Parsed SimpleXMLElement is missing attributes [closed]

Question

I have this small chunk of XML that I wish to parse in PHP. The Works tag is ok, I can parse all the attributes just fine. The problem I'm having is with the Doesnt tag, it seems that because it has text content that I can't access the attributes.

<Export id="123" apples="pears">
    <Works foo="bar" id="234"/>
    <Doesnt bar="foo" id="345">Stack Exchange</Doesnt>
</Export>

I run the following (very simple) code:

$plain = '<Export id="123" apples="pear....esnt></Export>'; // as above
$sxe = simplexml_load_string($plain);
$json = json_encode($sxe);
$native = json_decode($json);
print_r($sxe, true);
print_r($native, true);

And I end up with the following output:

SimpleXMLElement Object
(
    [@attributes] => Array
        (
            [id] => 123
            [apples] => pears
        )
    [Works] => SimpleXMLElement Object
        (
            [@attributes] => Array
                (
                    [foo] => bar
                    [id] => 234
                )
        )
    [Doesnt] => Stack Exchange
)

stdClass Object
(
    [@attributes] => stdClass Object
        (
            [id] => 123
            [apples] => pears
        )

    [Works] => stdClass Object
        (
            [@attributes] => stdClass Object
                (
                    [foo] => bar
                    [id] => 234
                )

        )

    [Doesnt] => Stack Exchange
)

As you can see, the SimpleXMLElement object and the stdClass object are both missing all the attributes for the <Doesnt> tag. Is there some work around or alternative means to get at them?

Answer 1

As @Gordon said, you don't get the whole picture with print_r and var_dump: SimpleXMLElement Object has some unusual properties, so these dumping functions would fail to represent its structure correctly.

Still, if you had used this instead:

$sxe = simplexml_load_string($plain);
var_dump($sxe->Doesnt);

... you would have seen these attributes and their values intact.

But look closely at the output:

object(SimpleXMLElement)[3]
  public '@attributes' => 
    array (size=2)
      'bar' => string 'foo' (length=3)
      'id' => string '345' (length=3)
  string 'Stack Exchange' (length=14)

It's kinda unusual, don't you think, to have string value just hanging up there - without any corresponding property owning it? But that's exactly what would make direct conversion problematic. JSON (as it is) deals with simple structures - objects and arrays - and this structure cannot be represented correctly in neither of them: you have at least to introduce some additional attribute to store the text content.

Still, this approach seems to be in demand (to some extent), and you're definitely not alone: here's an open ticket in PHP bug tracking system.