How to detect HTMLCollection/NodeList in JavaScript?

Question

I\'m not sure my current implementation is available all the time:

function isNodeList(nodes) {
    var result = Object.prototype.toString.call(nodes);
    /         


        

Answer 1

The following should return true, if nodes is of type NodeList

NodeList.prototype.isPrototypeOf(nodes)

@DavidSpector, for HTMLCollection you can similarly use :

HTMLCollection.prototype.isPrototypeOf(collection)

Answer 2

This answer is probably really really late, but....

if (nodes == '[object NodeList]') {
  // It's a nodeList
}

Answer 3

Here is how to test if an object is a NodeList in modern browsers:

if (nodes instanceof NodeList) {
  // It's a NodeList object
}

Answer 4

Check if variable is an HTMLcollection or a dom element

  var foo = document.getElementById('mydiv');
  var foo2 = document.getElementsByClassName('divCollection');
  console.log(foo instanceof HTMLElement);
  console.log(foo instanceof HTMLCollection);

Answer 5

I created a benchmark of all answers here to see, what is the best approve in speed. Turns out NodeList.prototype.isPrototypeOf(nodes) is by far the fastest. But in a normal use-case nodes instanceof NodeList would be fine too.

I personally would just not pick the isNodeList function, because its slow, custom and too much overhead.

Answer 6

I would structure the code differently:

function isNodeList(nodes) {
    var stringRepr = Object.prototype.toString.call(nodes);

    return typeof nodes === 'object' &&
        /^\[object (HTMLCollection|NodeList|Object)\]$/.test(stringRepr) &&
        (typeof nodes.length === 'number') &&
        (nodes.length === 0 || (typeof nodes[0] === "object" && nodes[0].nodeType > 0));
}

Notes:

• less return paths make easier-to-read code
• stick with one type of logic, if possible (i.e. use less negated checks)
• "item" is not mandatorily in a nodeList
• use hasOwnProperty() instead of in
• use square brackets to index into the list
• I don't think a try/catch is really necessary, but that might be wrong - you decide
• check for nodeType instead of tagName, as text nodes or comments do not have a name
• add more checks to the && chain if you see fit