How to convert last 4 bytes in an array to an integer? - javascript

If I have an Uint8Array array in JavaScript, how would I get the last four bytes and then convert that to an int? Using C# I would do something like this:
int count = BitConverter.ToInt32(array, array.Length - 4);
Is there an inequivalent way to do this using JavaScript?

Access the underlying ArrayBuffer and create a new TypedArray with a slice of its bytes:
var u8 = new Uint8Array([1,2,3,4,5,6]); // original array
var u32bytes = u8.buffer.slice(-4); // last four bytes as a new `ArrayBuffer`
var uint = new Uint32Array(u32bytes)[0];
If the TypedArray does not cover the entire buffer, you need to be a little trickier, but not much:
var startbyte = u8.byteOffset + u8.byteLength - Uint32Array.BYTES_PER_ELEMENT;
var u32bytes = u8.buffer.slice(startbyte, startbyte + Uint32Array.BYTES_PER_ELEMENT);
This works in both cases.
If the bytes you want fit in the alignment boundary of your underlying buffer for the datatype (e.g., you want the 32-bit value of bytes 4-8 of the underlying buffer), you can avoid copying the bytes with slice() and just supply a byteoffset to the view constructor, as in #Bergi's answer.
Below is a very-lightly-tested function that should get the scalar value of any offset you want. It will avoid copying if possible.
function InvalidArgument(msg) {
this.message = msg | null;
}
function scalarValue(buf_or_view, byteOffset, type) {
var buffer, bufslice, view, sliceLength = type.BYTES_PER_ELEMENT;
if (buf_or_view instanceof ArrayBuffer) {
buffer = buf_or_view;
if (byteOffset < 0) {
byteOffset = buffer.byteLength - byteOffset;
}
} else if (buf_or_view.buffer instanceof ArrayBuffer) {
view = buf_or_view;
buffer = view.buffer;
if (byteOffset < 0) {
byteOffset = view.byteOffset + view.byteLength + byteOffset;
} else {
byteOffset = view.byteOffset + byteOffset;
}
return scalarValue(buffer, view.byteOffset + byteOffset, type);
} else {
throw new InvalidArgument('buf_or_view must be ArrayBuffer or have a .buffer property');
}
// assert buffer instanceof ArrayBuffer
// assert byteOffset > 0
// assert byteOffset relative to entire buffer
try {
// try in-place first
// only works if byteOffset % slicelength === 0
return (new type(buffer, byteOffset, 1))[0]
} catch (e) {
// if this doesn't work, we need to copy the bytes (slice them out)
bufslice = buffer.slice(byteOffset, byteOffset + sliceLength);
return (new type(bufslice, 0, 1))[0]
}
}
You would use it like this:
// positive or negative byte offset
// relative to beginning or end *of a view*
100992003 === scalarValueAs(u8, -4, Uint32Array)
// positive or negative byte offset
// relative to the beginning or end *of a buffer*
100992003 === scalarValue(u8.buffer, -4, Uint32Array)

Do you have an example? I think this would do it:
var result = ((array[array.length - 1]) |
(array[array.length - 2] << 8) |
(array[array.length - 3] << 16) |
(array[array.length - 4] << 24));

Nowadays if you can live with IE 11+ / Chrome 49+ / Firefox 50+, then you can use DataView to make your life almost as easy as in C#:
var u8array = new Uint8Array([0xFF, 0xFF, 0xFF, 0xFF]); // -1
var view = new DataView(u8array.buffer)
console.log("result:" + view.getInt32());
Test it here: https://jsfiddle.net/3udtek18/1/

A little inelegant, but if you can do it manually based on the endianess.
Little endian:
var count = 0;
// assuming the array has at least four elements
for(var i = array.length - 1; i >= array.length - 4; i--)
{
count = count << 8 + array[i];
}
Big endian:
var count = 0;
// assuming the array has at least four elements
for(var i = array.length - 4; i <= array.length - 1 ; i++)
{
count = count << 8 + array[i];
}
This can be extended to other data lengths
Edit: Thanks to David for pointing out my typos

I'm surprised that the other answers don't use the native Buffer object, which provides a lot of this tooling in a simple, native library. It's possible that this library isn't widely used for bitpacking/unpacking simply because people don't think to check here and it took me a while to find it, too, but it's the right tool for bitpacking/unpacking in nodejs/javascript/typescript.
You can use it like so:
// Create a simple array with 5 elements. We'll pop the last 4 and you should expect the end value to be 1 because this is a little-endian array with all zeros other than the 1 in the littlest(?)-endian
const array = [0, 1, 0, 0, 0]
// get the last 4 elements of your array and convert it to a Buffer
const buffer = Buffer.from(array.slice(-4));
// Use the native Buffer type to read the object as an (U) unsigned (LE) little-endian 32 (32 bits) integer
const value = Buffer.readUInt32LE();
Or, more concisely:
const value = Buffer.from(array.slice(-4)).readUInt32LE();

It should be more efficient to just create an Uint32Array view on the same ArrayBuffer and accessing the 32-bit number directly:
var uint8array = new Uint8Array([1,2,3,4,5,6,7,8]);
var uint32array = new Uint32Array(
uint8array.buffer,
uint8array.byteOffset + uint8array.byteLength - 4,
1 // 4Bytes long
);
return uint32array[0];

var a = Uint8Array(6)
a.set([1,2,8,0,0,1])
i1 = a[a.length-4];
i2 = a[a.length-3];
i3 = a[a.length-2];
i4 = a[a.length-1];
console.log(i1<<24 | i2<<16 | i3<<8 | i4);

It's a shame there are not build in ways to do this.
I needed to read variables of variable sizes so based on Imortenson answer I've wrote this little function where p is read position and s is number of bytes to read:
function readUInt(arr, p, s) {
var r = 0;
for (var i = s-1; i >= 0; i--) {
r |= arr[p + i] << (i * 8);
} return r >>> 0;
}
var iable = readUint(arr, arr.length - 4, 4);

Related

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, '-');
}

Hashtable with 64-bit integers

I am looking to implement an O(1) lookup for binary quadkeys
I know that there is no true hashtable for javascript, instead using objects and o(1) lookup with their properties, but the issue there is that the keys are always converted to strings.
I suspect to have over a 10 million entries in memory, and if I have to rely on keys being strings and with the average quadkey string being 11.5 characters that would equate to (10 million entries * 11.5 length * 2 bytes) = 230,000,000 bytes or 230 MB.
Compared to storing as int64 (10 million entries * 8 bytes) = 80,000,000 bytes or 80 MB
I know that int64 isn't supported natively with javascript, but there are libraries that can assist with that in terms of doing the bitwise operations i'd want.
Now, even though there exists libraries that can deal with int64, they are ultimately objects that are not truly representing 8 bytes, so I believe I cannot use any int64 library in a hashtable, instead I was considering using 2-deep hash table with two int32. The first key would be the first 4 bytes, and the 2nd key being the last 4 bytes. It's not as ideal as 1 operation lookup to find a value, but 2 operations which is still good enough.
However, I feel this is not worth it if all keys are stored as strings, or the fact that all numbers are double-precision floating-point format numbers (8 bytes) so it each hash entry would then take up 16 bytes (two "int32" numbers).
My questions are:
1. If you store a number as the key to a property, will it take up 8
bytes of memory or will it convert to string and take up
length*2bytes?
Is there a way to encode binary quadkeys into the double-precision
floating-point format number standard that javascript has adopted,
even if the number makes no sense , the underlying bits serve a
purpose (unique identification of a construct).
PS: I am marking this with nodejs as there may exist libraries that could assist in my end goal
Edit 1:
Seems 1 is possible with Map() and node 0.12.x+
As far as number 2 I was able to use a int64 lib (bytebuffer) and convert a 64int to a buffer.
I wanted to just use the buffer as the key to Map() but it would not let me as the buffer was internally an object, which each instance acts as a new key to Map().
So I considered turning the buffer back into native type, a 64bit double.
Using readDoubleBE I read the buffer as a double, which represents my 64int binary and successfully lets me use it in a map and allows for O(1) lookup.
var ByteBuffer = require("bytebuffer");
var lookup = new Map();
var startStr = "123123738232777";
for(var i = 1; i <= 100000; i++) {
var longStr = startStr + i.toString();
var longVal = new ByteBuffer.Long.fromValue(longStr);
var buffer = new ByteBuffer().writeUint64(longVal).flip().toBuffer();
var doubleRepresentation = buffer.readDoubleBE();
lookup.set(doubleRepresentation, longStr);
}
console.log(exists("12312373823277744234")); // true
console.log(exists("123123738232777442341232332322")); // false
function exists(longStr) {
var longVal = new ByteBuffer.Long.fromValue(longStr);
var doubleRepresentation = new ByteBuffer().writeUint64(longVal).flip().toBuffer().readDoubleBE();
return lookup.has(doubleRepresentation);
}
The code is sloppy and there are probably shortcuts I am missing, so any suggestions/hints are welcomed.
Ideally I wanted to take advantage of bytebuffer's varints so that I can have even more savings with memory but I wasn't sure if that is possible in a Map, because I wasn't able to use a buffer as a key.
Edit 2:
Using memwatch-next I was able to see that max heap size was 497962856 bytes with this method with Set() whereas using a string in Set() was 1111082864 bytes. That is about 500MB vs 1GB, which is a far cry from 80mb and 230mb, not sure where the extra memory use is coming from. It's worth nothing for these memory tests I used Set over Map that way it should only store unique keys in the data structure. (Using Set as just a existence checker, where Map would serve as a lookup)
Because your keys are integers (and are unique) you could just use them as array indices. However, JS arrays are limited to contain max entries bounded by 32 or 64 bit integer depending on your platform.
To overcome this, you could use your two-step approach, but without using objects and their string hashes. You can store it something like this
store[0010][1000] = 'VALUE' - the item with binary key 00101000 is stored under index 0010 in first array and index 1000 in child array
In decimal, you're dealing with store[2][8] = 'VALUE', which is equivalent to doing store[40] = 'VALUE' in 64bit space
You get yourself a tree with all the properties you want:
You can easily look up by key (in two steps)
Your keys are integers
You're dealing with 32bit integers (or less, depending on how you chunk it)
The doubling in memory from Map to Set in your version of node comes from a bad implementation. Well, not "bad" per se just not fit for millions of entries. The simpler handling of Set gets bought with memory. There's no free lunch, as always, sorry.
Why do they use so much in general? They are supposed to handle any object and the method to be able to handle all of the possible varieties is really expensive. It is possible to optimize if all you have is of one kind but you have to check for it and in 99,99% of all cases it's not worth the hassle because the maps and sets and arrays are short, a couple of thousand entries at most. To be bland: developer time is expensive and better spend elsewhere. I could add: it's open source, do it yourself! but I know it's way easier said than done ;-)
You need to roll it yourself. You can use a Uint32Array for that and build a hash-table around it.
The Bing-Maps are encoded with strings of base-4 digits (at most 23) according to MS and the Quad-key description. Using the encoding of the latter (haven't read the former, so it might be wrong in the details) we can put it into two 32-bit integers:
function quadToInts(quad, zoom){
var high,low, qlen, i, c;
high = 0>>>0;
low = 0>>>0
zoom = zoom>>>0;
// checks & balances omitted!
qlen = quad.length;
for(i = 0; i < 16 && i < qlen ;i++){
c = parseInt(quad.charAt(i));
high |= c << (32-(i*2 + 2));
}
// max = 23 characters (says MS)
for(i = 0; i < 8 && i < qlen - 16 ;i++){
c = parseInt(quad.charAt(16 + i));
low |= c << (32-(i*2 + 2));
}
low |= zoom;
return [high,low];
}
And back
// obligatory https://graphics.stanford.edu/~seander/bithacks.html
function rev(v){
var s = 32;
var mask = (~0)>>>0;
while ((s >>>= 1) > 0) {
mask ^= (mask << s)>>>0;
v = ((v >>> s) & mask) | ((v << s) & (~mask)>>>0);
}
return v;
}
function intsToQuad(k1,k2){
var r, high, low, zoom, c, mask;
r = "";
mask = 0x3; // 0b11
high = k1>>>0;
low = k2>>>0;
zoom = low & (0x20 - 1);
low ^= zoom;
high = rev(high);
for(var i = 0;i<16;i++){
c = high & mask;
c = (c<<1 | c>>>1) & mask;
r += c.toString(10);
high >>>= 2;
}
low = rev(low);
for(var i = 0;i<16;i++){
c = low & mask;
c = (c<<1 | c>>>1) & mask;
r += c.toString(10);
low >>>= 2;
}
return [r,zoom];
}
(All quick hacks, please check before use! And the C&P devil might have had its hand in here, too)
A rough sketch for a hash-table base on the following hash-function
// shamelessly stolen from http://www.burtleburtle.net/bob/c/lookup3.c
function hashword(k1, // high word of 64 bit value
k2, // low word of 64 bit value
seed // the seed
){
var a,b,c;
var rot = function(x,k){
return (((x)<<(k)) | ((x)>>>(32-(k))));
};
/* Set up the internal state */
a = b = c = 0xdeadbeef + ((2<<2)>>>0) + seed>>>0;
if(arguments.length === 2){
seed = k1^k2;
}
b+=k2;
a+=k1;
c ^= b; c -= rot(b,14)>>>0;
a ^= c; a -= rot(c,11)>>>0;
b ^= a; b -= rot(a,25)>>>0;
c ^= b; c -= rot(b,16)>>>0;
a ^= c; a -= rot(c,4)>>>0;
b ^= a; b -= rot(a,14)>>>0;
c ^= b; c -= rot(b,24)>>>0;
return c;
}
function hashsize(N){
var highbit = function(n) {
var r = 0 >>> 0;
var m = n >>> 0;
while (m >>>= 1) {
r++;
}
return r;
};
return (1<<(highbit(N)+1))>>>0;
}
function hashmask(N){
return (hashsize(N)-1)>>>0;
}
And the (rather incomplete) code for table handling
/*
Room for 8-byte (64-bit) entries
Table pos. Array pos.
0 0 high, low
1 2 high, low
2 4 high, lowl
...
n n*2 high, low
*/
function HashTable(N){
var buf;
if(!N)
return null;
N = (N+1) * 2;
buf = new ArrayBuffer(hashsize(N) * 8);
this.table = new Uint32Array(buf);
this.mask = hashmask(N);
this.length = this.table.length;
}
HashTable.prototype.set = function(s,z){
var hash, quad, entry, check, i;
entry = null;
quad = quadToInts(s,z);
hash = hashword(quad[0],quad[1]);
entry = hash & this.mask;
check = entry * 2;
if(this.table[check] != 0 || this.table[check + 1] != 0){
//handle collisions here
console.log("collision in SET found")
return null;
} else {
this.table[check] = quad[0];
this.table[check + 1] = quad[1];
}
return entry;
};
HashTable.prototype.exists = function(s,z){
var hash, quad, entry, check, i;
entry = null;
quad = quadToInts(s,z);
hash = hashword(quad[0],quad[1]);
entry = hash & this.mask;
check = entry * 2;
if(this.table[check] != 0 || this.table[check + 1] != 0){
return entry;
}
return -1;
};
HashTable.prototype.get = function(index){
var entry = [0,0];
if(index > this.length)
return null;
entry[0] = this.table[index * 2];
entry[1] = this.table[index * 2 + 1];
return entry;
};
// short test
var ht = new HashTable(100);
ht.set("01231031020112310",17);
ht.set("11231031020112311",12);
ht.set("21231033020112310",1);
ht.set("31231031220112311321312",23);
var s = "";
for(var i=0;i<ht.table.length;i+=2){
if(ht.table[i] !== 0){
var e = intsToQuad(ht.table[i],ht.table[i+1]);
s += e[0] +", " +e[1] + "\n";
}
}
console.log(s)
Collisions should be rare, so a couple of short standard arrays would do to catch them. To handle it you need to add another byte to the 8 bytes for the two integers representing the Quad or, better, set the second integer to all ones (won't happen with a Quad) and the first to the position(s) in the collision array(s).
To add payload is a bit more complicated because you have only a fixed length to do so.
I have set the size of the table to the next higher power of two. That can be too much or even waaay too much and you might be tempted to adapt it, so be aware that the masking doesn't work as expected anymore you need to do a modulo instead.

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, '-');
}

How to use unit8array in javascript?

Accorting to this answer's comment, I would like to use Unit8Array to achieve the same thing:
var bin = parseInt('1000', 2); //=8
I mean:
inputs are:
length = 4
digits = [1,0,0,0]
and after some TypedArray magic the output will be:
8
Lets see a concrete example with (2 bits length):
counts = {};
var a = [[0,1],[0,0],[0,1],[0,0],[0,0],[1,1]];
for(var i in a){
var id = parseInt(a[i].join(''), 2);
if(!++counts[id])counts[id] = 1;
}
console.log(counts); //{0: 3, 1: 2, 3: 1}
my notes are here.
If you want more speed, handle allready the ids as integers, not array, and have bit changed with & and | operations :
counts = {};
var a = [1,0,1,0,0,3];
for(var i=0; i<a.length; i++ ){
var c = counts[a[i]] || 0 ;
counts[a[i]]= c+1 ;
}
console.log(counts); // { 0: 3, 1: 2, 3: 1 }
to set the k-th bit :
id |= 1 << k;
to clear the k-th bit :
id &= !(1 << k);
to swap the k-th bit :
id ^= (1 << k);
to read the k-th bit :
bitK = (id >> k ) & 1;
So you might write small functions to do those operations, with a 99% chance they will get
inlined by the javascript engine, it will be much faster than array handling + parseInt.
You might even write this code directly. Maybe in some case you will be able to
cache 1 << k, but anyway, shift instructions have a very little cost.
Another advantage is that you can handle several tests in a single instruction,
for instance :
var setRequiredMask = ( 1<<3 | 1<<5 ) ; // to test if bits 3 and 5 set
var clearRequiredMask = ( 1<<2 | 1 <<4 ) ; // to test if bit 2 and 4 cleared
var someRequiredMask = ( 1<<0 | 1 <<6 ) ; // to test if either bit 0 or bit 6 to be set
var satisfyConditions = ((id & setRequiredMask) == setRequiredMask) &&
((id & !clearRequiredMask) ==0 ) &&
(id & someRequiredMask) ;
Here we trade 7 access to memory to quite the same number of boolean operations.
If this test is used in a loop to filter an array, the speed gain is huge.
In the same way you can set / clear / swap several bits in a single instruction, test if one condition set is included in the other, test if they are orthogonal, .... all in a very fast way.
Only limit is that you can handle no more than 32 possibilities with this method.
Then you can store the ids in a typed a typed array, depending on the max number
of possibilities, use only a Int8Array or a Int32Array, but i am not sure the gain
in performance will be that interesting compared to [].
Syntax :
var arr = new Int8Array( _length_ );
or
var arr = new Int32Array( _length_ );
Are you trying to get this parseInt(string, radix)
radix An integer that represents the radix of the above mentioned
string.
Not sure, is like this somewhat you expect
var x = new Uint8Array([1,0,0,0]);
var a = x[0].toString()+x[1].toString()+x[2].toString()+x[3].toString();
console.log(parseInt(a, 2));
console.log(a.length);
Is this useful?
var a = 1 << 3; // shift the bits of 1 (0001) 3 places: 1000
if (a == 8) {
// this block will execute
}
Another example:
var i,
result = 0,
len = 4,
bits = [1, 0, 0, 0];
for (i = 0; i < len; ++i) {
result = (result << 1) + bits[i];
}
alert(result);
Keep in mind, parseInt is probably faster than this.

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.

Categories