Hi All,


I am trying to transform a 572 MB input.xml using a transform.xsl on Sun Solaris system(having more than 10 GB memory) in which I have used below for-each-group three times:

<xsl:for-each-group select="School/student" group-by="name/tokenize(., ';\s*')">
  <xsl:sort select="current-grouping-key()"/>
  <name count="{count(current-group())}">
    <xsl:value-of select="current-grouping-key()"/>
  </name>
</xsl:for-each-group>
I am using saxon9.0.0.3 processor for same. I tried it with this command-

java -jar saxon/saxon9.jar -o ./output_xml ./input_xml transform.xsl

and got following error:
Exception in thread "main" java.lang.OutOfMemoryError: Java heap space

Now I tried the same with this command-

java –Xmx1024m -jar saxon/saxon9.jar -o ./output_xml ./input_xml transform.xsl

and again got the same error:
Exception in thread "main" java.lang.OutOfMemoryError: Java heap space

Can I use any other command to resolve this error or Saxon cannot handle the 572 MB size of file? Is there any other processor which I can use?

Please help me.

Thanks,
Manish

I normally recommend allocating memory equal to five times the input document size - but you sometimes need more, sometimes less.

With Saxon-SA there are techniques for reducing the memory required: see

http://www.saxonica.com/documentatio...cs/serial.html

and

http://www.saxonica.com/documentatio...rojection.html

Michael Kay
http://www.saxonica.com/
Author, XSLT Programmer's Reference and XPath 2.0 Programmer's Reference

You might be missing a switch there I think (don't know if this will fix your problem). The -s switch is (I think) needed before the input.xml.

java -jar saxon/saxon9.jar -o output_xml -s input_xml transform.xsl

/- Sam Judson : Wrox Technical Editor -/

Thanks Michael and Sam for your quick reply.

I used the command below:
java -Xmx3870m -jar saxon/saxon9.jar -o ./output_xml -s ./input_xml transform.xsl
and got this error:

Exception in thread "main" java.lang.OutOfMemoryError
        at java.util.zip.ZipFile.open(Native Method)
        at java.util.zip.ZipFile.<init>(ZipFile.java:203)
        at java.util.jar.JarFile.<init>(JarFile.java:132)
        at java.util.jar.JarFile.<init>(JarFile.java:70)
        at sun.misc.URLClassPath$JarLoader.getJarFile(URLClas sPath.java:579)
        at sun.misc.URLClassPath$JarLoader.<init>(URLClassPat h.java:546)
        at sun.misc.URLClassPath$3.run(URLClassPath.java:324)
        at java.security.AccessController.doPrivileged(Native Method)
        at sun.misc.URLClassPath.getLoader(URLClassPath.java: 313)
        at sun.misc.URLClassPath.getLoader(URLClassPath.java: 290)
        at sun.misc.URLClassPath.getResource(URLClassPath.jav a:160)
        at java.net.URLClassLoader$1.run(URLClassLoader.java: 192)
        at java.security.AccessController.doPrivileged(Native Method)
        at java.net.URLClassLoader.findClass(URLClassLoader.j ava:188)
        at java.lang.ClassLoader.loadClass(ClassLoader.java:3 06)
        at sun.misc.Launcher$AppClassLoader.loadClass(Launche r.java:268)
        at java.lang.ClassLoader.loadClass(ClassLoader.java:2 51)
        at java.lang.ClassLoader.loadClassInternal(ClassLoade r.java:319)


then I tried with this command-
java -Xmx3500m -jar saxon/saxon9.jar -o ./output_xml -s ./input_xml transform.xsl

and got same previous error:
Exception in thread "main" java.lang.OutOfMemoryError: Java heap space

I don't know much about Saxon-SA,so I am going through the links suggested by Michael. Hope it will work for me.

Manish

Is it possible to construct smaller data files with the same structure? Try doing something smaller and see how the memory requirement increases with data size.

Your stack trace surprises me rather:

Exception in thread "main" java.lang.OutOfMemoryError
        at java.util.zip.ZipFile.open(Native Method)
        at java.util.zip.ZipFile.<init>(ZipFile.java:203)
        at java.util.jar.JarFile.<init>(JarFile.java:132)
        at java.util.jar.JarFile.<init>(JarFile.java:70)
        at sun.misc.URLClassPath$JarLoader.getJarFile(URLClas sPath.java:579)

because it's loading software at the time of the error. Usually you'll get an OutOfMemory error while building the tree representing the source document, which will show a very different kind of stack trace. That's not necessarily significant, just a bit odd.

Have you tried running with -t on the Saxon command line to see how far it gets before it fails?

Also, try running a trivial (identity) transform on the same source document. That should tell you whether the problem is in loading the source document into memory, or whether it's something that your stylesheet is doing (like creating huge amounts of temporary data in working storage).



Michael Kay
http://www.saxonica.com/
Author, XSLT Programmer's Reference and XPath 2.0 Programmer's Reference