Encoding issues with XMLWriter (PHP)

Question

Take this simple PHP code:

$xmlWriter = new XMLWriter();
$xmlWriter->openURI('php://output');
$xmlWriter->startDocument('1.0', 'utf-8');

$xmlWriter->writeElement('test', $data);

$xmlWriter->endDocument();
$xmlWriter->flush();

The XMLWriter class has a nice feature: it will convert any data you give to it to the output encoding. For example here it will convert $data to UTF-8 because I passed 'utf-8' in the startDocument function.

The problem is that in my case the content of $data comes from a database whose output format is UTF-8 and is therefore already in UTF-8. The XMLWriter probably thinks the data is in ISO-8859-1 and converts it again to UTF-8, and I get weird symbols where I should get accents.

Currently I'm using utf8_decode around each string coming from the database, which means I'm converting from UTF-8 to ISO-8859-1, and then XMLWriter turns it back into UTF-8.

This works but is not clean:

$xmlWriter->writeElement('test', utf8_decode($data));

Is there a cleaner solution ?

EDIT: showing a full example

$xmlWriter = new XMLWriter();
$xmlWriter->openURI('php://output');
$xmlWriter->startDocument('1.0', 'utf-8');
$xmlWriter->startElement('usersList');

$database = new PDO('mysql:host=localhost;dbname=xxxxx', 'xxxxx', 'xxxxx');
$database->exec('SET CHARACTER SET UTF8');
$database->exec('SET NAMES UTF8');
foreach ($database->query('SELECT name FROM usersList') as $user)
   $xmlWriter->writeElement('user', $user[0]);   // if the user's name is 'hervé' in the database, it will print 'hervé' instead

$xmlWriter->endElement();
$xmlWriter->endDocument();
$xmlWriter->flush();

Answer 1

I'm not sure where you got the idea that XMLWriter converts encodings. It doesn't. You must supply it with utf-8. It can output different encodings, but input strings must be utf-8.

One of two things may be going on here:

1. Whatever you are using to view your output document is interpreting the string as win-1252. If you are viewing your output in a browser, you may need to set the content-type header like so: header('Content-Type: application/xml; charset=UTF-8');
2. You stored your data in your database incorrectly, and your "é" is actually two unicode characters "é". Fixing this is difficult.