PHP group array by date with interval

Question

I have this array

  Array
  (
    [0] => Array
        (
            [date] => 2020-10-07 10:49:48
            [content] => 1
        )
    [1] => Array
        (
            [date] => 2020-10-08 13:49:48
            [content] => 2
        )
    [2] => Array
        (
            [date] => 2020-10-08 13:50:03
            [content] => 3
        )
  )

I want to get:

  Array
  (
    [0] => Array
        (
            [date] => 2020-10-07 10:49:48
            [content] => 1
        )
    [1] => Array
        (
            [date] => 2020-10-08 13:50:03
            [content] => 2, 3
        )
  )

Like you see, if date between elements is less than hour (for example), content of both elements should be grouped with the greater date value. I just don't understand how to do it properly

I got this, but idk why i have extra element in array

        foreach ($result as $key => $value) {
            foreach ($result as $key2 => $value2) {
                if (abs(strtotime($value['date_added']) - strtotime($value2['date_added'])) <= 3600 && $key != $key2) {
                    $result[$key]['content'] = $value['content'] . ',' . $value2['content'];
                    unset($result[$key2]);
                }
            }
            $merged_array[] = $result[$key];
        }

you wanna mean `date] => 2020-10-07 10:49:48 [content] => 1, 2` — Julien Baldy, Oct 08 '20 at 14:59
1 and 2 element have 1 minute difference, so they should be grouped — Bobby Redjeans, Oct 08 '20 at 15:01
ok sorry, but you regroupe [content] => 2, 3, where is 3 ? and where is [date] => 2020-10-08 13:49:48 ? — Julien Baldy, Oct 08 '20 at 15:07
do we need to check sequentially or in any order. Or what happens if 1st record is `7:00`, 2nd is `7:50` and 3rd is `8:30`. Should these be merged? — Dark Knight, Oct 08 '20 at 16:56
It depends on interval, if it's <= 3600 (1h) then only 1 and 2 should combined (7:00 and 7:50), if 2 hours then all 3 should be combined — Bobby Redjeans, Oct 08 '20 at 17:02
Hope this will work https://stackoverflow.com/questions/12706359/how-to-group-subarrays-by-a-column-value — Babitha, Oct 08 '20 at 17:04
you said you get one extra element. Could you show the output? — Adam, Oct 10 '20 at 10:07

score 1 · Answer 1 · answered Oct 10 '20 at 10:00

It is not clearly described which time differences should be used. The solution here always calculates the time differences to the highest date of the group.

The array will be sorted in descending order of the date. The highest date is set as $groupDate. A new $result array is created with a foreach loop. If the date difference to $groupDate is less than $intervalHour, then 'content' is accumulated. In the other case a new $groupDate is set. At the end the result array is sorted again according to increasing date.

$data = [
  ['date' => '2020-10-07 10:49:48', 'content' => 1],
  ['date' => '2020-10-08 13:49:48', 'content' => 2],
  ['date' => '2020-10-08 13:50:03', 'content' => 3],
  ['date' => '2020-10-08 14:50:04', 'content' => 4],
];

$intervalHour = 1.0;

usort($data,function($a,$b){return $b['date'] <=> $a['date'];});

$groupDate = date_create($data[0]['date']);
$content = $data[0]['content'];
$result = [$data[0]];
$k = 0;
foreach($data as $i => $row){
  if($i == 0) continue;
  $diff = date_create($row['date'])->diff($groupDate);
  $diffHour = $diff->days * 24 + $diff->h + $diff->i/60 + $diff->s/3600;  
  if($diffHour < $intervalHour) {
    $content = $row['content'].','.$content;
  }
  else {
    ++$k;
    $groupDate = date_create($row['date']);
    $result[$k]['date'] = $groupDate->format('Y-m-d H:i:s');
    $content = $row['content'];
  }
  $result[$k]['content'] = $content;
}

usort($result,function($a,$b){return $a['date'] <=> $b['date'];});

echo '<pre>',var_export($result);

Output with $intervalHour = 1.0

array (
  0 => 
  array (
    'date' => '2020-10-07 10:49:48',
    'content' => 1,
  ),
  1 => 
  array (
    'date' => '2020-10-08 13:50:03',
    'content' => '2,3',
  ),
  2 => 
  array (
    'date' => '2020-10-08 14:50:04',
    'content' => 4,
  ),
)

Output with $intervalHour = 2.0

array (
  0 => 
  array (
    'date' => '2020-10-07 10:49:48',
    'content' => 1,
  ),
  1 => 
  array (
    'date' => '2020-10-08 14:50:04',
    'content' => '2,3,4',
  ),
)

score 0 · Answer 2 · answered Oct 10 '20 at 10:42

The problem is that your unset command does not work. When you use a foreach on an array, a copy of that array will be created and iterated. If you want to iterate over an instance that you can change on the fly, you would need to pass an iterable object.

As an example, the following would output 12, despite the unset:

$x = [1, 2];

foreach ($x as $key => $value) {
    echo $value;
    unset($x[1]);
}

I think best approach is using OOP. But if you want to work only with forloops and arrays, it could be done like this:

  $usedKeys = [];
  $dateContent  =[];

  foreach ($result as $key => $value) {
      if (in_array($key, $usedKeys)) {
          continue;
      }
      $usedKeys[] = $key;
      $dateContent[$key]['content'] = [$value['content']];
      foreach ($result as $key2 => $value2) {
          if (in_array($key2, $usedKeys)) {
              continue;
          }

          if (abs(strtotime($value['date']) - strtotime($value2['date'])) <= 3600 && $key != $key2) {
              $dateContent[$key]['content'][] = $value2['content'];
              $usedKeys[] = $key2;
          }
      }
  }

  $dateContent = array_map(function ($value) {
      return implode(',', $value['content']);
  }, $dateContent);

PHP group array by date with interval

2 Answers2