XSLT transformation java.util.concurrent.CompletionException: java.lang.StackOverflowError

Question

I was trying to convert one XML to another using XSLT 2.0, normally it works, but when file size is large like 1-2 MB I am getting error like java.util.concurrent.CompletionException: java.lang.StackOverflowError

Below Java code is using for the transformation

public static String doTransformation(String request, File xslTransformer)
        throws TransformerException {
    TransformerFactory tf = TransformerFactory.newInstance();
    Transformer trans = tf.newTransformer(new StreamSource(xslTransformer));
    trans.setParameter("indent-elements", "yes");
    ByteArrayOutputStream outputStream = new ByteArrayOutputStream();
    InputStream inputStream = new ByteArrayInputStream(request.getBytes());
    trans.transform(new StreamSource(inputStream), new StreamResult(outputStream));
    return new String(outputStream.toByteArray());
}

Library used

<dependency>
      <groupId>xalan</groupId>
        <artifactId>xalan</artifactId>
     <version>2.7.1</version>
</dependency>

Any help would be appreciated.

Does the larger XML have more levels of nesting and does the XSLT have recursively called templates? Xalan has a command line interface as well I think so perhaps try running the samples from the command line first to check whether you can get an error message indicating which code/line in the XSLT causes the stack overflow. — Martin Honnen
– Martin Honnen, Commented Nov 2, 2020 at 14:55
Used string tokenize instead of split, which resolves the issue. Thank you. — Umaimath Thonikkadavan
– Umaimath Thonikkadavan, Commented Nov 3, 2020 at 13:30

Michael Kay · Accepted Answer · 2020-11-03 09:13:00Z

1

First, and probably irrelevant, but if you're using Xalan then you're not using XSLT 2.0.

The Java code you've shown is irrelevant, it's the XSLT code that matters.

A Stack Overflow exception is going to occur if your code uses recursion and the recursion goes too deep. The depth of recursion is quite likely to depend on the data volumetrics.

You need to find where this recursion occurs in the XSLT code (it's probably a named template that calls itself), and then understand the logic to find a different way of doing it. It may be doing something very simple like a text replacement that could be coded a different way, especially if you are able to move to XSLT 2.0+.

answered Nov 3, 2020 at 9:13

Michael Kay

165k11 gold badges97 silver badges173 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Umaimath Thonikkadavan Over a year ago

Thank you Michael, you were correct, there was logic to split long string, this I have changed to str:tokenize which resolved the issue.

Michael Kay Over a year ago

Please mark the answer as accepted (click the tick/check-mark) so people know the question has been answered.

Grzegorz Oledzki · Accepted Answer · 2020-11-03 09:18:36Z

1

StackOverflowError is not necessarily sign of an infinite recursion. It only says that the program while executing has ran out of memory for stack entries. That seems to confirm with your observations that smaller data files work, but larger don't.

What you need is to keep the program as is, but change the execution environment of it. Namely the stack size arguments to JVM. So that the program is given more memory for (each) stack while it's executing.

TLDR: Try java -xss 2048k $yourArguments or other values.

answered Nov 3, 2020 at 9:18

Grzegorz Oledzki

24.4k16 gold badges73 silver badges111 bronze badges

1 Comment

Umaimath Thonikkadavan Over a year ago

Used string tokenize instead of split, which resolves the issue. Thank you.

Collectives™ on Stack Overflow

XSLT transformation java.util.concurrent.CompletionException: java.lang.StackOverflowError

2 Answers 2

2 Comments

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related