Invalid xpath expression

[Dennis Zandvliet]

I'm using Xalan extension function to call a java function like this:

    
public void testNode(final Node formnode){

...
   Node node = XPathAPI.selectSingleNode(node, "//form");
...



}

But i keep getting this error, regardless of what xpath expression i use.

 java.lang.IllegalArgumentException: Invalid xpath expression: //form   

Any clues?

[Paul Sturrock]

Weird. That is valid XPath syntax. What happens if you try a differernt selector, //* for example?

[Dennis Zandvliet]

I'm sorry the errormessage is a bit misleading, it was thrown by wrapper method of

XPathAPI.selectSingleNode(parent, xpath);

This is the original error:

  	Unknown error in XPath

This selector //* also give the same error.


formnode.getNodeType() gives type ELEMENT;


I've also tried this:

		Document doc = formnode.getOwnerDocument();
		
			if ( doc == null){
				LOGGER.debug("xml node test 014");
			} else {
				LOGGER.debug("xml node test 014a");
				final Node nodetest = XMLUtil.selectSingleNode(doc, "//*");
				LOGGER.debug("xml node test 014b");
			}

Also give the same error:

Unknown error in XPath		

[Paul Clapham]

Can we clarify your throwaway remark about "Xalan extension function"?

If I understand this right, you have an XSL transformation. And from that you're calling Java methods. And from those Java methods you're calling this code that you have shown us. And the intent of the XPath is that it should be operating on the nodes of the XML document that is being transformed? Is that correct?

I can certainly see how that would be a stretch to make that scheme work, if that's actually what you are trying to do, but I don't have anything helpful to say about it at the moment. Except that you might have a look at just what it is you're trying to apply the XPath expression to. A little debugging, that is. Do you actually have a Node object? Things like that.

[Dennis Zandvliet]

Paul Clapham wrote:Can we clarify your throwaway remark about "Xalan extension function"?

If I understand this right, you have an XSL transformation. And from that you're calling Java methods. And from those Java methods you're calling this code that you have shown us. And the intent of the XPath is that it should be operating on the nodes of the XML document that is being transformed? Is that correct?

Yes, almost correct, the nodes are not from XML document itself but generated in the xslt and bind to a variable which is than passed to the java method.

Something like this:
When, in the Java method, I convert the node-tree to a string I get a xml document, with all the expected elements. So I'm sure the node is not empty.

Paul Clapham wrote:Can we clarify your throwaway remark about "Xalan extension function"?
Do you actually have a Node object? Things like that.

Yes, because when i do formnode.getNodeType() it gives type ELEMENT;
Also all the other methods work, like getChildNodes() etc.

[Paul Clapham]

Ah... you're generating the nodes in the XSLT. So perhaps they don't belong to any document, and calling getOwnerDocument returns something peculiar?

[Dennis Zandvliet]

Paul Clapham wrote:Ah... you're generating the nodes in the XSLT. So perhaps they don't belong to any document, and calling getOwnerDocument returns something peculiar?

Yes i was also thinking in this direction. But it this the case
1) How can i make them part of a proper created en initialised DOM document?
2) How is it possible that the nodes are not part of a document, because according to Sun's documentation nodes are always part of an document?
3) How come that after this call

Document doc = formnode.getOwnerDocument();

I can call al the other regular methods like getElementsByTagName() on this class?

I'ts very late here, i'm gonna try it tomorrow

Thanks