Hi,

While writing an XSL, is there any way of getting the position of the node in the original document?

By position I'm referring to char offset from the beginning of the file or any similar information like line and column number. I would like to embed this information in the output if possible.

Thanks in advance.

In "straight" XSLT no; the XML is presented to the XSL processor as an infoset, a tree of nodes, and not as a text document. However some processors have extensions that allow this information to come through. I know Saxon has this ability but I can't remember the setting that allows it.

--

Joe (Microsoft MVP - XML)

There's no standard way of doing this.

In Saxon, provided you use the -l switch or equivalent when parsing, you can access the line and column using the saxon:line-number() and saxon:column-number() extensions.

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

Thanks for the quick response, guys.

Sorry for not mentioning this earlier, but I'm using the transformation as part of a Java code and not executing it externally.

Let me try attacking this from a different angle: is there a way to uniquely identify the original XML element and use it in the output? In that case I can do some post-processing to get the location of the original node.

Saxon is a Java based XSLT processor, so I don't see how that should present you with a problem.

/- Sam Judson : Wrox Technical Editor -/

>Sorry for not mentioning this earlier, but I'm using the transformation as part of a Java code and not executing it externally.

There's an equivalent of the Saxon -l flag in the Java API: use setAttribute(LINE_NUMBERING) on the TransformerFactory. See http://www.saxonica.com/documentatio...formation.html
But again this is Saxon-specific.

>Let me try attacking this from a different angle: is there a way to uniquely identify the original XML element and use it in the output? In that case I can do some post-processing to get the location of the original node.

You might be better off preprocessing - put a SAX filter in the pipeline, and use it to add a lineNumber attribute to every element, which is then available to the XSLT code. This approach would work with any JAXP processor.

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

Thanks again, great suggestions.

BTW, any way this can be done in Xalan?

By the looks of the Xalan documentation they seem to provide a very similar function.

/- Sam Judson : Wrox Technical Editor -/