Reputation: 8621

Sorting multiple subarrays based on a total count from all sub arrays

I have some data which looks like this (reduced)

Here is a JSON string of the data (not reduced, numbers may differ slightly)

{"datasets":[{"label":"NEW","backgroundColor":"#37fdfd","data":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]},{"label":"Grade A","backgroundColor":"#76ef76","data":[9,14,12,0,4,17,13,0,10,0,18,18,12,13,13,4]},{"label":"Grade B","backgroundColor":"#f9f96d","data":[1,6,5,0,6,5,2,0,1,0,2,1,4,3,1,15]},{"label":"Grade C","backgroundColor":"#f3ca36","data":[3,0,1,0,2,0,0,0,0,0,1,1,0,0,0,0]},{"label":"Grade C","backgroundColor":"#f3ca36","data":[3,0,1,0,2,0,0,0,0,0,1,1,0,0,0,0]}],"labels":["User 0","User 1","User 2","User 3","User 4","User 5","User 6","User 7","User 8","User 9","User 10","User 11","User 12","User 13","User 14","User 15"]}

Each dataset has an array of data which has keys that directly relates to a key in the labels array. This is currently sorted in alphabetical order by the label.

This data structure is the structure required for Chart.js, which I am using to display a stacked bar chart on my webpage.

Essentially what I need to accomplish is to sort the data array for every user in the labels array based on the sum of each data set for that user. I also need to sort the labels array to be in the same order.

My original idea on how to achieve this is to create a temporary array, loop through all the data sets and add them to this temporary array in the order necessary, but I got stuck after calculating the total for each user. Here is my attempt:

$return = [];

foreach($calculated['labels'] as $key => &$name) {

    $total = 0;

    foreach($calculated['datasets'] as $dataset) {

        $total += $dataset['data'][$key];

    }

    echo "$name - $total<br>";

}

How can I sort my data and labels in descending order based on the total for each user from all datasets.

Here is my expected output for the reduced data above

The key in the labels array acts as a unique identifier for each user in each dataset data array.

Notice how each set of data inside of each dataset is in the same order, as is the labels array. Each set should be ordered by the total amount from all sets for each user, not necessarily the highest number in each dataset.

For clarification, each set of data in each dataset contains a list of values, the key for each value is directly related to the key for each user in the labels array. So in my example, we have User 0 who has the key "0". This user has a total of 23 from adding up the values from each dataset with the key "0".

Upvotes: 1

Answers (3)

mickmackusa

Reputation: 47904

I see this task as a perfect candidate for array_multisort(). Your synchronously sorted subarrays don't need to retain their initial keys like in u_mulder's output.

The first parameter must be the array of columnar sums, then the descending sort flag as the second parameter, then the labels subarray as a reference, then the dynamic number of data subarrays as references to the original array.

Code: (Demo)

$params = [[], SORT_DESC, &$array['labels']];
foreach ($array['datasets'] as ['data' => &$data]) {
    foreach ($data as $i => $d) {
        $params[0][$i] = ($params[0][$i] ?? 0) + $d;
    }
    $params[] = &$data;
}
array_multisort(...$params);
var_export($array);

Upvotes: 1

Mike 'Pomax' Kamermans

Reputation: 53598

This looks like a typical job for map(reduce).sort: map each element to an object with id, so you can preserve "which user this used to be" information, and total, the result of reducing data. Then sort with a custom sort function (a,b) => a.total - b.total.

E.g.

function map_total($user, $pos) {
  return array(
    "id" => $pos,
    "total" => array_sum($user.data)
  );
}

function cmp_total($a, $b) {
  return $a["total"] - $b["total"];
}

$mapped = array_map("map_total", $thing.dataset, array_keys($thing.dataset));
$sorted = usort($mapped, "cmp_total");

Upvotes: -1

u_mulder

Reputation: 54841

Complete solution:

// get array
$a = json_decode('{"datasets":[{"label":"NEW","backgroundColor":"#37fdfd","data":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]},{"label":"Grade A","backgroundColor":"#76ef76","data":[9,14,12,0,4,17,13,0,10,0,18,18,12,13,13,4]},{"label":"Grade B","backgroundColor":"#f9f96d","data":[1,6,5,0,6,5,2,0,1,0,2,1,4,3,1,15]},{"label":"Grade C","backgroundColor":"#f3ca36","data":[3,0,1,0,2,0,0,0,0,0,1,1,0,0,0,0]},{"label":"Grade C","backgroundColor":"#f3ca36","data":[3,0,1,0,2,0,0,0,0,0,1,1,0,0,0,0]}],"labels":["User 0","User 1","User 2","User 3","User 4","User 5","User 6","User 7","User 8","User 9","User 10","User 11","User 12","User 13","User 14","User 15"]}', true);

// get array of arrays with `data` key from each data set
$users = array_column($a['datasets'], 'data');

// tricky code to sum arrays
$sums = array_map('array_sum', array_map(null, ...$users));
// sort array with keeping keys
arsort($sums);

// we need flip so as `array_replace` will work as expected
$keys = array_flip(array_keys($sums));

// "sorting" `data` subarrays
foreach ($a['datasets'] as &$item) {
    $item['data'] = array_replace($keys, $item['data']);
}
// "sorting" `labels` subarray
$a['labels'] = array_replace($keys, $a['labels']);
// see the result
print_r($a);

Fiddle here https://3v4l.org/a7rPL

Upvotes: 1

Sorting multiple subarrays based on a total count from all sub arrays

Answers (3)

Related Questions