How to improve speed large xml validation against xsd in Java?

Question

I'm trying to validate a very XML (~200MB) against XSD. It's taking almost 3 hours. I'm not sure what am I doing wrong here?

    SchemaFactory sf = SchemaFactory.newInstance(W3C_XML_SCHEMA_NS_URI);
    Schema schema = sf.newSchema(new File(this.productExtraInfoXsd));

    DocumentBuilderFactory domFactory = DocumentBuilderFactory.newInstance();
    domFactory.setNamespaceAware(true);
    DocumentBuilder builder = domFactory.newDocumentBuilder();
    Document doc = builder.parse(new File(filePath));

    DOMSource domSource = new DOMSource(doc);
    DOMResult result = new DOMResult();

    Validator validator = schema.newValidator();
    validator.validate(domSource, result);

You might not be doing anything wrong. Run a profiler on your job and see if it's getting stuck in a particular place. If not, most likely it's just taking that long to process it. — chrylis -cautiouslyoptimistic-, Nov 04 '13 at 16:25
Run a profiler. Look for hotspots. If there aren't any, get a faster computer. — chrylis -cautiouslyoptimistic-, Nov 04 '13 at 16:28

constantlearner · Accepted Answer · 2013-11-04T16:57:32.640

3

check this article on XML unmarshalling from Marco Tedone see here. Based on his you can see Stax

XMLInputFactory xmlInputFactory = XMLInputFactory.newInstance();
XMLStreamReader xmlStreamReader = xmlInputFactory.createXMLStreamReader(fileInputStream);
Validator validator = schema.newValidator();
validator.validate(new StAXSource(xmlStreamReader));

edited Nov 04 '13 at 16:57

answered Nov 04 '13 at 16:35

constantlearner

5,157
7
42
64

According to what is written in the article, using StAX is slower than using "Pure JAXB unmarshalling" version, just using StAX will save memory. So I think that if @toy wants to speed up the validation process, this is not a good solution ... or am I missing something? – Paolo Nov 04 '13 at 16:52
1

@Paolo - I don't believe constantlearner meant `unmarshalling` in the JAXB sense. This is simply suggesting using a `StAXSource` instead of a `DOMSource`. – bdoughan Nov 04 '13 at 16:56
1

@BlaiseDoughan If so than ok, I have actually just found an old stackoverflow topic that is reporting right what you are saying ... – Paolo Nov 04 '13 at 17:02

score 3 · Answer 2 · edited May 23 '17 at 10:27

3

Have a look at this stackoverflow topic. Here is written that:

You should not use the DOMParser to validate a document (unless your goal is to create a document object model anyway). This will start creating DOM objects as it parses the document - wasteful if you aren't going to use them.

Maybe it will be useful!

edited May 23 '17 at 10:27

Community

1
1

answered Nov 04 '13 at 16:59

Paolo

1,641
11
15

1

+1. 3 hours does seem excessive for a 200Mb validation, though it depends a bit what's in the schema (e.g. expensive regular expressions). But the first step is to see how long it takes if you don't build a DOM at the same time. – Michael Kay Nov 05 '13 at 08:27

How to improve speed large xml validation against xsd in Java?

2 Answers2