We Keep Coding

JS reference multi-layered dynamic object string

Suppose I have a string that references a deep Javascript object, such as:
var string = 'response.basicInfo.gender';
I want to build a function that safely checks if that object exists, by splitting the string by .s and building the object, checking each level as it goes, and then dealing with the value of the object, if it actually exists.
var parseHelper = function(response, items) {
for (item in items) {
var parts = String(item).split('.');
for (var i = 0; i < parts.length; ++i) {
// ... etc. build the string
}
}
}
parseHelper(response, {
'basicInfo.gender': function(val){
return (val == 'M') ? 'Male' : (val == 'F') ? 'Female' : val;
},
})
While the above function is incomplete, let's suppose we use it to build the string and check if each exists:
// so let's suppose we build the string:
var builtString = "response['basicInfo']";
// Now we want to check if it exists
if (response['basicInfo']) {
// And if we are fine that it exists, go on to the next item
var builtString = "response['basicInfo']['gender']";
// etc.
}
I don't have a problem building that function, I just don't know how to evaluate a string like "response['basicInfo']['gender']" and turn it into an actual reference to the object. My only guess would be eval(), but eval is evil...
Update
I know you can reference a global object by going window['blah'], but this response object I want to reference is not in the global scope, so do I use this? And even if I can do this, how do I reference it with multiple layers?

Plus 1 to Bergi, who linked to a page with six links, one of which had an answer I adapted to solve the problem:
Convert JavaScript string in dot notation into an object reference
Here's the full solution.
// We want to run a parse function to convert
// the object response.basicInfo.gender (which is 'M')
// to 'Male', etc.
// Sets the value of a string representing a deep object.
setDeep: function(root, path, value) {
var parts = path.split('.'), obj = root;
for (var i = 0; i < parts.length - 1; ++i) {
obj = obj[parts[i]] || { };
}
obj[parts[parts.length - 1]] = value;
return obj;
},
// Gets the value of a string representing a deep object.
getDeep: function(root, path) {
var parts = path.split('.'), obj = root, target;
for (var i = 0; i < parts.length; ++i) {
target = obj[parts[i]];
if (typeof target == "undefined") return void 0;
obj = target;
}
return obj;
},
// Loops through multiple string representations
// of deep objects and runs the values through
// assigned parsing functions for each of them,
// returning the root object.
parseHelper: function(obj, items) {
for (item in items) {
var val = getDeep(obj, item);
var func = items[item];
if (val !== undefined) {
val = func(val);
}
setDeep(obj, item, val);
}
return obj;
},
// response.basicInfo.gender = 'M';
// response.foo.bar = true;
response = parseHelper(response, {
'basicInfo.gender': function(val){
return (val == 'M') ? 'Male' : (val == 'F') ? 'Female' : val;
},
'foo.bar': function(val) {
return (val) ? false : true;
},
});
// response.basicInfo.gender = 'Male';
// response.foo.bar = false;

Compare arrays as (multi-) sets

I'm looking for an efficient way to find out whether two arrays contain same amounts of equal elements (in the == sense), in any order:
foo = {/*some object*/}
bar = {/*some other object*/}
a = [1,2,foo,2,bar,2]
b = [bar,2,2,2,foo,1]
sameElements(a, b) --> true
PS. Note that pretty much every solution in the thread uses === and not == for comparison. This is fine for my needs though.

Update 5
I posted a new answer with a different approach.
Update
I extended the code to have the possibility of either checking by reference or equality
just pass true as second parameter to do a reference check.
Also I added the example to Brunos JSPerf
It runs at about 11 ops/s doing a reference check
I will comment the code as soon(!) as I get some spare time to explain it a bit more, but at the moment don't have the time for that, sry. Done
Update 2.
Like Bruno pointed out in the comments sameElements([NaN],[NaN]) yields false
In my opinion this is the correct behaviour as NaN is ambigious and should always lead to a false result,at least when comparing NaN.equals(NaN). But he had quite a good point.
Whether
[1,2,foo,bar,NaN,3] should be equal to [1,3,foo,NaN,bar,2] or not.
Ok.. honestly I'm a bit torn whether it should or not, so i added two flags.
Number.prototype.equal.NaN
If true
NaN.equals(NaN) //true
Array.prototype.equal.NaN
If true
[NaN].equals([NaN],true) //true
note this is only for reference checks. As a deep check would invoke Number.prototype.equals anyway
Update 3:
Dang i totally missed 2 lines in the sort function.
Added
r[0] = a._srt; //DANG i totally missed this line
r[1] = b._srt; //And this.
Line 105 in the Fiddle
Which is kind of important as it determines the consistent order of the Elements.
Update 4
I tried to optimize the sort function a bit, and managed to get it up to about 20 ops/s.
Below is the updated code, as well as the updated fiddle =)
Also i chose to mark the objects outside the sort function, it doesn't seem to make a performance difference anymore, and its more readable
Here is an approach using Object.defineProperty to add equals functions to
Array,Object,Number,String,Boolean's prototype to avoid typechecking in one function for
performance reasons. As we can recursively call .equals on any element.
But of course checking Objects for equality may cause performance issues in big Objects.
So if anyone feels unpleasant manipulating native prototypes, just do a type check and put it into one function
Object.defineProperty(Boolean.prototype, "equals", {
enumerable: false,
configurable: true,
value: function (c) {
return this == c; //For booleans simply return the equality
}
});
Object.defineProperty(Number.prototype, "equals", {
enumerable: false,
configurable: true,
value: function (c) {
if (Number.prototype.equals.NaN == true && isNaN(this) && c != c) return true; //let NaN equals NaN if flag set
return this == c; // else do a normal compare
}
});
Number.prototype.equals.NaN = false; //Set to true to return true for NaN == NaN
Object.defineProperty(String.prototype, "equals", {
enumerable: false,
configurable: true,
value: Boolean.prototype.equals //the same (now we covered the primitives)
});
Object.defineProperty(Object.prototype, "equals", {
enumerable: false,
configurable: true,
value: function (c, reference) {
if (true === reference) //If its a check by reference
return this === c; //return the result of comparing the reference
if (typeof this != typeof c) {
return false; //if the types don't match (Object equals primitive) immediately return
}
var d = [Object.keys(this), Object.keys(c)],//create an array with the keys of the objects, which get compared
f = d[0].length; //store length of keys of the first obj (we need it later)
if (f !== d[1].length) {//If the Objects differ in the length of their keys
return false; //immediately return
}
for (var e = 0; e < f; e++) { //iterate over the keys of the first object
if (d[0][e] != d[1][e] || !this[d[0][e]].equals(c[d[1][e]])) {
return false; //if either the key name does not match or the value does not match, return false. a call of .equal on 2 primitives simply compares them as e.g Number.prototype.equal gets called
}
}
return true; //everything is equal, return true
}
});
Object.defineProperty(Array.prototype, "equals", {
enumerable: false,
configurable: true,
value: function (c,reference) {
var d = this.length;
if (d != c.length) {
return false;
}
var f = Array.prototype.equals.sort(this.concat());
c = Array.prototype.equals.sort(c.concat(),f)
if (reference){
for (var e = 0; e < d; e++) {
if (f[e] != c[e] && !(Array.prototype.equals.NaN && f[e] != f[e] && c[e] != c[e])) {
return false;
}
}
} else {
for (var e = 0; e < d; e++) {
if (!f[e].equals(c[e])) {
return false;
}
}
}
return true;
}
});
Array.prototype.equals.NaN = false; //Set to true to allow [NaN].equals([NaN]) //true
Object.defineProperty(Array.prototype.equals,"sort",{
enumerable:false,
value:function sort (curr,prev) {
var weight = {
"[object Undefined]":6,
"[object Object]":5,
"[object Null]":4,
"[object String]":3,
"[object Number]":2,
"[object Boolean]":1
}
if (prev) { //mark the objects
for (var i = prev.length,j,t;i>0;i--) {
t = typeof (j = prev[i]);
if (j != null && t === "object") {
j._pos = i;
} else if (t !== "object" && t != "undefined" ) break;
}
}
curr.sort (sorter);
if (prev) {
for (var k = prev.length,l,t;k>0;k--) {
t = typeof (l = prev[k]);
if (t === "object" && l != null) {
delete l._pos;
} else if (t !== "object" && t != "undefined" ) break;
}
}
return curr;
function sorter (a,b) {
var tStr = Object.prototype.toString
var types = [tStr.call(a),tStr.call(b)]
var ret = [0,0];
if (types[0] === types[1] && types[0] === "[object Object]") {
if (prev) return a._pos - b._pos
else {
return a === b ? 0 : 1;
}
} else if (types [0] !== types [1]){
return weight[types[0]] - weight[types[1]]
}
return a>b?1:a<b?-1:0;
}
}
});
With this we can reduce the sameElements function to
function sameElements(c, d,referenceCheck) {
return c.equals(d,referenceCheck); //call .equals of Array.prototype.
}
Note. of course you could put all equal functions into the sameElements function, for the cost of the typechecking.
Now here are 3 examples: 1 with deep checking, 2 with reference checking.
var foo = {
a: 1,
obj: {
number: 2,
bool: true,
string: "asd"
},
arr: [1, 2, 3]
};
var bar = {
a: 1,
obj: {
number: 2,
bool: true,
string: "asd"
},
arr: [1, 2, 3]
};
var foobar = {
a: 1,
obj: {
number: 2,
bool: true,
string: "asd"
},
arr: [1, 2, 3, 4]
};
var a = [1, 2, foo, 2, bar, 2];
var b = [foo, 2, 2, 2, bar, 1];
var c = [bar, 2, 2, 2, bar, 1];
So these are the Arrays we compare. And the output is
Check a and b with references only.
console.log (sameElements ( a,b,true)) //true As they contain the same elements
Check b and c with references only
console.log (sameElements (b,c,true)) //false as c contains bar twice.
Check b and c deeply
console.log (sameElements (b,c,false)) //true as bar and foo are equal but not the same
Check for 2 Arrays containing NaN
Array.prototype.equals.NaN = true;
console.log(sameElements([NaN],[NaN],true)); //true.
Array.prototype.equals.NaN = false;
Demo on JSFiddle

You can implement the following algorithm:
If a and b do not have the same length:
Return false.
Otherwise:
Clone b,
For each item in a:
If the item exists in our clone of b:
Remove the item from our clone of b,
Otherwise:
Return false.
Return true.
With Javascript 1.6, you can use every() and indexOf() to write:
function sameElements(a, b)
{
if (a.length != b.length) {
return false;
}
var ourB = b.concat();
return a.every(function(item) {
var index = ourB.indexOf(item);
if (index < 0) {
return false;
} else {
ourB.splice(index, 1);
return true;
}
});
}
Note this implementation does not completely fulfill your requirements because indexOf() uses strict equality (===) internally. If you really want non-strict equality (==), you will have to write an inner loop instead.

Like this perhaps?
var foo = {}; var bar=[];
var a = [3,2,1,foo]; var b = [foo,1,2,3];
function comp(a,b)
{
// immediately discard if they are of different sizes
if (a.length != b.length) return false;
b = b.slice(0); // clone to keep original values after the function
a.forEach(function(e) {
var i;
if ((i = b.indexOf(e)) != -1)
b.splice(i, 1);
});
return !b.length;
}
comp(a,b);

UPDATE
As #Bergi and #thg435 point out my previous implementation was flawed so here is another implementation:
function sameElements(a, b) {
var objs = [];
// if length is not the same then must not be equal
if (a.length != b.length) return false;
// do an initial sort which will group types
a.sort();
b.sort();
for ( var i = 0; i < a.length; i++ ) {
var aIsPrimitive = isPrimitive(a[i]);
var bIsPrimitive = isPrimitive(b[i]);
// NaN will not equal itself
if( a[i] !== a[i] ) {
if( b[i] === b[i] ) {
return false;
}
}
else if (aIsPrimitive && bIsPrimitive) {
if( a[i] != b[i] ) return false;
}
// if not primitive increment the __count property
else if (!aIsPrimitive && !bIsPrimitive) {
incrementCountA(a[i]);
incrementCountB(b[i]);
// keep track on non-primitive objects
objs.push(i);
}
// if both types are not the same then this array
// contains different number of primitives
else {
return false;
}
}
var result = true;
for (var i = 0; i < objs.length; i++) {
var ind = objs[i];
// if __aCount and __bCount match then object exists same
// number of times in both arrays
if( a[ind].__aCount !== a[ind].__bCount ) result = false;
if( b[ind].__aCount !== b[ind].__bCount ) result = false;
// revert object to what it was
// before entering this function
delete a[ind].__aCount;
delete a[ind].__bCount;
delete b[ind].__aCount;
delete b[ind].__bCount;
}
return result;
}
// inspired by #Bergi's code
function isPrimitive(arg) {
return Object(arg) !== arg;
}
function incrementCountA(arg) {
if (arg.hasOwnProperty("__aCount")) {
arg.__aCount = arg.__aCount + 1;
} else {
Object.defineProperty(arg, "__aCount", {
enumerable: false,
value: 1,
writable: true,
configurable: true
});
}
}
function incrementCountB(arg) {
if (arg.hasOwnProperty("__bCount")) {
arg.__bCount = arg.__bCount + 1;
} else {
Object.defineProperty(arg, "__bCount", {
enumerable: false,
value: 1,
writable: true,
configurable: true
});
}
}
Then just call the function
sameElements( ["NaN"], [NaN] ); // false
// As "1" == 1 returns true
sameElements( [1],["1"] ); // true
sameElements( [1,2], [1,2,3] ); //false
The above implement actually defines a new property called "__count" that is used to keep track of non-primitive elements in both arrays. These are deleted before the function returns so as to leave the array elements as before.
Fiddle here
jsperf here.
The reason I changed the jsperf test case was that as #Bergi states the test arrays, especially the fact there were only 2 unique objects in the whole array is not representative of what we are testing for.
One other advantage of this implementation is that if you need to make it compatible with pre IE9 browsers instead of using the defineProperty to create a non-enumerable property you could just use a normal property.

Thanks everyone for sharing ideas! I've came up with the following
function sameElements(a, b) {
var hash = function(x) {
return typeof x + (typeof x == "object" ? a.indexOf(x) : x);
}
return a.map(hash).sort().join() == b.map(hash).sort().join();
}
This isn't the fastest solution, but IMO, most readable one so far.

i wasn't sure if "===" is ok, the question is a bit vauge...
if so, this is quite a bit faster and simpler than some other possible ways of doing it:
function isSame(a,b){
return a.length==b.length &&
a.filter(function(a){ return b.indexOf(a)!==-1 }).length == b.length;
}

Edit 2
1) Thanks to user2357112 for pointing out the Object.prototype.toString.call issue
this also showed, the reason it was that fast, that it didn't consider Arrays ...
I fixed the code,it should be working now :), unfortunately its now at about 59ops/s on chrome and 45ops/s on ff.
Fiddle and JSPerf is updated.
Edit
1)
I fixed the code, it supports mutliple variables referencing the same Object now.
A little bit slower than before, but still over 100ops/s on chrome.
2)
I tried using a bitmask instead of an array to keep multiple positions of the objects, but its nearly 15ops/s slow
3) As pointed ot in the comments i forgot to reset tmp after [[get]] is called fixed the code, the fiddle, and the perf.
So thanks to user2357112 with his Answer heres another approach using counting
var sameElements = (function () {
var f, of, objectFlagName;
of = objectFlagName = "__pos";
var tstr = function (o) {
var t = typeof o;
if (o === null)
t = "null";
return t
};
var types = {};
(function () {
var tmp = {};
Object.defineProperty(types, tstr(1), {
set: function (v) {
if (f)
tmp[v] = -~tmp[v];
else
tmp[v] = ~-tmp[v];
},
get: function () {
var ret = 1;
for (var k in tmp) {
ret &= !tmp[k];
}
tmp = {};
return ret;
}
});
})();
(function () {
var tmp = {};
Object.defineProperty(types, tstr(""), {
set: function (v) {
if (f) {
tmp[v] = -~tmp[v];
} else {
tmp[v] = ~-tmp[v];
}
},
get: function () {
var ret = 1;
for (var k in tmp) {
ret &= !tmp[k];
}
tmp = {};
return ret;
}
});
})();
(function () {
var tmp = [];
function add (v) {
tmp.push(v);
if (v[of]===undefined) {
v[of] = [tmp.length -1];
} else {
v[of].push(tmp.length -1)
}
}
Object.defineProperty(types, tstr({}), {
get: function () {
var ret = true;
for (var i = tmp.length - 1; i >= 0; i--) {
var c = tmp[i]
if (typeof c !== "undefined") {
ret = false
delete c[of]
}
}
tmp = [];
return ret;
},
set: function (v) {
var pos;
if (f) {
add (v);
} else if (!f && (pos = v[of]) !== void 0) {
tmp[pos.pop()] = undefined;
if (pos.length === 0)
delete v[of];
} else {
add (v);
}
}
});
}());
(function () {
var tmp = 0;
Object.defineProperty(types, tstr(undefined), {
get: function () {
var ret = !tmp;
tmp = 0;
return ret;
},
set: function () {
tmp += f ? 1 : -1;
}
});
})();
(function () {
var tmp = 0;
Object.defineProperty(types, tstr(null), {
get: function () {
var ret = !tmp;
tmp = 0;
return ret;
},
set: function () {
tmp += f ? 1 : -1;
}
});
})();
var tIt = [tstr(1), tstr(""), tstr({}), tstr(undefined), tstr(null)];
return function eq(a, b) {
f = true;
for (var i = a.length - 1; i >= 0; i--) {
var v = a[i];
types[tstr(v)] = v;
}
f = false;
for (var k = b.length - 1; k >= 0; k--) {
var w = b[k];
types[tstr(w)] = w;
}
var r = 1;
for (var l = 0, j; j = tIt[l]; l++) {
r &= types [j]
}
return !!r;
}
})()
Here is a JSFiddle and a JSPerf (it uses the same Arrays a and b as in the previous answers perf) with this code vs the Closure compiled
Heres the output. note: it doesn't support a deep comparison anymore, as is
var foo = {a:2}
var bar = {a:1};
var a = [1, 2, foo, 2, bar, 2];
var b = [foo, 2, 2, 2, bar, 1];
var c = [bar, 2, 2, 2, bar, 1];
console.log(sameElements([NaN],[NaN])); //true
console.log (sameElements ( a,b)) //true
console.log (sameElements (b,c)) //false

Using efficient lookup tables for the counts of the elements:
function sameElements(a) { // can compare any number of arrays
var map, maps = [], // counting booleans, numbers and strings
nulls = [], // counting undefined and null
nans = [], // counting nans
objs, counts, objects = [],
al = arguments.length;
// quick escapes:
if (al < 2)
return true;
var l0 = a.length;
if ([].slice.call(arguments).some(function(s) { return s.length != l0; }))
return false;
for (var i=0; i<al; i++) {
var multiset = arguments[i];
maps.push(map = {}); // better: Object.create(null);
objects.push({vals: objs=[], count: counts=[]});
nulls[i] = 0;
nans[i] = 0;
for (var j=0; j<l0; j++) {
var val = multiset[j];
if (val !== val)
nans[i]++;
else if (val === null)
nulls[i]++;
else if (Object(val) === val) { // non-primitive
var ind = objs.indexOf(val);
if (ind > -1)
counts[ind]++;
else
objs.push(val), counts.push(1);
} else { // booleans, strings and numbers do compare together
if (typeof val == "boolean")
val = +val;
if (val in map)
map[val]++;
else
map[val] = 1;
}
}
}
// testing if nulls and nans are the same everywhere
for (var i=1; i<al; i++)
if (nulls[i] != nulls[0] || nans[i] != nans[0])
return false;
// testing if primitives were the same everywhere
var map0 = maps[0];
for (var el in map0)
for (var i=1; i<al; i++) {
if (map0[el] !== maps[i][el])
return false;
delete maps[i][el];
}
for (var i=1; i<al; i++)
for (var el in maps[i])
return false;
// testing if objects were the same everywhere
var objs0 = objects[0].vals,
ol = objs0.length;
counts0 = objects[0].count;
for (var i=1; i<al; i++)
if (objects[i].count.length != ol)
return false;
for (var i=0; i<ol; i++)
for (var j=1; j<al; j++)
if (objects[j].count[ objects[j].vals.indexOf(objs0[i]) ] != counts0[i])
return false;
// else, the multisets are equal:
return true;
}
It still uses indexOf search amongst all objects, so if you have multisets with many different objects you might want to optimize that part as well. Have a look at Unique ID or object signature (and it's duplicate questions) for how to get lookup table keys for them. And if you don't have many primitive values in the multisets, you might just store them in arrays and sort those before comparing each item-by-item (like #Bruno did).
Disclaimer: This solution doesn't try to get the [[PrimitiveValue]] of objects, they will never be counted as equal to primitives (while == would do).
Here is the update on #Bruno's jsperf test of the answers, yet I guess only two objects (each of them present 500 times in the 10k array) and no duplicate primitive values are not representative.

Javascript HashTable use Object key

I want to create a hash table with Object keys that are not converted into String.
Some thing like this:
var object1 = new Object();
var object2 = new Object();
var myHash = new HashTable();
myHash.put(object1, "value1");
myHash.put(object2, "value2");
alert(myHash.get(object1), myHash.get(object2)); // I wish that it will print value1 value2
EDIT: See my answer for full solution

Here is a simple Map implementation that will work with any type of key, including object references, and it will not mutate the key in any way:
function Map() {
var keys = [], values = [];
return {
put: function (key, value) {
var index = keys.indexOf(key);
if(index == -1) {
keys.push(key);
values.push(value);
}
else {
values[index] = value;
}
},
get: function (key) {
return values[keys.indexOf(key)];
}
};
}
While this yields the same functionality as a hash table, it's not actually implemented using a hash function since it iterates over arrays and has a worst case performance of O(n). However, for the vast majority of sensible use cases this shouldn't be a problem at all. The indexOf function is implemented by the JavaScript engine and is highly optimized.

Here is a proposal:
function HashTable() {
this.hashes = {};
}
HashTable.prototype = {
constructor: HashTable,
put: function( key, value ) {
this.hashes[ JSON.stringify( key ) ] = value;
},
get: function( key ) {
return this.hashes[ JSON.stringify( key ) ];
}
};
The API is exactly as shown in your question.
You can't play with the reference in js however (so two empty objects will look like the same to the hashtable), because you have no way to get it. See this answer for more details: How to get javascript object references or reference count?
Jsfiddle demo: http://jsfiddle.net/HKz3e/
However, for the unique side of things, you could play with the original objects, like in this way:
function HashTable() {
this.hashes = {},
this.id = 0;
}
HashTable.prototype = {
constructor: HashTable,
put: function( obj, value ) {
obj.id = this.id;
this.hashes[ this.id ] = value;
this.id++;
},
get: function( obj ) {
return this.hashes[ obj.id ];
}
};
Jsfiddle demo: http://jsfiddle.net/HKz3e/2/
This means that your objects need to have a property named id that you won't use elsewhere. If you want to have this property as non-enumerable, I suggest you take a look at defineProperty (it's not cross-browser however, even with ES5-Shim, it doesn't work in IE7).
It also means you are limited on the number of items you can store in this hashtable. Limited to 253, that is.
And now, the "it's not going to work anywhere" solution: use ES6 WeakMaps. They are done exactly for this purpose: having objects as keys. I suggest you read MDN for more information: https://developer.mozilla.org/en/JavaScript/Reference/Global_Objects/WeakMap
It slightly differs from your API though (it's set and not put):
var myMap = new WeakMap(),
object1 = {},
object2 = {};
myMap.set( object1, 'value1' );
myMap.set( object2, 'value2' );
console.log( myMap.get( object1 ) ); // "value1"
console.log( myMap.get( object2 ) ); // "value2"
Jsfiddle demo with a weakmap shim: http://jsfiddle.net/Ralt/HKz3e/9/
However, weakmaps are implemented in FF and Chrome (only if you enable the "Experimental javascript features" flag in chrome however). There are shims available, like this one: https://gist.github.com/1269991. Use at your own risk.
You can also use Maps, they may more suit your needs, since you also need to store primitive values (strings) as keys. Doc, Shim.

I took #Florian Margaine's suggestion to higher level and came up with this:
function HashTable(){
var hash = new Object();
this.put = function(key, value){
if(typeof key === "string"){
hash[key] = value;
}
else{
if(key._hashtableUniqueId == undefined){
key._hashtableUniqueId = UniqueId.prototype.generateId();
}
hash[key._hashtableUniqueId] = value;
}
};
this.get = function(key){
if(typeof key === "string"){
return hash[key];
}
if(key._hashtableUniqueId == undefined){
return undefined;
}
return hash[key._hashtableUniqueId];
};
}
function UniqueId(){
}
UniqueId.prototype._id = 0;
UniqueId.prototype.generateId = function(){
return (++UniqueId.prototype._id).toString();
};
Usage
var map = new HashTable();
var object1 = new Object();
map.put(object1, "Cocakola");
alert(map.get(object1)); // Cocakola
//Overriding
map.put(object1, "Cocakola 2");
alert(map.get(object1)); // Cocakola 2
// String key is used as String
map.put("myKey", "MyValue");
alert(map.get("myKey")); // MyValue
alert(map.get("my".concat("Key"))); // MyValue
// Invalid keys
alert(map.get("unknownKey")); // undefined
alert(map.get(new Object())); // undefined

Here is a proposal, combining #Florian's solution with #Laurent's.
function HashTable() {
this.hashes = [];
}
HashTable.prototype = {
constructor: HashTable,
put: function( key, value ) {
this.hashes.push({
key: key,
value: value
});
},
get: function( key ) {
for( var i = 0; i < this.hashes.length; i++ ){
if(this.hashes[i].key == key){
return this.hashes[i].value;
}
}
}
};
It wont change your object in any way and it doesn't rely on JSON.stringify.

I know that I am a year late, but for all others who stumble upon this thread, I've written the ordered object stringify to JSON, that solves the above noted dilemma: http://stamat.wordpress.com/javascript-object-ordered-property-stringify/
Also I was playing with custom hash table implementations which is also related to the topic: http://stamat.wordpress.com/javascript-quickly-find-very-large-objects-in-a-large-array/
//SORT WITH STRINGIFICATION
var orderedStringify = function(o, fn) {
var props = [];
var res = '{';
for(var i in o) {
props.push(i);
}
props = props.sort(fn);
for(var i = 0; i < props.length; i++) {
var val = o[props[i]];
var type = types[whatis(val)];
if(type === 3) {
val = orderedStringify(val, fn);
} else if(type === 2) {
val = arrayStringify(val, fn);
} else if(type === 1) {
val = '"'+val+'"';
}
if(type !== 4)
res += '"'+props[i]+'":'+ val+',';
}
return res.substring(res, res.lastIndexOf(','))+'}';
};
//orderedStringify for array containing objects
var arrayStringify = function(a, fn) {
var res = '[';
for(var i = 0; i < a.length; i++) {
var val = a[i];
var type = types[whatis(val)];
if(type === 3) {
val = orderedStringify(val, fn);
} else if(type === 2) {
val = arrayStringify(val);
} else if(type === 1) {
val = '"'+val+'"';
}
if(type !== 4)
res += ''+ val+',';
}
return res.substring(res, res.lastIndexOf(','))+']';
}

Based on Peters answer, but with proper class design (not abusing closures), so the values are debuggable. Renamed from Map to ObjectMap, because Map is a builtin function. Also added the exists method:
ObjectMap = function() {
this.keys = [];
this.values = [];
}
ObjectMap.prototype.set = function(key, value) {
var index = this.keys.indexOf(key);
if (index == -1) {
this.keys.push(key);
this.values.push(value);
} else {
this.values[index] = value;
}
}
ObjectMap.prototype.get = function(key) {
return this.values[ this.keys.indexOf(key) ];
}
ObjectMap.prototype.exists = function(key) {
return this.keys.indexOf(key) != -1;
}
/*
TestObject = function() {}
testA = new TestObject()
testB = new TestObject()
om = new ObjectMap()
om.set(testA, true)
om.get(testB)
om.exists(testB)
om.exists(testA)
om.exists(testB)
*/

When you say you don't want your Object keys converted into Strings, I'm going to assume it's because you just don't want the entire code contents of your Objects being used as keys. This, of course, makes perfect sense.
While there is no "hash table" in Javascript per-se, you can accomplish what you're looking for by simply overriding your Object's prototype.toString and returning a valid key value that will be unique to each instance. One way to do this is with Symbol():
function Obj () {
this.symbol = Symbol() // Guaranteed to be unique to each instance
}
Obj.prototype.toString = function () {
return this.symbol // Return the unique Symbol, instead of Obj's stringified code
}
let a = new Obj()
let b = new Obj()
let table = {}
table[a] = 'A'
table[b] = 'B'
console.log(table) // {Symbol(): 'A', Symbol(): 'B'}
console.log(table[a]) // A
console.log(table[b]) // B

Using JSON.stringify() is completely awkward to me, and gives the client no real control over how their keys are uniquely identified. The objects that are used as keys should have a hashing function, but my guess is that in most cases overriding the toString() method, so that they will return unique strings, will work fine:
var myMap = {};
var myKey = { toString: function(){ return '12345' }};
var myValue = 6;
// same as myMap['12345']
myMap[myKey] = myValue;
Obviously, toString() should do something meaningful with the object's properties to create a unique string. If you want to enforce that your keys are valid, you can create a wrapper and in the get() and put() methods, add a check like:
if(!key.hasOwnProperty('toString')){
throw(new Error('keys must override toString()'));
}
But if you are going to go thru that much work, you may as well use something other than toString(); something that makes your intent more clear.
So a very simple proposal would be:
function HashTable() {
this.hashes = {};
}
HashTable.prototype = {
constructor: HashTable,
put: function( key, value ) {
// check that the key is meaningful,
// also will cause an error if primitive type
if( !key.hasOwnProperty( 'hashString' ) ){
throw( new Error( 'keys must implement hashString()' ) );
}
// use .hashString() because it makes the intent of the code clear
this.hashes[ key.hashString() ] = value;
},
get: function( key ) {
// check that the key is meaningful,
// also will cause an error if primitive type
if( !key.hasOwnProperty( 'hashString' ) ){
throw( new Error( 'keys must implement hashString()' ) );
}
// use .hashString() because it make the intent of the code clear
return this.hashes[ key.hashString() ];
}
};

Inspired by #florian, here's a way where the id doesn't need JSON.stringify:
'use strict';
module.exports = HashTable;
function HashTable () {
this.index = [];
this.table = [];
}
HashTable.prototype = {
constructor: HashTable,
set: function (id, key, value) {
var index = this.index.indexOf(id);
if (index === -1) {
index = this.index.length;
this.index.push(id);
this.table[index] = {};
}
this.table[index][key] = value;
},
get: function (id, key) {
var index = this.index.indexOf(id);
if (index === -1) {
return undefined;
}
return this.table[index][key];
}
};

I took #Ilya_Gazman solution and improved it by setting '_hashtableUniqueId' as a not enumerable property (it won't appear in JSON requests neither will be listed in for loops). Also removed UniqueId object, since it is enough using only HastTable function closure. For usage details please see Ilya_Gazman post
function HashTable() {
var hash = new Object();
return {
put: function (key, value) {
if(!HashTable.uid){
HashTable.uid = 0;
}
if (typeof key === "string") {
hash[key] = value;
} else {
if (key._hashtableUniqueId === undefined) {
Object.defineProperty(key, '_hashtableUniqueId', {
enumerable: false,
value: HashTable.uid++
});
}
hash[key._hashtableUniqueId] = value;
}
},
get: function (key) {
if (typeof key === "string") {
return hash[key];
}
if (key._hashtableUniqueId === undefined) {
return undefined;
}
return hash[key._hashtableUniqueId];
}
};
}

The best solution is to use WeakMap when you can (i.e. when you target browsers supporting it)
Otherwise you can use the following workaround (Typescript written and collision safe):
// Run this in the beginning of your app (or put it into a file you just import)
(enableObjectID)();
const uniqueId: symbol = Symbol('The unique id of an object');
function enableObjectID(): void {
if (typeof Object['id'] !== 'undefined') {
return;
}
let id: number = 0;
Object['id'] = (object: any) => {
const hasUniqueId: boolean = !!object[uniqueId];
if (!hasUniqueId) {
object[uniqueId] = ++id;
}
return object[uniqueId];
};
}
Then you can simply get a unique number for any object in your code (like if would have been for pointer address)
let objectA = {};
let objectB = {};
let dico = {};
dico[(<any>Object).id(objectA)] = "value1";
// or
dico[Object['id'](objectA);] = "value1";
// If you are not using typescript you don't need the casting
dico[Object.id(objectA)] = "value1"

I know I'm late, but here's a simple HashMap implementation:
Function.prototype.toJSON = Function.prototype.toString;
//taken from https://stackoverflow.com/questions/1249531/how-to-get-a-javascript-objects-class
function getNativeClass(obj) {
if (typeof obj === "undefined") return "undefined";
if (obj === null) return "null";
return Object.prototype.toString.call(obj).match(/^\[object\s(.*)\]$/)[1];
}
function globals() {
if (typeof global === "object") //node
return global;
return this;
}
function lookup(x) {
return globals()[x];
}
function getAnyClass(obj) {
if (typeof obj === "undefined") return "undefined";
if (obj === null) return "null";
return obj.constructor.name;
}
//taken from https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Errors/Cyclic_object_value#examples
var getCircularReplacer = () => {
const seen = new WeakSet();
return (key, value) => {
if (typeof value === "object" && value !== null) {
if (seen.has(value)) {
return "[Circular]";
}
seen.add(value);
}
return value;
};
};
function encode(x) {
if (typeof x === "object" && x !== null) {
var y = myClone(x);
x = Object.getPrototypeOf(x);
for (var i = 0; i < Object.getOwnPropertyNames(y).length; i++) { //Make enumerable
x[Object.getOwnPropertyNames(y)[i]] = y[Object.getOwnPropertyNames(y)[i]];
}
}
return getAnyClass(x) + " " + JSON.stringify(x, getCircularReplacer());
}
function decode(x) {
var a = x.split(" ").slice(1).join(" "); //OBJECT
if (typeof lookup(x.split(" ")[0])) {
return new (lookup(x.split(" ")[0]))(JSON.parse(a))
} else {
return JSON.parse(a);
}
}
//taken from https://github.com/feross/fromentries/blob/master/index.js
/*! fromentries. MIT License. Feross Aboukhadijeh <https://feross.org/opensource> */
function fromEntries(iterable) {
return [...iterable].reduce((obj, [key, val]) => {
obj[key] = val
return obj
}, {})
}
var toEnumerable = (obj) => {
return fromEntries(
Object.getOwnPropertyNames(obj).map(prop => [prop, obj[prop]])
);
};
//taken from https://stackoverflow.com/questions/41474986/how-to-clone-a-javascript-es6-class-instance
function myClone(instanceOfBlah) {
if (typeof instanceOfBlah !== "object" || !instanceOfBlah) { return instanceOfBlah; }
const clone = Object.assign({}, toEnumerable(instanceOfBlah));
const Blah = instanceOfBlah.constructor;
Object.setPrototypeOf(clone, Blah.prototype);
return clone;
}
function HashMap(a) {
if (typeof a === "undefined") {
a = [];
}
a = Array.from(a);
a = a.map((e) => [encode(e[0]), e[1]]);
this.a = a;
}
HashMap.from = function (a) {
var temp = myClone(a);
//convert to array
a = [];
for (var i = 0; i < Object.getOwnPropertyNames(temp).length; i++) {
a.push([Object.getOwnPropertyNames(temp)[i], temp[Object.getOwnPropertyNames(temp)[i]]]);
}
return new HashMap(a);
}
HashMap.prototype.put = function (x, y) {
this.a.push([encode(x), y]);
}
HashMap.prototype.get = function (x) {
var t1 = this.a.map((e) => e[0]);
return this.a[t1.indexOf(encode(x))][1];
}
HashMap.prototype.length = function () {
return this.a.length;
}
HashMap.prototype.toString = function () {
var result = [];
for (var i = 0; i < this.length(); i++) {
result.push(JSON.stringify(decode(this.a[i][0]), getCircularReplacer()) + " => " + this.a[i][1]);
}
return "HashMap {" + result + "}";
}
var foo = new HashMap();
foo.put("SQRT3", Math.sqrt(3));
foo.put({}, "bar");
console.log(foo.get({}));
console.log(foo.toString());
Note that it is ordered.
Methods:
put: Adds an item
get: Access an item
from (static): Convert from a JavaScript object
toString: Convert to string
Minified and without the test:
function getNativeClass(t){return void 0===t?"undefined":null===t?"null":Object.prototype.toString.call(t).match(/^\[object\s(.*)\]$/)[1]}function globals(){return"object"==typeof global?global:this}function lookup(t){return globals()[t]}function getAnyClass(t){return void 0===t?"undefined":null===t?"null":t.constructor.name}Function.prototype.toJSON=Function.prototype.toString;var getCircularReplacer=()=>{const t=new WeakSet;return(e,r)=>{if("object"==typeof r&&null!==r){if(t.has(r))return"[Circular]";t.add(r)}return r}};function encode(t){if("object"==typeof t&&null!==t){var e=myClone(t);t=Object.getPrototypeOf(t);for(var r=0;r<Object.getOwnPropertyNames(e).length;r++)t[Object.getOwnPropertyNames(e)[r]]=e[Object.getOwnPropertyNames(e)[r]]}return getAnyClass(t)+" "+JSON.stringify(t,getCircularReplacer())}function decode(t){var e=t.split(" ").slice(1).join(" ");return lookup(t.split(" ")[0]),new(lookup(t.split(" ")[0]))(JSON.parse(e))}function fromEntries(t){return[...t].reduce((t,[e,r])=>(t[e]=r,t),{})}var toEnumerable=t=>fromEntries(Object.getOwnPropertyNames(t).map(e=>[e,t[e]]));function myClone(t){if("object"!=typeof t||!t)return t;const e=Object.assign({},toEnumerable(t)),r=t.constructor;return Object.setPrototypeOf(e,r.prototype),e}function HashMap(t){void 0===t&&(t=[]),t=(t=Array.from(t)).map(t=>[encode(t[0]),t[1]]),this.a=t}HashMap.from=function(t){var e=myClone(t);t=[];for(var r=0;r<Object.getOwnPropertyNames(e).length;r++)t.push([Object.getOwnPropertyNames(e)[r],e[Object.getOwnPropertyNames(e)[r]]]);return new HashMap(t)},HashMap.prototype.put=function(t,e){this.a.push([encode(t),e])},HashMap.prototype.get=function(t){var e=this.a.map(t=>t[0]);return this.a[e.indexOf(encode(t))][1]},HashMap.prototype.length=function(){return this.a.length},HashMap.prototype.toString=function(){for(var t=[],e=0;e<this.length();e++)t.push(JSON.stringify(decode(this.a[e][0]),getCircularReplacer())+" => "+this.a[e][1]);return"HashMap {"+t+"}"};
Also, you can customize the encoder and decoder by changing encode and decode functions.
As in florian's answer, you can't play with the reference in js however (so two empty objects will look like the same to the hashtable).

class Dict{
constructor(){
this.keys = [];
this.values = [];
this.set = this.set.bind(this);
}
set(key, value){
this.keys.push(key);
this.values.push(value);
}
get(key){
return this.values[this.keys.indexOf(key)];
}
all(){
return this.keys.map((kk, ii)=>[kk, this.values[ii]]);
}
}
let d1 = new Dict();
let k1 = {1: 'a'};
d1.set(k1, 2);
console.log(d1.get(k1)); // 2
let k2 = {2: 'b'};
d1.set(k2, 3);
console.log(d1.all());
// [ [ { '1': 'a' }, 2 ], [ { '2': 'b' }, 3 ] ]

Just use the strict equality operator when looking up the object: ===
var objects = [];
objects.push(object1);
objects.push(object2);
objects[0] === object1; // true
objects[1] === object1; // false
The implementation will depend on how you store the objects in the HashTable class.

JavaScript Iterator Class

Do you know a JavaScript library that implements a generic Iterator class for collections (be it Arrays or some abstract Enumerable) with a full set of features, like the Google Common or the Apache Commons?
Edit: Enumerable#each is not an Iterator class. I'm looking for an Iterator, something that would let us write something like:
var iterator = new Iterator(myCollection);
for (var element = iterator.next(); iterator.hasNext(); element = iterator.next()) {
// iterator
}
Edit : mamoo reminded us of the Iterator implementation in Mozilla's Javascript 1.7. So the goal now is to find an implementation of this Iterator function in Javascript 1.5 (ECMA 4).
Edit2 : Why using an iterator when libraries (and ECMA 5) provide a each method? First, because each usually messes with this because the callback is call -ed (that's why each accepts a second argument in Prototype). Then, because people are much more familiar with the for(;;) construct than with the .each(callback) construct (at least, in my field). Lastly, because an iterator can iterate over plain objects (see JavaScript 1.7).
Edit3 : I accepted npup's anwser, but here is my shot at it :
function Iterator(o, keysOnly) {
if (!(this instanceof arguments.callee))
return new arguments.callee(o, keysOnly);
var index = 0, keys = [];
if (!o || typeof o != "object") return;
if ('splice' in o && 'join' in o) {
while(keys.length < o.length) keys.push(keys.length);
} else {
for (p in o) if (o.hasOwnProperty(p)) keys.push(p);
}
this.next = function next() {
if (index < keys.length) {
var key = keys[index++];
return keysOnly ? key : [key, o[key]];
} else throw { name: "StopIteration" };
};
this.hasNext = function hasNext() {
return index < keys.length;
};
}
var lang = { name: 'JavaScript', birthYear: 1995 };
var it = Iterator(lang);
while (it.hasNext()) {
alert(it.next());
}
//alert(it.next()); // A StopIteration exception is thrown
var langs = ['JavaScript', 'Python', 'C++'];
var it = Iterator(langs);
while (it.hasNext()) {
alert(it.next());
}
//alert(it.next()); // A StopIteration exception is thrown

Ok, the enumerable pattern is not a real iterator then.
Is this (below) useful for you? It conforms to the sematics you gave at least. As usual there are tradeoffs to be made here and there, and I didn't think very hard when deciding this time :).
And maybe you would like to be able to send in a number or two and iterate over a range in that way. But this could maybe be a start (there's support for iterating over hashes, arrays and strings).
It's a whole demo page which runs itself and does some debug output, but the (possibly) interesting stuff is in the
window.npup = (function() {
[...]
})();
spot.
Maybe it is just me who doesn't get it at all, but what would you use such a java-like Iterator for in a real situation?
Best
/npup
<html>
<head>
<title>untitled</title>
</head>
<body>
<ul id="output"></ul>
<script type="text/javascript">
window.log = (function (outputAreaId) {
var myConsole = document.getElementById(outputAreaId);
function createElem(color) {
var elem = document.createElement('li');
elem.style.color = color;
return elem;
}
function appendElem(elem) {
myConsole.appendChild(elem);
}
function debug(msg) {
var elem = createElem('#888');
elem.innerHTML = msg;
appendElem(elem);
}
function error(msg) {
var elem = createElem('#f88');
elem.innerHTML = msg;
appendElem(elem);
}
return {
debug: debug
, error: error
};
})('output');
window.npup = (function () {
// Array check as proposed by Mr. Crockford
function isArray(candidate) {
return candidate &&
typeof candidate==='object' &&
typeof candidate.length === 'number' &&
typeof candidate.splice === 'function' &&
!(candidate.propertyIsEnumerable('length'));
}
function dontIterate(collection) {
// put some checks chere for stuff that isn't iterable (yet)
return (!collection || typeof collection==='number' || typeof collection==='boolean');
}
function Iterator(collection) {
if (typeof collection==='string') {collection = collection.split('');}
if (dontIterate(collection)) {throw new Error('Oh you nasty man, I won\'t iterate over that ('+collection+')!');}
var arr = isArray(collection);
var idx = 0, top=0;
var keys = [], prop;
if (arr) {top = collection.length;}
else {for (prop in collection) {keys.push(prop);}}
this.next = function () {
if (!this.hasNext()) {throw new Error('Oh you nasty man. I have no more elements.');}
var elem = arr ? collection[idx] : {key:keys[idx], value:collection[keys[idx]]};
++idx;
return elem;
};
this.hasNext = function () {return arr ? idx<=top : idx<=keys.length;};
}
return {Iterator: Iterator};
})();
var element;
log.debug('--- Hash demo');
var o = {foo:1, bar:2, baz:3, bork:4, hepp: {a:1,b:2,c:3}, bluff:666, bluff2:777};
var iterator = new npup.Iterator(o);
for (element = iterator.next(); iterator.hasNext(); element = iterator.next()) {
log.debug('got elem from hash: '+element.key+' => '+element.value);
if (typeof element.value==='object') {
var i2 = new npup.Iterator(element.value);
for (var e2=i2.next(); i2.hasNext(); e2=i2.next()) {
log.debug(' # from inner hash: '+e2.key+' => '+e2.value);
}
}
}
log.debug('--- Array demo');
var a = [1,2,3,42,666,777];
iterator = new npup.Iterator(a);
for (element = iterator.next(); iterator.hasNext(); element = iterator.next()) {
log.debug('got elem from array: '+ element);
}
log.debug('--- String demo');
var s = 'First the pants, THEN the shoes!';
iterator = new npup.Iterator(s);
for (element = iterator.next(); iterator.hasNext(); element = iterator.next()) {
log.debug('got elem from string: '+ element);
}
log.debug('--- Emptiness demo');
try {
log.debug('Try to get next..');
var boogie = iterator.next();
}
catch(e) {
log.error('OW: '+e);
}
log.debug('--- Non iterables demo');
try{iterator = new npup.Iterator(true);} catch(e) {log.error('iterate over boolean: '+e);}
try{iterator = new npup.Iterator(6);} catch(e) {log.error('iterate over number: '+e);}
try{iterator = new npup.Iterator(null);} catch(e) {log.error('iterate over null: '+e);}
try{iterator = new npup.Iterator();} catch(e) {log.error('iterate over undefined: '+e);}
</script>
</body>
</html>

JQuery has the each() method:
http://api.jquery.com/jQuery.each/
but probably there's something similar even in other libraries such as Moo or Dojo.
Javascript 1.7 implements the Iterator function:
https://developer.mozilla.org/en/Core_JavaScript_1.5_Guide/Iterators_and_Generators

This is my attempt (jsfiddle) for ECMAScript 262 5th edition (aka Javascript). (Uses for example Object.keys and Array.isArray)
//Usage
b=Iterator(a);
while(b()){
console.log(b.value);
}
The code:
function Iterator(input,keys) {
// Input:
// input : object|array
// keys : array|undefined|boolean
function my() {
++my.index;
if (my.index >= my.keys.length) {
my.index = my.keys.length -1;
my.key = my.value = undefined;
return false;
}
my.key = my.useIndex ? my.index : my.keys[my.index];
my.value = my.input[my.key];
return my.index < my.keys.length;
}
if (input === null || typeof input !== 'object') {
throw new TypeError("'input' should be object|array");
}
if (
!Array.isArray(keys)
&& (typeof keys !== 'undefined')
&& (typeof keys !== 'boolean')
) {
throw new TypeError("'keys' should be array|boolean|undefined");
}
// Save a reference to the input object.
my.input = input;
if (Array.isArray(input)) {
//If the input is an array, set 'useIndex' to true if
//the internal index should be used as a key.
my.useIndex = !keys;
//Either create and use a list of own properties,
// or use the supplied keys
// or at last resort use the input (since useIndex is true in that
// case it is only used for the length)
my.keys = keys===true ? Object.keys(input) : keys || input;
} else {
my.useIndex = false;
my.keys = Array.isArray(keys) ? keys : Object.keys(input);
}
// Set index to before the first element.
my.index = -1;
return my;
}
Examples:
function Person(firstname, lastname, domain) {
this.firstname = firstname;
this.lastname = lastname;
this.domain = domain;
}
Person.prototype.type = 'Brillant';
var list = [
new Person('Paula','Bean','some.domain.name'),
new Person('John','Doe','another.domain.name'),
new Person('Johanna','Doe','yet.another.domain.name'),
];
var a,b;
var data_array = ['A','B','C','D','E','F'];
data_array[10]="Sparse";
console.log('Iterate over own keys in an object, unknown order');
a = Iterator(list[0]);
while(a()) console.log(" ",a.key, a.value);
console.log('Iterate over keys from anywhere, in specified order');
a = Iterator(list[0], ['lastname','firstname','type']);
while(a()) console.log(" ",a.key, a.value);
console.log('Iterate over all values in an array');
a = Iterator(list);
while(a()) console.log(a.key, a.value.firstname, a.value.lastname);
//Some abusing, that works for arrays (if the iterator.keys is modified
//it can also be used for objects)
console.log('Add more entries to the array, reusing the iterator...');
list.push(new Person('Another','Name','m.nu'));
while(a()) console.log(a.key, a.value.firstname, a.value.lastname);
console.log('Reset index and print everything again...');
a.index=-1; //Reset the index.
while(a()) console.log(a.key, a.value.firstname, a.value.lastname);
//With arrays, if setting 'keys' to true it will only print the
//elements that has values (If the array has more own enumerable values
//they too will be included)
console.log('Print sparce arrays...');
a = Iterator(data_array,true);
while(a()) console.log(a.key, a.value);

In the time since this question was asked JavaScript has added actual Iterators. Some built-in types, such as Array, Map, and String now have a default iteration behavior, but you can add your own to any object by including a next() function which returns one of two objects:
{done:true} /*or*/
{done:false, value:SOMEVALUE}
One way to access an object Iterator is with the:
for ( var of object ) { }
loop. Here is a (reasonably silly) example where we define an Iterator and then use it in such a loop to produce a string 1, 2, 3:
"use strict";
function count ( i ) {
let n = 0;
let I = {};
I[Symbol.iterator] = function() {
return { next: function() { return (n > i) ? {done:true}
: {done:false, value:n++} } } };
let s = "";
let c = "";
for ( let i of I ) { /* use the iterator we defined above */
s += c + i;
c = ", "
}
return s;
}
let s = count(3);
console.log(s);

Ive used LINQ to Javascript in a few projects.
http://jslinq.codeplex.com/Wikipage
var myList = [
{FirstName:"Chris",LastName:"Pearson"},
{FirstName:"Kate",LastName:"Johnson"},
{FirstName:"Josh",LastName:"Sutherland"},
{FirstName:"John",LastName:"Ronald"},
{FirstName:"Steve",LastName:"Pinkerton"}
];
var exampleArray = JSLINQ(myList)
.Where(function(item){ return item.FirstName == "Chris"; })
.OrderBy(function(item) { return item.FirstName; })
.Select(function(item){ return item.FirstName; });

I'm still a learner of js.class.
Though being close to Ruby, helps me.
http://jsclass.jcoglan.com/enumerable.html
MarkT

Since this hasn't been mention yet arrays have higher-order functions built in.
Map works like iterator that can only do a single pass.
[1,2,3,4,5].map( function(input){ console.log(input); } );
This code passes each element in the list into a function, in this case its a simple printer.
1
2
3
4
5

The $.param( ) inverse function in JavaScript / jQuery

Given the following form:
<form>
<input name="foo" value="bar">
<input name="hello" value="hello world">
</form>
I can use the $.param( .. ) construct to serialize the form:
$.param( $('form input') )
=> foo=bar&hello=hello+world
How can I deserialize the above String with JavaScript and get a hash back?
For example,
$.magicFunction("foo=bar&hello=hello+world")
=> {'foo' : 'bar', 'hello' : 'hello world'}
Reference: jQuery.param( obj ).

You should use jQuery BBQ's deparam function. It's well-tested and documented.

This is a slightly modified version of a function I wrote a while ago to do something similar.
var QueryStringToHash = function QueryStringToHash (query) {
var query_string = {};
var vars = query.split("&");
for (var i=0;i<vars.length;i++) {
var pair = vars[i].split("=");
pair[0] = decodeURIComponent(pair[0]);
pair[1] = decodeURIComponent(pair[1]);
// If first entry with this name
if (typeof query_string[pair[0]] === "undefined") {
query_string[pair[0]] = pair[1];
// If second entry with this name
} else if (typeof query_string[pair[0]] === "string") {
var arr = [ query_string[pair[0]], pair[1] ];
query_string[pair[0]] = arr;
// If third or later entry with this name
} else {
query_string[pair[0]].push(pair[1]);
}
}
return query_string;
};

How about this short functional approach?
function parseParams(str) {
return str.split('&').reduce(function (params, param) {
var paramSplit = param.split('=').map(function (value) {
return decodeURIComponent(value.replace(/\+/g, ' '));
});
params[paramSplit[0]] = paramSplit[1];
return params;
}, {});
}
Example:
parseParams("this=is&just=an&example") // Object {this: "is", just: "an", example: undefined}

My answer:
function(query){
var setValue = function(root, path, value){
if(path.length > 1){
var dir = path.shift();
if( typeof root[dir] == 'undefined' ){
root[dir] = path[0] == '' ? [] : {};
}
arguments.callee(root[dir], path, value);
}else{
if( root instanceof Array ){
root.push(value);
}else{
root[path] = value;
}
}
};
var nvp = query.split('&');
var data = {};
for( var i = 0 ; i < nvp.length ; i++ ){
var pair = nvp[i].split('=');
var name = decodeURIComponent(pair[0]);
var value = decodeURIComponent(pair[1]);
var path = name.match(/(^[^\[]+)(\[.*\]$)?/);
var first = path[1];
if(path[2]){
//case of 'array[level1]' || 'array[level1][level2]'
path = path[2].match(/(?=\[(.*)\]$)/)[1].split('][')
}else{
//case of 'name'
path = [];
}
path.unshift(first);
setValue(data, path, value);
}
return data;
}

I am using David Dorward's answer, and realized that it doesn't behave like PHP or Ruby on Rails how they parse the params:
1) a variable is only an array if it ends with [], such as ?choice[]=1&choice[]=12, not when it is ?a=1&a=2
2) when mulitple params exist with the same name, the later ones replaces the earlier ones, as on PHP servers (Ruby on Rails keep the first one and ignore the later ones), such as ?a=1&b=2&a=3
So modifying David's version, I have:
function QueryStringToHash(query) {
if (query == '') return null;
var hash = {};
var vars = query.split("&");
for (var i = 0; i < vars.length; i++) {
var pair = vars[i].split("=");
var k = decodeURIComponent(pair[0]);
var v = decodeURIComponent(pair[1]);
// If it is the first entry with this name
if (typeof hash[k] === "undefined") {
if (k.substr(k.length-2) != '[]') // not end with []. cannot use negative index as IE doesn't understand it
hash[k] = v;
else
hash[k.substr(0, k.length-2)] = [v];
// If subsequent entry with this name and not array
} else if (typeof hash[k] === "string") {
hash[k] = v; // replace it
// If subsequent entry with this name and is array
} else {
hash[k.substr(0, k.length-2)].push(v);
}
}
return hash;
};
which is tested fairly thoroughly.

I know this is an old thread, but maybe there is still some relevance in it?
Inspired by Jacky Li's good solution I tried a slight variation of my own with the objective to also be able to take care of arbitrary combinations of arrays and objects as input. I looked at how PHP would have done it and tried to get something "similar" going. Here is my code:
function getargs(str){
var ret={};
function build(urlnam,urlval,obj){ // extend the return object ...
var i,k,o=obj, x, rx=/\[([^\]]*)\]/g, idx=[urlnam.replace(rx,'')];
while (x=rx.exec(urlnam)) idx.push(x[1]);
while(true){
k=idx.shift();
if(k.trim()=='') {// key is empty: autoincremented index
if (o.constructor.name=='Array') k=o.length; // for Array
else if (o===obj ) {k=null} // for first level property name
else {k=-1; // for Object
for(i in o) if (+i>k) k=+i;
k++;
}
}
if(idx.length) {
// set up an array if the next key (idx[0]) appears to be
// numeric or empty, otherwise set up an object:
if (o[k]==null || typeof o[k]!='object') o[k]=isNaN(idx[0])?{}:[];
o=o[k]; // move on to the next level
}
else { // OK, time to store the urlval in its chosen place ...
// console.log('key',k,'val',urlval);
o[k]=urlval===""?null:urlval; break; // ... and leave the while loop.
}
}
return obj;
}
// ncnvt: is a flag that governs the conversion of
// numeric strings into numbers
var ncnvt=true,i,k,p,v,argarr=[],
ar=(str||window.location.search.substring(1)).split("&"),
l=ar.length;
for (i=0;i<l;i++) {if (ar[i]==="") continue;
p=ar[i].split("=");k=decodeURIComponent(p[0]);
v=p[1];v=(v!=null)?decodeURIComponent(v.replace(/\+/g,'%20')):'';
if (ncnvt && v.trim()>"" && !isNaN(v)) v-=0;
argarr.push([k,v]); // array: key-value-pairs of all arguments
}
for (i=0,l=argarr.length;i<l;i++) build(argarr[i][0],argarr[i][1],ret);
return ret;
}
If the function is called without the str-argument it will assume window.location.search.slice(1) as input.
Some examples:
['a=1&a=2', // 1
'x[y][0][z][]=1', // 2
'hello=[%22world%22]&world=hello', // 3
'a=1&a=2&&b&c=3&d=&=e&', // 4
'fld[2][]=2&fld[][]=3&fld[3][]=4&fld[]=bb&fld[]=cc', // 5
$.param({a:[[1,2],[3,4],{aa:'one',bb:'two'},[5,6]]}), // 6
'a[]=hi&a[]=2&a[3][]=7&a[3][]=99&a[]=13',// 7
'a[x]=hi&a[]=2&a[3][]=7&a[3][]=99&a[]=13'// 8
].map(function(v){return JSON.stringify(getargs(v));}).join('\n')
results in
{"a":2} // 1
{"x":{"y":[{"z":[1]}]}} // 2
{"hello":"[\"world\"]","world":"hello"} // 3
{"a":2,"b":null,"c":3,"d":null,"null":"e"} // 4 = { a: 2, b: null, c: 3, d: null, null: "e" }
{"fld":[null,null,[2],[3,4],"bb","cc"]} // 5
{"a":[[1,2],[3,4],{"aa":"one","bb":"two"},[5,6]]} // 6
{"a":["hi",2,null,[7,99],13]} // 7
{"a":{"0":2,"3":[7,99],"4":13,"x":"hi"}} // 8
Whereas Jacky Li's solution would produce the outer container for a as a plain object
{a:{"0":["1","2"],"1":["3","4"],"2":["5","6"]}} // 6: JackyLi's output
getargs() looks at the first given index for any level to determine whether this level will be an object (non-numeric index) or an array (numeric or empty), thus resulting in the output as shown in the listing bove (no. 6).
If the current object is an array then nulls get inserted wherever necessary to represent empty positions. Arrays are always consecutively numbered and 0-based).
Note, that in the example no. 8 the "autoincrement" for empty indices still works, even though we are dealing with an object now and not an array.
As far as I have tested it, my getargs() behaves pretty much identically to Chriss Roger's great jQuery $.deparam() plugin mentioned in the accepted answer. The main difference is that getargs runs without jQuery and that it does autoincrement in objects while $.deparam() will not do that:
JSON.stringify($.deparam('a[x]=hi&a[]=2&a[3][]=7&a[3][]=99&a[]=13').a);
results in
{"3":["7","99"],"x":"hi","undefined":"13"}
In $.deparam() the index [] is interpreted as an undefined instead of an autoincremented numerical index.

Here's how you could create a new jQuery function:
jQuery.unparam = function (value) {
var
// Object that holds names => values.
params = {},
// Get query string pieces (separated by &)
pieces = value.split('&'),
// Temporary variables used in loop.
pair, i, l;
// Loop through query string pieces and assign params.
for (i = 0, l = pieces.length; i < l; i++) {
pair = pieces[i].split('=', 2);
// Repeated parameters with the same name are overwritten. Parameters
// with no value get set to boolean true.
params[decodeURIComponent(pair[0])] = (pair.length == 2 ?
decodeURIComponent(pair[1].replace(/\+/g, ' ')) : true);
}
return params;
};

Thanks to him http://james.padolsey.com/javascript/parsing-urls-with-the-dom/
Pretty easy :D
function params_unserialize(p){
var ret = {},
seg = p.replace(/^\?/,'').split('&'),
len = seg.length, i = 0, s;
for (;i<len;i++) {
if (!seg[i]) { continue; }
s = seg[i].split('=');
ret[s[0]] = s[1];
}
return ret;}

This is really old question, but as i have coming - other people may coming to this post, and i want to a bit refresh this theme. Today no need to make custom solutions - there is URLSearchParams interface.
var paramsString = "q=URLUtils.searchParams&topic=api";
var searchParams = new URLSearchParams(paramsString);
//Iterate the search parameters.
for (let p of searchParams) {
console.log(p);
}
The only one limitation i know - this feature not supported in IE / Edge.

Here's my JavaScript implementation which I use in a server-side JScript ASP Classic page (demo):
// Transforms a query string in the form x[y][0][z][]=1 into {x:{y:[{z:[1]}]}}
function parseJQueryParams(p) {
var params = {};
var pairs = p.split('&');
for (var i=0; i<pairs.length; i++) {
var pair = pairs[i].split('=');
var indices = [];
var name = decodeURIComponent(pair[0]),
value = decodeURIComponent(pair[1]);
var name = name.replace(/\[([^\]]*)\]/g,
function(k, idx) { indices.push(idx); return ""; });
indices.unshift(name);
var o = params;
for (var j=0; j<indices.length-1; j++) {
var idx = indices[j];
var nextIdx = indices[j+1];
if (!o[idx]) {
if ((nextIdx == "") || (/^[0-9]+$/.test(nextIdx)))
o[idx] = [];
else
o[idx] = {};
}
o = o[idx];
}
idx = indices[indices.length-1];
if (idx == "") {
o.push(value);
}
else {
o[idx] = value;
}
}
return params;
}

I came up with this solution, which behaves like the .Net function HttpUtility.ParseQueryString.
In the result, the query string parameters are store in properties as lists of values, so that qsObj["param"] will be the same as calling GetValues("param") in .Net.
I hope you like it. JQuery not required.
var parseQueryString = function (querystring) {
var qsObj = new Object();
if (querystring) {
var parts = querystring.replace(/\?/, "").split("&");
var up = function (k, v) {
var a = qsObj[k];
if (typeof a == "undefined") {
qsObj[k] = [v];
}
else if (a instanceof Array) {
a.push(v);
}
};
for (var i in parts) {
var part = parts[i];
var kv = part.split('=');
if (kv.length == 1) {
var v = decodeURIComponent(kv[0] || "");
up(null, v);
}
else if (kv.length > 1) {
var k = decodeURIComponent(kv[0] || "");
var v = decodeURIComponent(kv[1] || "");
up(k, v);
}
}
}
return qsObj;
};
Here is how to use it:
var qsObj = parseQueryString("a=1&a=2&&b&c=3&d=&=e&");
To preview the result in the console juste type in:
JSON.stringify(qsObj)
Output:
"{"a":["1","2"],"null":["","b",""],"c":["3"],"d":[""],"":["e"]}"

There's a beautiful one-liner over at CSS-Tricks (original source from Nicholas Ortenzio):
function getQueryParameters(str) {
return (str || document.location.search).replace(/(^\?)/,'').split("&").map(function(n){return n = n.split("="),this[n[0]] = n[1],this}.bind({}))[0];
}
The really clever part is how it uses the anonymous function's this object, adding a key/value pair for each of the queries in the string. That said, there's some room for improvement. I've modified it a bit below, with the following changes:
Added handling of empty strings and non-string input.
Handled URI-encoded strings (%40->#, etc).
Removed the default use of document.location.search when the input was empty.
Changed the name, made it more readable, added comments.
function deparam(str) {
// Uses an empty 'this' to build up the results internally
function splitQuery(query) {
query = query.split('=').map(decodeURIComponent);
this[query[0]] = query[1];
return this;
}
// Catch bad input
if (!str || !(typeof str === 'string' || str instanceof String))
return {};
// Split the string, run splitQuery on each piece, and return 'this'
var queries = str.replace(/(^\?)/,'').split('&');
return queries.map(splitQuery.bind({}))[0];
}

use this :
// convert query string to json object
var queryString = "cat=3&sort=1&page=1";
queryString
.split("&")
.forEach((item) => {
const prop = item.split("=");
filter[prop[0]] = prop[1];
});
console.log(queryString);

This is my version in Coffeescript.
Also works for url like
http://localhost:4567/index.html?hello=[%22world%22]&world=hello#/home
getQueryString: (url)->
return null if typeof url isnt 'string' or url.indexOf("http") is -1
split = url.split "?"
return null if split.length < 2
path = split[1]
hash_pos = path.indexOf "#"
path = path[0...hash_pos] if hash_pos isnt -1
data = path.split "&"
ret = {}
for d in data
[name, val] = d.split "="
name = decodeURIComponent name
val = decodeURIComponent val
try
ret[name] = JSON.parse val
catch error
ret[name] = val
return ret

Here's a simple & compact one if you only want to quickly get the parameters from a GET request:
function httpGet() {
var a={},b,i,q=location.search.replace(/^\?/,"").split(/\&/);
for(i in q) if(q[i]) {b=q[i].split("=");if(b[0]) a[b[0]]=
decodeURIComponent(b[1]).replace(/\+/g," ");} return a;
}
It converts
something?aa=1&bb=2&cc=3
into an object like
{aa:1,bb:2,cc:3}

Creates a serialized representation of an array or object (can be used as URL query string for AJAX requests).
<button id='param'>GET</button>
<div id="show"></div>
<script>
$('#param').click(function () {
var personObj = new Object();
personObj.firstname = "vishal"
personObj.lastname = "pambhar";
document.getElementById('show').innerHTML=$.param(`personObj`));
});
</script>
output:firstname=vishal&lastname=pambhar

answers could use a bit of jQuery elegance:
(function($) {
var re = /([^&=]+)=?([^&]*)/g;
var decodeRE = /\+/g; // Regex for replacing addition symbol with a space
var decode = function (str) {return decodeURIComponent( str.replace(decodeRE, " ") );};
$.parseParams = function(query) {
var params = {}, e;
while ( e = re.exec(query) ) {
var k = decode( e[1] ), v = decode( e[2] );
if (k.substring(k.length - 2) === '[]') {
k = k.substring(0, k.length - 2);
(params[k] || (params[k] = [])).push(v);
}
else params[k] = v;
}
return params;
};
})(jQuery);
fork at https://gist.github.com/956897

You can use the function .serializeArray() (Link) of jQuery itself. This function returns an array of key-value pair. Result example:
[
{ name: "id", value: "1" },
{ name: "version", value: "100" }
]