I don't know how you are expecting array.remove(int) to behave. There are three possibilities I can think of that you might be wanting.

To remove an element of an array at an index i:

array.splice(i, 1);

If you want to remove every element with value number from the array:

for(var i = array.length - 1; i >= 0; i--) {
    if(array[i] === number) {
       array.splice(i, 1);
    }
}

If you just want to make the element at index i no longer exist, but you don't want the indexes of the other elements to change:

delete array[i];

Depends on whether you want to keep an empty spot or not.

If you do want an empty slot, delete is fine:

delete array[ index ];

If you don't, you should use the splice method:

array.splice( index, 1 );

And if you need the value of that item, you can just store the returned array's element:

var value = array.splice( index, 1 )[0];

In case you want to do it in some order, you can use array.pop() for the last one or array.shift() for the first one (and both return the value of the item too).

And if you don't know the index of the item, you can use array.indexOf( item ) to get it (in a if() to get one item or in a while() to get all of them). array.indexOf( item ) returns either the index or -1 if not found. 

A friend was having issues in Internet Explorer 8, and showed me what he did. I told him it was wrong, and he told me he got the answer here. The current top answer will not work in all browsers (Internet Explorer 8 for example), and it will only remove the first occurrence of the item.

Remove ALL instances from an array

  function remove(arr, item) {
      for(var i = arr.length; i--;) {
          if(arr[i] === item) {
              arr.splice(i, 1);
          }
      }
  }

It loops through the array backwards (since indices and length will change as items are removed) and removes the item if it's found. It works in all browsers.

Edited on 2016 october

• Do it simple, intuitive and explicit (https://en.wikipedia.org/wiki/Occam%27s_razor)
• Do it immutable (original array stay unchanged)
• Do it with standard JS functions, if your browser don't support them - use polyfill

In this code example I use "array.filter(...)" function to remove unwanted items from array, this function doesn't change the original array and creates a new one. If your browser don't support this function (e.g. IE before version 9, or Firefox before version 1.5), consider using the filter polyfill from Mozilla.

ECMAScript 5 code

var value = 3;

var arr = [1, 2, 3, 4, 5, 3];

arr = arr.filter(function(item) { 
    return item !== value;
});

console.log(arr); 
// [ 1, 2, 4, 5 ]

ECMAScript 6 code

let value = 3;

let arr = [1, 2, 3, 4, 5, 3];

arr = arr.filter(item => item !== value);

console.log(arr); 
// [ 1, 2, 4, 5 ]

An additional advantage of this method is that you can remove multiple items

['a', 'b', 'c', 'd'].filter((_, i) => i != 2 && i != 3) // makes ['a', 'b']

Whereas the delete/remove mutators update the array indexes after every delete/remove and, thus, invalidate the index array of elements that you want to remove.

The easiest way:

var index = array.indexOf(item);
if(index != -1)
    array.splice( index, 1 );

Be careful when you use delete for an array. It is good for deleting attributes of objects but not so good for arrays. It is better to use splice for arrays.

Keep in mind that when you use delete for an array you could get wrong results for anArray.length. In other words, delete would remove the element but wouldn't update the value of length property.

You can also expect to have holes in index numbers after using delete, e.g. you could end up with having indexes 1,3,4,8,9,11 and length as it was before using delete. In that case, all indexed for loops would crash, since indexes are no longer sequential.

If you are forced to use delete for some reason, then you should use for each loops when you need to loop through arrays.

Array.prototype.remByVal = function(val) {
    for (var i = 0; i < this.length; i++) {
        if (this[i] === val) {
            this.splice(i, 1);
            i--;
        }
    }
    return this;
}
//Call like
[1, 2, 3, 4].remByVal(3);

There is no need to use indexOf or splice. However, it performs better if you only want to remove one occurrence of an element.

Find and move (move):

function move(arr, val) {
  var j = 0;
  for (var i = 0, l = arr.length; i < l; i++) {
    if (arr[i] !== val) {
      arr[j++] = arr[i];
    }
  }
  arr.length = j;
}

Use indexOf and splice (indexof):

function indexof(arr, val) {
  var i;
  while ((i = arr.indexOf(val)) != -1) {
    arr.splice(i, 1);
  }
}

Use only splice (splice):

function splice(arr, val) {
  for (var i = arr.length; i--;) {
    if (arr[i] === val) {
      arr.splice(i, 1);
    }
  }
}

Run-times on nodejs for array with 1000 elements (average over 10000 runs):

indexof is approximately 10x slower than move. Even if improved by removing the call to indexOf in splice it performs much worse than move.

Remove all occurrences:
    move 0.0048 ms
    indexof 0.0463 ms
    splice 0.0359 ms

Remove first occurrence:
    move_one 0.0041 ms
    indexof_one 0.0021 ms

John Resig posted a good implementation:

// Array Remove - By John Resig (MIT Licensed)
Array.prototype.remove = function(from, to) {
  var rest = this.slice((to || from) + 1 || this.length);
  this.length = from < 0 ? this.length + from : from;
  return this.push.apply(this, rest);
};

If you don’t want to extend a global object, you can do something like the following, instead:

// Array Remove - By John Resig (MIT Licensed)
Array.remove = function(array, from, to) {
    var rest = array.slice((to || from) + 1 || array.length);
    array.length = from < 0 ? array.length + from : from;
    return array.push.apply(array, rest);
};

But the main reason I am posting this is to warn users against the alternative implementation suggested in the comments on that page (Dec 14, 2007):

Array.prototype.remove = function(from, to){
  this.splice(from, (to=[0,from||1,++to-from][arguments.length])<0?this.length+to:to);
  return this.length;
};

It seems to work well at first, but through a painful process I discovered it fails when trying to remove the second to last element in an array. For example, if you have a 10-element array and you try to remove the 9th element with this:

myArray.remove(8);

You end up with an 8-element array. Don't know why but I confirmed John's original implementation doesn't have this problem.

You can do it easily with filter method:

function remove(arrOriginal, elementToRemove){
    return arrOriginal.filter(function(el){return el !== elementToRemove});
}
console.log( remove([1, 2, 1, 0, 3, 1, 4], 1) );

This removes all elements from the array and also works faster then combination of slice and indexOf

Underscore.js can be used to solve issues with multiple browsers. It uses in-build browser methods if present. If they are absent like in the case of older IE it uses its own custom methods.

Simple example to remove elements from array (from the website) -

_.without([1, 2, 1, 0, 3, 1, 4], 0, 1); // => [2, 3, 4]

Too old to reply, but may it help someone, by providing a predicate instead of a value.

NOTE: it will update the passed array, and return affected rows

Usage

var removed = helper.removeOne(arr, row => row.id === 5 );

var removed = helper.remove(arr, row => row.name.startsWith('BMW'));

Definition

var helper = {

    // Remove and return the first occurrence     

    removeOne: function(array, predicate) {
        for (var i = 0; i < array.length; i++) {
            if (predicate(array[i])) {
                return array.splice(i, 1);
            }
        }
    },

    // Remove and return all occurrences  

    remove: function(array, predicate) {
        var removed = [];

        for (var i = 0; i < array.length;) {

            if (predicate(array[i])) {
                removed.push(array.splice(i, 1));
                continue;
            }

            i++;                
        }

        return removed;
    }
};

If you want a new array with the deleted positions removed, you can always delete the specific element and filter out the array. It might need an extension of the array object for browsers that don't implement the filter method but in the long term its easier since all you do is this:

var my_array = [1,2,3,4,5,6];
delete my_array[4];
console.log(my_array.filter(function(a){return typeof a !== 'undefined';}));

Should display [1, 2, 3, 4, 6]

Check out this code. It works in every major browser.

remove_item = function (arr, value) {
    var b = '';
    for (b in arr) {
        if (arr[b] === value) {
            arr.splice(b, 1);
            break;
        }
    }
    return arr;
}

Call this function

remove_item(array,value);

I'm pretty new to JavaScript and needed this functionality. I merely wrote this:

function removeFromArray(array, item, index) {
  while((index = array.indexOf(item)) > -1) {
    array.splice(index, 1);
  }
}

Then when I want to use it:

//Set-up some dummy data
var dummyObj = {name:"meow"};
var dummyArray = [dummyObj, "item1", "item1", "item2"];

//Remove the dummy data
removeFromArray(dummyArray, dummyObj);
removeFromArray(dummyArray, "item2");

Output - As expected. ["item1", "item1"]

You may have different needs than I, so you can easily modify it to suit them. I hope this helps someone.

You can use lodash _.pull (mutate array), _.pullAt (mutate array) or _.without (does't mutate array),

var array1 = ['a', 'b', 'c', 'd']
_.pull(array1, 'c')
console.log(array1) // ['a', 'b', 'd']

var array2 = ['e', 'f', 'g', 'h']
_.pullAt(array2, 0)
console.log(array2) // ['f', 'g', 'h']

var array3 = ['i', 'j', 'k', 'l']
var newArray = _.without(array3, 'i') // ['j', 'k', 'l']
console.log(array3) // ['i', 'j', 'k', 'l']

If you have complex objects in the array you can use filters? In situations where $.inArray or array.splice is not as easy to use. Especially if the objects are perhaps shallow in the array.

E.g. if you have an object with an Id field and you want the object removed from an array:

this.array = this.array.filter(function(element, i) {
    return element.id !== idToRemove;
});

A more modern, ECMAScript 2015 (formerly known as Harmony or ES 6) approach. Given:

const items = [1, 2, 3, 4];
const index = 2;

Then:

items.filter((x, i) => i !== index);

Yielding:

[1, 2, 4]

You can use Babel and a polyfill service to ensure this is well supported across browsers.

This gist here will solve your problem, and also deletes all occurrences of the argument instead of just 1 (or a specified value).

Array.prototype.destroy = function(obj){
    // Return null if no objects were found and removed
    var destroyed = null;

    for(var i = 0; i < this.length; i++){

        // Use while-loop to find adjacent equal objects
        while(this[i] === obj){

            // Remove this[i] and store it within destroyed
            destroyed = this.splice(i, 1)[0];
        }
    }

    return destroyed;
}

Usage:

var x = [1, 2, 3, 3, true, false, undefined, false];

x.destroy(3);         // => 3
x.destroy(false);     // => false
x;                    // => [1, 2, true, undefined]

x.destroy(true);      // => true
x.destroy(undefined); // => undefined
x;                    // => [1, 2]

x.destroy(3);         // => null
x;                    // => [1, 2]

Use jQuery's InArray:

A = [1, 2, 3, 4, 5, 6];
A.splice($.inArray(3, A), 1);
//It will return A=[1, 2, 4, 5, 6]`   

Note: inArray will return -1, if the element was not found.

I know there are a lot of answers already, but many of them seem to over complicate the problem. Here is a simple, recursive way of removing all instances of a key - calls self until index isn't found. Yes, it only works in browsers with indexOf, but it's simple and can be easily polyfilled.

Stand-alone function

function removeAll(array, key){
    var index = array.indexOf(key);

    if(index === -1) return;

    array.splice(index, 1);
    removeAll(array,key);
}

Prototype method

Array.prototype.removeAll = function(key){
    var index = this.indexOf(key);

    if(index === -1) return;

    this.splice(index, 1);
    this.removeAll(key);
}

I also ran in the situation where I had to remove an element from Array. .indexOf was not working in IE* so sharing my working jQuery.inArray() solution.

var index = jQuery.inArray(val,arr);
if (index > -1) {
    arr.splice(index, 1);
    //console.log(arr);
}

You can do a backward loop to make sure not to screw up the indexes, if there are multiple instances of the element.

var myElement = "chocolate";
var myArray = ['chocolate', 'poptart', 'poptart', 'poptart', 'chocolate', 'poptart', 'poptart', 'chocolate'];

/* Important code */
for (var i = myArray.length - 1; i >= 0; i--) {
    if (myArray[i] == myElement) myArray.splice(i, 1);
}

Live Demo

  Array.prototype.removeItem = function(a) {
            for (i = 0; i < this.length; i++) {
                if (this[i] == a) {
                    for (i2 = i; i2 < this.length - 1; i2++) {
                        this[i2] = this[i2 + 1];
                    }
                    this.length = this.length - 1
                    return;
                }
            }
        }

    var recentMovies = ['Iron Man', 'Batman', 'Superman', 'Spiderman'];
    recentMovies.removeItem('Superman');

In CoffeeScript:

my_array.splice(idx, 1) for ele, idx in my_array when ele is this_value

I think many of the JavaScript instructions are not well thought out for functional programming. Splice returns the deleted element where most of the time you need the reduced array. This is bad.

Imagine you are doing a recursive call and have to pass an array with one less item, probably without the current indexed item. Or imagine you are doing another recursive call and has to pass an array with an element pushed.

In neither of these cases you can do myRecursiveFunction(myArr.push(c)) or myRecursiveFunction(myArr.splice(i,1)). The first idiot will in fact pass the length of the array and the second idiot will pass the deleted element as a parameter.

So what I do in fact... For deleting an array element and passing the resulting to a function as a parameter at the same time I do as follows

myRecursiveFunction(myArr.slice(0,i).concat(a.slice(i+1)))

When it comes to push that's more silly... I do like,

myRecursiveFunction((myArr.push(c),myArr))

I believe in a proper functional language a method mutating the object it's called upon must return a reference to the very object as a result.

ES6 & without mutation: (October 2016)

const removeByIndex = (list, index) =>
  [
    ...list.slice(0, index),
    ...list.slice(index + 1)
  ];

Then :

removeByIndex([33,22,11,44],1) //=> [33,11,44]

Create new array:

var my_array = new Array();

Add elements to this array:

my_array.push("element1");

The function indexOf (Returns index or -1 when not found) :

var indexOf = function(needle) 
{
    if(typeof Array.prototype.indexOf === 'function') // newer browsers
    {
        indexOf = Array.prototype.indexOf;
    } 
    else // older browsers
    {
        indexOf = function(needle) 
        {
            var index = -1;

            for(var i = 0; i < this.length; i++) 
            {
                if(this[i] === needle) 
                {
                    index = i;
                    break;
                }
            }
            return index;
        };
    }

    return indexOf.call(this, needle);
};

Check index of this element (tested with firefox and IE8+):

var index = indexOf.call(my_array, "element1");

Remove 1 element located at index from the array

my_array.splice(index, 1);

You can use ES6.

var array=['1','2','3','4','5','6'] var index = array.filter((value)=>value!='3');

Output :

["1", "2", "4", "5", "6"]