XSLT 3.0 by example

On Stackoverflow there was a question to merge two documents with the same XML structure by copying text data from elements in a secondary input document to empty elements in the main input. I think this can be nicely done by using the path() function on empty elements together with the xsl:evaluate instruction (that is new in XSLT 3.0) to simply select and copy the text data from the secondary input document <xsl:template match="*[not(has-children())]"> <xsl:copy> <div class="merged"> <xsl:evaluate context-item="$doc2" xpath="path() || '/text()'"></xsl:evaluate> </div> </xsl:copy> </xsl:template> Full gist is The posted stylesheet works fine with Saxon 9.7 and later editions supporting xsl:evaluate (PE and EE) and with current versions of Altova XMLSpy/Raptor.

In this post I will explore how the new snapshot function in XSLT 3.0 makes life easier with tasks like extracting sub trees from a document, for instance to split up a file into several, each containing a sub tree from the original file. So let's assume we have some input data as follows: and we want to split it up into several documents, each containing a page element in its sub tree, for instance: With XSLT 2.0 we would have to reconstruct the sub tree a page element is contained in, now with XSLT 3.0's snapshot function it is as easy as: So we simply process each page element and in the matching template we create a result document containing the sub tree of the page by using         <xsl:result-document href="pages/page{position()}.xml">             <xsl:copy-of select="root(snapshot())"/>         </xsl:result-document> It is also easy to adapt the stylesheet to work with streaming , to make our apply-templates