PHP xml to array - how to get rid of empty tags?

Question

There is the same problem at xml to array - remove empty array php Dont know how you handle this. I mean how can i get an answer to a question that is not mine and asked > 2 years ago. So im asking here my own question:

Simple script:

$xml
    = '<?xml version="1.0"?>
       <Envelope>
           <foo>
               <bar>
                   <baz>Hello</baz>
                   <bat/>
               </bar> 
           </foo>
           <foo>
               <bar>
                   <baz>Hello Again</baz>
                   <bat></bat>
               </bar>
           </foo>
           <foo>
               <bar>
                   <baz>Hello Again</baz>
                   <bat> </bat>
               </bar>
           </foo>
       </Envelope>';

$xml = new \SimpleXMLElement(
    $xml,
    LIBXML_NOBLANKS | LIBXML_NOEMPTYTAG | LIBXML_NOCDATA
);
$array = json_decode(json_encode((array)$xml), true);
// [
//     'foo' => [
//         0 => [
//             'bar' => [
//                 'baz' => 'Hello',
//                 'bat' => [], <<-- how to get this to NULL
//             ],
//         ],
//         1 => [
//             'bar' => [
//                 'baz' => 'Hello Again',
//                 'bat' => [], <<-- how to get this to NULL
//             ],
//         ],
//         2 => [
//             'bar' => [
//                 'baz' => 'Hello Again',
//                 'bat' => [   <<-- how to get this to NULL
//                     0 => ' ',     or at least to value of " " without array
//                 ],
//             ],
//         ],
//     ],
// ];

As you can see there is an empty <bat/> tag and a whitespace in the last <bat> </bat> tag.

I would like to get those to null in the array.

I tried the following but this works for the first level only ofc:

$data = (array)$xml;
foreach ($data as &$item) {
    if (
        $item instanceof \SimpleXMLElement
        and $item->count() === 0
    ) {
        // is a object(SimpleXMLElement)#1 (0) {}
        $item = null; 
    }
}

I tried and failed doing this recursively.

Also tried RecursiveIteratorIterator but failed.

But there must be a way to get those offset to null.

Anybody done this before?

EDIT

Solved. See https://stackoverflow.com/a/55733384/3411766

Every time I see someone trying to write a generic XML-to-array function, particularly when it begins with a hack like $array = json_decode(json_encode((array)$xml), true);, I despair. Rather than jumping through hoops to special case all the situations where your generic algorithm gives the "wrong" answer, use an XML parser (like SimpleXML) to access the data you actually need, and create an array (or a pre-defined object) that makes sense for your actual application. — IMSoP
– IMSoP, Commented Apr 17, 2019 at 9:39

cottton · Accepted Answer · 2019-04-17 17:50:37Z

Found it out my self. Took a while but works perfectly.

/** 
 * @param array|\SimpleXMLElement[]|\SimpleXMLElement $data .
 *
 * @return array
 */
protected function emptyNodesToNull($data)
{
    if ($data instanceof \SimpleXMLElement and $data->count() === 0) {
        // is empty object like
        //  SimpleXMLElement::__set_state(array())
        //  which was f.e. a <foo/> tag
        // or
        //  SimpleXMLElement::__set_state(array(0 => ' ',))
        //  which was f.e. a <foo> </foo> (with white space only)
        return null;
    }
    $data = (array)$data;
    foreach ($data as &$value) {
        if (is_array($value) or $value instanceof \SimpleXMLElement) {
            $value = $this->emptyNodesToNull($value);
        } else {
            // $value is the actual value of a node.
            // Could do further checks here.
        }
    }
    return $data;
}

My tests did exactly what i expected

and returns imo exactly what you can expect from a xmlToArray method.

I mean we wont be able to handle attributes, but this is not the requirement.

Test:

    $xml
        = '<?xml version="1.0"?>
   <Envelope>
       <a/><!-- expecting null -->
       <foo>
           <b/><!-- expecting null -->
           <bar>
               <baz>Hello</baz>

               <!-- expecting here an array of 2 x null -->
               <c/>
               <c/>

           </bar> 
       </foo>
       <foo>
           <bar>
               <baz>Hello Again</baz>
               <d>    </d><!-- expecting null -->
               <item>
                   <firstname>Foo</firstname>
                   <email></email><!-- expecting null -->
                   <telephone/><!-- expecting null -->
                   <lastname>Bar</lastname>
               </item>
               <item>
                   <firstname>Bar</firstname>
                   <email>0</email><!-- expecting value 0 (zero) -->
                   <telephone/><!-- expecting null -->
                   <lastname>Baz</lastname>
               </item>

               <!-- expecting array of values 1, 2 null, 4 -->
               <number>1</number>
               <number>2</number>
               <number></number>
               <number>4</number>
           </bar>
       </foo>
   </Envelope>';

$xml = new \SimpleXMLElement($xml);
$array = $class::emptyNodesToNull($xml);

Returns:

[
    'Envelope' => [
        'a'   => null,
        'foo' => [
            0 => [
                'b'   => null,
                'bar' => [
                    'baz' => 'Hello',
                    'c'   => [
                        0 => null,
                        1 => null,
                    ],
                ],
            ],
            1 => [
                'bar' => [
                    'baz'    => 'Hello Again',
                    'd'      => null,
                    'item'   => [
                        0 => [
                            'firstname' => 'Foo',
                            'email'     => null,
                            'telephone' => null,
                            'lastname'  => 'Bar',
                        ],
                        1 => [
                            'firstname' => 'Bar',
                            'email'     => '0',
                            'telephone' => null,
                            'lastname'  => 'Baz',
                        ],
                    ],
                    'number' => [
                        0 => '1',
                        1 => '2',
                        2 => null,
                        3 => '4',
                    ],
                ],
            ],
        ],
    ],
];

Ezequiel Esnaola · Accepted Answer · 2019-04-16 19:08:45Z

0

You can use XPath with the predicate not(node()) to select all elements that do not have child nodes.

<?php

$doc = new DOMDocument;
$doc->preserveWhiteSpace = false;
$doc->loadxml('<?xml version="1.0"?>
       <Envelope>
           <foo>
               <bar>
                   <baz>Hello</baz>
                   <bat/>
               </bar>
           </foo>
           <foo>
               <bar>
                   <baz>Hello Again</baz>
                   <bat></bat>
               </bar>
           </foo>
           <foo>
               <bar>
                   <baz>Hello Again</baz>
                   <bat></bat>
               </bar>
           </foo>
       </Envelope>');

$xpath = new DOMXPath($doc);

foreach( $xpath->query('//*[not(node())]') as $node ) {
    $node->parentNode->removeChild($node);
}

$doc->formatOutput = true;
echo $doc->savexml();

Print:

<?xml version="1.0"?>
<Envelope>
  <foo>
    <bar>
      <baz>Hello</baz>
    </bar>
  </foo>
  <foo>
    <bar>
      <baz>Hello Again</baz>
    </bar>
  </foo>
  <foo>
    <bar>
      <baz>Hello Again</baz>
    </bar>
  </foo>
</Envelope>

Regards!

answered Apr 16, 2019 at 19:08

Ezequiel Esnaola

603 bronze badges

2 Comments

Nigel Ren Over a year ago

Not sure if the fact that the last element <bat> </bat> has a space in it in OP's version.

cottton Over a year ago

Same here like @NigelRen answer: this removes the offset. As posted i would like to get them to be null. So {offset} is not set but exists. Same as for the xml (exists but empty)

Collectives™ on Stack Overflow

PHP xml to array - how to get rid of empty tags?

2 Answers 2

Comments

2 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

2 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related