Remove a child with a specific attribute, in SimpleXML for PHP

I have several identical elements with different attributes that I'm accessing with SimpleXML:

<data>
    <seg id="A1"/>
    <seg id="A5"/>
    <seg id="A12"/>
    <seg id="A29"/>
    <seg id="A30"/>
</data>

I need to remove a specific seg element, with an id of "A12", how can I do this? I've tried looping through the seg elements and unsetting the specific one, but this doesn't work, the elements remain.

foreach($doc->seg as $seg)
{
    if($seg['id'] == 'A12')
    {
        unset($seg);
    }
}

Answers


While SimpleXML provides a way to remove XML nodes, its modification capabilities are somewhat limited. One other solution is to resort to using the DOM extension. dom_import_simplexml() will help you with converting your SimpleXMLElement into a DOMElement.

Just some example code (tested with PHP 5.2.5):

$data='<data>
    <seg id="A1"/>
    <seg id="A5"/>
    <seg id="A12"/>
    <seg id="A29"/>
    <seg id="A30"/>
</data>';
$doc=new SimpleXMLElement($data);
foreach($doc->seg as $seg)
{
    if($seg['id'] == 'A12') {
        $dom=dom_import_simplexml($seg);
        $dom->parentNode->removeChild($dom);
    }
}
echo $doc->asXml();

outputs

<?xml version="1.0"?>
<data><seg id="A1"/><seg id="A5"/><seg id="A29"/><seg id="A30"/></data>

By the way: selecting specific nodes is much more simple when you use XPath (SimpleXMLElement->xpath):

$segs=$doc->xpath('//seq[@id="A12"]');
if (count($segs)>=1) {
    $seg=$segs[0];
}
// same deletion procedure as above

Contrary to popular belief in the existing answers, each Simplexml element node can be removed from the document just by itself and unset(). The point in case is just that you need to understand how SimpleXML actually works.

First locate the element you want to remove:

list($element) = $doc->xpath('/*/seg[@id="A12"]');

Then remove the element represented in $element you unset its self-reference:

unset($element[0]);

This works because the first element of any element is the element itself in Simplexml (self-reference). This has to do with its magic nature, numeric indices are representing the elements in any list (e.g. parent->children), and even the single child is such a list.

Non-numeric string indices represent attributes (in array-access) or child-element(s) (in property-access).

Therefore numeric indecies in property-access like:

unset($element->{0});

work as well.

Naturally with that xpath example, it is rather straight forward (in PHP 5.4):

unset($doc->xpath('/*/seg[@id="A12"]')[0][0]);

The full example code (Demo):

<?php
/**
 * Remove a child with a specific attribute, in SimpleXML for PHP
 * @link http://stackoverflow.com/a/16062633/367456
 */

$data=<<<DATA
<data>
    <seg id="A1"/>
    <seg id="A5"/>
    <seg id="A12"/>
    <seg id="A29"/>
    <seg id="A30"/>
</data>
DATA;


$doc = new SimpleXMLElement($data);

unset($doc->xpath('seg[@id="A12"]')[0]->{0});

$doc->asXml('php://output');

Output:

<?xml version="1.0"?>
<data>
    <seg id="A1"/>
    <seg id="A5"/>

    <seg id="A29"/>
    <seg id="A30"/>
</data>

Just unset the node:

$str = <<<STR
<a>
  <b>
    <c>
    </c>
  </b>
</a>
STR;

$xml = simplexml_load_string($str);
unset($xml –> a –> b –> c); // this would remove node c
echo $xml –> asXML(); // xml document string without node c

This code was taken from How to delete / remove nodes in SimpleXML.


I believe Stefan's answer is right on. If you want to remove only one node (rather than all matching nodes), here is another example:

//Load XML from file (or it could come from a POST, etc.)
$xml = simplexml_load_file('fileName.xml');

//Use XPath to find target node for removal
$target = $xml->xpath("//seg[@id=$uniqueIdToDelete]");

//If target does not exist (already deleted by someone/thing else), halt
if(!$target)
return; //Returns null

//Import simpleXml reference into Dom & do removal (removal occurs in simpleXML object)
$domRef = dom_import_simplexml($target[0]); //Select position 0 in XPath array
$domRef->parentNode->removeChild($domRef);

//Format XML to save indented tree rather than one line and save
$dom = new DOMDocument('1.0');
$dom->preserveWhiteSpace = false;
$dom->formatOutput = true;
$dom->loadXML($xml->asXML());
$dom->save('fileName.xml');

Note that sections Load XML... (first) and Format XML... (last) could be replaced with different code depending on where your XML data comes from and what you want to do with the output; it is the sections in between that find a node and remove it.

In addition, the if statement is only there to ensure that the target node exists before trying to move it. You could choose different ways to handle or ignore this case.


If you extend the base SimpleXMLElement class, you can use this method:

class MyXML extends SimpleXMLElement {

    public function find($xpath) {
        $tmp = $this->xpath($xpath);
        return isset($tmp[0])? $tmp[0]: null;
    }

    public function remove() {
        $dom = dom_import_simplexml($this);
        return $dom->parentNode->removeChild($dom);
    }

}

// Example: removing the <bar> element with id = 1
$foo = new MyXML('<foo><bar id="1"/><bar id="2"/></foo>');
$foo->find('//bar[@id="1"]')->remove();
print $foo->asXML(); // <foo><bar id="2"/></foo>

This work for me:

$data = '<data>
<seg id="A1"/>
<seg id="A5"/>
<seg id="A12"/>
<seg id="A29"/>
<seg id="A30"/></data>';

$doc = new SimpleXMLElement($data);

$segarr = $doc->seg;

$count = count($segarr);

$j = 0;

for ($i = 0; $i < $count; $i++) {

    if ($segarr[$j]['id'] == 'A12') {
        unset($segarr[$j]);
        $j = $j - 1;
    }
    $j = $j + 1;
}

echo $doc->asXml();

For future reference, deleting nodes with SimpleXML can be a pain sometimes, especially if you don't know the exact structure of the document. That's why I have written SimpleDOM, a class that extends SimpleXMLElement to add a few convenience methods.

For instance, deleteNodes() will delete all nodes matching a XPath expression. And if you want to delete all nodes with the attribute "id" equal to "A5", all you have to do is:

// don't forget to include SimpleDOM.php
include 'SimpleDOM.php';

// use simpledom_load_string() instead of simplexml_load_string()
$data = simpledom_load_string(
    '<data>
        <seg id="A1"/>
        <seg id="A5"/>
        <seg id="A12"/>
        <seg id="A29"/>
        <seg id="A30"/>
    </data>'
);

// and there the magic happens
$data->deleteNodes('//seg[@id="A5"]');

To remove/keep nodes with certain attribute value or falling into array of attribute values you can extend SimpleXMLElement class like this (most recent version in my GitHub Gist):

class SimpleXMLElementExtended extends SimpleXMLElement
{    
    /**
    * Removes or keeps nodes with given attributes
    *
    * @param string $attributeName
    * @param array $attributeValues
    * @param bool $keep TRUE keeps nodes and removes the rest, FALSE removes nodes and keeps the rest 
    * @return integer Number o affected nodes
    *
    * @example: $xml->o->filterAttribute('id', $products_ids); // Keeps only nodes with id attr in $products_ids
    * @see: http://stackoverflow.com/questions/17185959/simplexml-remove-nodes
    */
    public function filterAttribute($attributeName = '', $attributeValues = array(), $keepNodes = TRUE)
    {       
        $nodesToRemove = array();

        foreach($this as $node)
        {
            $attributeValue = (string)$node[$attributeName];

            if ($keepNodes)
            {
                if (!in_array($attributeValue, $attributeValues)) $nodesToRemove[] = $node;
            }
            else
            { 
                if (in_array($attributeValue, $attributeValues)) $nodesToRemove[] = $node;
            }
        }

        $result = count($nodesToRemove);

        foreach ($nodesToRemove as $node) {
            unset($node[0]);
        }

        return $result;
    }
}

Then having your $doc XML you can remove your <seg id="A12"/> node calling:

$data='<data>
    <seg id="A1"/>
    <seg id="A5"/>
    <seg id="A12"/>
    <seg id="A29"/>
    <seg id="A30"/>
</data>';

$doc=new SimpleXMLElementExtended($data);
$doc->seg->filterAttribute('id', ['A12'], FALSE);

or remove multiple <seg /> nodes:

$doc->seg->filterAttribute('id', ['A1', 'A12', 'A29'], FALSE);

For keeping only <seg id="A5"/> and <seg id="A30"/> nodes and removing the rest:

$doc->seg->filterAttribute('id', ['A5', 'A30'], TRUE);

There is a way to remove a child element via SimpleXml. The code looks for a element, and does nothing. Otherwise it adds the element to a string. It then writes out the string to a file. Also note that the code saves a backup before overwriting the original file.

$username = $_GET['delete_account'];
echo "DELETING: ".$username;
$xml = simplexml_load_file("users.xml");

$str = "<?xml version=\"1.0\"?>
<users>";
foreach($xml->children() as $child){
  if($child->getName() == "user") {
      if($username == $child['name']) {
        continue;
    } else {
        $str = $str.$child->asXML();
    }
  }
}
$str = $str."
</users>";
echo $str;

$xml->asXML("users_backup.xml");
$myFile = "users.xml";
$fh = fopen($myFile, 'w') or die("can't open file");
fwrite($fh, $str);
fclose($fh);

A new idea: simple_xml works as a array.

We can search for the indexes of the "array" we want to delete, and then, use the unset() function to delete this array indexes. My example:

$pos=$this->xml->getXMLUser();
$i=0; $array_pos=array();
foreach($this->xml->doc->users->usr[$pos]->u_cfg_root->profiles->profile as $profile) {
    if($profile->p_timestamp=='0') { $array_pos[]=$i; }
    $i++;
}
//print_r($array_pos);
for($i=0;$i<count($array_pos);$i++) {
    unset($this->xml->doc->users->usr[$pos]->u_cfg_root->profiles->profile[$array_pos[$i]]);
}

Even though SimpleXML doesn't have a detailed way to remove elements, you can remove elements from SimpleXML by using PHP's unset(). The key to doing this is managing to target the desired element. At least one way to do the targeting is using the order of the elements. First find out the order number of the element you want to remove (for example with a loop), then remove the element:

$target = false;
$i = 0;
foreach ($xml->seg as $s) {
  if ($s['id']=='A12') { $target = $i; break; }
  $i++;
}
if ($target !== false) {
  unset($xml->seg[$target]);
}

You can even remove multiple elements with this, by storing the order number of target items in an array. Just remember to do the removal in a reverse order (array_reverse($targets)), because removing an item naturally reduces the order number of the items that come after it.

Admittedly, it's a bit of a hackaround, but it seems to work fine.


I was also strugling with this issue and the answer is way easier than those provided over here. you can just look for it using xpath and unset it it the following method:

unset($XML->xpath("NODESNAME[@id='test']")[0]->{0});

this code will look for a node named "NODESNAME" with the id attribute "test" and remove the first occurence.

remember to save the xml using $XML->saveXML(...);


Since I encountered the same fatal error as Gerry and I'm not familiar with DOM, I decided to do it like this:

$item = $xml->xpath("//seg[@id='A12']");
$page = $xml->xpath("/data");
$id = "A12";

if (  count($item)  &&  count($page) ) {
    $item = $item[0];
    $page = $page[0];

     // find the numerical index within ->children().
    $ch = $page->children();
    $ch_as_array = (array) $ch;

    if (  count($ch_as_array)  &&  isset($ch_as_array['seg'])  ) {
        $ch_as_array = $ch_as_array['seg'];
        $index_in_array = array_search($item, $ch_as_array);
        if (  ($index_in_array !== false)
          &&  ($index_in_array !== null)
          &&  isset($ch[$index_in_array])
          &&  ($ch[$index_in_array]['id'] == $id)  ) {

             // delete it!
            unset($ch[$index_in_array]);

            echo "<pre>"; var_dump($xml); echo "</pre>";
        }
    }  // end of ( if xml object successfully converted to array )
}  // end of ( valid item  AND  section )

Idea about helper functions is from one of the comments for DOM on php.net and idea about using unset is from kavoir.com. For me this solution finally worked:

function Myunset($node)
{
 unsetChildren($node);
 $parent = $node->parentNode;
 unset($node);
}

function unsetChildren($node)
{
 while (isset($node->firstChild))
 {
 unsetChildren($node->firstChild);
 unset($node->firstChild);
 }
}

using it: $xml is SimpleXmlElement

Myunset($xml->channel->item[$i]);

The result is stored in $xml, so don’t worry about assigning it to any variable.


With FluidXML you can use XPath to select the elements to remove.

$doc = fluidify($doc);

$doc->remove('//*[@id="A12"]');

https://github.com/servo-php/fluidxml


The XPath //*[@id="A12"] means:

  • in any point of the document (//)
  • every node (*)
  • with the attribute id equal to A12 ([@id="A12"]).

If you want to cut list of similar (not unique) child elements, for example items of RSS feed, you could use this code:

for ( $i = 9999; $i > 10; $i--) {
    unset($xml->xpath('/rss/channel/item['. $i .']')[0]->{0});
}

It will cut tail of RSS to 10 elements. I tried to remove with

for ( $i = 10; $i < 9999; $i ++ ) {
    unset($xml->xpath('/rss/channel/item[' . $i . ']')[0]->{0});
}

But it works somehow randomly and cuts only some of the elements.


Your initial approach was right, but you forgot one little thing about foreach. It doesn't work on the original array/object, but creates a copy of each element as it iterates, so you did unset the copy. Use reference like this:

foreach($doc->seg as &$seg) 
{
    if($seg['id'] == 'A12')
    {
        unset($seg);
    }
}

Need Your Help

Skipping Incompatible Libraries at compile

c++ 64bit compilation 32-bit conflicting-libraries

When I try to compile a copy of my project on my local machine, I get an error stating that it 's skipping over incompatible libraries. This isn't the case when I'm messing around with the live ve...