Question

I have a table similar to:

ID  |  NAME    |  GROUP_ID
1   | rand_str |  4
2   | rand_str |  2
3   | rand_str |  3
4   | rand_str |  0
5   | rand_str |  3
6   | rand_str |  1
7   | rand_str |  3
8   | rand_str |  3
9   | rand_str |  1
10  | rand_str |  0

ID is unique ID, name valie is not null and contains varchar value and group_id is a max 2 digit positive number.

Let’s say that the query returns this table.

What I want is, using PHP, show the results grouped by their group_id and if possible, ordering the group id by the most numbered to least numbered and “0” being always the last no matter how populated it is.

so end result:

group id 3 exists 3 times, so

1. 1- id_3, id_5, id_7 and id_8 (group_id 3)
2. 2- id_6, id_9 (group_id 1)
3. 3- either group id 4 or 2 since both contain 1 item each
4. 4- group_id 0, even though it has 2 items -> id_4, id_10

How can I achieve this?

Thank you.

Answer 1

Use a subquery to determine the count of each group (and therefore its order), then join that with your data to retrieve the records:

SELECT my_table.* FROM my_table JOIN (
  SELECT   GROUP_ID,
           IF(GROUP_ID = 0, -1, COUNT(*)) AS rank
  FROM     my_table
  GROUP BY GROUP_ID
) AS t USING (GROUP_ID)
ORDER BY t.rank DESC

See it on sqlfiddle.

You would then loop over the results in PHP, keeping track of the GROUP_ID of the last record and comparing with the current one to see whether you’re now in a new group:

$last = null;
echo "<table>";
while ($row = $stmt->fetch()) {
  if ($row['GROUP_ID'] !== $last) {
    if ($last !== null) echo "</tr>";
    echo "<tr><th scope='row'>Group " . html_entities($row['GROUP_ID']) . "</th>";
    $last = $row['GROUP_ID'];
  }
  echo "<td>ID " . html_entities($row['ID']) . "</td>";
}
echo "</tr></table>";