loading...

Use $ & $$ Instead of document.querySelector/All in JavaScript without jQuery

mrahmadawais profile image Ahmad Awais ⚡️ ・1 min read

I started writing JavaScript code because of WordPress back in 2007. JavaScript was not what it is today. It would behave differently on different browsers and there were too many hacks around basic stuff to keep up with all the changes all the time.

So, many of us decided to rely on jQuery — a simple JavaScript library with one single form of syntax that just worked everywhere in all the browsers.

Fast forward to 2019, as a full-time JavaScript Developer Advocate — I advocate modern JavaScript. Because it's super awesome. Though at times I miss the simplicity of jQuery in things like you'd just need a $ sign to pick up an element and do stuff to it.

Now with JavaScript, I find myself doing document.querySelector multiple times in an application. Well, guess what, there's an easy way to just bind that $ sign to your document's document.querySelector.

gif

Here's how you go about it.

const $ = document.querySelector.bind(document);
const $$ = document.querySelectorAll.bind(document);

Now you can use the following:


// Change the background color of a class.
$('.class').style.background="#BADA55";

// Change the inner HTML of an ID.
$('#id').innerHTML="<span>Cool beans!</span>";

// Select all images on the webpage.
$$('img')

// Print the image addresses for all the images on a webpage.
$$('img').forEach(img => console.log(img.src))

Use your code for good and have fun! :)
Peace! ✌️

Posted on by:

mrahmadawais profile

Ahmad Awais ⚡️

@mrahmadawais

Edutianer VSCode.pro ❯ Google Dev Expert ❯ Award-winning Dev Advocate ❯ Node.js Community Committee Member ❯ WordPress Core Dev ❯ TEDx Speaker ❯ Hilarious tech comedian! ❯ Loves @MaedahBatool & 🍕

Discussion

markdown guide
 

If you work on a distributed team or with legacy code, and considering the potential for conflict (mostly with jQuery but some other libraries, and any code where someone may have used $ for document.getElementById(), as used to be common), I think it'd be a good idea to use trivially longer identifiers like this:

const $q = document.querySelector.bind(document);
const $qa = document.querySelectorAll.bind(document);

This is similar to a common method of using jQuery in noConflict mode, e.g.:

const $j = jQuery.noConflict();

It could be judiciously extended for similar methods if you use them often enough.

 

I have been using
$ = function(e) {return document.getElementById(e)}
for years. Now I started using
$ = function(e) {
switch (e[0]) {
case '.': return document.getElementsByClassName(e.substring(1, e.length)); break;
case '#': return document.getElementById(e.substring(1, e.length)); break;
default: return document.getElementsByTagName(e);
}
}

 

Watch out that querySelectorAll returns a NodeList object, whereas $$ in the console returns an array.

So if you're used to call all the sweet array prototype methods, you might fall short!

Solution: wrap everything in Array.from.

 
 

Chrome has its own Console Utilities API Reference where $ and $$ are offered regardless, among other utils, unless the global scope has been overwritten with something else (i.e. jQuery).

The API for $ and $$ is more jQuery-ish, and if you'd like to reproduce it its like:

const $ = (css, parent = document) => parent.querySelector(css);
const $$ = (css, parent = document) => parent.querySelectorAll(css);

To avoid needing polyfills for forEach or Symbol.iterator on NodeList collections, you can also go ahead and convert the static collection as Array so that all methods, including filter and map can be used.

const $$ = (css, parent = document) =>
  Array.from(parent.querySelectorAll(css));

If you don't have Array.from in your debug session, or you are after a thin layer that works across browsers, you can go for the following:

function $(css, parent) {
  return (parent || document).querySelector(css);
}

function $$(css, parent) {
  var nodes = (parent || document).querySelectorAll(css);
  return Array.prototype.slice.call(nodes, 0);
}
 

This is my take on it. It checks whether I'm asking for an id or not, and if I am, it returns just that one element. If not, the NodeList is made into an array.

const $ = (q, d = document) => 
  /#\S+$/.test(q)                       // check if query asks for ID
  ? d.querySelector.bind(d)(q)          // if so, return one element
  : [...d.querySelectorAll.bind(d)(q)]  // else, return all elements in an array.
 

Instead of using 2 different constants, what about checking the length of the selector first like this:

const $ = s => document.querySelectorAll.bind(document)(s).length > 1 ? document.querySelectorAll.bind(document)(s) : document.querySelector.bind(document)(s);

or something like that to make it much more similar to jQuery $?

 

From a performance perspective, that would double the amount of work for every selector. Going from jQuery to Vanilla JS is mostly about being more performant, so that looks like an anti-pattern.

 

Can you explain how would that be the double amount of work? It's just one extra call.

It's 2 calls instead of one, so it's double.

 

Ugh, I had to replace jQuery in our app with a work-a-like jQuery that I wrote. It was THE WORST, because I had to learn jQuery, then I learned all about jQuery's internals, then I had to figure out the subset of jQuery that we depended upon that did not have all the internal state that jQuery had (did everyone know jQuery has a ton of internal state?), then I had to put up with all the upset other JavaScript developers who complained about my not-jQuery not behaving like jQuery when they added new code that relied on jQuery-isms that I hadn't implemented because the prior versions of the code hadn't relied upon those facilities (so I didn't bother to implement them).

What a royal pain in the keester.

I wish I had the foresight to just do your $ and $$ bindings and tell all the other devs to suck it up. Your solution is GENIUS.

 

Could you explain the purpose of creating your jQuery-like library? I don't understand what you gained, compared to just continuing to use jQuery.

 

The lawyers at my company would not allow us to ship or use third party libraries.

 

I'm using document.querySelectorAll for all elements. Even for id elements.

Actually, you can use querySelectorAll for id elements. It will work for one element. There must not be multiple elements in a document that have the same id value.

const $ = document.querySelectorAll.bind(document)

const testEl = $("#test")

testEl.forEach(el => {
    console.log(el)
})
 
const $ = (selector, base = document) => {
   let elements = base.querySelectorAll(selector);
   return (elements.length == 1) ? elements[0] : elements;
}
 

I'm interested in your thoughts a little deeper on this:

"I advocate modern JavaScript. Because it's super awesome." - Why would you advocate modern JavaScript ahead of jQuery beyond it being awesome. What do you feel is the main selling point is that can't be accomplished in other ways?

 

This looks cool.

with you example could you still do this though?

const username = 'johndoe'
const age = 32

console.log(`The name of the user is ${username} and his age is ${age}.`)

 

Yep, it would look like this:

const username = 'johndoe'
const age = 32

console.log(`The name of the user is ${username} and his age is ${age}.`)
console.log(`Element is: ${$('#article-reaction-actions')}.`)

The outer $ in the template string will start the evaluation. the $ inside of ${} will refer to the global $.

Hope this clears it up!

 

now that I think about it, I guess this would be a scope issue

 

Things like this are AMAZING so long as you don't have to support any legacy systems. It's nice to see that the feature set built by jQuery is essentially built into the browser directly now, and a simple wrapper gives us similar functionality.

 

I already did this in some projects that's cool !
But one point to take care is about performance.
Because querySelector do a lot more jobs thant getElementsById or class so if you dont care about perf you solution is perfect

 
 

one question.
i always did it like this:

const $ = s => document.querySelector(s)

what are the differences?

 
 

I made a module exactly for these purposes just recently: gimme back my dollars! It has a very distinctive name.

 

Great tip Ahmad.

Also, in Chrome dev tools, $ and $$ are already bound by default to querySelector and querySelectorAll respectively.

 

Isn't $ and $$ already binded to querySelector and querySelectorAll from the start though? Which means that binding it again seems unnecessary.

or I suppose the default binding is recent..?

 

...and finally we returned to jQuery

 
 

Great tip Ahmad.
Thank you for sharing.