The XML parser converts the text of an XML document into UTF-8, even if you have set the character encoding of the XML, for example as a second parameter of the DOMDocument constructor. After parsing the XML with the load() command all its texts have been converted to UTF-8.
In case you append text nodes with special characters (e. g. Umlaut) to your XML document you should therefore use utf8_encode() with your text to convert it into UTF-8 before you append the text to the document. Otherwise you will get an error message like "output conversion failed due to conv error" at the save() command. See example below:
<?php
// Text to insert into XML below
$txt = "a text with special characters like 'ä', 'ß', 'Ü' etc.";
// Create Instance of DOMDocument
$dom = new DOMDocument;
// Load XML file
// Was created before with DOMDocument('1.0', 'iso-8859-1')
$dom = $dom->load("file.xml");
// Find the parent node
$parent = $dom->documentElement;
// Create Instance of DomXPath
$xpath = new DomXPath($dom);
// new node will be inserted before this node
$next = $xpath->query("//parentnode/childnode");
// Create the new element
$new_elem = $dom->createElement('new_elem');
// Insert the new element
$parent->insertBefore($new_elem, $next->item(0));
// DOMXML = utf-8! (will be converted to iso-8859-1 only at 'save()')
// prevents error message "output conversion failed due to conv error" at 'save()'
$txt = utf8_encode($txt);
// Create new text node with utf-8 encoded string
$nodetext = $dom->createTextNode("$txt");
// Append text node to new element
$nodetext = $new_elem->appendChild($nodetext);
// save
$dom->save("file.xml");
?>
Hope this helps someone.
siegparr
DOMDocument::save
(PHP 5)
DOMDocument::save — Sauvegarde l'arbre interne XML dans un fichier
Description
int DOMDocument::save
( string
$filename
[, int $options
] )Crée un document XML depuis la représentation DOM. Cette fonction est habituellement appelée après la construction d'un nouveau document DOM, comme dans l'exemple ci-dessous.
Liste de paramètres
-
filename -
Le chemin où sera sauvegardé le document XML.
-
options -
Options additionnelles. Actuellement, seul LIBXML_NOEMPTYTAG est supporté.
Valeurs de retour
Retourne le nombre d'octets écrit ou FALSE si une erreur survient.
Historique
| Version | Description |
|---|---|
| 5.1.0 |
Ajout du paramètre options
|
Exemples
Exemple #1 Sauvegarde d'un arbre DOM dans un fichier
<?php
$doc = new DOMDocument('1.0');
// nous voulons un bel affichage
$doc->formatOutput = true;
$root = $doc->createElement('book');
$root = $doc->appendChild($root);
$title = $doc->createElement('title');
$title = $root->appendChild($title);
$text = $doc->createTextNode('This is the title');
$text = $title->appendChild($text);
echo 'Écrit : ' . $doc->save("/tmp/test.xml") . ' bytes'; // Écrit : 72 bytes
?>
Voir aussi
- DOMDocument::save()
- DOMDocument::load() - Charge du XML depuis un fichier
- DOMDocument::loadXML() - Charge du XML depuis une chaîne de caractères
siegparr at NOSPAM dot web dot de ¶
6 years ago
ss at littlerain dot com ¶
4 years ago
In addition to:
DOMDocument->formatOutput = true
DOMDocument->preserveWhiteSpace = false
...you also have to ensure you don't add text nodes as siblings of element nodes, or formatOutput won't work.
chenel324 at gmail dot com ¶
4 years ago
It took me forever to discover that
DOMDocument->formatOutput = true
will only have an effect on documents that are loaded from disk if one also sets
DOMDocument->preserveWhiteSpace = false ....
Hope this saves somebody a headache.
jon at mysql dot com ¶
5 years ago
If you're working with an existing XML file that contains an entity definition with an external reference like this one:
[
<!ENTITY % all.entities SYSTEM "all-entities.ent">
%all.entities;
]>
Be warned that DOMDocument->save() will expand this to include all definitions in all referenced entities file(s) when it writes the file. In other words, the same section of your saved file would now look something like this:
<!ENTITY % all.entities SYSTEM "all-entities.ent">
<!--
This file names all the entity files needed by .xml files in the
current directory. All ENTITY declarations should be given
first, followed by references to the those entities.
--><!ENTITY % fixedchars.entities SYSTEM "../common/fixedchars.ent">
<!ENTITY % urls.entities SYSTEM "../refman-common/urls.ent">
<!ENTITY % phrases.entities SYSTEM "../common/phrases.ent">
<!ENTITY % ndb.entities SYSTEM "ndb.en.ent">
<!ENTITY minus "–"><!-- 2013 is actually Unicode for ndash -->
<!ENTITY Oslash "Ø">
<!ENTITY macr "¯">
<!ENTITY auml "ä">
[etc.]
I've not found a way to keep this from occurring.
