OutOfMemory (the correct one).

[Wim Lambrecht]

Hi,
We get an OutOfMemoryException during the run of the code at the bottom of this message. The XML-file to parse is about 72M big. After 30 sec it runs out of memory. Maybe it has something to do with the parser Xalan vs. Xerces ...
Any ideas ? TIA
wil

public static void main(String[] args) {
    System.out.println("begin");
    FileWriter pwos = null; 
    try {
        TransformerFactory tFactory = TransformerFactory.newInstance();
        String stylesheet = "mysql.xsl";
        String sourceId = "fdfakt.xml";
        File pricesHTML = new File("output.txt");
        pwos = new FileWriter(pricesHTML);
        Transformer transformer = tFactory.newTransformer(new StreamSource(stylesheet));
        transformer.transform(new StreamSource(sourceId), new StreamResult(pwos));
        System.out.println("einde");
    } catch (Exception e) {
        e.printStackTrace();
    } finally {
        try {
            System.out.println("finally closing the PrintWriter-os.");
            pwos.close();
        } catch(Exception e) {
            System.out.println("Can't close pwos.");
        }
    }
}

EDIT: I took the privilege of adding CODE tags for your code snippet and removed the duplicate post.
[ October 06, 2003: Message edited by: Lasse Koskela ]

[Christophe Grosjean]

I may be wrong but :
72M is quite huge... The transformation requires the building of a DOM-like structure of the XML (in Xalan, I think that they have a "light" implementation of DOM), so it takes memory !
If your XSL is simple, perhaps a solution is to parse your XML with SAX and to output the content in your text file (doing the XSL job in the java code instead).
Christophe

[Wim Lambrecht]

Do i need a SAXTransformerFactory instead of a TransformerFactory then ?
TIA. wil.

[Christophe Grosjean]

No. In fact, what I'm suggesting is to avoid using Transformer and XSL.
Perhaps you can do the transformation by yourself by :
1) parse the XML input file with a SAX parser
2) sets to the SAX parser an XMLFilter as ContentHandler
3) sets an XMLWriter as the parent for the XMLFilter
4) The XMLFilter set in (2) must be a subclass of XMLFitlerImpl in which the event methods have been overriden (do specific job that was in your XSL, for instance in the startElement()).
But :
* this works if your output format is XML
* if not (HTML, text...) do not use XMLFilter and XMLWriter, but simply create an implementation of ContentHandler directly writing the output data in a File.
Questions :
* Did you try your transformation with a smaller version of your document ? Was it working ?
* Is your transformation complex ? An example ?
Christophe

[Wim Lambrecht]

A quick reply: with a smaller document it works perfect. I will reply later on ...