Removing php array duplicates with regard to a few specific indexes - php

Question

Suppose I have a multi-dimensional array with 100's of sub-arrays. The sub-arrays always have at least 3 indexes, but can have more. I'd like to remove from the big array all the sub-arrays which are duplicates of others in regards to those three indexes. Example with an array that only has two sub-arrays:
array(array(0 => 'a', 1=> 'b', 2 => 'c', 3 => 'd'), array(0 => 'a', 1=> 'b', 2=> 'c', 3=> 'z'))
one of the sub-arrays would be removed, because the first 3 indexes match even though the 4th does not.
I'm looking for the most elegant/efficient solution.

Answer 1

/**
* Create Unique Arrays using an md5 hash
*
* #param array $array
* #return array
*/
function arrayUnique($array, $preserveKeys = false)
{
$arrayRewrite = array();
$arrayHashes = array();
foreach($array as $key => $item) {
$hash = md5(serialize($item));
if (!isset($arrayHashes[$hash])) {
$arrayHashes[$hash] = $hash;
if ($preserveKeys) {
$arrayRewrite[$key] = $item;
} else {
$arrayRewrite[] = $item;
}
}
}
return $arrayRewrite;
}
$uniqueArray = arrayUnique($array);
var_dump($uniqueArray);
FROM: http://www.phpdevblog.net/2009/01/using-array-unique-with-multidimensional-arrays.html
Removed comments to give people incentive to visit site - I've used this on a few occasions.
Hope that helps!
EDIT: although not a solution to this particular problem in that you require matching the first 3 indexes, it is still nontheless a very good solution to the general question: how do I use array_unique() on a multidimensional array.
If somebody could pop along and edit for your purposes, all the better!

Answer 2

Zenph gets it 90% right, but he wanted to only look at the first 3 elements as unique. You can use the function below in conjunction with Zenph's code right before the serialize to only look at the first three elements.
function firstThree($array)
{
$retArray = array();
array_push($retArray, $array[1], $array[2], $array[3]);
return $retArray;
}

Answer 3

This will do the trick. Not the most elegant example due to the free function with a static variable, but you can make it more elegant by using a lambda or an instance of a class for the callback target.
function filter($subArray) {
static $seenKeys = null;
if ($seenKeys === null) {
$seekKeys = array();
}
// I 'm just selecting the "three first" indexes here,
// you can change it to better suit your needs
$thisKey = serialize(array_slice($subArray, 0, 3));
if (isset($seenKeys[$thisKey])) {
return false;
}
else {
return $seenKeys[$thisKey] = true;
}
}
$result = array_filter($inputArray, 'filter');
I think this example is as fast as you can go in PHP without making assumptions about the type and/or values of the first three items in each subarray. If any such assumption can be made, then at the very least the serialize call can be replaced with something more appropriate. I imagine that this would speed up the process quite a bit.