Convert string to unique numeric value in php

Question

I needed a unique internal int to use that would represent several random content types that I couldn't hard code as class constants or config options. Yet I didn't want to refer to "2" when I could use the easier to follow term "post". So though up the a way to get a INT value from a string which seems to work fairly collision free for a couple items.

Is there a way to do this in php rather than what I built?

$strings = array('post', 'comment', 'blog', 'article', 'forum', 'news', 'page');

foreach( $strings as $string ) {

    print $string . ' = ';

    $string = str_split($string);
    $sum = 0;
    foreach( $string as $char ) {
     $sum += ord($char);
    }

    var_dump( $sum );

    print '<br />';
}

This has nothing to do with user input so don't worry about the obvious flaws in the design.

:EDIT:

I need a numeric index for storing in a database for fast look ups. i.e. to tell "posts" from "articles" or "news". Yet I don't know what those 1-8 types of content will be called so I can't hard code those as constants in the app. Therefore, the best thing I could figure would be to create a numeric version of each word (which means checking to make sure two words don't sum up to the same number).

If this were for a system that would have more than 8 words then this method is doomed to failure do to the high probability of sum collisions. Also, if I had a choice in knowing what words would be used then this would also be terrible design as class constants would work much better.

Answer 1

$strings = array('post' => 0, 'comment' => 1, 'blog' => 2, 'article' => 3, 'forum' => 4, 'news' => 5, 'page' => 6);

$s = 'comment';
$s_as_int = $strings[$s];

Answer 2

You can't use a single unique integer, since two different integers have a very high chance of adding up to the same result. 3+3 = 6; 2+2+2 = 6;

You are better off using a non-reversable hash. This is more of a unique sequence, then the sum of the sequence.

ie: md5(), sha1(), sha256() etc.

You can also use PHP5's spl_object_hash() which assigns a unique hash to every Object. http://us3.php.net/manual/en/function.spl-object-hash.php

ps: if you needed integers, you'd be using really large integers. Thus you can just concatenate the integers in your example, instead of adding them.

Answer 3

<?php
$strings = array('post', 'comment', 'blog', 'article', 'forum', 'news', 'page');

foreach( $strings as $string )
{
    echo $string . ' = ' . base_convert(md5($string), 16, 10) . '<br />';
}
?>

Answer 4

I think you're looking to achieve an analog to an "enumerated type" like in C, C++, C#, Pascal, Java, etc. You don't mention what type of database you are using, but MySQL makes this easy to accomplish by using the ENUM type.

Answer 5

You could use crc32, which is a checksumming function that simply returns a 4 byte integer.

Answer 6

Couldn't you just array_flip() or array_search() your dictionary?

Or store your dictionary in the database as well in another table called dictionary that looks something like this (MySQL)

CREATE TABLE `dictionary` (
  `id` INT NOT NULL PRIMARY KEY AUTO_INCREMENT,
  `word` VARCHAR(25),
  UNIQUE `word` (`word`)
)