Why is 'for(var item in list)' with arrays considered bad practice in JavaScript?

Question

Given a simple zero based, numerically indexed array:

var list = ['Foo', 'Bar', 'Baz'];

Many times, I have noticed that when someone suggests looping through variables in an array like this:

for(var item in list) { ... }

...there's almost certainly someone suggesting that that's bad practice and suggests an alternative approach:

var count = list.length;

for(var i = 0; i < count; i++) {
    var item = list[i];
    ...
}

What's the reasoning for not using the simpler version above and to use the second example instead?

Answer 1

First, the order of the loop is undefined for a for...in loop, so there's no guarantee the properties will be iterated in the order you want.

Second, for...in iterates over all enumerable properties of an object, including those inherited from its prototype. In the case of arrays, this could affect you if your code or any library included in your page has augmented the prototype of Array, which can be a genuinely useful thing to do:

Array.prototype.remove = function(val) {
    // Irrelevant implementation details
};

var a = ["a", "b", "c"];

for (var i in a) {
    console.log(i);
}

// Logs 0, 1, 2, "remove" (though not necessarily in that order)

Answer 2

Add 'list.foo = bar;' and try to use simple "for". If you don't use some libraries(like prototypeJs) and don't add any new properties to array object - you can use simple for-statement.

Answer 3

If you use for/in like that, item enumerates through string values "0", "1", ..., so not the actual objects in the list. So the the 'item' in the first snippet is more like the i in the second snippet,not the item. Furthermore string values are enumerated where you'd expect numbers. And you get in trouble when you properties to the list, like array.ID = "a123", as they will get enumerated also.

But with these downsides, I still think the syntax is very useful, if your team is aware of what it does.

Answer 4

Speed?

for(..;..;..) loop proved to be 36 times faster than for .. in when I tested it here.

Link courtesy this SO answer