Saxon

Sorry hete the Expression in code

parentNode.getUniquePath()

sorry, sorry; I hope this will be the last time

uniquePath example:

 /M_UTILMD/G_SG4[1]/S_DTM[4]/D_2380

The main challenge with getting a usable path expression for a node is what to do about namespaces. I've no idea how the DOM4J getUniquePath() method solves that problem.

Saxon offers the XPath 3.0 function fn:path(), which solves that issue using EQNames.

If you want to invoke that directly from Java, at the s9api level, you could do XdmFunctionItem.getSystemFunction(....).call(....);

Hi Michael,

to make it clear: I got a program where I used dom4j.

With dom4j I can select nodes using the following expression/example

ListNodeList = document.selectNodes("/M_UTILMD/G_SG4/S_DTM/C_C507/D_2005");

This expression provides a NodeList. Now I can loop over this NodeList to get the unique paths:

			for (Iterator iterator = NodeList .iterator(); iterator.hasNext();) {
				Node node = (Node) iterator.next();
				 System.out.println(node.getUniquePath());

			}

The output will look like:

/M_UTILMD[1]/G_SG4[1]/S_DTM[1]/C_C507[1]/D_2005[1]
/M_UTILMD[1]/G_SG4[1]/S_DTM[2]/C_C507[1]/D_2005[1]
/M_UTILMD[1]/G_SG4[1]/S_DTM[3]/C_C507[1]/D_2005[1]
/M_UTILMD[1]/G_SG4[2]/S_DTM[1]/C_C507[1]/D_2005[1]
/M_UTILMD[1]/G_SG4[3]/S_DTM[1]/C_C507[1]/D_2005[1]

Here I'd like to know the analog expression in SAXON.

Regards Mario

Well, I was trying to explain, I can't see from the DOM4J Javadoc what the precise specification of DOM4J's getUniquePath() is, and that makes it hard to suggest an alternative way of getting the same result. In particular, what does getUniquePath() do when it finds elements that are in a namespace?

How are you using these paths? Are they for human consumption, or are they for evaluation using an XPath engine? (that choice has a profound effect on how you handle namespaces).

There's nothing directly analogous in Saxon's Java API, but if you can tell me exactly what's needed, then I'm sure we can devise an XPath expression that delivers it. If you want a path that is suitable for evaluation by an XPath engine, then the XPath 3.0 fn:path() function meets the requirement, although it won't give you exactly the same output as the getUniquePath() method. If you want something intended for a human reader, then a recursive XQuery function like this might be the answer:

declare function local:getUniquePath($n as node()) as xs:string {
   concat(
     if ($n/parent::*) then local:getUniquePath($n/parent::*) else "",
     name($n),
     if ($n/preceding-sibling::*[name() = name($n)]) then concat('[', count($n/preceding-sibling::*[name() = name($n)]+1, ']' else ""
   )
};

Hi Michael,

many thanks. I got a solution:

		try {
			XPathSelector xpathSel = xPathCompiler.compile(<<>>>+ "/path(.)").load();
			xpathSel.setContextItem(xdmdocument);

			// Do not delete: For debugging
			for (XdmValue val : xpathSel.evaluate()) {
				if (excelRow.number.equalsIgnoreCase(StopAtLine)) {
					String UniquePath = val.toString().replace("Q{}", "");
					System.out.println(UniquePath);
				}
			}

Regards Mario