We Keep Coding

How to generate trillions of random IDs quickly [duplicate]

How do I create GUIDs (globally-unique identifiers) in JavaScript? The GUID / UUID should be at least 32 characters and should stay in the ASCII range to avoid trouble when passing them around.
I'm not sure what routines are available on all browsers, how "random" and seeded the built-in random number generator is, etc.

[Edited 2021-10-16 to reflect latest best-practices for producing RFC4122-compliant UUIDs]
Most readers here will want to use the uuid module. It is well-tested and supported.
The crypto.randomUUID() function is an emerging standard that is supported in Node.js and an increasing number of browsers. However because new browser APIs are restricted to secure contexts this method is only available to pages served locally (localhost or 127.0.0.1) or over HTTPS. If you're interested in seeing this restriction lifted for crypto.randomUUID() you can follow this GitHub issue.
If neither of those work for you, there is this method (based on the original answer to this question):
function uuidv4() {
return ([1e7]+-1e3+-4e3+-8e3+-1e11).replace(/[018]/g, c =>
(c ^ crypto.getRandomValues(new Uint8Array(1))[0] & 15 >> c / 4).toString(16)
);
}
console.log(uuidv4());
Note: The use of any UUID generator that relies on Math.random() is strongly discouraged (including snippets featured in previous versions of this answer) for reasons best explained here. TL;DR: solutions based on Math.random() do not provide good uniqueness guarantees.

UUIDs (Universally Unique IDentifier), also known as GUIDs (Globally Unique IDentifier), according to RFC 4122, are identifiers designed to provide certain uniqueness guarantees.
While it is possible to implement RFC-compliant UUIDs in a few lines of JavaScript code (e.g., see #broofa's answer, below) there are several common pitfalls:
Invalid id format (UUIDs must be of the form "xxxxxxxx-xxxx-Mxxx-Nxxx-xxxxxxxxxxxx", where x is one of [0-9, a-f] M is one of [1-5], and N is [8, 9, a, or b]
Use of a low-quality source of randomness (such as Math.random)
Thus, developers writing code for production environments are encouraged to use a rigorous, well-maintained implementation such as the uuid module.

I really like how clean Broofa's answer is, but it's unfortunate that poor implementations of Math.random leave the chance for collision.
Here's a similar RFC4122 version 4 compliant solution that solves that issue by offsetting the first 13 hex numbers by a hex portion of the timestamp, and once depleted offsets by a hex portion of the microseconds since pageload. That way, even if Math.random is on the same seed, both clients would have to generate the UUID the exact same number of microseconds since pageload (if high-perfomance time is supported) AND at the exact same millisecond (or 10,000+ years later) to get the same UUID:
function generateUUID() { // Public Domain/MIT
var d = new Date().getTime();//Timestamp
var d2 = ((typeof performance !== 'undefined') && performance.now && (performance.now()*1000)) || 0;//Time in microseconds since page-load or 0 if unsupported
return 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx'.replace(/[xy]/g, function(c) {
var r = Math.random() * 16;//random number between 0 and 16
if(d > 0){//Use timestamp until depleted
r = (d + r)%16 | 0;
d = Math.floor(d/16);
} else {//Use microseconds since page-load if supported
r = (d2 + r)%16 | 0;
d2 = Math.floor(d2/16);
}
return (c === 'x' ? r : (r & 0x3 | 0x8)).toString(16);
});
}
var onClick = function(){
document.getElementById('uuid').textContent = generateUUID();
}
onClick();
#uuid { font-family: monospace; font-size: 1.5em; }
<p id="uuid"></p>
<button id="generateUUID" onclick="onClick();">Generate UUID</button>
Here's a fiddle to test.
Modernized snippet for ES6
const generateUUID = () => {
let
d = new Date().getTime(),
d2 = ((typeof performance !== 'undefined') && performance.now && (performance.now() * 1000)) || 0;
return 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx'.replace(/[xy]/g, c => {
let r = Math.random() * 16;
if (d > 0) {
r = (d + r) % 16 | 0;
d = Math.floor(d / 16);
} else {
r = (d2 + r) % 16 | 0;
d2 = Math.floor(d2 / 16);
}
return (c == 'x' ? r : (r & 0x7 | 0x8)).toString(16);
});
};
const onClick = (e) => document.getElementById('uuid').textContent = generateUUID();
document.getElementById('generateUUID').addEventListener('click', onClick);
onClick();
#uuid { font-family: monospace; font-size: 1.5em; }
<p id="uuid"></p>
<button id="generateUUID">Generate UUID</button>

broofa's answer is pretty slick, indeed - impressively clever, really... RFC4122 compliant, somewhat readable, and compact. Awesome!
But if you're looking at that regular expression, those many replace() callbacks, toString()'s and Math.random() function calls (where he's only using four bits of the result and wasting the rest), you may start to wonder about performance. Indeed, joelpt even decided to toss out an RFC for generic GUID speed with generateQuickGUID.
But, can we get speed and RFC compliance? I say, YES! Can we maintain readability? Well... Not really, but it's easy if you follow along.
But first, my results, compared to broofa, guid (the accepted answer), and the non-rfc-compliant generateQuickGuid:
Desktop Android
broofa: 1617ms 12869ms
e1: 636ms 5778ms
e2: 606ms 4754ms
e3: 364ms 3003ms
e4: 329ms 2015ms
e5: 147ms 1156ms
e6: 146ms 1035ms
e7: 105ms 726ms
guid: 962ms 10762ms
generateQuickGuid: 292ms 2961ms
- Note: 500k iterations, results will vary by browser/CPU.
So by my 6th iteration of optimizations, I beat the most popular answer by over 12 times, the accepted answer by over 9 times, and the fast-non-compliant answer by 2-3 times. And I'm still RFC 4122 compliant.
Interested in how? I've put the full source on http://jsfiddle.net/jcward/7hyaC/3/ and on https://jsben.ch/xczxS
For an explanation, let's start with broofa's code:
function broofa() {
return 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx'.replace(/[xy]/g, function(c) {
var r = Math.random()*16|0, v = c == 'x' ? r : (r&0x3|0x8);
return v.toString(16);
});
}
console.log(broofa())
So it replaces x with any random hexadecimal digit, y with random data (except forcing the top two bits to 10 per the RFC spec), and the regex doesn't match the - or 4 characters, so he doesn't have to deal with them. Very, very slick.
The first thing to know is that function calls are expensive, as are regular expressions (though he only uses 1, it has 32 callbacks, one for each match, and in each of the 32 callbacks it calls Math.random() and v.toString(16)).
The first step toward performance is to eliminate the RegEx and its callback functions and use a simple loop instead. This means we have to deal with the - and 4 characters whereas broofa did not. Also, note that we can use String Array indexing to keep his slick String template architecture:
function e1() {
var u='',i=0;
while(i++<36) {
var c='xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx'[i-1],r=Math.random()*16|0,v=c=='x'?r:(r&0x3|0x8);
u+=(c=='-'||c=='4')?c:v.toString(16)
}
return u;
}
console.log(e1())
Basically, the same inner logic, except we check for - or 4, and using a while loop (instead of replace() callbacks) gets us an almost 3X improvement!
The next step is a small one on the desktop but makes a decent difference on mobile. Let's make fewer Math.random() calls and utilize all those random bits instead of throwing 87% of them away with a random buffer that gets shifted out each iteration. Let's also move that template definition out of the loop, just in case it helps:
function e2() {
var u='',m='xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx',i=0,rb=Math.random()*0xffffffff|0;
while(i++<36) {
var c=m[i-1],r=rb&0xf,v=c=='x'?r:(r&0x3|0x8);
u+=(c=='-'||c=='4')?c:v.toString(16);rb=i%8==0?Math.random()*0xffffffff|0:rb>>4
}
return u
}
console.log(e2())
This saves us 10-30% depending on platform. Not bad. But the next big step gets rid of the toString function calls altogether with an optimization classic - the look-up table. A simple 16-element lookup table will perform the job of toString(16) in much less time:
function e3() {
var h='0123456789abcdef';
var k='xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx';
/* same as e4() below */
}
function e4() {
var h=['0','1','2','3','4','5','6','7','8','9','a','b','c','d','e','f'];
var k=['x','x','x','x','x','x','x','x','-','x','x','x','x','-','4','x','x','x','-','y','x','x','x','-','x','x','x','x','x','x','x','x','x','x','x','x'];
var u='',i=0,rb=Math.random()*0xffffffff|0;
while(i++<36) {
var c=k[i-1],r=rb&0xf,v=c=='x'?r:(r&0x3|0x8);
u+=(c=='-'||c=='4')?c:h[v];rb=i%8==0?Math.random()*0xffffffff|0:rb>>4
}
return u
}
console.log(e4())
The next optimization is another classic. Since we're only handling four bits of output in each loop iteration, let's cut the number of loops in half and process eight bits in each iteration. This is tricky since we still have to handle the RFC compliant bit positions, but it's not too hard. We then have to make a larger lookup table (16x16, or 256) to store 0x00 - 0xFF, and we build it only once, outside the e5() function.
var lut = []; for (var i=0; i<256; i++) { lut[i] = (i<16?'0':'')+(i).toString(16); }
function e5() {
var k=['x','x','x','x','-','x','x','-','4','x','-','y','x','-','x','x','x','x','x','x'];
var u='',i=0,rb=Math.random()*0xffffffff|0;
while(i++<20) {
var c=k[i-1],r=rb&0xff,v=c=='x'?r:(c=='y'?(r&0x3f|0x80):(r&0xf|0x40));
u+=(c=='-')?c:lut[v];rb=i%4==0?Math.random()*0xffffffff|0:rb>>8
}
return u
}
console.log(e5())
I tried an e6() that processes 16-bits at a time, still using the 256-element LUT, and it showed the diminishing returns of optimization. Though it had fewer iterations, the inner logic was complicated by the increased processing, and it performed the same on desktop, and only ~10% faster on mobile.
The final optimization technique to apply - unroll the loop. Since we're looping a fixed number of times, we can technically write this all out by hand. I tried this once with a single random variable, r, that I kept reassigning, and performance tanked. But with four variables assigned random data up front, then using the lookup table, and applying the proper RFC bits, this version smokes them all:
var lut = []; for (var i=0; i<256; i++) { lut[i] = (i<16?'0':'')+(i).toString(16); }
function e7()
{
var d0 = Math.random()*0xffffffff|0;
var d1 = Math.random()*0xffffffff|0;
var d2 = Math.random()*0xffffffff|0;
var d3 = Math.random()*0xffffffff|0;
return lut[d0&0xff]+lut[d0>>8&0xff]+lut[d0>>16&0xff]+lut[d0>>24&0xff]+'-'+
lut[d1&0xff]+lut[d1>>8&0xff]+'-'+lut[d1>>16&0x0f|0x40]+lut[d1>>24&0xff]+'-'+
lut[d2&0x3f|0x80]+lut[d2>>8&0xff]+'-'+lut[d2>>16&0xff]+lut[d2>>24&0xff]+
lut[d3&0xff]+lut[d3>>8&0xff]+lut[d3>>16&0xff]+lut[d3>>24&0xff];
}
console.log(e7())
Modualized: http://jcward.com/UUID.js - UUID.generate()
The funny thing is, generating 16 bytes of random data is the easy part. The whole trick is expressing it in string format with RFC compliance, and it's most tightly accomplished with 16 bytes of random data, an unrolled loop and lookup table.
I hope my logic is correct -- it's very easy to make a mistake in this kind of tedious bit work. But the outputs look good to me. I hope you enjoyed this mad ride through code optimization!
Be advised: my primary goal was to show and teach potential optimization strategies. Other answers cover important topics such as collisions and truly random numbers, which are important for generating good UUIDs.

Use:
let uniqueId = Date.now().toString(36) + Math.random().toString(36).substring(2);
document.getElementById("unique").innerHTML =
Math.random().toString(36).substring(2) + (new Date()).getTime().toString(36);
<div id="unique">
</div>
If IDs are generated more than 1 millisecond apart, they are 100% unique.
If two IDs are generated at shorter intervals, and assuming that the random method is truly random, this would generate IDs that are 99.99999999999999% likely to be globally unique (collision in 1 of 10^15).
You can increase this number by adding more digits, but to generate 100% unique IDs you will need to use a global counter.
If you need RFC compatibility, this formatting will pass as a valid version 4 GUID:
let u = Date.now().toString(16) + Math.random().toString(16) + '0'.repeat(16);
let guid = [u.substr(0,8), u.substr(8,4), '4000-8' + u.substr(13,3), u.substr(16,12)].join('-');
let u = Date.now().toString(16)+Math.random().toString(16)+'0'.repeat(16);
let guid = [u.substr(0,8), u.substr(8,4), '4000-8' + u.substr(13,3), u.substr(16,12)].join('-');
document.getElementById("unique").innerHTML = guid;
<div id="unique">
</div>
The above code follow the intention, but not the letter of the RFC. Among other discrepancies it's a few random digits short. (Add more random digits if you need it) The upside is that this is really fast :)
You can test validity of your GUID here

Here's some code based on RFC 4122, section 4.4 (Algorithms for Creating a UUID from Truly Random or Pseudo-Random Number).
function createUUID() {
// http://www.ietf.org/rfc/rfc4122.txt
var s = [];
var hexDigits = "0123456789abcdef";
for (var i = 0; i < 36; i++) {
s[i] = hexDigits.substr(Math.floor(Math.random() * 0x10), 1);
}
s[14] = "4"; // bits 12-15 of the time_hi_and_version field to 0010
s[19] = hexDigits.substr((s[19] & 0x3) | 0x8, 1); // bits 6-7 of the clock_seq_hi_and_reserved to 01
s[8] = s[13] = s[18] = s[23] = "-";
var uuid = s.join("");
return uuid;
}

This is the fastest GUID-like string generator method in the format XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX. It does not generate a standard-compliant GUID.
Ten million executions of this implementation take just 32.5 seconds, which is the fastest I've ever seen in a browser (the only solution without loops/iterations).
The function is as simple as:
/**
* Generates a GUID string.
* #returns {string} The generated GUID.
* #example af8a8416-6e18-a307-bd9c-f2c947bbb3aa
* #author Slavik Meltser.
* #link http://slavik.meltser.info/?p=142
*/
function guid() {
function _p8(s) {
var p = (Math.random().toString(16)+"000000000").substr(2,8);
return s ? "-" + p.substr(0,4) + "-" + p.substr(4,4) : p ;
}
return _p8() + _p8(true) + _p8(true) + _p8();
}
To test the performance, you can run this code:
console.time('t');
for (var i = 0; i < 10000000; i++) {
guid();
};
console.timeEnd('t');
I'm sure most of you will understand what I did there, but maybe there is at least one person that will need an explanation:
The algorithm:
The Math.random() function returns a decimal number between 0 and 1 with 16 digits after the decimal fraction point (for
example 0.4363923368509859).
Then we take this number and convert
it to a string with base 16 (from the example above we'll get
0.6fb7687f).
Math.random().toString(16).
Then we cut off the 0. prefix (0.6fb7687f =>
6fb7687f) and get a string with eight hexadecimal
characters long.
(Math.random().toString(16).substr(2,8).
Sometimes the Math.random() function will return
shorter number (for example 0.4363), due to zeros at the end (from the example above, actually the number is 0.4363000000000000). That's why I'm appending to this string "000000000" (a string with nine zeros) and then cutting it off with substr() function to make it nine characters exactly (filling zeros to the right).
The reason for adding exactly nine zeros is because of the worse case scenario, which is when the Math.random() function will return exactly 0 or 1 (probability of 1/10^16 for each one of them). That's why we needed to add nine zeros to it ("0"+"000000000" or "1"+"000000000"), and then cutting it off from the second index (third character) with a length of eight characters. For the rest of the cases, the addition of zeros will not harm the result because it is cutting it off anyway.
Math.random().toString(16)+"000000000").substr(2,8).
The assembly:
The GUID is in the following format XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX.
I divided the GUID into four pieces, each piece divided into two types (or formats): XXXXXXXX and -XXXX-XXXX.
Now I'm building the GUID using these two types to assemble the GUID with call four pieces, as follows: XXXXXXXX -XXXX-XXXX -XXXX-XXXX XXXXXXXX.
To differ between these two types, I added a flag parameter to a pair creator function _p8(s), the s parameter tells the function whether to add dashes or not.
Eventually we build the GUID with the following chaining: _p8() + _p8(true) + _p8(true) + _p8(), and return it.
Link to this post on my blog
Enjoy! :-)

Here is a totally non-compliant but very performant implementation to generate an ASCII-safe GUID-like unique identifier.
function generateQuickGuid() {
return Math.random().toString(36).substring(2, 15) +
Math.random().toString(36).substring(2, 15);
}
Generates 26 [a-z0-9] characters, yielding a UID that is both shorter and more unique than RFC compliant GUIDs. Dashes can be trivially added if human-readability matters.
Here are usage examples and timings for this function and several of this question's other answers. The timing was performed under Chrome m25, 10 million iterations each.
>>> generateQuickGuid()
"nvcjf1hs7tf8yyk4lmlijqkuo9"
"yq6gipxqta4kui8z05tgh9qeel"
"36dh5sec7zdj90sk2rx7pjswi2"
runtime: 32.5s
>>> GUID() // John Millikin
"7a342ca2-e79f-528e-6302-8f901b0b6888"
runtime: 57.8s
>>> regexGuid() // broofa
"396e0c46-09e4-4b19-97db-bd423774a4b3"
runtime: 91.2s
>>> createUUID() // Kevin Hakanson
"403aa1ab-9f70-44ec-bc08-5d5ac56bd8a5"
runtime: 65.9s
>>> UUIDv4() // Jed Schmidt
"f4d7d31f-fa83-431a-b30c-3e6cc37cc6ee"
runtime: 282.4s
>>> Math.uuid() // broofa
"5BD52F55-E68F-40FC-93C2-90EE069CE545"
runtime: 225.8s
>>> Math.uuidFast() // broofa
"6CB97A68-23A2-473E-B75B-11263781BBE6"
runtime: 92.0s
>>> Math.uuidCompact() // broofa
"3d7b7a06-0a67-4b67-825c-e5c43ff8c1e8"
runtime: 229.0s
>>> bitwiseGUID() // jablko
"baeaa2f-7587-4ff1-af23-eeab3e92"
runtime: 79.6s
>>>> betterWayGUID() // Andrea Turri
"383585b0-9753-498d-99c3-416582e9662c"
runtime: 60.0s
>>>> UUID() // John Fowler
"855f997b-4369-4cdb-b7c9-7142ceaf39e8"
runtime: 62.2s
Here is the timing code.
var r;
console.time('t');
for (var i = 0; i < 10000000; i++) {
r = FuncToTest();
};
console.timeEnd('t');

From sagi shkedy's technical blog:
function generateGuid() {
var result, i, j;
result = '';
for(j=0; j<32; j++) {
if( j == 8 || j == 12 || j == 16 || j == 20)
result = result + '-';
i = Math.floor(Math.random()*16).toString(16).toUpperCase();
result = result + i;
}
return result;
}
There are other methods that involve using an ActiveX control, but stay away from these!
I thought it was worth pointing out that no GUID generator can guarantee unique keys (check the Wikipedia article). There is always a chance of collisions. A GUID simply offers a large enough universe of keys to reduce the change of collisions to almost nil.

Here is a combination of the top voted answer, with a workaround for Chrome's collisions:
generateGUID = (typeof(window.crypto) != 'undefined' &&
typeof(window.crypto.getRandomValues) != 'undefined') ?
function() {
// If we have a cryptographically secure PRNG, use that
// https://stackoverflow.com/questions/6906916/collisions-when-generating-uuids-in-javascript
var buf = new Uint16Array(8);
window.crypto.getRandomValues(buf);
var S4 = function(num) {
var ret = num.toString(16);
while(ret.length < 4){
ret = "0"+ret;
}
return ret;
};
return (S4(buf[0])+S4(buf[1])+"-"+S4(buf[2])+"-"+S4(buf[3])+"-"+S4(buf[4])+"-"+S4(buf[5])+S4(buf[6])+S4(buf[7]));
}
:
function() {
// Otherwise, just use Math.random
// https://stackoverflow.com/questions/105034/how-to-create-a-guid-uuid-in-javascript/2117523#2117523
return 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx'.replace(/[xy]/g, function(c) {
var r = Math.random()*16|0, v = c == 'x' ? r : (r&0x3|0x8);
return v.toString(16);
});
};
It is on jsbin if you want to test it.

Here's a solution dated Oct. 9, 2011 from a comment by user jed at https://gist.github.com/982883:
UUIDv4 = function b(a){return a?(a^Math.random()*16>>a/4).toString(16):([1e7]+-1e3+-4e3+-8e3+-1e11).replace(/[018]/g,b)}
This accomplishes the same goal as the current highest-rated answer, but in 50+ fewer bytes by exploiting coercion, recursion, and exponential notation. For those curious how it works, here's the annotated form of an older version of the function:
UUIDv4 =
function b(
a // placeholder
){
return a // if the placeholder was passed, return
? ( // a random number from 0 to 15
a ^ // unless b is 8,
Math.random() // in which case
* 16 // a random number from
>> a/4 // 8 to 11
).toString(16) // in hexadecimal
: ( // or otherwise a concatenated string:
[1e7] + // 10000000 +
-1e3 + // -1000 +
-4e3 + // -4000 +
-8e3 + // -80000000 +
-1e11 // -100000000000,
).replace( // replacing
/[018]/g, // zeroes, ones, and eights with
b // random hex digits
)
}

You can use node-uuid. It provides simple, fast generation of RFC4122 UUIDS.
Features:
Generate RFC4122 version 1 or version 4 UUIDs
Runs in Node.js and browsers.
Cryptographically strong random # generation on supporting platforms.
Small footprint (Want something smaller? Check this out!)
Install Using NPM:
npm install uuid
Or using uuid via a browser:
Download Raw File (uuid v1): https://raw.githubusercontent.com/kelektiv/node-uuid/master/v1.js
Download Raw File (uuid v4): https://raw.githubusercontent.com/kelektiv/node-uuid/master/v4.js
Want even smaller? Check this out: https://gist.github.com/jed/982883
Usage:
// Generate a v1 UUID (time-based)
const uuidV1 = require('uuid/v1');
uuidV1(); // -> '6c84fb90-12c4-11e1-840d-7b25c5ee775a'
// Generate a v4 UUID (random)
const uuidV4 = require('uuid/v4');
uuidV4(); // -> '110ec58a-a0f2-4ac4-8393-c866d813b8d1'
// Generate a v5 UUID (namespace)
const uuidV5 = require('uuid/v5');
// ... using predefined DNS namespace (for domain names)
uuidV5('hello.example.com', v5.DNS)); // -> 'fdda765f-fc57-5604-a269-52a7df8164ec'
// ... using predefined URL namespace (for, well, URLs)
uuidV5('http://example.com/hello', v5.URL); // -> '3bbcee75-cecc-5b56-8031-b6641c1ed1f1'
// ... using a custom namespace
const MY_NAMESPACE = '(previously generated unique uuid string)';
uuidV5('hello', MY_NAMESPACE); // -> '90123e1c-7512-523e-bb28-76fab9f2f73d'
ECMAScript 2015 (ES6):
import uuid from 'uuid/v4';
const id = uuid();

var uuid = function() {
var buf = new Uint32Array(4);
window.crypto.getRandomValues(buf);
var idx = -1;
return 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx'.replace(/[xy]/g, function(c) {
idx++;
var r = (buf[idx>>3] >> ((idx%8)*4))&15;
var v = c == 'x' ? r : (r&0x3|0x8);
return v.toString(16);
});
};
This version is based on Briguy37's answer and some bitwise operators to extract nibble sized windows from the buffer.
It should adhere to the RFC Type 4 (random) schema, since I had problems last time parsing non-compliant UUIDs with Java's UUID.

This creates a version 4 UUID (created from pseudo random numbers):
function uuid()
{
var chars = '0123456789abcdef'.split('');
var uuid = [], rnd = Math.random, r;
uuid[8] = uuid[13] = uuid[18] = uuid[23] = '-';
uuid[14] = '4'; // version 4
for (var i = 0; i < 36; i++)
{
if (!uuid[i])
{
r = 0 | rnd()*16;
uuid[i] = chars[(i == 19) ? (r & 0x3) | 0x8 : r & 0xf];
}
}
return uuid.join('');
}
Here is a sample of the UUIDs generated:
682db637-0f31-4847-9cdf-25ba9613a75c
97d19478-3ab2-4aa1-b8cc-a1c3540f54aa
2eed04c9-2692-456d-a0fd-51012f947136

One line solution using Blobs.
window.URL.createObjectURL(new Blob([])).substring(31);
The value at the end (31) depends on the length of the URL.
EDIT:
A more compact and universal solution, as suggested by rinogo:
URL.createObjectURL(new Blob([])).substr(-36);

Simple JavaScript module as a combination of best answers in this question.
var crypto = window.crypto || window.msCrypto || null; // IE11 fix
var Guid = Guid || (function() {
var EMPTY = '00000000-0000-0000-0000-000000000000';
var _padLeft = function(paddingString, width, replacementChar) {
return paddingString.length >= width ? paddingString : _padLeft(replacementChar + paddingString, width, replacementChar || ' ');
};
var _s4 = function(number) {
var hexadecimalResult = number.toString(16);
return _padLeft(hexadecimalResult, 4, '0');
};
var _cryptoGuid = function() {
var buffer = new window.Uint16Array(8);
crypto.getRandomValues(buffer);
return [_s4(buffer[0]) + _s4(buffer[1]), _s4(buffer[2]), _s4(buffer[3]), _s4(buffer[4]), _s4(buffer[5]) + _s4(buffer[6]) + _s4(buffer[7])].join('-');
};
var _guid = function() {
var currentDateMilliseconds = new Date().getTime();
return 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx'.replace(/[xy]/g, function(currentChar) {
var randomChar = (currentDateMilliseconds + Math.random() * 16) % 16 | 0;
currentDateMilliseconds = Math.floor(currentDateMilliseconds / 16);
return (currentChar === 'x' ? randomChar : (randomChar & 0x7 | 0x8)).toString(16);
});
};
var create = function() {
var hasCrypto = crypto != 'undefined' && crypto !== null,
hasRandomValues = typeof(window.crypto.getRandomValues) != 'undefined';
return (hasCrypto && hasRandomValues) ? _cryptoGuid() : _guid();
};
return {
newGuid: create,
empty: EMPTY
};
})();
// DEMO: Create and show GUID
console.log('1. New Guid: ' + Guid.newGuid());
// DEMO: Show empty GUID
console.log('2. Empty Guid: ' + Guid.empty);
Usage:
Guid.newGuid()
"c6c2d12f-d76b-5739-e551-07e6de5b0807"
Guid.empty
"00000000-0000-0000-0000-000000000000"

The version below is an adaptation of broofa's answer, but updated to include a "true" random function that uses crypto libraries where available, and the Alea() function as a fallback.
Math.log2 = Math.log2 || function(n){ return Math.log(n) / Math.log(2); }
Math.trueRandom = (function() {
var crypt = window.crypto || window.msCrypto;
if (crypt && crypt.getRandomValues) {
// If we have a crypto library, use it
var random = function(min, max) {
var rval = 0;
var range = max - min;
if (range < 2) {
return min;
}
var bits_needed = Math.ceil(Math.log2(range));
if (bits_needed > 53) {
throw new Exception("We cannot generate numbers larger than 53 bits.");
}
var bytes_needed = Math.ceil(bits_needed / 8);
var mask = Math.pow(2, bits_needed) - 1;
// 7776 -> (2^13 = 8192) -1 == 8191 or 0x00001111 11111111
// Create byte array and fill with N random numbers
var byteArray = new Uint8Array(bytes_needed);
crypt.getRandomValues(byteArray);
var p = (bytes_needed - 1) * 8;
for(var i = 0; i < bytes_needed; i++ ) {
rval += byteArray[i] * Math.pow(2, p);
p -= 8;
}
// Use & to apply the mask and reduce the number of recursive lookups
rval = rval & mask;
if (rval >= range) {
// Integer out of acceptable range
return random(min, max);
}
// Return an integer that falls within the range
return min + rval;
}
return function() {
var r = random(0, 1000000000) / 1000000000;
return r;
};
} else {
// From https://web.archive.org/web/20120502223108/http://baagoe.com/en/RandomMusings/javascript/
// Johannes Baagøe <baagoe#baagoe.com>, 2010
function Mash() {
var n = 0xefc8249d;
var mash = function(data) {
data = data.toString();
for (var i = 0; i < data.length; i++) {
n += data.charCodeAt(i);
var h = 0.02519603282416938 * n;
n = h >>> 0;
h -= n;
h *= n;
n = h >>> 0;
h -= n;
n += h * 0x100000000; // 2^32
}
return (n >>> 0) * 2.3283064365386963e-10; // 2^-32
};
mash.version = 'Mash 0.9';
return mash;
}
// From http://baagoe.com/en/RandomMusings/javascript/
function Alea() {
return (function(args) {
// Johannes Baagøe <baagoe#baagoe.com>, 2010
var s0 = 0;
var s1 = 0;
var s2 = 0;
var c = 1;
if (args.length == 0) {
args = [+new Date()];
}
var mash = Mash();
s0 = mash(' ');
s1 = mash(' ');
s2 = mash(' ');
for (var i = 0; i < args.length; i++) {
s0 -= mash(args[i]);
if (s0 < 0) {
s0 += 1;
}
s1 -= mash(args[i]);
if (s1 < 0) {
s1 += 1;
}
s2 -= mash(args[i]);
if (s2 < 0) {
s2 += 1;
}
}
mash = null;
var random = function() {
var t = 2091639 * s0 + c * 2.3283064365386963e-10; // 2^-32
s0 = s1;
s1 = s2;
return s2 = t - (c = t | 0);
};
random.uint32 = function() {
return random() * 0x100000000; // 2^32
};
random.fract53 = function() {
return random() +
(random() * 0x200000 | 0) * 1.1102230246251565e-16; // 2^-53
};
random.version = 'Alea 0.9';
random.args = args;
return random;
}(Array.prototype.slice.call(arguments)));
};
return Alea();
}
}());
Math.guid = function() {
return 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx'.replace(/[xy]/g, function(c) {
var r = Math.trueRandom() * 16 | 0,
v = c == 'x' ? r : (r & 0x3 | 0x8);
return v.toString(16);
});
};

JavaScript project on GitHub - https://github.com/LiosK/UUID.js
UUID.js The RFC-compliant UUID generator for JavaScript.
See RFC 4122 http://www.ietf.org/rfc/rfc4122.txt.
Features Generates RFC 4122 compliant UUIDs.
Version 4 UUIDs (UUIDs from random numbers) and version 1 UUIDs
(time-based UUIDs) are available.
UUID object allows a variety of access to the UUID including access to
the UUID fields.
Low timestamp resolution of JavaScript is compensated by random
numbers.

// RFC 4122
//
// A UUID is 128 bits long
//
// String representation is five fields of 4, 2, 2, 2, and 6 bytes.
// Fields represented as lowercase, zero-filled, hexadecimal strings, and
// are separated by dash characters
//
// A version 4 UUID is generated by setting all but six bits to randomly
// chosen values
var uuid = [
Math.random().toString(16).slice(2, 10),
Math.random().toString(16).slice(2, 6),
// Set the four most significant bits (bits 12 through 15) of the
// time_hi_and_version field to the 4-bit version number from Section
// 4.1.3
(Math.random() * .0625 /* 0x.1 */ + .25 /* 0x.4 */).toString(16).slice(2, 6),
// Set the two most significant bits (bits 6 and 7) of the
// clock_seq_hi_and_reserved to zero and one, respectively
(Math.random() * .25 /* 0x.4 */ + .5 /* 0x.8 */).toString(16).slice(2, 6),
Math.random().toString(16).slice(2, 14)].join('-');

Added in: v15.6.0, v14.17.0 there is a built-in crypto.randomUUID() function.
import * as crypto from "crypto";
const uuid = crypto.randomUUID();
In the browser, crypto.randomUUID() is currently supported in Chromium 92+ and Firefox 95+.

For those wanting an RFC 4122 version 4 compliant solution with speed considerations (few calls to Math.random()):
var rand = Math.random;
function UUID() {
var nbr, randStr = "";
do {
randStr += (nbr = rand()).toString(16).substr(3, 6);
} while (randStr.length < 30);
return (
randStr.substr(0, 8) + "-" +
randStr.substr(8, 4) + "-4" +
randStr.substr(12, 3) + "-" +
((nbr*4|0)+8).toString(16) + // [89ab]
randStr.substr(15, 3) + "-" +
randStr.substr(18, 12)
);
}
console.log( UUID() );
The above function should have a decent balance between speed and randomness.

I wanted to understand broofa's answer, so I expanded it and added comments:
var uuid = function () {
return 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx'.replace(
/[xy]/g,
function (match) {
/*
* Create a random nibble. The two clever bits of this code:
*
* - Bitwise operations will truncate floating point numbers
* - For a bitwise OR of any x, x | 0 = x
*
* So:
*
* Math.random * 16
*
* creates a random floating point number
* between 0 (inclusive) and 16 (exclusive) and
*
* | 0
*
* truncates the floating point number into an integer.
*/
var randomNibble = Math.random() * 16 | 0;
/*
* Resolves the variant field. If the variant field (delineated
* as y in the initial string) is matched, the nibble must
* match the mask (where x is a do-not-care bit):
*
* 10xx
*
* This is achieved by performing the following operations in
* sequence (where x is an intermediate result):
*
* - x & 0x3, which is equivalent to x % 3
* - x | 0x8, which is equivalent to x + 8
*
* This results in a nibble between 8 inclusive and 11 exclusive,
* (or 1000 and 1011 in binary), all of which satisfy the variant
* field mask above.
*/
var nibble = (match == 'y') ?
(randomNibble & 0x3 | 0x8) :
randomNibble;
/*
* Ensure the nibble integer is encoded as base 16 (hexadecimal).
*/
return nibble.toString(16);
}
);
};

ES6 sample
const guid=()=> {
const s4=()=> Math.floor((1 + Math.random()) * 0x10000).toString(16).substring(1);
return `${s4() + s4()}-${s4()}-${s4()}-${s4()}-${s4() + s4() + s4()}`;
}

I adjusted my own UUID/GUID generator with some extras here.
I'm using the following Kybos random number generator to be a bit more cryptographically sound.
Below is my script with the Mash and Kybos methods from baagoe.com excluded.
//UUID/Guid Generator
// use: UUID.create() or UUID.createSequential()
// convenience: UUID.empty, UUID.tryParse(string)
(function(w){
// From http://baagoe.com/en/RandomMusings/javascript/
// Johannes Baagøe <baagoe#baagoe.com>, 2010
//function Mash() {...};
// From http://baagoe.com/en/RandomMusings/javascript/
//function Kybos() {...};
var rnd = Kybos();
//UUID/GUID Implementation from http://frugalcoder.us/post/2012/01/13/javascript-guid-uuid-generator.aspx
var UUID = {
"empty": "00000000-0000-0000-0000-000000000000"
,"parse": function(input) {
var ret = input.toString().trim().toLowerCase().replace(/^[\s\r\n]+|[\{\}]|[\s\r\n]+$/g, "");
if ((/[a-f0-9]{8}\-[a-f0-9]{4}\-[a-f0-9]{4}\-[a-f0-9]{4}\-[a-f0-9]{12}/).test(ret))
return ret;
else
throw new Error("Unable to parse UUID");
}
,"createSequential": function() {
var ret = new Date().valueOf().toString(16).replace("-","")
for (;ret.length < 12; ret = "0" + ret);
ret = ret.substr(ret.length-12,12); //only least significant part
for (;ret.length < 32;ret += Math.floor(rnd() * 0xffffffff).toString(16));
return [ret.substr(0,8), ret.substr(8,4), "4" + ret.substr(12,3), "89AB"[Math.floor(Math.random()*4)] + ret.substr(16,3), ret.substr(20,12)].join("-");
}
,"create": function() {
var ret = "";
for (;ret.length < 32;ret += Math.floor(rnd() * 0xffffffff).toString(16));
return [ret.substr(0,8), ret.substr(8,4), "4" + ret.substr(12,3), "89AB"[Math.floor(Math.random()*4)] + ret.substr(16,3), ret.substr(20,12)].join("-");
}
,"random": function() {
return rnd();
}
,"tryParse": function(input) {
try {
return UUID.parse(input);
} catch(ex) {
return UUID.empty;
}
}
};
UUID["new"] = UUID.create;
w.UUID = w.Guid = UUID;
}(window || this));

The native URL.createObjectURL is generating an UUID. You can take advantage of this.
function uuid() {
const url = URL.createObjectURL(new Blob())
const [id] = url.toString().split('/').reverse()
URL.revokeObjectURL(url)
return id
}

The better way:
function(
a, b // Placeholders
){
for( // Loop :)
b = a = ''; // b - result , a - numeric variable
a++ < 36; //
b += a*51&52 // If "a" is not 9 or 14 or 19 or 24
? // return a random number or 4
(
a^15 // If "a" is not 15,
? // generate a random number from 0 to 15
8^Math.random() *
(a^20 ? 16 : 4) // unless "a" is 20, in which case a random number from 8 to 11,
:
4 // otherwise 4
).toString(16)
:
'-' // In other cases, (if "a" is 9,14,19,24) insert "-"
);
return b
}
Minimized:
function(a,b){for(b=a='';a++<36;b+=a*51&52?(a^15?8^Math.random()*(a^20?16:4):4).toString(16):'-');return b}

The following is simple code that uses crypto.getRandomValues(a) on supported browsers (Internet Explorer 11+, iOS 7+, Firefox 21+, Chrome, and Android Chrome).
It avoids using Math.random(), because that can cause collisions (for example 20 collisions for 4000 generated UUIDs in a real situation by Muxa).
function uuid() {
function randomDigit() {
if (crypto && crypto.getRandomValues) {
var rands = new Uint8Array(1);
crypto.getRandomValues(rands);
return (rands[0] % 16).toString(16);
} else {
return ((Math.random() * 16) | 0).toString(16);
}
}
var crypto = window.crypto || window.msCrypto;
return 'xxxxxxxx-xxxx-4xxx-8xxx-xxxxxxxxxxxx'.replace(/x/g, randomDigit);
}
Notes:
Optimised for code readability, not speed, so it is suitable for, say, a few hundred UUIDs per second. It generates about 10000 uuid() per second in Chromium on my laptop using http://jsbin.com/fuwigo/1 to measure performance.
It only uses 8 for "y" because that simplifies code readability (y is allowed to be 8, 9, A, or B).

If you just need a random 128 bit string in no particular format, you can use:
function uuid() {
return crypto.getRandomValues(new Uint32Array(4)).join('-');
}
Which will return something like 2350143528-4164020887-938913176-2513998651.

I couldn't find any answer that uses a single 16-octet TypedArray and a DataView, so I think the following solution for generating a version 4 UUID per the RFC will stand on its own here:
const uuid4 = () => {
const ho = (n, p) => n.toString(16).padStart(p, 0); /// Return the hexadecimal text representation of number `n`, padded with zeroes to be of length `p`
const data = crypto.getRandomValues(new Uint8Array(16)); /// Fill the buffer with random data
data[6] = (data[6] & 0xf) | 0x40; /// Patch the 6th byte to reflect a version 4 UUID
data[8] = (data[8] & 0x3f) | 0x80; /// Patch the 8th byte to reflect a variant 1 UUID (version 4 UUIDs are)
const view = new DataView(data.buffer); /// Create a view backed by a 16-byte buffer
return `${ho(view.getUint32(0), 8)}-${ho(view.getUint16(4), 4)}-${ho(view.getUint16(6), 4)}-${ho(view.getUint16(8), 4)}-${ho(view.getUint32(10), 8)}${ho(view.getUint16(14), 4)}`; /// Compile the canonical textual form from the array data
};
I prefer it because:
it only relies on functions available to the standard ECMAScript platform, where possible -- which is all but one procedure
it only uses a single buffer, minimizing copying of data, which should in theory yield performance advantages
At the time of writing this, getRandomValues is not something implemented for the crypto object in Node.js. However, it has the equivalent randomBytes function which may be used instead.

Just another more readable variant with just two mutations.
function uuid4()
{
function hex (s, b)
{
return s +
(b >>> 4 ).toString (16) + // high nibble
(b & 0b1111).toString (16); // low nibble
}
let r = crypto.getRandomValues (new Uint8Array (16));
r[6] = r[6] >>> 4 | 0b01000000; // Set type 4: 0100
r[8] = r[8] >>> 3 | 0b10000000; // Set variant: 100
return r.slice ( 0, 4).reduce (hex, '' ) +
r.slice ( 4, 6).reduce (hex, '-') +
r.slice ( 6, 8).reduce (hex, '-') +
r.slice ( 8, 10).reduce (hex, '-') +
r.slice (10, 16).reduce (hex, '-');
}

Is there a way to create a unique ID using Node.js without additional modules? [duplicate]

How do I create GUIDs (globally-unique identifiers) in JavaScript? The GUID / UUID should be at least 32 characters and should stay in the ASCII range to avoid trouble when passing them around.
I'm not sure what routines are available on all browsers, how "random" and seeded the built-in random number generator is, etc.

[Edited 2021-10-16 to reflect latest best-practices for producing RFC4122-compliant UUIDs]
Most readers here will want to use the uuid module. It is well-tested and supported.
The crypto.randomUUID() function is an emerging standard that is supported in Node.js and an increasing number of browsers. However because new browser APIs are restricted to secure contexts this method is only available to pages served locally (localhost or 127.0.0.1) or over HTTPS. If you're interested in seeing this restriction lifted for crypto.randomUUID() you can follow this GitHub issue.
If neither of those work for you, there is this method (based on the original answer to this question):
function uuidv4() {
return ([1e7]+-1e3+-4e3+-8e3+-1e11).replace(/[018]/g, c =>
(c ^ crypto.getRandomValues(new Uint8Array(1))[0] & 15 >> c / 4).toString(16)
);
}
console.log(uuidv4());
Note: The use of any UUID generator that relies on Math.random() is strongly discouraged (including snippets featured in previous versions of this answer) for reasons best explained here. TL;DR: solutions based on Math.random() do not provide good uniqueness guarantees.

UUIDs (Universally Unique IDentifier), also known as GUIDs (Globally Unique IDentifier), according to RFC 4122, are identifiers designed to provide certain uniqueness guarantees.
While it is possible to implement RFC-compliant UUIDs in a few lines of JavaScript code (e.g., see #broofa's answer, below) there are several common pitfalls:
Invalid id format (UUIDs must be of the form "xxxxxxxx-xxxx-Mxxx-Nxxx-xxxxxxxxxxxx", where x is one of [0-9, a-f] M is one of [1-5], and N is [8, 9, a, or b]
Use of a low-quality source of randomness (such as Math.random)
Thus, developers writing code for production environments are encouraged to use a rigorous, well-maintained implementation such as the uuid module.

I really like how clean Broofa's answer is, but it's unfortunate that poor implementations of Math.random leave the chance for collision.
Here's a similar RFC4122 version 4 compliant solution that solves that issue by offsetting the first 13 hex numbers by a hex portion of the timestamp, and once depleted offsets by a hex portion of the microseconds since pageload. That way, even if Math.random is on the same seed, both clients would have to generate the UUID the exact same number of microseconds since pageload (if high-perfomance time is supported) AND at the exact same millisecond (or 10,000+ years later) to get the same UUID:
function generateUUID() { // Public Domain/MIT
var d = new Date().getTime();//Timestamp
var d2 = ((typeof performance !== 'undefined') && performance.now && (performance.now()*1000)) || 0;//Time in microseconds since page-load or 0 if unsupported
return 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx'.replace(/[xy]/g, function(c) {
var r = Math.random() * 16;//random number between 0 and 16
if(d > 0){//Use timestamp until depleted
r = (d + r)%16 | 0;
d = Math.floor(d/16);
} else {//Use microseconds since page-load if supported
r = (d2 + r)%16 | 0;
d2 = Math.floor(d2/16);
}
return (c === 'x' ? r : (r & 0x3 | 0x8)).toString(16);
});
}
var onClick = function(){
document.getElementById('uuid').textContent = generateUUID();
}
onClick();
#uuid { font-family: monospace; font-size: 1.5em; }
<p id="uuid"></p>
<button id="generateUUID" onclick="onClick();">Generate UUID</button>
Here's a fiddle to test.
Modernized snippet for ES6
const generateUUID = () => {
let
d = new Date().getTime(),
d2 = ((typeof performance !== 'undefined') && performance.now && (performance.now() * 1000)) || 0;
return 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx'.replace(/[xy]/g, c => {
let r = Math.random() * 16;
if (d > 0) {
r = (d + r) % 16 | 0;
d = Math.floor(d / 16);
} else {
r = (d2 + r) % 16 | 0;
d2 = Math.floor(d2 / 16);
}
return (c == 'x' ? r : (r & 0x7 | 0x8)).toString(16);
});
};
const onClick = (e) => document.getElementById('uuid').textContent = generateUUID();
document.getElementById('generateUUID').addEventListener('click', onClick);
onClick();
#uuid { font-family: monospace; font-size: 1.5em; }
<p id="uuid"></p>
<button id="generateUUID">Generate UUID</button>

broofa's answer is pretty slick, indeed - impressively clever, really... RFC4122 compliant, somewhat readable, and compact. Awesome!
But if you're looking at that regular expression, those many replace() callbacks, toString()'s and Math.random() function calls (where he's only using four bits of the result and wasting the rest), you may start to wonder about performance. Indeed, joelpt even decided to toss out an RFC for generic GUID speed with generateQuickGUID.
But, can we get speed and RFC compliance? I say, YES! Can we maintain readability? Well... Not really, but it's easy if you follow along.
But first, my results, compared to broofa, guid (the accepted answer), and the non-rfc-compliant generateQuickGuid:
Desktop Android
broofa: 1617ms 12869ms
e1: 636ms 5778ms
e2: 606ms 4754ms
e3: 364ms 3003ms
e4: 329ms 2015ms
e5: 147ms 1156ms
e6: 146ms 1035ms
e7: 105ms 726ms
guid: 962ms 10762ms
generateQuickGuid: 292ms 2961ms
- Note: 500k iterations, results will vary by browser/CPU.
So by my 6th iteration of optimizations, I beat the most popular answer by over 12 times, the accepted answer by over 9 times, and the fast-non-compliant answer by 2-3 times. And I'm still RFC 4122 compliant.
Interested in how? I've put the full source on http://jsfiddle.net/jcward/7hyaC/3/ and on https://jsben.ch/xczxS
For an explanation, let's start with broofa's code:
function broofa() {
return 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx'.replace(/[xy]/g, function(c) {
var r = Math.random()*16|0, v = c == 'x' ? r : (r&0x3|0x8);
return v.toString(16);
});
}
console.log(broofa())
So it replaces x with any random hexadecimal digit, y with random data (except forcing the top two bits to 10 per the RFC spec), and the regex doesn't match the - or 4 characters, so he doesn't have to deal with them. Very, very slick.
The first thing to know is that function calls are expensive, as are regular expressions (though he only uses 1, it has 32 callbacks, one for each match, and in each of the 32 callbacks it calls Math.random() and v.toString(16)).
The first step toward performance is to eliminate the RegEx and its callback functions and use a simple loop instead. This means we have to deal with the - and 4 characters whereas broofa did not. Also, note that we can use String Array indexing to keep his slick String template architecture:
function e1() {
var u='',i=0;
while(i++<36) {
var c='xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx'[i-1],r=Math.random()*16|0,v=c=='x'?r:(r&0x3|0x8);
u+=(c=='-'||c=='4')?c:v.toString(16)
}
return u;
}
console.log(e1())
Basically, the same inner logic, except we check for - or 4, and using a while loop (instead of replace() callbacks) gets us an almost 3X improvement!
The next step is a small one on the desktop but makes a decent difference on mobile. Let's make fewer Math.random() calls and utilize all those random bits instead of throwing 87% of them away with a random buffer that gets shifted out each iteration. Let's also move that template definition out of the loop, just in case it helps:
function e2() {
var u='',m='xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx',i=0,rb=Math.random()*0xffffffff|0;
while(i++<36) {
var c=m[i-1],r=rb&0xf,v=c=='x'?r:(r&0x3|0x8);
u+=(c=='-'||c=='4')?c:v.toString(16);rb=i%8==0?Math.random()*0xffffffff|0:rb>>4
}
return u
}
console.log(e2())
This saves us 10-30% depending on platform. Not bad. But the next big step gets rid of the toString function calls altogether with an optimization classic - the look-up table. A simple 16-element lookup table will perform the job of toString(16) in much less time:
function e3() {
var h='0123456789abcdef';
var k='xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx';
/* same as e4() below */
}
function e4() {
var h=['0','1','2','3','4','5','6','7','8','9','a','b','c','d','e','f'];
var k=['x','x','x','x','x','x','x','x','-','x','x','x','x','-','4','x','x','x','-','y','x','x','x','-','x','x','x','x','x','x','x','x','x','x','x','x'];
var u='',i=0,rb=Math.random()*0xffffffff|0;
while(i++<36) {
var c=k[i-1],r=rb&0xf,v=c=='x'?r:(r&0x3|0x8);
u+=(c=='-'||c=='4')?c:h[v];rb=i%8==0?Math.random()*0xffffffff|0:rb>>4
}
return u
}
console.log(e4())
The next optimization is another classic. Since we're only handling four bits of output in each loop iteration, let's cut the number of loops in half and process eight bits in each iteration. This is tricky since we still have to handle the RFC compliant bit positions, but it's not too hard. We then have to make a larger lookup table (16x16, or 256) to store 0x00 - 0xFF, and we build it only once, outside the e5() function.
var lut = []; for (var i=0; i<256; i++) { lut[i] = (i<16?'0':'')+(i).toString(16); }
function e5() {
var k=['x','x','x','x','-','x','x','-','4','x','-','y','x','-','x','x','x','x','x','x'];
var u='',i=0,rb=Math.random()*0xffffffff|0;
while(i++<20) {
var c=k[i-1],r=rb&0xff,v=c=='x'?r:(c=='y'?(r&0x3f|0x80):(r&0xf|0x40));
u+=(c=='-')?c:lut[v];rb=i%4==0?Math.random()*0xffffffff|0:rb>>8
}
return u
}
console.log(e5())
I tried an e6() that processes 16-bits at a time, still using the 256-element LUT, and it showed the diminishing returns of optimization. Though it had fewer iterations, the inner logic was complicated by the increased processing, and it performed the same on desktop, and only ~10% faster on mobile.
The final optimization technique to apply - unroll the loop. Since we're looping a fixed number of times, we can technically write this all out by hand. I tried this once with a single random variable, r, that I kept reassigning, and performance tanked. But with four variables assigned random data up front, then using the lookup table, and applying the proper RFC bits, this version smokes them all:
var lut = []; for (var i=0; i<256; i++) { lut[i] = (i<16?'0':'')+(i).toString(16); }
function e7()
{
var d0 = Math.random()*0xffffffff|0;
var d1 = Math.random()*0xffffffff|0;
var d2 = Math.random()*0xffffffff|0;
var d3 = Math.random()*0xffffffff|0;
return lut[d0&0xff]+lut[d0>>8&0xff]+lut[d0>>16&0xff]+lut[d0>>24&0xff]+'-'+
lut[d1&0xff]+lut[d1>>8&0xff]+'-'+lut[d1>>16&0x0f|0x40]+lut[d1>>24&0xff]+'-'+
lut[d2&0x3f|0x80]+lut[d2>>8&0xff]+'-'+lut[d2>>16&0xff]+lut[d2>>24&0xff]+
lut[d3&0xff]+lut[d3>>8&0xff]+lut[d3>>16&0xff]+lut[d3>>24&0xff];
}
console.log(e7())
Modualized: http://jcward.com/UUID.js - UUID.generate()
The funny thing is, generating 16 bytes of random data is the easy part. The whole trick is expressing it in string format with RFC compliance, and it's most tightly accomplished with 16 bytes of random data, an unrolled loop and lookup table.
I hope my logic is correct -- it's very easy to make a mistake in this kind of tedious bit work. But the outputs look good to me. I hope you enjoyed this mad ride through code optimization!
Be advised: my primary goal was to show and teach potential optimization strategies. Other answers cover important topics such as collisions and truly random numbers, which are important for generating good UUIDs.

Use:
let uniqueId = Date.now().toString(36) + Math.random().toString(36).substring(2);
document.getElementById("unique").innerHTML =
Math.random().toString(36).substring(2) + (new Date()).getTime().toString(36);
<div id="unique">
</div>
If IDs are generated more than 1 millisecond apart, they are 100% unique.
If two IDs are generated at shorter intervals, and assuming that the random method is truly random, this would generate IDs that are 99.99999999999999% likely to be globally unique (collision in 1 of 10^15).
You can increase this number by adding more digits, but to generate 100% unique IDs you will need to use a global counter.
If you need RFC compatibility, this formatting will pass as a valid version 4 GUID:
let u = Date.now().toString(16) + Math.random().toString(16) + '0'.repeat(16);
let guid = [u.substr(0,8), u.substr(8,4), '4000-8' + u.substr(13,3), u.substr(16,12)].join('-');
let u = Date.now().toString(16)+Math.random().toString(16)+'0'.repeat(16);
let guid = [u.substr(0,8), u.substr(8,4), '4000-8' + u.substr(13,3), u.substr(16,12)].join('-');
document.getElementById("unique").innerHTML = guid;
<div id="unique">
</div>
The above code follow the intention, but not the letter of the RFC. Among other discrepancies it's a few random digits short. (Add more random digits if you need it) The upside is that this is really fast :)
You can test validity of your GUID here

Here's some code based on RFC 4122, section 4.4 (Algorithms for Creating a UUID from Truly Random or Pseudo-Random Number).
function createUUID() {
// http://www.ietf.org/rfc/rfc4122.txt
var s = [];
var hexDigits = "0123456789abcdef";
for (var i = 0; i < 36; i++) {
s[i] = hexDigits.substr(Math.floor(Math.random() * 0x10), 1);
}
s[14] = "4"; // bits 12-15 of the time_hi_and_version field to 0010
s[19] = hexDigits.substr((s[19] & 0x3) | 0x8, 1); // bits 6-7 of the clock_seq_hi_and_reserved to 01
s[8] = s[13] = s[18] = s[23] = "-";
var uuid = s.join("");
return uuid;
}

This is the fastest GUID-like string generator method in the format XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX. It does not generate a standard-compliant GUID.
Ten million executions of this implementation take just 32.5 seconds, which is the fastest I've ever seen in a browser (the only solution without loops/iterations).
The function is as simple as:
/**
* Generates a GUID string.
* #returns {string} The generated GUID.
* #example af8a8416-6e18-a307-bd9c-f2c947bbb3aa
* #author Slavik Meltser.
* #link http://slavik.meltser.info/?p=142
*/
function guid() {
function _p8(s) {
var p = (Math.random().toString(16)+"000000000").substr(2,8);
return s ? "-" + p.substr(0,4) + "-" + p.substr(4,4) : p ;
}
return _p8() + _p8(true) + _p8(true) + _p8();
}
To test the performance, you can run this code:
console.time('t');
for (var i = 0; i < 10000000; i++) {
guid();
};
console.timeEnd('t');
I'm sure most of you will understand what I did there, but maybe there is at least one person that will need an explanation:
The algorithm:
The Math.random() function returns a decimal number between 0 and 1 with 16 digits after the decimal fraction point (for
example 0.4363923368509859).
Then we take this number and convert
it to a string with base 16 (from the example above we'll get
0.6fb7687f).
Math.random().toString(16).
Then we cut off the 0. prefix (0.6fb7687f =>
6fb7687f) and get a string with eight hexadecimal
characters long.
(Math.random().toString(16).substr(2,8).
Sometimes the Math.random() function will return
shorter number (for example 0.4363), due to zeros at the end (from the example above, actually the number is 0.4363000000000000). That's why I'm appending to this string "000000000" (a string with nine zeros) and then cutting it off with substr() function to make it nine characters exactly (filling zeros to the right).
The reason for adding exactly nine zeros is because of the worse case scenario, which is when the Math.random() function will return exactly 0 or 1 (probability of 1/10^16 for each one of them). That's why we needed to add nine zeros to it ("0"+"000000000" or "1"+"000000000"), and then cutting it off from the second index (third character) with a length of eight characters. For the rest of the cases, the addition of zeros will not harm the result because it is cutting it off anyway.
Math.random().toString(16)+"000000000").substr(2,8).
The assembly:
The GUID is in the following format XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX.
I divided the GUID into four pieces, each piece divided into two types (or formats): XXXXXXXX and -XXXX-XXXX.
Now I'm building the GUID using these two types to assemble the GUID with call four pieces, as follows: XXXXXXXX -XXXX-XXXX -XXXX-XXXX XXXXXXXX.
To differ between these two types, I added a flag parameter to a pair creator function _p8(s), the s parameter tells the function whether to add dashes or not.
Eventually we build the GUID with the following chaining: _p8() + _p8(true) + _p8(true) + _p8(), and return it.
Link to this post on my blog
Enjoy! :-)

Here is a totally non-compliant but very performant implementation to generate an ASCII-safe GUID-like unique identifier.
function generateQuickGuid() {
return Math.random().toString(36).substring(2, 15) +
Math.random().toString(36).substring(2, 15);
}
Generates 26 [a-z0-9] characters, yielding a UID that is both shorter and more unique than RFC compliant GUIDs. Dashes can be trivially added if human-readability matters.
Here are usage examples and timings for this function and several of this question's other answers. The timing was performed under Chrome m25, 10 million iterations each.
>>> generateQuickGuid()
"nvcjf1hs7tf8yyk4lmlijqkuo9"
"yq6gipxqta4kui8z05tgh9qeel"
"36dh5sec7zdj90sk2rx7pjswi2"
runtime: 32.5s
>>> GUID() // John Millikin
"7a342ca2-e79f-528e-6302-8f901b0b6888"
runtime: 57.8s
>>> regexGuid() // broofa
"396e0c46-09e4-4b19-97db-bd423774a4b3"
runtime: 91.2s
>>> createUUID() // Kevin Hakanson
"403aa1ab-9f70-44ec-bc08-5d5ac56bd8a5"
runtime: 65.9s
>>> UUIDv4() // Jed Schmidt
"f4d7d31f-fa83-431a-b30c-3e6cc37cc6ee"
runtime: 282.4s
>>> Math.uuid() // broofa
"5BD52F55-E68F-40FC-93C2-90EE069CE545"
runtime: 225.8s
>>> Math.uuidFast() // broofa
"6CB97A68-23A2-473E-B75B-11263781BBE6"
runtime: 92.0s
>>> Math.uuidCompact() // broofa
"3d7b7a06-0a67-4b67-825c-e5c43ff8c1e8"
runtime: 229.0s
>>> bitwiseGUID() // jablko
"baeaa2f-7587-4ff1-af23-eeab3e92"
runtime: 79.6s
>>>> betterWayGUID() // Andrea Turri
"383585b0-9753-498d-99c3-416582e9662c"
runtime: 60.0s
>>>> UUID() // John Fowler
"855f997b-4369-4cdb-b7c9-7142ceaf39e8"
runtime: 62.2s
Here is the timing code.
var r;
console.time('t');
for (var i = 0; i < 10000000; i++) {
r = FuncToTest();
};
console.timeEnd('t');

From sagi shkedy's technical blog:
function generateGuid() {
var result, i, j;
result = '';
for(j=0; j<32; j++) {
if( j == 8 || j == 12 || j == 16 || j == 20)
result = result + '-';
i = Math.floor(Math.random()*16).toString(16).toUpperCase();
result = result + i;
}
return result;
}
There are other methods that involve using an ActiveX control, but stay away from these!
I thought it was worth pointing out that no GUID generator can guarantee unique keys (check the Wikipedia article). There is always a chance of collisions. A GUID simply offers a large enough universe of keys to reduce the change of collisions to almost nil.

Here is a combination of the top voted answer, with a workaround for Chrome's collisions:
generateGUID = (typeof(window.crypto) != 'undefined' &&
typeof(window.crypto.getRandomValues) != 'undefined') ?
function() {
// If we have a cryptographically secure PRNG, use that
// https://stackoverflow.com/questions/6906916/collisions-when-generating-uuids-in-javascript
var buf = new Uint16Array(8);
window.crypto.getRandomValues(buf);
var S4 = function(num) {
var ret = num.toString(16);
while(ret.length < 4){
ret = "0"+ret;
}
return ret;
};
return (S4(buf[0])+S4(buf[1])+"-"+S4(buf[2])+"-"+S4(buf[3])+"-"+S4(buf[4])+"-"+S4(buf[5])+S4(buf[6])+S4(buf[7]));
}
:
function() {
// Otherwise, just use Math.random
// https://stackoverflow.com/questions/105034/how-to-create-a-guid-uuid-in-javascript/2117523#2117523
return 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx'.replace(/[xy]/g, function(c) {
var r = Math.random()*16|0, v = c == 'x' ? r : (r&0x3|0x8);
return v.toString(16);
});
};
It is on jsbin if you want to test it.

Here's a solution dated Oct. 9, 2011 from a comment by user jed at https://gist.github.com/982883:
UUIDv4 = function b(a){return a?(a^Math.random()*16>>a/4).toString(16):([1e7]+-1e3+-4e3+-8e3+-1e11).replace(/[018]/g,b)}
This accomplishes the same goal as the current highest-rated answer, but in 50+ fewer bytes by exploiting coercion, recursion, and exponential notation. For those curious how it works, here's the annotated form of an older version of the function:
UUIDv4 =
function b(
a // placeholder
){
return a // if the placeholder was passed, return
? ( // a random number from 0 to 15
a ^ // unless b is 8,
Math.random() // in which case
* 16 // a random number from
>> a/4 // 8 to 11
).toString(16) // in hexadecimal
: ( // or otherwise a concatenated string:
[1e7] + // 10000000 +
-1e3 + // -1000 +
-4e3 + // -4000 +
-8e3 + // -80000000 +
-1e11 // -100000000000,
).replace( // replacing
/[018]/g, // zeroes, ones, and eights with
b // random hex digits
)
}

You can use node-uuid. It provides simple, fast generation of RFC4122 UUIDS.
Features:
Generate RFC4122 version 1 or version 4 UUIDs
Runs in Node.js and browsers.
Cryptographically strong random # generation on supporting platforms.
Small footprint (Want something smaller? Check this out!)
Install Using NPM:
npm install uuid
Or using uuid via a browser:
Download Raw File (uuid v1): https://raw.githubusercontent.com/kelektiv/node-uuid/master/v1.js
Download Raw File (uuid v4): https://raw.githubusercontent.com/kelektiv/node-uuid/master/v4.js
Want even smaller? Check this out: https://gist.github.com/jed/982883
Usage:
// Generate a v1 UUID (time-based)
const uuidV1 = require('uuid/v1');
uuidV1(); // -> '6c84fb90-12c4-11e1-840d-7b25c5ee775a'
// Generate a v4 UUID (random)
const uuidV4 = require('uuid/v4');
uuidV4(); // -> '110ec58a-a0f2-4ac4-8393-c866d813b8d1'
// Generate a v5 UUID (namespace)
const uuidV5 = require('uuid/v5');
// ... using predefined DNS namespace (for domain names)
uuidV5('hello.example.com', v5.DNS)); // -> 'fdda765f-fc57-5604-a269-52a7df8164ec'
// ... using predefined URL namespace (for, well, URLs)
uuidV5('http://example.com/hello', v5.URL); // -> '3bbcee75-cecc-5b56-8031-b6641c1ed1f1'
// ... using a custom namespace
const MY_NAMESPACE = '(previously generated unique uuid string)';
uuidV5('hello', MY_NAMESPACE); // -> '90123e1c-7512-523e-bb28-76fab9f2f73d'
ECMAScript 2015 (ES6):
import uuid from 'uuid/v4';
const id = uuid();

var uuid = function() {
var buf = new Uint32Array(4);
window.crypto.getRandomValues(buf);
var idx = -1;
return 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx'.replace(/[xy]/g, function(c) {
idx++;
var r = (buf[idx>>3] >> ((idx%8)*4))&15;
var v = c == 'x' ? r : (r&0x3|0x8);
return v.toString(16);
});
};
This version is based on Briguy37's answer and some bitwise operators to extract nibble sized windows from the buffer.
It should adhere to the RFC Type 4 (random) schema, since I had problems last time parsing non-compliant UUIDs with Java's UUID.

This creates a version 4 UUID (created from pseudo random numbers):
function uuid()
{
var chars = '0123456789abcdef'.split('');
var uuid = [], rnd = Math.random, r;
uuid[8] = uuid[13] = uuid[18] = uuid[23] = '-';
uuid[14] = '4'; // version 4
for (var i = 0; i < 36; i++)
{
if (!uuid[i])
{
r = 0 | rnd()*16;
uuid[i] = chars[(i == 19) ? (r & 0x3) | 0x8 : r & 0xf];
}
}
return uuid.join('');
}
Here is a sample of the UUIDs generated:
682db637-0f31-4847-9cdf-25ba9613a75c
97d19478-3ab2-4aa1-b8cc-a1c3540f54aa
2eed04c9-2692-456d-a0fd-51012f947136

One line solution using Blobs.
window.URL.createObjectURL(new Blob([])).substring(31);
The value at the end (31) depends on the length of the URL.
EDIT:
A more compact and universal solution, as suggested by rinogo:
URL.createObjectURL(new Blob([])).substr(-36);

Simple JavaScript module as a combination of best answers in this question.
var crypto = window.crypto || window.msCrypto || null; // IE11 fix
var Guid = Guid || (function() {
var EMPTY = '00000000-0000-0000-0000-000000000000';
var _padLeft = function(paddingString, width, replacementChar) {
return paddingString.length >= width ? paddingString : _padLeft(replacementChar + paddingString, width, replacementChar || ' ');
};
var _s4 = function(number) {
var hexadecimalResult = number.toString(16);
return _padLeft(hexadecimalResult, 4, '0');
};
var _cryptoGuid = function() {
var buffer = new window.Uint16Array(8);
crypto.getRandomValues(buffer);
return [_s4(buffer[0]) + _s4(buffer[1]), _s4(buffer[2]), _s4(buffer[3]), _s4(buffer[4]), _s4(buffer[5]) + _s4(buffer[6]) + _s4(buffer[7])].join('-');
};
var _guid = function() {
var currentDateMilliseconds = new Date().getTime();
return 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx'.replace(/[xy]/g, function(currentChar) {
var randomChar = (currentDateMilliseconds + Math.random() * 16) % 16 | 0;
currentDateMilliseconds = Math.floor(currentDateMilliseconds / 16);
return (currentChar === 'x' ? randomChar : (randomChar & 0x7 | 0x8)).toString(16);
});
};
var create = function() {
var hasCrypto = crypto != 'undefined' && crypto !== null,
hasRandomValues = typeof(window.crypto.getRandomValues) != 'undefined';
return (hasCrypto && hasRandomValues) ? _cryptoGuid() : _guid();
};
return {
newGuid: create,
empty: EMPTY
};
})();
// DEMO: Create and show GUID
console.log('1. New Guid: ' + Guid.newGuid());
// DEMO: Show empty GUID
console.log('2. Empty Guid: ' + Guid.empty);
Usage:
Guid.newGuid()
"c6c2d12f-d76b-5739-e551-07e6de5b0807"
Guid.empty
"00000000-0000-0000-0000-000000000000"

The version below is an adaptation of broofa's answer, but updated to include a "true" random function that uses crypto libraries where available, and the Alea() function as a fallback.
Math.log2 = Math.log2 || function(n){ return Math.log(n) / Math.log(2); }
Math.trueRandom = (function() {
var crypt = window.crypto || window.msCrypto;
if (crypt && crypt.getRandomValues) {
// If we have a crypto library, use it
var random = function(min, max) {
var rval = 0;
var range = max - min;
if (range < 2) {
return min;
}
var bits_needed = Math.ceil(Math.log2(range));
if (bits_needed > 53) {
throw new Exception("We cannot generate numbers larger than 53 bits.");
}
var bytes_needed = Math.ceil(bits_needed / 8);
var mask = Math.pow(2, bits_needed) - 1;
// 7776 -> (2^13 = 8192) -1 == 8191 or 0x00001111 11111111
// Create byte array and fill with N random numbers
var byteArray = new Uint8Array(bytes_needed);
crypt.getRandomValues(byteArray);
var p = (bytes_needed - 1) * 8;
for(var i = 0; i < bytes_needed; i++ ) {
rval += byteArray[i] * Math.pow(2, p);
p -= 8;
}
// Use & to apply the mask and reduce the number of recursive lookups
rval = rval & mask;
if (rval >= range) {
// Integer out of acceptable range
return random(min, max);
}
// Return an integer that falls within the range
return min + rval;
}
return function() {
var r = random(0, 1000000000) / 1000000000;
return r;
};
} else {
// From https://web.archive.org/web/20120502223108/http://baagoe.com/en/RandomMusings/javascript/
// Johannes Baagøe <baagoe#baagoe.com>, 2010
function Mash() {
var n = 0xefc8249d;
var mash = function(data) {
data = data.toString();
for (var i = 0; i < data.length; i++) {
n += data.charCodeAt(i);
var h = 0.02519603282416938 * n;
n = h >>> 0;
h -= n;
h *= n;
n = h >>> 0;
h -= n;
n += h * 0x100000000; // 2^32
}
return (n >>> 0) * 2.3283064365386963e-10; // 2^-32
};
mash.version = 'Mash 0.9';
return mash;
}
// From http://baagoe.com/en/RandomMusings/javascript/
function Alea() {
return (function(args) {
// Johannes Baagøe <baagoe#baagoe.com>, 2010
var s0 = 0;
var s1 = 0;
var s2 = 0;
var c = 1;
if (args.length == 0) {
args = [+new Date()];
}
var mash = Mash();
s0 = mash(' ');
s1 = mash(' ');
s2 = mash(' ');
for (var i = 0; i < args.length; i++) {
s0 -= mash(args[i]);
if (s0 < 0) {
s0 += 1;
}
s1 -= mash(args[i]);
if (s1 < 0) {
s1 += 1;
}
s2 -= mash(args[i]);
if (s2 < 0) {
s2 += 1;
}
}
mash = null;
var random = function() {
var t = 2091639 * s0 + c * 2.3283064365386963e-10; // 2^-32
s0 = s1;
s1 = s2;
return s2 = t - (c = t | 0);
};
random.uint32 = function() {
return random() * 0x100000000; // 2^32
};
random.fract53 = function() {
return random() +
(random() * 0x200000 | 0) * 1.1102230246251565e-16; // 2^-53
};
random.version = 'Alea 0.9';
random.args = args;
return random;
}(Array.prototype.slice.call(arguments)));
};
return Alea();
}
}());
Math.guid = function() {
return 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx'.replace(/[xy]/g, function(c) {
var r = Math.trueRandom() * 16 | 0,
v = c == 'x' ? r : (r & 0x3 | 0x8);
return v.toString(16);
});
};

JavaScript project on GitHub - https://github.com/LiosK/UUID.js
UUID.js The RFC-compliant UUID generator for JavaScript.
See RFC 4122 http://www.ietf.org/rfc/rfc4122.txt.
Features Generates RFC 4122 compliant UUIDs.
Version 4 UUIDs (UUIDs from random numbers) and version 1 UUIDs
(time-based UUIDs) are available.
UUID object allows a variety of access to the UUID including access to
the UUID fields.
Low timestamp resolution of JavaScript is compensated by random
numbers.

// RFC 4122
//
// A UUID is 128 bits long
//
// String representation is five fields of 4, 2, 2, 2, and 6 bytes.
// Fields represented as lowercase, zero-filled, hexadecimal strings, and
// are separated by dash characters
//
// A version 4 UUID is generated by setting all but six bits to randomly
// chosen values
var uuid = [
Math.random().toString(16).slice(2, 10),
Math.random().toString(16).slice(2, 6),
// Set the four most significant bits (bits 12 through 15) of the
// time_hi_and_version field to the 4-bit version number from Section
// 4.1.3
(Math.random() * .0625 /* 0x.1 */ + .25 /* 0x.4 */).toString(16).slice(2, 6),
// Set the two most significant bits (bits 6 and 7) of the
// clock_seq_hi_and_reserved to zero and one, respectively
(Math.random() * .25 /* 0x.4 */ + .5 /* 0x.8 */).toString(16).slice(2, 6),
Math.random().toString(16).slice(2, 14)].join('-');

Added in: v15.6.0, v14.17.0 there is a built-in crypto.randomUUID() function.
import * as crypto from "crypto";
const uuid = crypto.randomUUID();
In the browser, crypto.randomUUID() is currently supported in Chromium 92+ and Firefox 95+.

For those wanting an RFC 4122 version 4 compliant solution with speed considerations (few calls to Math.random()):
var rand = Math.random;
function UUID() {
var nbr, randStr = "";
do {
randStr += (nbr = rand()).toString(16).substr(3, 6);
} while (randStr.length < 30);
return (
randStr.substr(0, 8) + "-" +
randStr.substr(8, 4) + "-4" +
randStr.substr(12, 3) + "-" +
((nbr*4|0)+8).toString(16) + // [89ab]
randStr.substr(15, 3) + "-" +
randStr.substr(18, 12)
);
}
console.log( UUID() );
The above function should have a decent balance between speed and randomness.

I wanted to understand broofa's answer, so I expanded it and added comments:
var uuid = function () {
return 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx'.replace(
/[xy]/g,
function (match) {
/*
* Create a random nibble. The two clever bits of this code:
*
* - Bitwise operations will truncate floating point numbers
* - For a bitwise OR of any x, x | 0 = x
*
* So:
*
* Math.random * 16
*
* creates a random floating point number
* between 0 (inclusive) and 16 (exclusive) and
*
* | 0
*
* truncates the floating point number into an integer.
*/
var randomNibble = Math.random() * 16 | 0;
/*
* Resolves the variant field. If the variant field (delineated
* as y in the initial string) is matched, the nibble must
* match the mask (where x is a do-not-care bit):
*
* 10xx
*
* This is achieved by performing the following operations in
* sequence (where x is an intermediate result):
*
* - x & 0x3, which is equivalent to x % 3
* - x | 0x8, which is equivalent to x + 8
*
* This results in a nibble between 8 inclusive and 11 exclusive,
* (or 1000 and 1011 in binary), all of which satisfy the variant
* field mask above.
*/
var nibble = (match == 'y') ?
(randomNibble & 0x3 | 0x8) :
randomNibble;
/*
* Ensure the nibble integer is encoded as base 16 (hexadecimal).
*/
return nibble.toString(16);
}
);
};

ES6 sample
const guid=()=> {
const s4=()=> Math.floor((1 + Math.random()) * 0x10000).toString(16).substring(1);
return `${s4() + s4()}-${s4()}-${s4()}-${s4()}-${s4() + s4() + s4()}`;
}

I adjusted my own UUID/GUID generator with some extras here.
I'm using the following Kybos random number generator to be a bit more cryptographically sound.
Below is my script with the Mash and Kybos methods from baagoe.com excluded.
//UUID/Guid Generator
// use: UUID.create() or UUID.createSequential()
// convenience: UUID.empty, UUID.tryParse(string)
(function(w){
// From http://baagoe.com/en/RandomMusings/javascript/
// Johannes Baagøe <baagoe#baagoe.com>, 2010
//function Mash() {...};
// From http://baagoe.com/en/RandomMusings/javascript/
//function Kybos() {...};
var rnd = Kybos();
//UUID/GUID Implementation from http://frugalcoder.us/post/2012/01/13/javascript-guid-uuid-generator.aspx
var UUID = {
"empty": "00000000-0000-0000-0000-000000000000"
,"parse": function(input) {
var ret = input.toString().trim().toLowerCase().replace(/^[\s\r\n]+|[\{\}]|[\s\r\n]+$/g, "");
if ((/[a-f0-9]{8}\-[a-f0-9]{4}\-[a-f0-9]{4}\-[a-f0-9]{4}\-[a-f0-9]{12}/).test(ret))
return ret;
else
throw new Error("Unable to parse UUID");
}
,"createSequential": function() {
var ret = new Date().valueOf().toString(16).replace("-","")
for (;ret.length < 12; ret = "0" + ret);
ret = ret.substr(ret.length-12,12); //only least significant part
for (;ret.length < 32;ret += Math.floor(rnd() * 0xffffffff).toString(16));
return [ret.substr(0,8), ret.substr(8,4), "4" + ret.substr(12,3), "89AB"[Math.floor(Math.random()*4)] + ret.substr(16,3), ret.substr(20,12)].join("-");
}
,"create": function() {
var ret = "";
for (;ret.length < 32;ret += Math.floor(rnd() * 0xffffffff).toString(16));
return [ret.substr(0,8), ret.substr(8,4), "4" + ret.substr(12,3), "89AB"[Math.floor(Math.random()*4)] + ret.substr(16,3), ret.substr(20,12)].join("-");
}
,"random": function() {
return rnd();
}
,"tryParse": function(input) {
try {
return UUID.parse(input);
} catch(ex) {
return UUID.empty;
}
}
};
UUID["new"] = UUID.create;
w.UUID = w.Guid = UUID;
}(window || this));

The native URL.createObjectURL is generating an UUID. You can take advantage of this.
function uuid() {
const url = URL.createObjectURL(new Blob())
const [id] = url.toString().split('/').reverse()
URL.revokeObjectURL(url)
return id
}

The better way:
function(
a, b // Placeholders
){
for( // Loop :)
b = a = ''; // b - result , a - numeric variable
a++ < 36; //
b += a*51&52 // If "a" is not 9 or 14 or 19 or 24
? // return a random number or 4
(
a^15 // If "a" is not 15,
? // generate a random number from 0 to 15
8^Math.random() *
(a^20 ? 16 : 4) // unless "a" is 20, in which case a random number from 8 to 11,
:
4 // otherwise 4
).toString(16)
:
'-' // In other cases, (if "a" is 9,14,19,24) insert "-"
);
return b
}
Minimized:
function(a,b){for(b=a='';a++<36;b+=a*51&52?(a^15?8^Math.random()*(a^20?16:4):4).toString(16):'-');return b}

The following is simple code that uses crypto.getRandomValues(a) on supported browsers (Internet Explorer 11+, iOS 7+, Firefox 21+, Chrome, and Android Chrome).
It avoids using Math.random(), because that can cause collisions (for example 20 collisions for 4000 generated UUIDs in a real situation by Muxa).
function uuid() {
function randomDigit() {
if (crypto && crypto.getRandomValues) {
var rands = new Uint8Array(1);
crypto.getRandomValues(rands);
return (rands[0] % 16).toString(16);
} else {
return ((Math.random() * 16) | 0).toString(16);
}
}
var crypto = window.crypto || window.msCrypto;
return 'xxxxxxxx-xxxx-4xxx-8xxx-xxxxxxxxxxxx'.replace(/x/g, randomDigit);
}
Notes:
Optimised for code readability, not speed, so it is suitable for, say, a few hundred UUIDs per second. It generates about 10000 uuid() per second in Chromium on my laptop using http://jsbin.com/fuwigo/1 to measure performance.
It only uses 8 for "y" because that simplifies code readability (y is allowed to be 8, 9, A, or B).

If you just need a random 128 bit string in no particular format, you can use:
function uuid() {
return crypto.getRandomValues(new Uint32Array(4)).join('-');
}
Which will return something like 2350143528-4164020887-938913176-2513998651.

I couldn't find any answer that uses a single 16-octet TypedArray and a DataView, so I think the following solution for generating a version 4 UUID per the RFC will stand on its own here:
const uuid4 = () => {
const ho = (n, p) => n.toString(16).padStart(p, 0); /// Return the hexadecimal text representation of number `n`, padded with zeroes to be of length `p`
const data = crypto.getRandomValues(new Uint8Array(16)); /// Fill the buffer with random data
data[6] = (data[6] & 0xf) | 0x40; /// Patch the 6th byte to reflect a version 4 UUID
data[8] = (data[8] & 0x3f) | 0x80; /// Patch the 8th byte to reflect a variant 1 UUID (version 4 UUIDs are)
const view = new DataView(data.buffer); /// Create a view backed by a 16-byte buffer
return `${ho(view.getUint32(0), 8)}-${ho(view.getUint16(4), 4)}-${ho(view.getUint16(6), 4)}-${ho(view.getUint16(8), 4)}-${ho(view.getUint32(10), 8)}${ho(view.getUint16(14), 4)}`; /// Compile the canonical textual form from the array data
};
I prefer it because:
it only relies on functions available to the standard ECMAScript platform, where possible -- which is all but one procedure
it only uses a single buffer, minimizing copying of data, which should in theory yield performance advantages
At the time of writing this, getRandomValues is not something implemented for the crypto object in Node.js. However, it has the equivalent randomBytes function which may be used instead.

Just another more readable variant with just two mutations.
function uuid4()
{
function hex (s, b)
{
return s +
(b >>> 4 ).toString (16) + // high nibble
(b & 0b1111).toString (16); // low nibble
}
let r = crypto.getRandomValues (new Uint8Array (16));
r[6] = r[6] >>> 4 | 0b01000000; // Set type 4: 0100
r[8] = r[8] >>> 3 | 0b10000000; // Set variant: 100
return r.slice ( 0, 4).reduce (hex, '' ) +
r.slice ( 4, 6).reduce (hex, '-') +
r.slice ( 6, 8).reduce (hex, '-') +
r.slice ( 8, 10).reduce (hex, '-') +
r.slice (10, 16).reduce (hex, '-');
}

NewId vs NewSequentialId when used in combination with another uuid source (javascript)

I need to create UID's on my my server side code, as well as on client side code (in the browser). I currently use NewID() as my default value, but when creating objects the client side (in browser), I use uuid.js. Am I more likely to get collisions with NewSequentialId() as my default value (which will be used when objects are created server side)?
FYI here is the uuid.js code, as I can't recall where I downloaded it.
// uuid.js
//
// Copyright (c) 2010-2012 Robert Kieffer
// MIT License - http://opensource.org/licenses/mit-license.php
(function() {
var _global = this;
// Unique ID creation requires a high quality random # generator. We feature
// detect to determine the best RNG source, normalizing to a function that
// returns 128-bits of randomness, since that's what's usually required
var _rng;
// Node.js crypto-based RNG - http://nodejs.org/docs/v0.6.2/api/crypto.html
//
// Moderately fast, high quality
if (typeof(require) == 'function') {
try {
var _rb = require('crypto').randomBytes;
_rng = _rb && function() {return _rb(16);};
} catch(e) {}
}
if (!_rng && _global.crypto && crypto.getRandomValues) {
// WHATWG crypto-based RNG - http://wiki.whatwg.org/wiki/Crypto
//
// Moderately fast, high quality
var _rnds8 = new Uint8Array(16);
_rng = function whatwgRNG() {
crypto.getRandomValues(_rnds8);
return _rnds8;
};
}
if (!_rng) {
// Math.random()-based (RNG)
//
// If all else fails, use Math.random(). It's fast, but is of unspecified
// quality.
var _rnds = new Array(16);
_rng = function() {
for (var i = 0, r; i < 16; i++) {
if ((i & 0x03) === 0) r = Math.random() * 0x100000000;
_rnds[i] = r >>> ((i & 0x03) << 3) & 0xff;
}
return _rnds;
};
}
// Buffer class to use
var BufferClass = typeof(Buffer) == 'function' ? Buffer : Array;
// Maps for number <-> hex string conversion
var _byteToHex = [];
var _hexToByte = {};
for (var i = 0; i < 256; i++) {
_byteToHex[i] = (i + 0x100).toString(16).substr(1);
_hexToByte[_byteToHex[i]] = i;
}
// **`parse()` - Parse a UUID into it's component bytes**
function parse(s, buf, offset) {
var i = (buf && offset) || 0, ii = 0;
buf = buf || [];
s.toLowerCase().replace(/[0-9a-f]{2}/g, function(oct) {
if (ii < 16) { // Don't overflow!
buf[i + ii++] = _hexToByte[oct];
}
});
// Zero out remaining bytes if string was short
while (ii < 16) {
buf[i + ii++] = 0;
}
return buf;
}
// **`unparse()` - Convert UUID byte array (ala parse()) into a string**
function unparse(buf, offset) {
var i = offset || 0, bth = _byteToHex;
return bth[buf[i++]] + bth[buf[i++]] +
bth[buf[i++]] + bth[buf[i++]] + '-' +
bth[buf[i++]] + bth[buf[i++]] + '-' +
bth[buf[i++]] + bth[buf[i++]] + '-' +
bth[buf[i++]] + bth[buf[i++]] + '-' +
bth[buf[i++]] + bth[buf[i++]] +
bth[buf[i++]] + bth[buf[i++]] +
bth[buf[i++]] + bth[buf[i++]];
}
// **`v1()` - Generate time-based UUID**
//
// Inspired by https://github.com/LiosK/UUID.js
// and http://docs.python.org/library/uuid.html
// random #'s we need to init node and clockseq
var _seedBytes = _rng();
// Per 4.5, create and 48-bit node id, (47 random bits + multicast bit = 1)
var _nodeId = [
_seedBytes[0] | 0x01,
_seedBytes[1], _seedBytes[2], _seedBytes[3], _seedBytes[4], _seedBytes[5]
];
// Per 4.2.2, randomize (14 bit) clockseq
var _clockseq = (_seedBytes[6] << 8 | _seedBytes[7]) & 0x3fff;
// Previous uuid creation time
var _lastMSecs = 0, _lastNSecs = 0;
// See https://github.com/broofa/node-uuid for API details
function v1(options, buf, offset) {
var i = buf && offset || 0;
var b = buf || [];
options = options || {};
var clockseq = options.clockseq != null ? options.clockseq : _clockseq;
// UUID timestamps are 100 nano-second units since the Gregorian epoch,
// (1582-10-15 00:00). JSNumbers aren't precise enough for this, so
// time is handled internally as 'msecs' (integer milliseconds) and 'nsecs'
// (100-nanoseconds offset from msecs) since unix epoch, 1970-01-01 00:00.
var msecs = options.msecs != null ? options.msecs : new Date().getTime();
// Per 4.2.1.2, use count of uuid's generated during the current clock
// cycle to simulate higher resolution clock
var nsecs = options.nsecs != null ? options.nsecs : _lastNSecs + 1;
// Time since last uuid creation (in msecs)
var dt = (msecs - _lastMSecs) + (nsecs - _lastNSecs)/10000;
// Per 4.2.1.2, Bump clockseq on clock regression
if (dt < 0 && options.clockseq == null) {
clockseq = clockseq + 1 & 0x3fff;
}
// Reset nsecs if clock regresses (new clockseq) or we've moved onto a new
// time interval
if ((dt < 0 || msecs > _lastMSecs) && options.nsecs == null) {
nsecs = 0;
}
// Per 4.2.1.2 Throw error if too many uuids are requested
if (nsecs >= 10000) {
throw new Error('uuid.v1(): Can\'t create more than 10M uuids/sec');
}
_lastMSecs = msecs;
_lastNSecs = nsecs;
_clockseq = clockseq;
// Per 4.1.4 - Convert from unix epoch to Gregorian epoch
msecs += 12219292800000;
// `time_low`
var tl = ((msecs & 0xfffffff) * 10000 + nsecs) % 0x100000000;
b[i++] = tl >>> 24 & 0xff;
b[i++] = tl >>> 16 & 0xff;
b[i++] = tl >>> 8 & 0xff;
b[i++] = tl & 0xff;
// `time_mid`
var tmh = (msecs / 0x100000000 * 10000) & 0xfffffff;
b[i++] = tmh >>> 8 & 0xff;
b[i++] = tmh & 0xff;
// `time_high_and_version`
b[i++] = tmh >>> 24 & 0xf | 0x10; // include version
b[i++] = tmh >>> 16 & 0xff;
// `clock_seq_hi_and_reserved` (Per 4.2.2 - include variant)
b[i++] = clockseq >>> 8 | 0x80;
// `clock_seq_low`
b[i++] = clockseq & 0xff;
// `node`
var node = options.node || _nodeId;
for (var n = 0; n < 6; n++) {
b[i + n] = node[n];
}
return buf ? buf : unparse(b);
}
// **`v4()` - Generate random UUID**
// See https://github.com/broofa/node-uuid for API details
function v4(options, buf, offset) {
// Deprecated - 'format' argument, as supported in v1.2
var i = buf && offset || 0;
if (typeof(options) == 'string') {
buf = options == 'binary' ? new BufferClass(16) : null;
options = null;
}
options = options || {};
var rnds = options.random || (options.rng || _rng)();
// Per 4.4, set bits for version and `clock_seq_hi_and_reserved`
rnds[6] = (rnds[6] & 0x0f) | 0x40;
rnds[8] = (rnds[8] & 0x3f) | 0x80;
// Copy bytes to buffer, if provided
if (buf) {
for (var ii = 0; ii < 16; ii++) {
buf[i + ii] = rnds[ii];
}
}
return buf || unparse(rnds);
}
// Export public API
var uuid = v4;
uuid.v1 = v1;
uuid.v4 = v4;
uuid.parse = parse;
uuid.unparse = unparse;
uuid.BufferClass = BufferClass;
if (typeof define === 'function' && define.amd) {
// Publish as AMD module
define(function() {return uuid;});
} else if (typeof(module) != 'undefined' && module.exports) {
// Publish as node.js module
module.exports = uuid;
} else {
// Publish as global (in browsers)
var _previousRoot = _global.uuid;
// **`noConflict()` - (browser only) to reset global 'uuid' var**
uuid.noConflict = function() {
_global.uuid = _previousRoot;
return uuid;
};
_global.uuid = uuid;
}
}).call(this);

This is actually a pretty interesting question, with a number of levels to it.
First, it's worth noting that uuid.js supports two different forms of id. uuid.v4() creates IDs using random numbers, while uuid.v1() creates IDs based on timestamps. The "version" of the id is actually encoded in the id itself, which guarantees that in theory no v4 id will ever collide with a v1 id. That's part of RFC4122, the UUID specification.
It's also worth noting that for v1 ids, each id source is supposed to have a unique "node id", also encoded in the id, that guarantees the uniqueness of id sequence created by that source. For id sources that have access to a guaranteed-unique value (e.g. a device's MAC address) this works well. However uuid.js doesn't have access to such a value and, thus, generates a random value for it's node id. This introduces the risk of it generating a node id that matches the one used by your server. The node id is 48-bit value, meaning the chance of a node id collision are 281,474,976,710,656:1. So, theres a chance, but it's pretty damn low.
... but none of that matters!
It turns out that even though NewSequentialID() produces IDs that are superficially similar to v1 IDs, Microsoft for whatever reasons decided to swap the various fields within the ID around, breaking RFC4122 compatability. What this means is that, depending on the sequence number, the IDs may or may not look like valid v1 ids, or valid v4 ids, or simply invalid UUIDs. I.e. using NewSequentialID() throws a wrench into the works if you want to talk about the possibility of uuid collision.
I'm not sure there's a simple way to quantify the risk of collision given this last issue. At the end of the day, UUIDs are 128-bit values, meaning there's a HUGE numberspace to draw from. For all but the most demanding of requirements you're probably okay. But there will be an increased risk of collision compared to what you'd have if you used an RFC-compliant UUID source.
[FWIW, your uuid.js comes from the node-uuid project (*cough* said the author).]

How do I create bit array in Javascript?

What is the best way of implementing a bit array in JavaScript?

Here's one I whipped up:
UPDATE - something about this class had been bothering me all day - it wasn't size based - creating a BitArray with N slots/bits was a two step operation - instantiate, resize. Updated the class to be size based with an optional second paramter for populating the size based instance with either array values or a base 10 numeric value.
(Fiddle with it here)
/* BitArray DataType */
// Constructor
function BitArray(size, bits) {
// Private field - array for our bits
this.m_bits = new Array();
//.ctor - initialize as a copy of an array of true/false or from a numeric value
if (bits && bits.length) {
for (var i = 0; i < bits.length; i++)
this.m_bits.push(bits[i] ? BitArray._ON : BitArray._OFF);
} else if (!isNaN(bits)) {
this.m_bits = BitArray.shred(bits).m_bits;
}
if (size && this.m_bits.length != size) {
if (this.m_bits.length < size) {
for (var i = this.m_bits.length; i < size; i++) {
this.m_bits.push(BitArray._OFF);
}
} else {
for(var i = size; i > this.m_bits.length; i--){
this.m_bits.pop();
}
}
}
}
/* BitArray PUBLIC INSTANCE METHODS */
// read-only property - number of bits
BitArray.prototype.getLength = function () { return this.m_bits.length; };
// accessor - get bit at index
BitArray.prototype.getAt = function (index) {
if (index < this.m_bits.length) {
return this.m_bits[index];
}
return null;
};
// accessor - set bit at index
BitArray.prototype.setAt = function (index, value) {
if (index < this.m_bits.length) {
this.m_bits[index] = value ? BitArray._ON : BitArray._OFF;
}
};
// resize the bit array (append new false/0 indexes)
BitArray.prototype.resize = function (newSize) {
var tmp = new Array();
for (var i = 0; i < newSize; i++) {
if (i < this.m_bits.length) {
tmp.push(this.m_bits[i]);
} else {
tmp.push(BitArray._OFF);
}
}
this.m_bits = tmp;
};
// Get the complimentary bit array (i.e., 01 compliments 10)
BitArray.prototype.getCompliment = function () {
var result = new BitArray(this.m_bits.length);
for (var i = 0; i < this.m_bits.length; i++) {
result.setAt(i, this.m_bits[i] ? BitArray._OFF : BitArray._ON);
}
return result;
};
// Get the string representation ("101010")
BitArray.prototype.toString = function () {
var s = new String();
for (var i = 0; i < this.m_bits.length; i++) {
s = s.concat(this.m_bits[i] === BitArray._ON ? "1" : "0");
}
return s;
};
// Get the numeric value
BitArray.prototype.toNumber = function () {
var pow = 0;
var n = 0;
for (var i = this.m_bits.length - 1; i >= 0; i--) {
if (this.m_bits[i] === BitArray._ON) {
n += Math.pow(2, pow);
}
pow++;
}
return n;
};
/* STATIC METHODS */
// Get the union of two bit arrays
BitArray.getUnion = function (bitArray1, bitArray2) {
var len = BitArray._getLen(bitArray1, bitArray2, true);
var result = new BitArray(len);
for (var i = 0; i < len; i++) {
result.setAt(i, BitArray._union(bitArray1.getAt(i), bitArray2.getAt(i)));
}
return result;
};
// Get the intersection of two bit arrays
BitArray.getIntersection = function (bitArray1, bitArray2) {
var len = BitArray._getLen(bitArray1, bitArray2, true);
var result = new BitArray(len);
for (var i = 0; i < len; i++) {
result.setAt(i, BitArray._intersect(bitArray1.getAt(i), bitArray2.getAt(i)));
}
return result;
};
// Get the difference between to bit arrays
BitArray.getDifference = function (bitArray1, bitArray2) {
var len = BitArray._getLen(bitArray1, bitArray2, true);
var result = new BitArray(len);
for (var i = 0; i < len; i++) {
result.setAt(i, BitArray._difference(bitArray1.getAt(i), bitArray2.getAt(i)));
}
return result;
};
// Convert a number into a bit array
BitArray.shred = function (number) {
var bits = new Array();
var q = number;
do {
bits.push(q % 2);
q = Math.floor(q / 2);
} while (q > 0);
return new BitArray(bits.length, bits.reverse());
};
/* BitArray PRIVATE STATIC CONSTANTS */
BitArray._ON = 1;
BitArray._OFF = 0;
/* BitArray PRIVATE STATIC METHODS */
// Calculate the intersection of two bits
BitArray._intersect = function (bit1, bit2) {
return bit1 === BitArray._ON && bit2 === BitArray._ON ? BitArray._ON : BitArray._OFF;
};
// Calculate the union of two bits
BitArray._union = function (bit1, bit2) {
return bit1 === BitArray._ON || bit2 === BitArray._ON ? BitArray._ON : BitArray._OFF;
};
// Calculate the difference of two bits
BitArray._difference = function (bit1, bit2) {
return bit1 === BitArray._ON && bit2 !== BitArray._ON ? BitArray._ON : BitArray._OFF;
};
// Get the longest or shortest (smallest) length of the two bit arrays
BitArray._getLen = function (bitArray1, bitArray2, smallest) {
var l1 = bitArray1.getLength();
var l2 = bitArray2.getLength();
return l1 > l2 ? smallest ? l2 : l1 : smallest ? l2 : l1;
};
CREDIT TO #Daniel Baulig for asking for the refactor from quick and dirty to prototype based.

I don't know about bit arrays, but you can make byte arrays easy with new features.
Look up typed arrays. I've used these in both Chrome and Firefox. The important one is Uint8Array.
To make an array of 512 uninitialized bytes:
var arr = new UintArray(512);
And accessing it (the sixth byte):
var byte = arr[5];
For node.js, use Buffer (server-side).
EDIT:
To access individual bits, use bit masks.
To get the bit in the one's position, do num & 0x1

The Stanford Javascript Crypto Library (SJCL) provides a Bit Array implementation and can convert different inputs (Hex Strings, Byte Arrays, etc.) to Bit Arrays.
Their code is public on GitHub: bitwiseshiftleft/sjcl. So if you lookup bitArray.js, you can find their bit array implementation.
A conversion from bytes to bits can be found here.

Something like this is as close as I can think of. Saves bit arrays as 32 bit numbers, and has a standard array backing it to handle larger sets.
class bitArray {
constructor(length) {
this.backingArray = Array.from({length: Math.ceil(length/32)}, ()=>0)
this.length = length
}
get(n) {
return (this.backingArray[n/32|0] & 1 << n % 32) > 0
}
on(n) {
this.backingArray[n/32|0] |= 1 << n % 32
}
off(n) {
this.backingArray[n/32|0] &= ~(1 << n % 32)
}
toggle(n) {
this.backingArray[n/32|0] ^= 1 << n % 32
}
forEach(callback) {
this.backingArray.forEach((number, container)=>{
const max = container == this.backingArray.length-1 ? this.length%32 : 32
for(let x=0; x<max; x++) {
callback((number & 1<<x)>0, 32*container+x)
}
})
}
}
let bits = new bitArray(10)
bits.get(2) //false
bits.on(2)
bits.get(2) //true
bits.forEach(console.log)
/* outputs:
false
false
true
false
false
false
false
false
false
false
*/
bits.toggle(2)
bits.forEach(console.log)
/* outputs:
false
false
false
false
false
false
false
false
false
false
*/
bits.toggle(0)
bits.toggle(1)
bits.toggle(2)
bits.off(2)
bits.off(3)
bits.forEach(console.log)
/* outputs:
true
true
false
false
false
false
false
false
false
false
*/

2022
As can be seen from past answers and comments, the question of "implementing a bit array" can be understood in two different (non-exclusive) ways:
an array that takes 1-bit in memory for each entry
an array on which bitwise operations can be applied
As #beatgammit points out, ecmascript specifies typed arrays, but bit arrays are not part of it. I have just published #bitarray/typedarray, an implementation of typed arrays for bits, that emulates native typed arrays and takes 1 bit in memory for each entry.
Because it reproduces the behaviour of native typed arrays, it does not include any bitwise operations though. So, I have also published #bitarray/es6, which extends the previous with bitwise operations.
I wouldn't debate what is the best way of implementing bit array, as per the asked question, because "best" could be argued at length, but those are certainly some way of implementing bit arrays, with the benefit that they behave like native typed arrays.
import BitArray from "#bitarray/es6"
const bits1 = BitArray.from("11001010");
const bits2 = BitArray.from("10111010");
for (let bit of bits1.or(bits2)) console.log(bit) // 1 1 1 1 1 0 1 0

You can easily do that by using bitwise operators. It's quite simple.
Let's try with the number 75.
Its representation in binary is 100 1011. So, how do we obtain each bit from the number?
You can use an AND "&" operator to select one bit and set the rest of them to 0. Then with a Shift operator, you remove the rest of 0 that doesn't matter at the moment.
Example:
Let's do an AND operation with 4 (000 0010)
0100 1011 & 0000 0010 => 0000 0010
Now we need to filter the selected bit, in this case, was the second-bit reading right to left.
0000 0010 >> 1 => 1
The zeros on the left are no representative. So the output will be the bit we selected, in this case, the second one.
var word=75;
var res=[];
for(var x=7; x>=0; x--){
res.push((word&Math.pow(2,x))>>x);
}
console.log(res);
The output:
Expected:
In case you need more than a simple number, you can apply the same function for a byte. Let's say you have a file with multiple bytes. So, you can decompose that file in a ByteArray, then each byte in the array in a BitArray.
Good luck!

#Commi's implementation is what I ended up using .
I believe there is a bug in this implementation. Bits on every 31st boundary give the wrong result. (ie when index is (32 * index - 1), so 31, 63, 95 etc.
I fixed it in the get() method by replacing > 0 with != 0.
get(n) {
return (this.backingArray[n/32|0] & 1 << n % 32) != 0
}
The reason for the bug is that the ints are 32-bit signed. Shifting 1 left by 31 gets you a negative number. Since the check is for >0, this will be false when it should be true.
I wrote a program to prove the bug before, and the fix after. Will post it running out of space.
for (var i=0; i < 100; i++) {
var ar = new bitArray(1000);
ar.on(i);
for(var j=0;j<1000;j++) {
// we should have TRUE only at one position and that is "i".
// if something is true when it should be false or false when it should be true, then report it.
if(ar.get(j)) {
if (j != i) console.log('we got a bug at ' + i);
}
if (!ar.get(j)) {
if (j == i) console.log('we got a bug at ' + i);
}
}
}

2022
We can implement a BitArray class which behaves similar to TypedArrays by extending DataView. However in order to avoid the cost of trapping the direct accesses to the numerical properties (the indices) by using a Proxy, I believe it's best to stay in DataView domain. DataView is preferable to TypedArrays these days anyway as it's performance is highly improved in recent V8 versions (v7+).
Just like TypedArrays, BitArray will have a predetermined length at construction time. I just include a few methods in the below snippet. The popcnt property very efficiently returns the total number of 1s in BitArray. Unlike normal arrays popcnt is a highly sought after functionality for BitArrays. So much so that Web Assembly and even modern CPU's have a dedicated pop count instruction. Apart from these you can easily add methods like .forEach(), .map() etc. if need be.
class BitArray extends DataView{
constructor(n,ab){
if (n > 1.5e10) throw new Error("BitArray size can not exceed 1.5e10");
super(ab instanceof ArrayBuffer ? ab
: new ArrayBuffer(Number((BigInt(n + 31) & ~31n) >> 3n))); // Sets ArrayBuffer.byteLength to multiples of 4 bytes (32 bits)
}
get length(){
return this.buffer.byteLength << 3;
}
get popcount(){
var m1 = 0x55555555,
m2 = 0x33333333,
m4 = 0x0f0f0f0f,
h01 = 0x01010101,
pc = 0,
x;
for (var i = 0, len = this.buffer.byteLength >> 2; i < len; i++){
x = this.getUint32(i << 2);
x -= (x >> 1) & m1; //put count of each 2 bits into those 2 bits
x = (x & m2) + ((x >> 2) & m2); //put count of each 4 bits into those 4 bits
x = (x + (x >> 4)) & m4; //put count of each 8 bits into those 8 bits
pc += (x * h01) >> 56;
}
return pc;
}
// n >> 3 is Math.floor(n/8)
// n & 7 is n % 8
and(bar){
var len = Math.min(this.buffer.byteLength,bar.buffer.byteLength),
res = new BitArray(len << 3);
for (var i = 0; i < len; i += 4) res.setUint32(i,this.getUint32(i) & bar.getUint32(i));
return res;
}
at(n){
return this.getUint8(n >> 3) & (1 << (n & 7)) ? 1 : 0;
}
or(bar){
var len = Math.min(this.buffer.byteLength,bar.buffer.byteLength),
res = new BitArray(len << 3);
for (var i = 0; i < len; i += 4) res.setUint32(i,this.getUint32(i) | bar.getUint32(i));
return res;
}
not(){
var len = this.buffer.byteLength,
res = new BitArray(len << 3);
for (var i = 0; i < len; i += 4) res.setUint32(i,~(this.getUint32(i) >> 0));
return res;
}
reset(n){
this.setUint8(n >> 3, this.getUint8(n >> 3) & ~(1 << (n & 7)));
}
set(n){
this.setUint8(n >> 3, this.getUint8(n >> 3) | (1 << (n & 7)));
}
slice(a = 0, b = this.length){
return new BitArray(b-a,this.buffer.slice(a >> 3, b >> 3));
}
toggle(n){
this.setUint8(n >> 3, this.getUint8(n >> 3) ^ (1 << (n & 7)));
}
toString(){
return new Uint8Array(this.buffer).reduce((p,c) => p + ((BigInt(c)* 0x0202020202n & 0x010884422010n) % 1023n).toString(2).padStart(8,"0"),"");
}
xor(bar){
var len = Math.min(this.buffer.byteLength,bar.buffer.byteLength),
res = new BitArray(len << 3);
for (var i = 0; i < len; i += 4) res.setUint32(i,this.getUint32(i) ^ bar.getUint32(i));
return res;
}
}
Just do like
var u = new BitArray(12);
I hope it helps.

Probably [definitely] not the most efficient way to do this, but a string of zeros and ones can be parsed as a number as a base 2 number and converted into a hexadecimal number and finally a buffer.
const bufferFromBinaryString = (binaryRepresentation = '01010101') =>
Buffer.from(
parseInt(binaryRepresentation, 2).toString(16), 'hex');
Again, not efficient; but I like this approach because of the relative simplicity.

Thanks for a wonderfully simple class that does just what I need.
I did find a couple of edge-case bugs while testing:
get(n) {
return (this.backingArray[n/32|0] & 1 << n % 32) != 0
// test of > 0 fails for bit 31
}
forEach(callback) {
this.backingArray.forEach((number, container)=>{
const max = container == this.backingArray.length-1 && this.length%32
? this.length%32 : 32;
// tricky edge-case: at length-1 when length%32 == 0,
// need full 32 bits not 0 bits
for(let x=0; x<max; x++) {
callback((number & 1<<x)!=0, 32*container+x) // see fix in get()
}
})
My final implementation fixed the above bugs and changed the backArray to be a Uint8Array instead of Array, which avoids signed int bugs.

How to convert decimal to hexadecimal in JavaScript

How do you convert decimal values to their hexadecimal equivalent in JavaScript?

Convert a number to a hexadecimal string with:
hexString = yourNumber.toString(16);
And reverse the process with:
yourNumber = parseInt(hexString, 16);

If you need to handle things like bit fields or 32-bit colors, then you need to deal with signed numbers. The JavaScript function toString(16) will return a negative hexadecimal number which is usually not what you want. This function does some crazy addition to make it a positive number.
function decimalToHexString(number)
{
if (number < 0)
{
number = 0xFFFFFFFF + number + 1;
}
return number.toString(16).toUpperCase();
}
console.log(decimalToHexString(27));
console.log(decimalToHexString(48.6));

The code below will convert the decimal value d to hexadecimal. It also allows you to add padding to the hexadecimal result. So 0 will become 00 by default.
function decimalToHex(d, padding) {
var hex = Number(d).toString(16);
padding = typeof (padding) === "undefined" || padding === null ? padding = 2 : padding;
while (hex.length < padding) {
hex = "0" + hex;
}
return hex;
}

function toHex(d) {
return ("0"+(Number(d).toString(16))).slice(-2).toUpperCase()
}

For completeness, if you want the two's-complement hexadecimal representation of a negative number, you can use the zero-fill-right shift >>> operator. For instance:
> (-1).toString(16)
"-1"
> ((-2)>>>0).toString(16)
"fffffffe"
There is however one limitation: JavaScript bitwise operators treat their operands as a sequence of 32 bits, that is, you get the 32-bits two's complement.

With padding:
function dec2hex(i) {
return (i+0x10000).toString(16).substr(-4).toUpperCase();
}

The accepted answer did not take into account single digit returned hexadecimal codes. This is easily adjusted by:
function numHex(s)
{
var a = s.toString(16);
if ((a.length % 2) > 0) {
a = "0" + a;
}
return a;
}
and
function strHex(s)
{
var a = "";
for (var i=0; i<s.length; i++) {
a = a + numHex(s.charCodeAt(i));
}
return a;
}
I believe the above answers have been posted numerous times by others in one form or another. I wrap these in a toHex() function like so:
function toHex(s)
{
var re = new RegExp(/^\s*(\+|-)?((\d+(\.\d+)?)|(\.\d+))\s*$/);
if (re.test(s)) {
return '#' + strHex( s.toString());
}
else {
return 'A' + strHex(s);
}
}
Note that the numeric regular expression came from 10+ Useful JavaScript Regular Expression Functions to improve your web applications efficiency.
Update: After testing this thing several times I found an error (double quotes in the RegExp), so I fixed that. HOWEVER! After quite a bit of testing and having read the post by almaz - I realized I could not get negative numbers to work.
Further - I did some reading up on this and since all JavaScript numbers are stored as 64 bit words no matter what - I tried modifying the numHex code to get the 64 bit word. But it turns out you can not do that. If you put "3.14159265" AS A NUMBER into a variable - all you will be able to get is the "3", because the fractional portion is only accessible by multiplying the number by ten(IE:10.0) repeatedly. Or to put that another way - the hexadecimal value of 0xF causes the floating point value to be translated into an integer before it is ANDed which removes everything behind the period. Rather than taking the value as a whole (i.e.: 3.14159265) and ANDing the floating point value against the 0xF value.
So the best thing to do, in this case, is to convert the 3.14159265 into a string and then just convert the string. Because of the above, it also makes it easy to convert negative numbers because the minus sign just becomes 0x26 on the front of the value.
So what I did was on determining that the variable contains a number - just convert it to a string and convert the string. This means to everyone that on the server side you will need to unhex the incoming string and then to determine the incoming information is numeric. You can do that easily by just adding a "#" to the front of numbers and "A" to the front of a character string coming back. See the toHex() function.
Have fun!
After another year and a lot of thinking, I decided that the "toHex" function (and I also have a "fromHex" function) really needed to be revamped. The whole question was "How can I do this more efficiently?" I decided that a to/from hexadecimal function should not care if something is a fractional part but at the same time it should ensure that fractional parts are included in the string.
So then the question became, "How do you know you are working with a hexadecimal string?". The answer is simple. Use the standard pre-string information that is already recognized around the world.
In other words - use "0x". So now my toHex function looks to see if that is already there and if it is - it just returns the string that was sent to it. Otherwise, it converts the string, number, whatever. Here is the revised toHex function:
/////////////////////////////////////////////////////////////////////////////
// toHex(). Convert an ASCII string to hexadecimal.
/////////////////////////////////////////////////////////////////////////////
toHex(s)
{
if (s.substr(0,2).toLowerCase() == "0x") {
return s;
}
var l = "0123456789ABCDEF";
var o = "";
if (typeof s != "string") {
s = s.toString();
}
for (var i=0; i<s.length; i++) {
var c = s.charCodeAt(i);
o = o + l.substr((c>>4),1) + l.substr((c & 0x0f),1);
}
return "0x" + o;
}
This is a very fast function that takes into account single digits, floating point numbers, and even checks to see if the person is sending a hex value over to be hexed again. It only uses four function calls and only two of those are in the loop. To un-hex the values you use:
/////////////////////////////////////////////////////////////////////////////
// fromHex(). Convert a hex string to ASCII text.
/////////////////////////////////////////////////////////////////////////////
fromHex(s)
{
var start = 0;
var o = "";
if (s.substr(0,2).toLowerCase() == "0x") {
start = 2;
}
if (typeof s != "string") {
s = s.toString();
}
for (var i=start; i<s.length; i+=2) {
var c = s.substr(i, 2);
o = o + String.fromCharCode(parseInt(c, 16));
}
return o;
}
Like the toHex() function, the fromHex() function first looks for the "0x" and then it translates the incoming information into a string if it isn't already a string. I don't know how it wouldn't be a string - but just in case - I check. The function then goes through, grabbing two characters and translating those in to ASCII characters. If you want it to translate Unicode, you will need to change the loop to going by four(4) characters at a time. But then you also need to ensure that the string is NOT divisible by four. If it is - then it is a standard hexadecimal string. (Remember the string has "0x" on the front of it.)
A simple test script to show that -3.14159265, when converted to a string, is still -3.14159265.
<?php
echo <<<EOD
<html>
<head><title>Test</title>
<script>
var a = -3.14159265;
alert( "A = " + a );
var b = a.toString();
alert( "B = " + b );
</script>
</head>
<body>
</body>
</html>
EOD;
?>
Because of how JavaScript works in respect to the toString() function, all of those problems can be eliminated which before were causing problems. Now all strings and numbers can be converted easily. Further, such things as objects will cause an error to be generated by JavaScript itself. I believe this is about as good as it gets. The only improvement left is for W3C to just include a toHex() and fromHex() function in JavaScript.

Without the loop:
function decimalToHex(d) {
var hex = Number(d).toString(16);
hex = "000000".substr(0, 6 - hex.length) + hex;
return hex;
}
// Or "#000000".substr(0, 7 - hex.length) + hex;
// Or whatever
// *Thanks to MSDN
Also isn't it better not to use loop tests that have to be evaluated?
For example, instead of:
for (var i = 0; i < hex.length; i++){}
have
for (var i = 0, var j = hex.length; i < j; i++){}

Combining some of these good ideas for an RGB-value-to-hexadecimal function (add the # elsewhere for HTML/CSS):
function rgb2hex(r,g,b) {
if (g !== undefined)
return Number(0x1000000 + r*0x10000 + g*0x100 + b).toString(16).substring(1);
else
return Number(0x1000000 + r[0]*0x10000 + r[1]*0x100 + r[2]).toString(16).substring(1);
}

Constrained/padded to a set number of characters:
function decimalToHex(decimal, chars) {
return (decimal + Math.pow(16, chars)).toString(16).slice(-chars).toUpperCase();
}

For anyone interested, here's a JSFiddle comparing most of the answers given to this question.
And here's the method I ended up going with:
function decToHex(dec) {
return (dec + Math.pow(16, 6)).toString(16).substr(-6)
}
Also, bear in mind that if you're looking to convert from decimal to hex for use in CSS as a color data type, you might instead prefer to extract the RGB values from the decimal and use rgb().
For example (JSFiddle):
let c = 4210330 // your color in decimal format
let rgb = [(c & 0xff0000) >> 16, (c & 0x00ff00) >> 8, (c & 0x0000ff)]
// Vanilla JS:
document.getElementById('some-element').style.color = 'rgb(' + rgb + ')'
// jQuery:
$('#some-element').css('color', 'rgb(' + rgb + ')')
This sets #some-element's CSS color property to rgb(64, 62, 154).

var number = 3200;
var hexString = number.toString(16);
The 16 is the radix and there are 16 values in a hexadecimal number :-)

function dec2hex(i)
{
var result = "0000";
if (i >= 0 && i <= 15) { result = "000" + i.toString(16); }
else if (i >= 16 && i <= 255) { result = "00" + i.toString(16); }
else if (i >= 256 && i <= 4095) { result = "0" + i.toString(16); }
else if (i >= 4096 && i <= 65535) { result = i.toString(16); }
return result
}

If you want to convert a number to a hexadecimal representation of an RGBA color value, I've found this to be the most useful combination of several tips from here:
function toHexString(n) {
if(n < 0) {
n = 0xFFFFFFFF + n + 1;
}
return "0x" + ("00000000" + n.toString(16).toUpperCase()).substr(-8);
}

AFAIK comment 57807 is wrong and should be something like:
var hex = Number(d).toString(16);
instead of
var hex = parseInt(d, 16);
function decimalToHex(d, padding) {
var hex = Number(d).toString(16);
padding = typeof (padding) === "undefined" || padding === null ? padding = 2 : padding;
while (hex.length < padding) {
hex = "0" + hex;
}
return hex;
}

And if the number is negative?
Here is my version.
function hexdec (hex_string) {
hex_string=((hex_string.charAt(1)!='X' && hex_string.charAt(1)!='x')?hex_string='0X'+hex_string : hex_string);
hex_string=(hex_string.charAt(2)<8 ? hex_string =hex_string-0x00000000 : hex_string=hex_string-0xFFFFFFFF-1);
return parseInt(hex_string, 10);
}

As the accepted answer states, the easiest way to convert from decimal to hexadecimal is var hex = dec.toString(16). However, you may prefer to add a string conversion, as it ensures that string representations like "12".toString(16) work correctly.
// Avoids a hard-to-track-down bug by returning `c` instead of `12`
(+"12").toString(16);
To reverse the process you may also use the solution below, as it is even shorter.
var dec = +("0x" + hex);
It seems to be slower in Google Chrome and Firefox, but is significantly faster in Opera. See http://jsperf.com/hex-to-dec.

I'm doing conversion to hex string in a pretty large loop, so I tried several techniques in order to find the fastest one. My requirements were to have a fixed-length string as a result, and encode negative values properly (-1 => ff..f).
Simple .toString(16) didn't work for me since I needed negative values to be properly encoded. The following code is the quickest I've tested so far on 1-2 byte values (note that symbols defines the number of output symbols you want to get, that is for 4-byte integer it should be equal to 8):
var hex = ['0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 'a', 'b', 'c', 'd', 'e', 'f'];
function getHexRepresentation(num, symbols) {
var result = '';
while (symbols--) {
result = hex[num & 0xF] + result;
num >>= 4;
}
return result;
}
It performs faster than .toString(16) on 1-2 byte numbers and slower on larger numbers (when symbols >= 6), but still should outperform methods that encode negative values properly.

Converting hex color numbers to hex color strings:
A simple solution with toString and ES6 padStart for converting hex color numbers to hex color strings.
const string = `#${color.toString(16).padStart(6, '0')}`;
For example:
0x000000 will become #000000
0xFFFFFF will become #FFFFFF
Check this example in a fiddle here

How to convert decimal to hexadecimal in JavaScript
I wasn't able to find a brutally clean/simple decimal to hexadecimal conversion that didn't involve a mess of functions and arrays ... so I had to make this for myself.
function DecToHex(decimal) { // Data (decimal)
length = -1; // Base string length
string = ''; // Source 'string'
characters = [ '0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 'A', 'B', 'C', 'D', 'E', 'F' ]; // character array
do { // Grab each nibble in reverse order because JavaScript has no unsigned left shift
string += characters[decimal & 0xF]; // Mask byte, get that character
++length; // Increment to length of string
} while (decimal >>>= 4); // For next character shift right 4 bits, or break on 0
decimal += 'x'; // Convert that 0 into a hex prefix string -> '0x'
do
decimal += string[length];
while (length--); // Flip string forwards, with the prefixed '0x'
return (decimal); // return (hexadecimal);
}
/* Original: */
D = 3678; // Data (decimal)
C = 0xF; // Check
A = D; // Accumulate
B = -1; // Base string length
S = ''; // Source 'string'
H = '0x'; // Destination 'string'
do {
++B;
A& = C;
switch(A) {
case 0xA: A='A'
break;
case 0xB: A='B'
break;
case 0xC: A='C'
break;
case 0xD: A='D'
break;
case 0xE: A='E'
break;
case 0xF: A='F'
break;
A = (A);
}
S += A;
D >>>= 0x04;
A = D;
} while(D)
do
H += S[B];
while (B--)
S = B = A = C = D; // Zero out variables
alert(H); // H: holds hexadecimal equivalent

You can do something like this in ECMAScript 6:
const toHex = num => (num).toString(16).toUpperCase();

If you are looking for converting Large integers i.e. Numbers greater than Number.MAX_SAFE_INTEGER -- 9007199254740991, then you can use the following code
const hugeNumber = "9007199254740991873839" // Make sure its in String
const hexOfHugeNumber = BigInt(hugeNumber).toString(16);
console.log(hexOfHugeNumber)

To sum it all up;
function toHex(i, pad) {
if (typeof(pad) === 'undefined' || pad === null) {
pad = 2;
}
var strToParse = i.toString(16);
while (strToParse.length < pad) {
strToParse = "0" + strToParse;
}
var finalVal = parseInt(strToParse, 16);
if ( finalVal < 0 ) {
finalVal = 0xFFFFFFFF + finalVal + 1;
}
return finalVal;
}
However, if you don't need to convert it back to an integer at the end (i.e. for colors), then just making sure the values aren't negative should suffice.

I haven't found a clear answer, without checks if it is negative or positive, that uses two's complement (negative numbers included). For that, I show my solution to one byte:
((0xFF + number +1) & 0x0FF).toString(16);
You can use this instruction to any number bytes, only you add FF in respective places. For example, to two bytes:
((0xFFFF + number +1) & 0x0FFFF).toString(16);
If you want cast an array integer to string hexadecimal:
s = "";
for(var i = 0; i < arrayNumber.length; ++i) {
s += ((0xFF + arrayNumber[i] +1) & 0x0FF).toString(16);
}

In case you're looking to convert to a 'full' JavaScript or CSS representation, you can use something like:
numToHex = function(num) {
var r=((0xff0000&num)>>16).toString(16),
g=((0x00ff00&num)>>8).toString(16),
b=(0x0000ff&num).toString(16);
if (r.length==1) { r = '0'+r; }
if (g.length==1) { g = '0'+g; }
if (b.length==1) { b = '0'+b; }
return '0x'+r+g+b; // ('#' instead of'0x' for CSS)
};
var dec = 5974678;
console.log( numToHex(dec) ); // 0x5b2a96

This is based on Prestaul and Tod's solutions. However, this is a generalisation that accounts for varying size of a variable (e.g. Parsing signed value from a microcontroller serial log).
function decimalToPaddedHexString(number, bitsize)
{
let byteCount = Math.ceil(bitsize/8);
let maxBinValue = Math.pow(2, bitsize)-1;
/* In node.js this function fails for bitsize above 32bits */
if (bitsize > 32)
throw "number above maximum value";
/* Conversion to unsigned form based on */
if (number < 0)
number = maxBinValue + number + 1;
return "0x"+(number >>> 0).toString(16).toUpperCase().padStart(byteCount*2, '0');
}
Test script:
for (let n = 0 ; n < 64 ; n++ ) {
let s=decimalToPaddedHexString(-1, n);
console.log(`decimalToPaddedHexString(-1,${(n+"").padStart(2)}) = ${s.padStart(10)} = ${("0b"+parseInt(s).toString(2)).padStart(34)}`);
}
Test results:
decimalToPaddedHexString(-1, 0) = 0x0 = 0b0
decimalToPaddedHexString(-1, 1) = 0x01 = 0b1
decimalToPaddedHexString(-1, 2) = 0x03 = 0b11
decimalToPaddedHexString(-1, 3) = 0x07 = 0b111
decimalToPaddedHexString(-1, 4) = 0x0F = 0b1111
decimalToPaddedHexString(-1, 5) = 0x1F = 0b11111
decimalToPaddedHexString(-1, 6) = 0x3F = 0b111111
decimalToPaddedHexString(-1, 7) = 0x7F = 0b1111111
decimalToPaddedHexString(-1, 8) = 0xFF = 0b11111111
decimalToPaddedHexString(-1, 9) = 0x01FF = 0b111111111
decimalToPaddedHexString(-1,10) = 0x03FF = 0b1111111111
decimalToPaddedHexString(-1,11) = 0x07FF = 0b11111111111
decimalToPaddedHexString(-1,12) = 0x0FFF = 0b111111111111
decimalToPaddedHexString(-1,13) = 0x1FFF = 0b1111111111111
decimalToPaddedHexString(-1,14) = 0x3FFF = 0b11111111111111
decimalToPaddedHexString(-1,15) = 0x7FFF = 0b111111111111111
decimalToPaddedHexString(-1,16) = 0xFFFF = 0b1111111111111111
decimalToPaddedHexString(-1,17) = 0x01FFFF = 0b11111111111111111
decimalToPaddedHexString(-1,18) = 0x03FFFF = 0b111111111111111111
decimalToPaddedHexString(-1,19) = 0x07FFFF = 0b1111111111111111111
decimalToPaddedHexString(-1,20) = 0x0FFFFF = 0b11111111111111111111
decimalToPaddedHexString(-1,21) = 0x1FFFFF = 0b111111111111111111111
decimalToPaddedHexString(-1,22) = 0x3FFFFF = 0b1111111111111111111111
decimalToPaddedHexString(-1,23) = 0x7FFFFF = 0b11111111111111111111111
decimalToPaddedHexString(-1,24) = 0xFFFFFF = 0b111111111111111111111111
decimalToPaddedHexString(-1,25) = 0x01FFFFFF = 0b1111111111111111111111111
decimalToPaddedHexString(-1,26) = 0x03FFFFFF = 0b11111111111111111111111111
decimalToPaddedHexString(-1,27) = 0x07FFFFFF = 0b111111111111111111111111111
decimalToPaddedHexString(-1,28) = 0x0FFFFFFF = 0b1111111111111111111111111111
decimalToPaddedHexString(-1,29) = 0x1FFFFFFF = 0b11111111111111111111111111111
decimalToPaddedHexString(-1,30) = 0x3FFFFFFF = 0b111111111111111111111111111111
decimalToPaddedHexString(-1,31) = 0x7FFFFFFF = 0b1111111111111111111111111111111
decimalToPaddedHexString(-1,32) = 0xFFFFFFFF = 0b11111111111111111111111111111111
Thrown: 'number above maximum value'
Note: Not too sure why it fails above 32 bitsize

rgb(255, 255, 255) // returns FFFFFF
rgb(255, 255, 300) // returns FFFFFF
rgb(0,0,0) // returns 000000
rgb(148, 0, 211) // returns 9400D3
function rgb(...values){
return values.reduce((acc, cur) => {
let val = cur >= 255 ? 'ff' : cur <= 0 ? '00' : Number(cur).toString(16);
return acc + (val.length === 1 ? '0'+val : val);
}, '').toUpperCase();
}

Arbitrary precision
This solution take on input decimal string, and return hex string. A decimal fractions are supported. Algorithm
split number to sign (s), integer part (i) and fractional part (f) e.g for -123.75 we have s=true, i=123, f=75
integer part to hex:
if i='0' stop
get modulo: m=i%16 (in arbitrary precision)
convert m to hex digit and put to result string
for next step calc integer part i=i/16 (in arbitrary precision)
fractional part
count fractional digits n
multiply k=f*16 (in arbitrary precision)
split k to right part with n digits and put them to f, and left part with rest of digits and put them to d
convert d to hex and add to result.
finish when number of result fractional digits is enough
// #param decStr - string with non-negative integer
// #param divisor - positive integer
function dec2HexArbitrary(decStr, fracDigits=0) {
// Helper: divide arbitrary precision number by js number
// #param decStr - string with non-negative integer
// #param divisor - positive integer
function arbDivision(decStr, divisor)
{
// algorithm https://www.geeksforgeeks.org/divide-large-number-represented-string/
let ans='';
let idx = 0;
let temp = +decStr[idx];
while (temp < divisor) temp = temp * 10 + +decStr[++idx];
while (decStr.length > idx) {
ans += (temp / divisor)|0 ;
temp = (temp % divisor) * 10 + +decStr[++idx];
}
if (ans.length == 0) return "0";
return ans;
}
// Helper: calc module of arbitrary precision number
// #param decStr - string with non-negative integer
// #param mod - positive integer
function arbMod(decStr, mod) {
// algorithm https://www.geeksforgeeks.org/how-to-compute-mod-of-a-big-number/
let res = 0;
for (let i = 0; i < decStr.length; i++)
res = (res * 10 + +decStr[i]) % mod;
return res;
}
// Helper: multiply arbitrary precision integer by js number
// #param decStr - string with non-negative integer
// #param mult - positive integer
function arbMultiply(decStr, mult) {
let r='';
let m=0;
for (let i = decStr.length-1; i >=0 ; i--) {
let n = m+mult*(+decStr[i]);
r= (i ? n%10 : n) + r
m= n/10|0;
}
return r;
}
// dec2hex algorithm starts here
let h= '0123456789abcdef'; // hex 'alphabet'
let m= decStr.match(/-?(.*?)\.(.*)?/) || decStr.match(/-?(.*)/); // separate sign,integer,ractional
let i= m[1].replace(/^0+/,'').replace(/^$/,'0'); // integer part (without sign and leading zeros)
let f= (m[2]||'0').replace(/0+$/,'').replace(/^$/,'0'); // fractional part (without last zeros)
let s= decStr[0]=='-'; // sign
let r=''; // result
if(i=='0') r='0';
while(i!='0') { // integer part
r=h[arbMod(i,16)]+r;
i=arbDivision(i,16);
}
if(fracDigits) r+=".";
let n = f.length;
for(let j=0; j<fracDigits; j++) { // frac part
let k= arbMultiply(f,16);
f = k.slice(-n);
let d= k.slice(0,k.length-n);
r+= d.length ? h[+d] : '0';
}
return (s?'-':'')+r;
}
// -----------
// TESTS
// -----------
let tests = [
["0",2],
["000",2],
["123",0],
["-123",0],
["00.000",2],
["255.75",5],
["-255.75",5],
["127.999",32],
];
console.log('Input Standard Abitrary');
tests.forEach(t=> {
let nonArb = (+t[0]).toString(16).padEnd(17,' ');
let arb = dec2HexArbitrary(t[0],t[1]);
console.log(t[0].padEnd(10,' '), nonArb, arb);
});
// Long Example (40 digits after dot)
let example = "123456789012345678901234567890.09876543210987654321"
console.log(`\nLong Example:`);
console.log('dec:',example);
console.log('hex: ',dec2HexArbitrary(example,40));

The problem basically how many padding zeros to expect.
If you expect string 01 and 11 from Number 1 and 17. it's better to use Buffer as a bridge, with which number is turn into bytes, and then the hex is just an output format of it. And the bytes organization is well controlled by Buffer functions, like writeUInt32BE, writeInt16LE, etc.
import { Buffer } from 'buffer';
function toHex(n) { // 4byte
const buff = Buffer.alloc(4);
buff.writeInt32BE(n);
return buff.toString('hex');
}
> toHex(1)
'00000001'
> toHex(17)
'00000011'
> toHex(-1)
'ffffffff'
> toHex(-1212)
'fffffb44'
> toHex(1212)
'000004bc'

Here's my solution:
hex = function(number) {
return '0x' + Math.abs(number).toString(16);
}
The question says: "How to convert decimal to hexadecimal in JavaScript". While, the question does not specify that the hexadecimal string should begin with a 0x prefix, anybody who writes code should know that 0x is added to hexadecimal codes to distinguish hexadecimal codes from programmatic identifiers and other numbers (1234 could be hexadecimal, decimal, or even octal).
Therefore, to correctly answer this question, for the purpose of script-writing, you must add the 0x prefix.
The Math.abs(N) function converts negatives to positives, and as a bonus, it doesn't look like somebody ran it through a wood-chipper.
The answer I wanted, would have had a field-width specifier, so we could for example show 8/16/32/64-bit values the way you would see them listed in a hexadecimal editing application. That, is the actual, correct answer.