Question

Specifically, I want to convert the XML structure into this format for the database so that I can use modified preorder tree traversal:

_{(source: sitepointstatic.com)}

My XML structure looks something like this:

<?xml version="1.0" standalone="yes"?>
<products>
    <node>
      <name>Top Membership</name>
      <node>
        <name>Middle Membership</name>
        <node>
          <name>Bottom Membership</name>
          <node>
            <name>Some content</name>
            <node>
              <name>Specific content</name>
            </node>
          </node>
        </node>
      </node>
    </node>

I’ve made a PHP script to traverse the XML structure and dump me out the names, I’m just having trouble working out the left and right values. I think I probably have to change my traversal logic so that it’s followed down to a single leaf each time, then goes back up and starts again, rather than bottoming each entire branch each time. If I could do this then I think it would be easier to calculate left and right.

My PHP script so far:

<?php

$node = new SimpleXMLElement(file_get_contents('products.xml'));

echo '<pre>';

function getRowData($node, $depth)
{  
  //Not a leaf
  if(isset($node->node))
  {
    echo $node->name."\t\t\t($depth) parent, children: ".count($node->children())."\n";

    foreach($node->node as $n)
      getRowData($n, $depth + 1);
  }
  else //It's a leaf
    echo $node->name."\t\t\t($depth) leaf\n";


}

getRowData($node->node, 1);

echo '</pre>';

?>

I’m using this SitePoint article for reference http://www.sitepoint.com/hierarchical-data-database-2/

Edit:

Revised PHP script which is closer to a solution (and outputs in a better format with more (possibly relevant) numbers):

<?php

$node = new SimpleXMLElement(file_get_contents('products.xml'));

echo '<table border="1">';
  echo '
    <tr>
      <th>Name</th>
      <th>Depth</th>
      <th>Child num</th>
      <th>Children</th>
      <th>Ancestors</th>
      <th>Total siblings</th>
      <th>"Left"</th>
      <th>Parent "Left"</th>
    </tr>';

function getRowData($node, $depth = 1, $child_num = 1, $prior_nodes = 0, $sibling_total = 0, $parent_left = 0)
{  

    echo '<tr>';
  echo '
          <td>'.$node->name."</td>
          <td>$depth</td>
          <td>$child_num</td>
          <td>".(count($node->children()) - 1)."</td>
          <td>$prior_nodes</td>
          <td>$sibling_total</td>";

  $left = $parent_left + ($child_num == 1 ? 1 : ($child_num * 2) - 1);


          echo "<td>$left</td>";
          echo "<td>$parent_left</td>";
  echo '</tr>';
  $child_num = 1;

  foreach($node->node as $n)
  {
    getRowData(
            $n, 
            $depth + 1, 
            $child_num++, 
            $prior_nodes + (count($node->children()) - 1), 
            (count($node->children()) - 1),
            $left
    );
  }
}

getRowData($node->node);

echo '</table>';

?>

Answer 1

The general approach here is to create an adjacency list (a tree), then
traverse the tree assigning lft on your way down and rgt on your way up.

Assuming this xml:

<?xml version="1.0" standalone="yes"?>
<products>
<node><name>Top Membership</name>
  <node><name>Middle Membership</name>
    <node><name>Bottom Membership 1</name></node>
    <node><name>Bottom Membership 2</name>
      <node><name>Some content</name>
        <node><name>Specific content</name></node>
      </node>
    </node>
  </node>
</node>
</products>

The following code should do what you need.

$root = simplexml_load_string($xml);

function nested_set($parent, $rows=array(), $counter=1) {
    foreach ($parent->node as $node) {
        $row = array(
            'title'   => (string) $node->name,
            'parent' => (string) $parent->name,
            'lft'    => $counter++,
            'rgt'    => null,
        );
        list($rows, $counter) = nested_set($node, $rows, $counter);
        $row['rgt'] = $counter++;
        $rows[] = $row;
    }
    return array($rows, $counter);
}

function print_rows($rows) {
    $sep = '+'.str_repeat('-', 22).'+'.str_repeat('-', 22).'+'.str_repeat('-', 5).'+'.str_repeat('-', 5).'+'."\n";
    echo $sep;
    echo vsprintf("| %-20s | %-20s | %3s | %3s |\n", array_keys($rows[0]));
    echo $sep;
    foreach ($rows as $row) {
        echo vsprintf("| %-20s | %-20s | %3d | %3d |\n", array_values($row));
    }
    echo $sep;
}

list($rows, $counter) = nested_set($root);
// rows will be in reverse-traversal order
print_rows($rows);

Output will be:

+----------------------+----------------------+-----+-----+
| title                | parent               | lft | rgt |
+----------------------+----------------------+-----+-----+
| Bottom Membership 1  | Middle Membership    |   3 |   4 |
| Specific content     | Some content         |   7 |   8 |
| Some content         | Bottom Membership 2  |   6 |   9 |
| Bottom Membership 2  | Middle Membership    |   5 |  10 |
| Middle Membership    | Top Membership       |   2 |  11 |
| Top Membership       |                      |   1 |  12 |
+----------------------+----------------------+-----+-----+

For more information on nested sets, see Joe Celko’s work:

• Trees in SQL
• The chapter on trees and hierarchies in SQL for Smarties (I have the first edition–it’s very valuable.)
• His book Trees and Hierarchies in SQL for Smarties