Grouping array based on the key value similarity - php

Question

Let's say I have an array like that:
$data[0]['name'] = 'product 1 brandX';
$data[0]['id_product'] = '77777777';
$data[1]['name'] = 'brandX product 1';
$data[1]['id_product'] = '77777777';
$data[2]['name'] = 'brandX product 1 RED';
$data[2]['id_product'] = '77777777';
$data[3]['name'] = 'product 1 brandX';
$data[3]['id_product'] = '';
$data[4]['name'] = 'product 2 brandY';
$data[4]['id_product'] = '8888888';
$data[5]['name'] = 'product 2 brandY RED';
$data[5]['id_product'] = '';
I am trying to group them by their similarities (name or id_product).
That would be the expected final array:
$uniques[0]['name'] = 'product 1 brandX'; //The smallest name for the product
$uniques[0]['count'] = 4; //Entry which has all the words of the smallest name or the same id_product
$uniques[0]['name'] = 'product 2 brandY';
$uniques[0]['count'] = 2;
That's what I tried so far:
foreach ($data as $t) {
if (!isset($uniques[$t['id_product']]['name']) || mb_strlen($uniques[$t['id_product']]['name']) > mb_strlen($t['name'])) {
$uniques[$t['id_product']]['name'] = $t['name'];
$uniques[$t['id_product']]['count']++;
}
}
But I can't be based on the id_product because sometimes it will be the same product but one will have the id and the other one no. I have to check the name also, but couldn't manage to get it done.

Answer 1

I'm basing my answer on two assumptions about how products should be grouped:
Although id_product can be missing, where it is present, it is
correct and sufficient to match two products; and
For two product names to match, the longest name (name with the most
words) must contain all of the words in the shortest name (name with
the fewest words).
Given these assumptions, here is a function to determine if two individual products match (products should be grouped together) and a helper function to get words from names:
function productsMatch(array $product1, array $product2)
{
if (
!empty($product1['id_product'])
&& !empty($product2['id_product'])
&& $product1['id_product'] === $product2['id_product']
) {
// match based on id_product
return true;
}
$words1 = getWordsFromProduct($product1);
$words2 = getWordsFromProduct($product2);
$min_word_count = min(count($words1), count($words2));
$match_word_count = count(array_intersect_key($words1, $words2));
if ($min_word_count >= 1 && $match_word_count === $min_word_count) {
// match based on name similarity
return true;
}
// no match
return false;
}
function getWordsFromProduct(array $product)
{
$name = mb_strtolower($product['name']);
preg_match_all('/\S+/', $name, $matches);
$words = array_flip($matches[0]);
return $words;
}
This function can be used to group the products:
function groupProducts(array $data)
{
$groups = array();
foreach ($data as $product1) {
foreach ($groups as $key => $products) {
foreach ($products as $product2) {
if (productsMatch($product1, $product2)) {
$groups[$key][] = $product1;
continue 3; // foreach ($data as $product1)
}
}
}
$groups[] = array($product1);
}
return $groups;
}
And then this function can be used to extract the shortest name and count for each group:
function uniqueProducts(array $groups)
{
$uniques = array();
foreach ($groups as $products) {
$shortest_name = '';
$shortest_length = PHP_INT_MAX;
$count = 0;
foreach ($products as $product) {
$length = mb_strlen($product['name']);
if ($length < $shortest_length) {
$shortest_name = $product['name'];
$shortest_length = $length;
}
$count++;
}
$uniques[] = array(
'name' => $shortest_name,
'count' => $count,
);
}
return $uniques;
}
So combining all 4 functions, you can get the uniques as follows (tested with php 5.6):
$data[0]['name'] = 'product 1 brandX';
$data[0]['id_product'] = '77777777';
$data[1]['name'] = 'brandX product 1';
$data[1]['id_product'] = '77777777';
$data[2]['name'] = 'brandX product 1 RED';
$data[2]['id_product'] = '77777777';
$data[3]['name'] = 'product 1 brandX';
$data[3]['id_product'] = '';
$data[4]['name'] = 'product 2 brandY';
$data[4]['id_product'] = '8888888';
$data[5]['name'] = 'product 2 brandY RED';
$data[5]['id_product'] = '';
$groups = groupProducts($data);
$uniques = uniqueProducts($groups);
var_dump($uniques);
Which gives output:
array(2) {
[0]=>
array(2) {
["name"]=>
string(16) "product 1 brandX"
["count"]=>
int(4)
}
[1]=>
array(2) {
["name"]=>
string(16) "product 2 brandY"
["count"]=>
int(2)
}
}

Answer 2

I don't think this will solve your problem, but it may get you moving forward again
$data = [];
$data[0]['name'] = 'product 1 brandX';
$data[0]['id_product'] = '77777777';
$data[1]['name'] = 'brandX product 1';
$data[1]['id_product'] = '77777777';
$data[2]['name'] = 'brandX product 1 RED';
$data[2]['id_product'] = '77777777';
$data[3]['name'] = 'product 1 brandX';
$data[3]['id_product'] = '';
$data[4]['name'] = 'product 2 brandY';
$data[4]['id_product'] = '8888888';
$data[5]['name'] = 'product 2 brandY RED';
$data[5]['id_product'] = '';
$data = collect($data);
$tallies = [
'brand_x' => 0,
'brand_y' => 0,
'other' => 0
];
$unique = $data->unique(function ($item) use (&$tallies){
switch(true){
case(strpos($item['name'], 'brandX') !== false):
$tallies['brand_x']++;
return 'product X';
break;
case(strpos($item['name'], 'brandY') !== false):
$tallies['brand_y']++;
return 'product Y';
break;
default:
$tallies['other']++;
return 'other';
break;
}
});
print_r($unique);
print_r($tallies);

Answer 3

I think the best way to solve this is to use a unique product_id, but if you want create unique keys by finding similarities in the name field you can use preg_split to transform names to arrays and then use array_diff to find an array of difference. 2 names are considered unique if their difference count is lower than 2. I create this function it returns similar names in $arr or false if not found:
function get_similare_key($arr, $name) {
$names = preg_split("/\s+/", $name);
// get similaire key from $arr
foreach( $arr as $key => $value ) {
$key_names = preg_split("/\s+/", $key);
$diff = array_diff($key_names, $names);
if ( count($diff) <= 1 ) {
return $key;
}
}
return false;
}
Here is a working demo here