50

Is there any way to create unique keys like those used in YouTube video urls (ex: https://www.youtube.com/watch?v=nWChTnkVdKE)?

PowerOfM
  • 303
  • 4
  • 8
kusanagi
  • 14,296
  • 20
  • 86
  • 111
  • 2
    Take a look at this tutorial: - [Create short IDs with PHP - Like Youtube or TinyURL](http://kevin.vanzonneveld.net/techblog/article/create_short_ids_with_php_like_youtube_or_tinyurl/) – Sarfraz Oct 21 '10 at 18:49
  • If you have integer IDs in your database, you can simply encode them using [`(new Id())->encode($id)`](https://github.com/delight-im/PHP-IDs). – caw Nov 20 '19 at 15:24
  • See Sean Coates' Blog: - [Arbitrary Incrementer in PHP](http://seancoates.com/blogs/arbitrary-incrementer-in-php) > I wanted the URL shortener to make the shortest possible URLs. To keep the number of characters in a URL short, I had to increase the set of characters that could comprise a key. and also the article linked within: - [manufacturing flic.kr style photo URLs](http://www.flickr.com/groups/api/discuss/72157616713786392/) On a sidenote, this [has likely been answered before](https://stackoverflow.com/search?q=url+shortener+php) – Gordon Oct 21 '10 at 20:08

10 Answers10

43

The idea is to convert a unique integer (such as current time) in an other mathematical base.

A very simple way in PHP :

// With this precision (microsecond) ID will looks like '2di2adajgq6h'

// From PHP 7.4.0 this is needed, otherwise a warning is displayed
$cleanNumber = preg_replace( '/[^0-9]/', '', microtime(false) );
$id = base_convert($cleanNumber, 10, 36);


// With less precision (second) ID will looks like 'niu7pj'

$id = base_convert(time(), 10, 36);
Peter VARGA
  • 4,780
  • 3
  • 39
  • 75
Diantre
  • 446
  • 5
  • 2
  • Love it! Simple and efficient! – iamtankist Jan 13 '16 at 14:15
  • Why is `36` a valid base? – gskema Apr 19 '16 at 10:38
  • 3
    @gskema Because it's 0-9 then A-Z – Muqito Apr 27 '16 at 00:43
  • 8
    Just a note. Do not use this for things like verification/activation codes that are sent via email/sms. They are predictable and could be brute forced. – BugHunterUK Nov 28 '16 at 17:41
  • 1
    Thanks! To avoid getting the same IDs (e.g for multiple executions) I'd go for a random `int`: `base_convert(rand(1000000000,PHP_INT_MAX), 10, 36)` – Ismail Jan 09 '17 at 17:06
  • @SamuelElh: should be random_int() now but appreciate! – Adrian Föder Mar 15 '17 at 09:37
  • 2
    Just to be very explicit regarding @BugHunterUK's comment - and this is critical - this is **not** cryptographically strong, it is **not** random and it **does** give away the timestamp when generation occurred. That said, for any context which doesn't need those, **it's great**. It's not far off the `uniqid()` function. – wally Aug 15 '17 at 08:00
  • I'm getting this error on PHP 7.4 "Deprecated: Invalid characters passed for attempted conversion, these have been ignored in...." – Alvaro Oct 22 '20 at 15:17
  • @Alvaro I fixed the code so it runs now also under PHP 7.4 – Peter VARGA Nov 11 '20 at 12:27
7

A small PHP class to generate YouTube-like hashes from one or many numbers. Use hashids when you do not want to expose your database ids to the user.

Source: https://github.com/ivanakimov/hashids.php

Junaid Atari
  • 533
  • 7
  • 17
6

Here is a small function that generates unique key randomly each time. It has very fewer chances to repeat same unique ID.

function uniqueKey($limit = 10) {

    $characters = '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ';

    $randstring = '';

    for ($i = 0; $i < $limit; $i++) {

        $randstring .= $characters[rand(0, strlen($characters))];
    }

    return $randstring;
}
Ostap Brehin
  • 3,240
  • 3
  • 25
  • 28
Bhaskar Bhatt
  • 1,399
  • 13
  • 19
5

Use whichever you like :-)

// Generates Alphanumeric Output

function generateRandomID() {
    // http://mohnish.in
    $required_length = 11;
    $limit_one = rand();
    $limit_two = rand();
    $randomID = substr(uniqid(sha1(crypt(md5(rand(min($limit_one, $limit_two), max($limit_one, $limit_two)))))), 0, $required_length);
    return $randomID;
}

// Generates only alphabetic output

function anotherRandomIDGenerator() {
    // Copyright: http://snippets.dzone.com/posts/show/3123
    $len = 8;
    $base='ABCDEFGHKLMNOPQRSTWXYZabcdefghjkmnpqrstwxyz';
    $max=strlen($base)-1;
    $activatecode='';
    mt_srand((double)microtime()*1000000);
    while (strlen($activatecode)<$len+1)
    $activatecode.=$base{mt_rand(0,$max)};
    return $activatecode;
}
MT.
  • 1,915
  • 3
  • 16
  • 19
5

For me best algorithm is this: Create Youtube-Like IDs With PHP/Python/Javascript/Java/SQL

<?php
/**
 * Translates a number to a short alhanumeric version
 *
 * Translated any number up to 9007199254740992
 * to a shorter version in letters e.g.:
 * 9007199254740989 --> PpQXn7COf
 *
 * specifiying the second argument true, it will
 * translate back e.g.:
 * PpQXn7COf --> 9007199254740989
 *
 * this function is based on any2dec && dec2any by
 * fragmer[at]mail[dot]ru
 * see: http://nl3.php.net/manual/en/function.base-convert.php#52450
 *
 * If you want the alphaID to be at least 3 letter long, use the
 * $pad_up = 3 argument
 *
 * In most cases this is better than totally random ID generators
 * because this can easily avoid duplicate ID's.
 * For example if you correlate the alpha ID to an auto incrementing ID
 * in your database, you're done.
 *
 * The reverse is done because it makes it slightly more cryptic,
 * but it also makes it easier to spread lots of IDs in different
 * directories on your filesystem. Example:
 * $part1 = substr($alpha_id,0,1);
 * $part2 = substr($alpha_id,1,1);
 * $part3 = substr($alpha_id,2,strlen($alpha_id));
 * $destindir = "/".$part1."/".$part2."/".$part3;
 * // by reversing, directories are more evenly spread out. The
 * // first 26 directories already occupy 26 main levels
 *
 * more info on limitation:
 * - http://blade.nagaokaut.ac.jp/cgi-bin/scat.rb/ruby/ruby-talk/165372
 *
 * if you really need this for bigger numbers you probably have to look
 * at things like: http://theserverpages.com/php/manual/en/ref.bc.php
 * or: http://theserverpages.com/php/manual/en/ref.gmp.php
 * but I haven't really dugg into this. If you have more info on those
 * matters feel free to leave a comment.
 *
 * The following code block can be utilized by PEAR's Testing_DocTest
 * <code>
 * // Input //
 * $number_in = 2188847690240;
 * $alpha_in  = "SpQXn7Cb";
 *
 * // Execute //
 * $alpha_out  = alphaID($number_in, false, 8);
 * $number_out = alphaID($alpha_in, true, 8);
 *
 * if ($number_in != $number_out) {
 *   echo "Conversion failure, ".$alpha_in." returns ".$number_out." instead of the ";
 *   echo "desired: ".$number_in."\n";
 * }
 * if ($alpha_in != $alpha_out) {
 *   echo "Conversion failure, ".$number_in." returns ".$alpha_out." instead of the ";
 *   echo "desired: ".$alpha_in."\n";
 * }
 *
 * // Show //
 * echo $number_out." => ".$alpha_out."\n";
 * echo $alpha_in." => ".$number_out."\n";
 * echo alphaID(238328, false)." => ".alphaID(alphaID(238328, false), true)."\n";
 *
 * // expects:
 * // 2188847690240 => SpQXn7Cb
 * // SpQXn7Cb => 2188847690240
 * // aaab => 238328
 *
 * </code>
 *
 * @author  Kevin van Zonneveld &lt;kevin@vanzonneveld.net>
 * @author  Simon Franz
 * @author  Deadfish
 * @author  SK83RJOSH
 * @copyright 2008 Kevin van Zonneveld (http://kevin.vanzonneveld.net)
 * @license   http://www.opensource.org/licenses/bsd-license.php New BSD Licence
 * @version   SVN: Release: $Id: alphaID.inc.php 344 2009-06-10 17:43:59Z kevin $
 * @link    http://kevin.vanzonneveld.net/
 *
 * @param mixed   $in   String or long input to translate
 * @param boolean $to_num  Reverses translation when true
 * @param mixed   $pad_up  Number or boolean padds the result up to a specified length
 * @param string  $pass_key Supplying a password makes it harder to calculate the original ID
 *
 * @return mixed string or long
 */
function alphaID($in, $to_num = false, $pad_up = false, $pass_key = null)
{
  $out   =   '';
  $index = 'abcdefghijklmnopqrstuvwxyz0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ';
  $base  = strlen($index);

  if ($pass_key !== null) {
    // Although this function's purpose is to just make the
    // ID short - and not so much secure,
    // with this patch by Simon Franz (http://blog.snaky.org/)
    // you can optionally supply a password to make it harder
    // to calculate the corresponding numeric ID

    for ($n = 0; $n < strlen($index); $n++) {
      $i[] = substr($index, $n, 1);
    }

    $pass_hash = hash('sha256',$pass_key);
    $pass_hash = (strlen($pass_hash) < strlen($index) ? hash('sha512', $pass_key) : $pass_hash);

    for ($n = 0; $n < strlen($index); $n++) {
      $p[] =  substr($pass_hash, $n, 1);
    }

    array_multisort($p, SORT_DESC, $i);
    $index = implode($i);
  }

  if ($to_num) {
    // Digital number  <<--  alphabet letter code
    $len = strlen($in) - 1;

    for ($t = $len; $t >= 0; $t--) {
      $bcp = bcpow($base, $len - $t);
      $out = $out + strpos($index, substr($in, $t, 1)) * $bcp;
    }

    if (is_numeric($pad_up)) {
      $pad_up--;

      if ($pad_up > 0) {
        $out -= pow($base, $pad_up);
      }
    }
  } else {
    // Digital number  -->>  alphabet letter code
    if (is_numeric($pad_up)) {
      $pad_up--;

      if ($pad_up > 0) {
        $in += pow($base, $pad_up);
      }
    }

    for ($t = ($in != 0 ? floor(log($in, $base)) : 0); $t >= 0; $t--) {
      $bcp = bcpow($base, $t);
      $a   = floor($in / $bcp) % $base;
      $out = $out . substr($index, $a, 1);
      $in  = $in - ($a * $bcp);
    }
  }

  return $out;
}
Ernestas Stankevičius
  • 2,420
  • 2
  • 24
  • 30
  • That's the most relevant answer to the question - Youtube is using the same style alphanumeric case-sensitive IDs. Other answers suggest alternatives to that - good ones, but still alternatives (not the same). It's important to mention that shortlink services also use this style because the case sensitivity allows more combinations of shorter IDs and long numbers. – Dimitar Atanasov Apr 13 '21 at 14:29
2

I was looking for a very quick and simple solution (that was NOT intended to be secure in any way) that would replicate the actual hash code upon regeneration (unlike PHP's microtime unique id). This one did the trick.

http://php.net/manual/en/book.hash.php

So:

for($i=0;$i<count($myArray);$i++) {
    $hashCode = "#".hash('crc32b', $myArray[$i]."-".$i);        
}

Yields:

#179507fa
#8e9c5640
#f99b66d6
#67fff375
#10f8c3e3

etc

I won't pretend to know how all the different hash algorithms are used but for non-secure uses (like link #anchors) this is handy.

Patryk
  • 22,602
  • 44
  • 128
  • 244
cbmtrx
  • 591
  • 8
  • 16
2

all the methods above are good but make sure you know and not assume the generated string is unique. What I have done in the past is made a recursive function that checks the string against my database and if it is unique it returns the value, otherwise it just runs again. This will hardly happen though depending how long your unique string is. Its just good to know its unique.

Darren
  • 10,631
  • 8
  • 42
  • 64
1
<?php
function keygen(){
    $chars = "bcdfghjklmnpqrstvwxyz";
    $chars .= "BCDFGHJKLMNPQRSTVWXYZ";
    $chars .= "0123456789";
    while(1){
        $key = '';
        srand((double)microtime()*1000000);
        for($i = 0; $i < 10; $i++){
            $key .= substr($chars,(rand()%(strlen($chars))), 1);
        }
        break;
    }
    return $key;
}

//echo keygen();
?>
madhushankarox
  • 1,448
  • 1
  • 13
  • 17
  • 4
    When providing code that solves the problem, it is best to also give at least a short explanation of how it works so that folks reading won't have to mentally parse it line by line to understand the differences. – Fluffeh Sep 27 '12 at 11:09
1

From comments on here:

<?php 
function generateRandStr($length){ 
      $randstr = ""; 
      for($i=0; $i<$length; $i++){ 
         $randnum = mt_rand(0,61); 
         if($randnum < 10){ 
            $randstr .= chr($randnum+48); 
         }else if($randnum < 36){ 
            $randstr .= chr($randnum+55); 
         }else{ 
            $randstr .= chr($randnum+61); 
         } 
      } 
      return $randstr; 
   } 
?> 

Simply use: 
generateRandStr(10); 

Sample output: $%29zon(4f

You can mess around with this function to generate just alphanumeric, or just alphabetic characters.

Chetan
  • 46,743
  • 31
  • 106
  • 145
1

For most purposes uniqid will suffice, if you need to make sure there's absolutely no clash, more convoluted measures are necessary.

Wrikken
  • 69,272
  • 8
  • 97
  • 136