Validate XML using a custom DTD in PHP

Asked 19/9, 2008 at 13:46 Answered 12/9, 2011 at 14:15

Is there a way (without installing any libraries) of validating XML using a custom DTD in PHP?

Sabinasabine answered 19/9, 2008 at 13:46 Comment(2)

So, just to clarify - does "custom DTD" mean "DTD which is independent/different from any DTD which may be specified in the content of the XML file"? – Dorcia 11/9, 2011 at 1:56

See #1274673 – Mulligatawny 10/10, 2012 at 11:21

Take a look at PHP's DOM, especially DOMDocument::schemaValidate and DOMDocument::validate.

The example for DOMDocument::validate is fairly simple:

<?php
$dom = new DOMDocument;
$dom->Load('book.xml');
if ($dom->validate()) {
    echo "This document is valid!\n";
}
?>

Patric answered 19/9, 2008 at 13:50 Comment(6)

the only way to get the validation error is to use a custom error handler. really ugly. php sucks at error handling – Jeffereyjefferies 2/4, 2009 at 11:29

uk3.php.net/manual/en/domdocument.schemavalidate.php#62032 looks like there is a better way than a custom error handler – Jeffereyjefferies 2/4, 2009 at 11:47

@Andrei - It certainly helps to see validation errors displayed properly, so it is a win call libxml_use_internal_errors(true) before validation and libxml_get_errors() after a failure. – Dorcia 9/9, 2011 at 16:39

@Patric - I don't think this really answers the original question because "book.xml" will simply be validated against whatever DTD is specified in the content of book.xml and not a "custom" DTD specified by the caller at runtime. – Dorcia 9/9, 2011 at 16:43

FYI there is a bug in PHP with DOMDocument::validate() bugs.php.net/bug.php?id=48080 – Cellulose 8/3, 2013 at 12:8

This validates if an xml is in the correct format as an xml.. But I think he is asking for how to validate against custom ruls like DTD? I mean you need to check if the xml elements namings tags, correspond to the correct xml element names? – Christmann 6/11, 2014 at 17:20

If you have the dtd in a string, you can validate against it by using a data wrapper for the dtd:

$xml = '<?xml version="1.0"?>
        <!DOCTYPE note SYSTEM "note.dtd">
        <note>
            <to>Tove</to>
            <from>Jani</from>
            <heading>Reminder</heading>
            <body>Don\'t forget me this weekend!</body>
        </note>';

$dtd = '<!ELEMENT note (to,from,heading,body)>
        <!ELEMENT to (#PCDATA)>
        <!ELEMENT from (#PCDATA)>
        <!ELEMENT heading (#PCDATA)>
        <!ELEMENT body (#PCDATA)>';


$root = 'note';

$systemId = 'data://text/plain;base64,'.base64_encode($dtd);

$old = new DOMDocument;
$old->loadXML($xml);

$creator = new DOMImplementation;
$doctype = $creator->createDocumentType($root, null, $systemId);
$new = $creator->createDocument(null, null, $doctype);
$new->encoding = "utf-8";

$oldNode = $old->getElementsByTagName($root)->item(0);
$newNode = $new->importNode($oldNode, true);
$new->appendChild($newNode);

if (@$new->validate()) {
    echo "Valid";
} else {
    echo "Not valid";
}

Gatekeeper answered 30/6, 2011 at 9:48 Comment(4)

So why does this code produce output "Not valid"? Trapping errors from libxml I see the following: <b>Error 517</b>: Could not load the external subset "data://text/plain;base64,PCFFTEVNRU5UIG5vdGUgKHRvLGZyb20saGVhZGluZyxib2R5KT4KICAgICAgICA8IUVMRU1FTlQgdG8gKCNQQ0RBVEEpPgogICAgICAgIDwhRUxFTUVOVCBmcm9tICgjUENEQVRBKT4KICAgICAgICA8IUVMRU1FTlQgaGVhZGluZyAoI1BDREFUQSk+CiAgICAgICAgPCFFTEVNRU5UIGJvZHkgKCNQQ0RBVEEpPg==" on line <b>0</b> – Dorcia 9/9, 2011 at 16:15

I wish I could downvote this for broken code (or at least revoke my upvote). – Dorcia 9/9, 2011 at 16:34

The problem which I'm having with the above code appears to be in the createDocumentType() call, which generates the DOCTYPE element. This is what I want (for the example): <!DOCTYPE note [<!ELEMENT note (to,from,heading,body)> <!ELEMENT to (#PCDATA)> ... <!ELEMENT body (#PCDATA)>]> but this is what I get: <!DOCTYPE note SYSTEM "data://text/plain;base64,PCFFTEVNRU5UIG5vdGUgKHRvLGZyb20saGVhZGluZyxib2R5KT4KICAgICAgICA8IUVMRU1FTlQgdG8gKCNQQ0RBVEEpPgogICAgICAgIDwhRUxFTUVOVCBmcm9tICgjUENEQVRBKT4KICAgICAgICA8IUVMRU1FTlQgaGVhZGluZyAoI1BDREFUQSk+CiAgICAgICAgPCFFTEVNRU5UIGJvZHkgKCNQQ0RBVEEpPg=="> – Dorcia 9/9, 2011 at 17:19

looks more like the code has been copied over from here and then combined with the data wrapper. @Peter: External subset loading might be disabled on your configuration, it does work. – Dibbuk 18/10, 2014 at 17:3

My interpretation of the original question is that we have an "on board" XML file that we want to validate against an "on board" DTD file. So here's how I would implement the "interpolate a local DTD inside the DOCTYPE element" idea expressed in comments by both Soren and PayamRWD:

public function validate($xml_realpath, $dtd_realpath=null) {
    $xml_lines = file($xml_realpath);
    $doc = new DOMDocument;
    if ($dtd_realpath) {
        // Inject DTD inside DOCTYPE line:
        $dtd_lines = file($dtd_realpath);
        $new_lines = array();
        foreach ($xml_lines as $x) {
            // Assume DOCTYPE SYSTEM "blah blah" format:
            if (preg_match('/DOCTYPE/', $x)) {
                $y = preg_replace('/SYSTEM "(.*)"/', " [\n" . implode("\n", $dtd_lines) . "\n]", $x);
                $new_lines[] = $y;
            } else {
                $new_lines[] = $x;
            }
        }
        $doc->loadXML(implode("\n", $new_lines));
    } else {
        $doc->loadXML(implode("\n", $xml_lines));
    }
    // Enable user error handling
    libxml_use_internal_errors(true);
    if (@$doc->validate()) {
        echo "Valid!\n";
    } else {
        echo "Not valid:\n";
        $errors = libxml_get_errors();
        foreach ($errors as $error) {
            print_r($error, true);
        }
    }
}

Note that error handling has been suppressed for brevity, and there may be a better/more general way to handle the interpolation. But I have actually used this code with real data, and it works with PHP version 5.2.17.

Dorcia answered 12/9, 2011 at 14:15 Comment(0)

Trying to complete "owenmarshall" answer:

in xml-validator.php:

add html, header, body, ...

<?php

$dom = new DOMDocument; <br/>
$dom->Load('template-format.xml');<br/>
if ($dom->validate()) { <br/>
    echo "This document is valid!\n"; <br/>
}

?>

template-format.xml:

<?xml version="1.0" encoding="utf-8"?>

<!-- DTD to Validate against (format example) -->

<!DOCTYPE template-format [  <br/>
  <!ELEMENT template-format (template)>  <br/>
  <!ELEMENT template (background-color, color, font-size, header-image)>  <br/>
  <!ELEMENT background-color   (#PCDATA)>  <br/>
  <!ELEMENT color (#PCDATA)>  <br/>
  <!ELEMENT font-size (#PCDATA)>  <br/>
  <!ELEMENT header-image (#PCDATA)>  <br/>
]>

<!-- XML example -->

<template-format>

<template>

<background-color>&lt;/background-color>  <br/>
<color>&lt;/color>  <br/>
<font-size>&lt;/font-size>  <br/>
<header-image>&lt;/header-image>  <br/>

</template> 

</template-format>

Afrikander answered 3/3, 2011 at 16:11 Comment(2)

Same here, you don't load the DTD anywhere. – Tendance 24/3, 2011 at 9:33

In his example he interpolated the DTD locally inside the DOCTYPE element (that's what Soren's code is trying to do, but it doesn't seem to work). – Dorcia 9/9, 2011 at 18:24

Trying to complete "owenmarshall" answer:

Recommended topics

Hot tags