[Solved] What do querySelectorAll and getElementsBy* methods return?

Question

Your getElementById code works since IDs have to be unique and thus the function always returns exactly one element (or null if none was found).

However, the methods
getElementsByClassName,
getElementsByName,
getElementsByTagName, and
getElementsByTagNameNS
return an iterable collection of elements.

The method names provide the hint: getElement implies singular, whereas getElements implies plural.

The method querySelector also returns a single element, and querySelectorAll returns an iterable collection.

The iterable collection can either be a NodeList or an HTMLCollection.

getElementsByName and querySelectorAll are both specified to return a NodeList; the other getElementsBy* methods are specified to return an HTMLCollection, but please note that some browser versions implement this differently.

Both of these collection types don’t offer the same properties that Elements, Nodes, or similar types offer; that’s why reading style off of document.getElements…(…) fails.
In other words: a NodeList or an HTMLCollection doesn’t have a style; only an Element has a style.

These “array-like” collections are lists that contain zero or more elements, which you need to iterate over, in order to access them.
While you can iterate over them similarly to an array, note that they are different from Arrays.

In modern browsers, you can convert these iterables to a proper Array with Array.from; then you can use forEach and other Array methods, e.g. iteration methods:

Array.from(document.getElementsByClassName("myElement"))
  .forEach((element) => element.style.size = "100px");

In old browsers that don’t support Array.from or the iteration methods, you can still use Array.prototype.slice.call.
Then you can iterate over it like you would with a real array:

var elements = Array.prototype.slice
    .call(document.getElementsByClassName("myElement"));

for(var i = 0; i < elements.length; ++i){
  elements[i].style.size = "100px";
}

You can also iterate over the NodeList or HTMLCollection itself, but be aware that in most circumstances, these collections are live (MDN docs, DOM spec), i.e. they are updated as the DOM changes.
So if you insert or remove elements as you loop, make sure to not accidentally skip over some elements or create an infinite loop.
MDN documentation should always note if a method returns a live collection or a static one.

For example, a NodeList offers some iteration methods such as forEach in modern browsers:

document.querySelectorAll(".myElement")
  .forEach((element) => element.style.size = "100px");

A simple for loop can also be used:

var elements = document.getElementsByClassName("myElement");

for(var i = 0; i < elements.length; ++i){
  elements[i].style.size = "100px";
}

Aside: .childNodes yields a live NodeList and .children yields a live HTMLCollection, so these two getters also need to be handled carefully.

There are some libraries like jQuery which make DOM querying a bit shorter and create a layer of abstraction over “one element” and “a collection of elements”:

$(".myElement").css("size", "100px");

Accepted Answer