Output array of XML to plain XML

Question

I have a script which collects 4 URL's (XML) using CURL and returns an array with 4 items each ocntaining the results of the URL's.

Here is the array:

array(3) {
  [0]=>
  string(41772) "<?xml version="1.0" encoding="UTF-8"?>
<statuses type="array">
<status>
  <created_at>Tue Mar 30 20:58:53 +0000 2010</created_at>
  <id>11328253513</id>
  <text>...</text>
  <source...</source>
  <truncated>false</truncated>
  <in_reply_to_status_id></in_reply_to_status_id>
  <in_reply_to_user_id></in_reply_to_user_id>
  <favorited>true</favorited>
  <in_reply_to_screen_name></in_reply_to_screen_name>
  <user>
    <id>1...</id>
    <name>....</name>
</status>
</statuses>
"
  [1]=>
  string(20630) "<?xml version="1.0" encoding="UTF-8"?>
<statuses type="array">
<status>
  <created_at>Sun Feb 28 14:12:30 +0000 2010</created_at>
  <id>...</id>
  <text>...</text>
  <source>&lt;a

etc...

How can I easily output the XML from the array? I also need to combine the 3 XML structures into one where <statuses> begins and ends in between each array.

Answer 1

A possible idea would be to :

• Create a new $destination instance of DOMDocument
  • Initialize it with a <statuses> node
• For each of your 3 XML strings :
  • load it to another instance of DOMDocument : $currentDocument
  • find the <status> node, with $currentDocument->getElementsByTagName or an equivalent
  • import the <status> node you've just found to the $destination document, with $destination->importNode
• When the loop over each XML string is finished, the $destination document should contain what you wanted, and you can save it, using $destination->saveXML

And here's a quick example of code that could help you understand what I meant :

First, here's the array of XML strings -- I've made them much shorter, but the idea is the same that what you have :

$strings = array(
  '<?xml version="1.0" encoding="UTF-8"?>
    <statuses type="array"><status>
      <id>ID 1</id>
    </status></statuses>',
  '<?xml version="1.0" encoding="UTF-8"?>
    <statuses type="array"><status>
      <id>ID 2</id>
    </status></statuses>',
  '<?xml version="1.0" encoding="UTF-8"?>
    <statuses type="array"><status>
      <id>ID 3</id>
    </status></statuses>',
);

Let's create the destination document, and put a <statuses> tag in it :

$destination = new DOMDocument();
$destination->formatOutput = true;
$destinationStatuses = $destination->createElement('statuses');
$destination->appendChild($destinationStatuses);

Now, we loop over the 3 XML strings :

foreach ($strings as $str) {
  $current = new DOMDocument();
  $current->loadXML($str);
  $currentStatuses = $current->getElementsByTagName('status');
  foreach ($currentStatuses as $currentStatus) {
    $destinationStatus = $destination->importNode($currentStatus, true);
    $destinationStatuses->appendChild($destinationStatus);
  }
}

For each string, we :

• Load it to a new DOMDocument
• Find the <status> tag(s)
• For each <status> tag, import it to the destination document
• And add it to its <statuses> tag

And, finally, if we output the content of the new document :

echo '<pre>' . htmlspecialchars($destination->saveXML()) . '</pre>';

We get :

<?xml version="1.0"?>
<statuses>
  <status>
      <id>ID 1</id>
    </status>
  <status>
      <id>ID 2</id>
    </status>
  <status>
      <id>ID 3</id>
    </status>
</statuses>

i.e. our three <status> from the three original strings have been merged into a single XML Document ;-)

Answer 2

For a simple merger like this you can also just do:

$xml = implode('', $theArray); 
$xml = str_replace(array('<?xml version="1.0" encoding="UTF-8"?>',
                         '<statuses type="array">',
                         '</statuses>'), '', $xml); 

$xml = '<?xml version="1.0" encoding="UTF-8"?>'
     . '<statuses type="array">'
     . $xml
     . '</statuses>';

Note: untested, but basically all it does is glue all XML documents into one, then remove the XML Prologs and all root nodes, so only the status nodes remain. These are then wrapped into a valid XML skeleton, e.g. prolog and root node again. Done.

If you want to work with the DOM Nodes afterwards, using DOM would be more reliable, since DOM knows what nodes are, while string functions have no clue about them. If you decide to work with DOM, consider loading the documents with load instead of CURL, like shown in Pascal's example - or use the above first and then load that document from the string with loadXml().

Whatever you decide to use though, don't use Regex. That is the road to madness.