select DOM elements by ID, tag name, or class name

annotated.js

// inspired by @techpriester: https://gist.github.com/988627

function(
  a,                         // take a simple selector like "name", "#name", or ".name", and
  b                          // an optional context, and
){
  a = a.match(/^(\W)?(.*)/); // split the selector into name and symbol.
  return(                    // return an element or list, from within the scope of
    b                        // the passed context
    || document              // or document,
  )[
    "getElement" + (         // obtained by the appropriate method calculated by
      a[1]
        ? a[1] == "#"      
          ? "ById"           // the node by ID,
          : "sByClassName"   // the nodes by class name, or
        : "sByTagName"       // the nodes by tag name,
    )
  ](
    a[2]                     // called with the name.
  )
}

index.js

function(a,b){a=a.match(/^(\W)?(.*)/);return(b||document)["getElement"+(a[1]?a[1]=="#"?"ById":"sByClassName":"sByTagName")](a[2])}

LICENSE.txt

            DO WHAT THE FUCK YOU WANT TO PUBLIC LICENSE
                    Version 2, December 2004

 Copyright (C) 2011 Jed Schmidt <http://jed.is>

 Everyone is permitted to copy and distribute verbatim or modified
 copies of this license document, and changing it is allowed as long
 as the name is changed.

            DO WHAT THE FUCK YOU WANT TO PUBLIC LICENSE
   TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION

  0. You just DO WHAT THE FUCK YOU WANT TO.

package.json

{
  "name": "cssSelect",

  "description": "select DOM elements by ID, tag name, or class name",

  "keywords": [
    "dom",
    "css",
    "id",
    "class",
    "tag"
  ]
}

Seems like that will fail on .foo-bar. You should .shift() off the indicator, and then use a.join('') instead of a[1]. More bytes, but more correct.

Good point. Fixed for a total cost of 1 byte.

Btw.: The ID selector will only work on document.

Another btw.: why use match when exec will do the same? How does the saying go: a byte saved...

Other than that: older IEs don't support .getElementsByClassName. That's gonna be one for the wishlist, ain't it?

@atk yep that getElementsByClassName just bit me while working on a Windows Gadget. I'm going to put a check around my copy, make it return null or something. Crazy thing is, I was selecting by id "#", so how did it end up with that by class name function anyways. Hmm...

I ran a few tests. Turns out the logic to decide which getElement(s)By method to use, wrongly drops to "sByClassName" for # selector.

If we isolated the logic into a function called "m". Then the function would return:

m('#foo')
getElementsByClassName

m('.foo')
getElementsByClassName

m('foo')
getElementsByTagName

It works well for class and tag selectors but erroneously returns the class-getter for an id selector.

My proposed fix just adds an explicit check for the class selector:

"getElement" + (         // obtained by the appropriate method calculated by
    a[1] == "#"      
      ? "ById"           // the node by ID,
      : a[1] == '.' 
        ? "sByClassName" // the nodes by class name, or
        : "sByTagName"   // the nodes by tag name
)

Isolating this into a method, "m", and testing it yields the following results:

m('#foo')
getElementById

m('.foo')
getElementsByClassName

m('foo')
getElementsByTagName

Tests/specs around this snippet would have easily caught this. Let's add them.

I hope you merge this fix in. And thanks for the share @jed, this is a sweet little helper!

For anyone who really wants/needs the "getElementsByClassName" support in IE7 or below (really? why!?! Oh you are working on a Windows Gadget you say, ok then... but the blood of innocent web standards is on your head) check out http://ejohn.org/blog/getelementsbyclassname-speed-comparison/

thanks @rudylattae. looks like some removed parens around (b=a[1]) caused a bad order of operations. it's fixed, so thanks for pointing this out!

Awesome, the fix works brilliantly!

I would advice you about that IE<9 doesn't support getElementsByClassName (source: caniuse.com) contrary to rudylattae comment.

You can save another byte by removing ^. Due to how the regular expressions is executed this does not make a difference in your case. And by the way, I tried to create getElementsByClassName in 140 bytes, which is missing in Internet Explorer <9.0, but currently it still is 162 bytes.