PHP simplexml_load_string() "eats" some tags

Question

After hours of debugging my code i find out that simplexml_load_string() seem to silently remove all sub-entities within nodes values. To demonstrate the problem i made a little PHP script:

#!/usr/bin/php
<?php
$xml = <<<XML
<?xml version="1.0" encoding="UTF-8"?>
<channel>
  <element name="headline">
    <p>Line 1<br class="HardReturn"/>Line 2</p>
    <p>This is <b>bold</b> text</p>
  </element>
</channel>
XML;

libxml_use_internal_errors(true);
$xml = simplexml_load_string($xml, "SimpleXMLElement");
$errors = libxml_get_errors();
print_r($errors);
var_dump($xml);

When run this in a shell i get this output:

Array
(
)
object(SimpleXMLElement)#1 (1) {
  ["element"]=>
  object(SimpleXMLElement)#2 (2) {
    ["@attributes"]=>
    array(1) {
      ["name"]=>
      string(8) "headline"
    }
    ["p"]=>
    array(2) {
      [0]=>
      string(12) "Line 1Line 2"
      [1]=>
      string(13) "This is  text"
    }
  }
}

Any idea what i do wrong?

Jack Fleeting · Accepted Answer · 2021-12-06 00:20:11Z

1

Not sure why var_dump skips the <bold> tags, but he <bold> tags are still there. To prove it, try

echo $xml->asXML() ."\r\n";

The output should be exactly the same as your input. Also try

$target = $xml->xpath("//b");
echo $target[0];

And the output is

bold

answered Dec 6, 2021 at 0:20

Jack Fleeting

25k6 gold badges27 silver badges49 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

HeckMeck Over a year ago

I don't get why the bolds are missing when i coerce it to string? I expected the node to have child-nodes of type type bold.

Jack Fleeting Over a year ago

What do you mean by "when i coerce it to string"?

IMSoP · Accepted Answer · 2021-12-09 16:47:04Z

SimpleXML hasn't "eaten" the nodes, it just isn't very good at showing them in debug outputs like print_r. The SimpleXML API is optimised for applications of XML as a tree of data, rather than text with markup like you have here.

In a more complex API such as the DOM, your first p element would be shown as having three child nodes:

A text node containing "Line 1"
A br element with an attribute class, value HardReturn
A text node containing "Line 2"

However, SimpleXML doesn't have a separate object for "text nodes", instead it has an accessor for the combined text content of an element, when you cast to string. In this case, that combines the two text nodes, giving "Line 1Line 2". You can still access the br element, you just can't see where it fits relative to the text.

echo "String content: ", (string)$xml->element->p[0], "\n";
echo "Child element attribute: ", (string)$xml->element->p[0]->br['class'], "\n";

String content: Line 1Line 2
Child element attribute: HardReturn

With the second p element, there's a whole tree of nodes to represent:

A text node contained "This is "
A child element b
- A text node inside the b element with the text "bold"
A text node (in the p element) with the text " text"

Since the string representation only captures directly contained text nodes, the combined text content of the p element is just "This is " + " text".

echo "String content: ", (string)$xml->element->p[1], "\n";
echo "Child element string content: ", (string)$xml->element->p[1]->b, "\n";

String content: This is text
Child element string content: bold

In both cases, we can see that all the nodes are still there, and get out the original XML:

echo "First paragraph: ", $xml->element->p[0]->asXML(), "\n";
echo "Second paragraph: ", $xml->element->p[1]->asXML(), "\n";

First paragraph: Line 1 Line 2
Second paragraph: This is bold text

To work directly with the text nodes, you need to use a different API, such as DOM; or, since this example appears to actually be HTML, look for an HTML-oriented library.

Collectives™ on Stack Overflow

PHP simplexml_load_string() "eats" some tags

2 Answers 2

2 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related