utility $$

View source on GitHub →

$$(cssRule...) → [Element…]

Takes an arbitrary number of CSS selectors (strings) and returns a document-order array of extended DOM elements that match any of them.

Sometimes the usual tools from your DOM arsenal -- document.getElementById encapsulated by $, getElementsByTagName and even Prototype's very own getElementsByClassName extensions -- just aren't enough to quickly find elements or collections of elements. If you know the DOM tree structure, you can simply resort to CSS selectors to get the job done.

Quick examples
// -> all DIVs in the document. Same as document.getElementsByTagName('div').
// Nice addition, the elements you're getting back are already extended!
// -> same as $('contents'), only it returns an array anyway (even though IDs must
// be unique within a document).
// -> all LI elements with class 'faux'

The $$ function searches the entire document. For selector queries on more specific sections of a document, use Element.select.

Supported CSS syntax

The $$ function does not rely on the browser's internal CSS parsing capabilities (otherwise, we'd be in cross-browser trouble...), and therefore offers a consistent set of selectors across all supported browsers.

Supported in v1.5.0
  • Type selector: tag names, as in div.
  • Descendant selector: the space(s) between other selectors, as in #a li.
  • Attribute selectors: the full CSS 2.1 set of [attr], [attr=value], [attr~=value] and [attr|=value]. It also supports [attr!=value]. If the value you're matching against includes a space, be sure to enclose the value in quotation marks ([title="Hello World!"]).
  • Class selector: CSS class names, as in .highlighted or .example.wrong.
  • ID selector: as in #item1.
Supported from v1.5.1

Virtually all of CSS3 is supported, with the exception of pseudo-elements (like ::first-letter) and some pseudo-classes (like :hover). Some examples of new selectors that can be used in 1.5.1:

  • Child selector: selects immediate descendants, as in #a > li.
  • Attribute selectors: all attribute operators are supported, including ~= (matches part of a space-delimited attribute value, like rel or class); ^= (matches the beginning of a value); $= (matches the end of a value); and *= (matches any part of the value).
  • The :not pseudo-class, as in #a *:not(li) (matches all descendants of #a that aren't LIs).
  • All the :nth, :first, and :last pseudo-classes. Examples include tr:nth-child(even) (all even table rows), li:first-child (the first item in any list), or p:nth-last-of-type(3) (the third-to-last paragraph on the page).
  • The :empty pseudo-class (for selecting elements without children or text content).
  • The :enabled, :disabled, and :checked pseudo-classes (for use with form controls).
$$('#contents a[rel]');
// -> all links inside the element of ID "contents" with a rel attribute
// -> all links with a href attribute of value "#" (eyeew!)
 $$('#navbar a', '#sidebar a');
// -> all links within the elements of ID "navbar" or "sidebar"

With version 1.5.1 and above you can do various types of advanced selectors:

// -> all links, excluding those whose rel attribute contains the word "nofollow"
 $$('table tbody > tr:nth-child(even)');
// -> all even rows within all table bodies
// -> all DIVs without content (i.e., whitespace-only)