loop-comparison2.
As an example of the problems that can occur when using for (var i in a), when I use that when the mootools library is included in the project, I get all these values for i:
0
1
2
3
$family
$constructor
each
clone
clean
invoke
associate
link
contains
append
getLast
getRandom
include
combine
erase
empty
flatten
pick
hexToRgb
rgbToHex
which appears to be a bunch of methods that mootools has added to the array object.
I don't know across browsers, but in my test with Firefox there is. for (i=0; etc...) is much faster. Here is a jsfiddle example that shows the difference. http://jsfiddle.net/pseudosavant/VyRH3/
Add to that the problems that you can encounter with (for i in etc) when the Array object has been prototype (perhaps in a library), you should always use for (i=0; etc...) for looping over arrays.
(for i in etc) should only ever be used on objects.
I was looking for a way to add max/min functions to JavaScript's Array class, which seemed to be a solved problem: JavaScript: min & max Array values?. However, when I tried using that, I started getting errors from my code. It turns out that this approach doesn't work with loops.
for(i in myArray) { console.log(i) }
prints out
1
2
3
max
min
Is there another approach I can use?
The accepted solution solves your immediate problem, but extending core objects is generally a bad idea. What happens when you include a library later that uses for..in? Or when you forget months later and use the wrong approach in a different section of your code?
Another option is to wrap and extend. Create a new type that uses an instance of Array as its prototype:
function ArrayThing() {}
ArrayThing.prototype = new Array();
Now you've got an object that you can extend without affecting Array:
ArrayThing.prototype.max = function() {
return Math.max.apply(null, [].slice.call(this, 0))
}
ArrayThing.prototype.min = function() {
return Math.min.apply(null, [].slice.call(this, 0))
}
var list = new ArrayThing();
// standard array methods still work
list.push(5);
list.push(22);
list.push(0);
list.push(-14);
list.length // => 4
// as do your new custom methods
list.max() // => 22
list.min() // => -14
This won't work in every situation, but unless you're sure you really, really need an Array, this is a useful alternative.
The for...in loop is used for looping through properties of an object. If you want to get the values from your array, you can do this:
for (var i = 0; i < myArray.length; i++)
{
console.log(myArray[i])
}
for in is a common Javascript trap. Instead of behaving like a foreach from other languages, it actualy enumerates all properties in a given object.
Since Javascript Arrays happen to have a property for each index using for in works sometimes, but as you have seen, it also enumerates any other properties you add. Another issue is that the for in is not guaranteed to go through the properties in any particular order, so your results can vary depending on which browser/runtime you use.
It is safer, then, to just use a boring for loop instead. There are many for loop idioms in Javascript, so I will list some:
Regular for loop:
for(i=0; i<arr.length; i++){
Regular loop, caching the length:
for(i=0, n=arr.length; i<n; i++){
Loops over arrays of objects/NodeLists:
for(i=0; obj=arr[i]; i++){ //this works as long as the array has no falsy values
foo(obj)
You need to apply a check using hasOwnProperty.
However this needs to applied wherever you are looping.
i.e:
for(i in myArray)
{
if(arr.hasOwnProperty(i))
{
console.log(i);
}
}
There's a more modern (IE9+) way to do this now:
var g = [];
Object.defineProperty(g, "log", {
enumerable: false,
configurable: false,
writable: false,
value: function(){ console.log(this); }
});
g.push(5);
g.push(9);
var t;
for (t in g){
console.log(g[t]);
}
prints 5 and 9 but does not list the "log" function
g.log() -> echos [5,9]
The key to this working is being able to flag a property as "enumerable: false" with marks the property as something that shouldn't be iterated over.
More on Object.defineProperty here:
https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Object/defineProperty
Is there a more efficient way to convert an HTMLCollection to an Array, other than iterating through the contents of said collection and manually pushing each item into an array?
var arr = Array.prototype.slice.call( htmlCollection )
will have the same effect using "native" code.
Edit
Since this gets a lot of views, note (per #oriol's comment) that the following more concise expression is effectively equivalent:
var arr = [].slice.call(htmlCollection);
But note per #JussiR's comment, that unlike the "verbose" form, it does create an empty, unused, and indeed unusable array instance in the process. What compilers do about this is outside the programmer's ken.
Edit
Since ECMAScript 2015 (ES 6) there is also Array.from:
var arr = Array.from(htmlCollection);
Edit
ECMAScript 2015 also provides the spread operator, which is functionally equivalent to Array.from (although note that Array.from supports a mapping function as the second argument).
var arr = [...htmlCollection];
I've confirmed that both of the above work on NodeList.
A performance comparison for the mentioned methods: http://jsben.ch/h2IFA
not sure if this is the most efficient, but a concise ES6 syntax might be:
let arry = [...htmlCollection]
Edit: Another one, from Chris_F comment:
let arry = Array.from(htmlCollection)
I saw a more concise method of getting Array.prototype methods in general that works just as well. Converting an HTMLCollection object into an Array object is demonstrated below:
[].slice.call( yourHTMLCollectionObject );
And, as mentioned in the comments, for old browsers such as IE7 and earlier, you simply have to use a compatibility function, like:
function toArray(x) {
for(var i = 0, a = []; i < x.length; i++)
a.push(x[i]);
return a
}
I know this is an old question, but I felt the accepted answer was a little incomplete; so I thought I'd throw this out there FWIW.
For a cross browser implementation I'd sugguest you look at prototype.js $A function
copyed from 1.6.1:
function $A(iterable) {
if (!iterable) return [];
if ('toArray' in Object(iterable)) return iterable.toArray();
var length = iterable.length || 0, results = new Array(length);
while (length--) results[length] = iterable[length];
return results;
}
It doesn't use Array.prototype.slice probably because it isn't available on every browser. I'm afraid the performance is pretty bad as there a the fall back is a javascript loop over the iterable.
This works in all browsers including earlier IE versions.
var arr = [];
[].push.apply(arr, htmlCollection);
Since jsperf is still down at the moment, here is a jsfiddle that compares the performance of different methods. https://jsfiddle.net/qw9qf48j/
To convert array-like to array in efficient way we can make use of the jQuery makeArray :
makeArray: Convert an array-like object into a true JavaScript array.
Usage:
var domArray = jQuery.makeArray(htmlCollection);
A little extra:
If you do not want to keep reference to the array object (most of the time HTMLCollections are dynamically changes so its better to copy them into another array, This example pay close attention to performance:
var domDataLength = domData.length //Better performance, no need to calculate every iteration the domArray length
var resultArray = new Array(domDataLength) // Since we know the length its improves the performance to declare the result array from the beginning.
for (var i = 0 ; i < domDataLength ; i++) {
resultArray[i] = domArray[i]; //Since we already declared the resultArray we can not make use of the more expensive push method.
}
What is array-like?
HTMLCollection is an "array-like" object, the array-like objects are similar to array's object but missing a lot of its functionally definition:
Array-like objects look like arrays. They have various numbered
elements and a length property. But that’s where the similarity stops.
Array-like objects do not have any of Array’s functions, and for-in
loops don’t even work!
This is my personal solution, based on the information here (this thread):
var Divs = new Array();
var Elemns = document.getElementsByClassName("divisao");
try {
Divs = Elemns.prototype.slice.call(Elemns);
} catch(e) {
Divs = $A(Elemns);
}
Where $A was described by Gareth Davis in his post:
function $A(iterable) {
if (!iterable) return [];
if ('toArray' in Object(iterable)) return iterable.toArray();
var length = iterable.length || 0, results = new Array(length);
while (length--) results[length] = iterable[length];
return results;
}
If browser supports the best way, ok, otherwise will use the cross browser.
I suppose that calling Array.prototype functions on instances of HTMLCollection is a much better option than converting collections to arrays (e.g.,[...collection] or Array.from(collection)), because in the latter case a collection is unnecessarily implicitly iterated and a new array object is created, and this eats up additional resources. Array.prototype iterating functions can be safely called upon objects with consecutive numeric keys starting from [0] and a length property with a valid number value of such keys' quantity (including, e.g., instances of HTMLCollection and FileList), so it's a reliable way. Also, if there is a frequent need in such operations, an empty array [] can be used for quick access to Array.prototype functions; or a shortcut for Array.prototype can be created instead. A runnable example:
const _ = Array.prototype;
const collection = document.getElementById('ol').children;
alert(_.reduce.call(collection, (acc, { textContent }, i) => {
return acc += `${i+1}) ${textContent}` + '\n';
}, ''));
<ol id="ol">
<li>foo</li>
<li>bar</li>
<li>bat</li>
<li>baz</li>
</ol>
Sometimes, Even You have written code the correct way, But still it doesn't work properly.
var allbuttons = document.getElementsByTagName("button");
console.log(allbuttons);
var copyAllButtons = [];
for (let i = 0; i < allbuttons.length; i++) {
copyAllButtons.push(allbuttons[i]);
}
console.log(copyAllButtons);
you get empty array.
Like, This
HTMLCollection []
[]
Console_javascript
For Solving this problem, You have to add link of javascript file after body tag in html file.
<script src="./script.js"></script>
As you can see below,
html_file
Final Output
HTMLCollection(6) [button.btn.btn-dark.click-me, button.btn.btn-dark.reset, button#b, button#b, button#b, button#b, b: button#b]
(6) [button.btn.btn-dark.click-me, button.btn.btn-dark.reset, button#b, button#b, button#b, button#b]