xslt how to read the document-node() - xslt-2.0

Question

I have a xml file in which one of the element has the CDATA as the value. I put the CDATA value into a variable which I can see is value type of document-node(1) when i debug my code from oXygen. How do I iterate the document-node()?
copy can give me a new xml file. but what I need is not a new file. I only need to read certain nodes and generate a report based on the values on those nodes. so I directly copy the CDATA to my variable and thought I can manipulate it.
I tried to use substring to read the variable things but failed.
I tried to use document(variable) to open the variable but Oxygen give me the debug-error of FODC0002:I/O error reported by xml parser processing file.
here the file is my variable which looks like a xml file
I did google search for the error but only got bench of non-closed questions like Oxygen throw I/O error when use document().
Would anybody let me know what's going wrong? or give me a better solution?
I also tried parse-xml() but I got the following error from Saxon:
F[Saxon-EE9.5.1.5] the processing instruction target matching "[xX][mM][lL]" is not allowed
F[Saxon-EE9.5.1.5] FODC0006: First argument to parse-xml() is not a well formed and namespace-well-formed XML document.
my code to use parse-xml is as below:
<xsl:template match="data"
<xsl:for-each select="parse-xml(root/outsideData)//nodeLevel1/nodeLevel2">
Could anyone give me a sample about how to use parse-xml()? I did google search but didn't find useful samples.
Thanks very much!
A piece of my data is like the following:
<root>
<outsideData id="123">
<child1 key="124375438"/>
<![CDATA[ <?xml version=1.0 encoding="UTF-8"?><insideData xmlns:xlink="http://www.w3.org/1999/xlink">
<nodeLevel1>
<nodeLevel21>packing</nodeLevel21>
<nodeLevel22 ref="12343-454/560" xlink:href="URN:X-MN:DD%3FM=B888%26SDC=A%26CH=79% .../>
</nodeLevel1>
]]>
</outsideData>
</root>
I want to get the inside CDATA <nodeLevel22> #ref and #xlink which will get DD-FM-B888-26-79
My variables are:
<xsl:for-each select="/root/outsideData">
<xsl:variable name="insideData">
<xsl:value-of select="." disable-output-escaping="yes"/>
</xsl:variable>
<xsl:variable name="Data">
<xsl:value-of
select="normalize-space(substring-after($insideData,'?>'))"
disable-output-escaping="yes"/>
</xsl:variable>
</xsl:foreach>
From the debug I can see that the variable insideData and Data are both value type of document-node(1)
Martin's solution works for me very well :)
But I'm still wondering why the following doesn't work:
<xsl:variable name="insideData">
<xsl:value-of select="." disable-output-escaping="yes"/>
</xsl:variable>
<ref>
<xsl:value-of select="substring-before(substring-after($insideData, '<nodeLevel22 ref'),>/>')"/>
</ref>
Here I got empty <ref/>

Answer 1

If you do <xsl:variable name="varName"><xsl:value-of select="..."/><xsl:variable> then you are creating a temporary document fragment that contains a single text with the string contents of the item(s) selected in the value-of. That does not make sense in most cases, doing <xsl:variable name="varName" select="..."/> is usually sufficient.
As for parsing the contents of the outsideData element with parse-xml, there is indeed not only the escaped XML document inside that element but white space as well, thus if you try to parse the contents as XML you get that error as white space before the XML declaration is not allowed. The whole approach of stuffing the XML into a CDATA section with an element with mixed contents is flawed in my view, if you want to store escaped XML into a CDATA then you should make sure that you use a single element that contains nothing but the CDATA section which then only contains the XML markup with no leading white space.
If you can't change the creation of the input data then you will need to make sure you pass in only that part of the string contents of the element to parse-xml that is a well-formed XML document, so you need some way to strip the white space before the XML declaration doing e.g.
<xsl:for-each select="/root/outsideData">
<xsl:variable name="xml-string" select="replace(., '^\s+', '')"/>
<xsl:variable name="xml-doc" select="parse-xml($xml-string)"/>
<!-- now output data e.g. -->
<xsl:value-of select="$xml-doc//nodeLevel1/nodeLevel22/#ref"/>
...
</xsl:for-each>
Untested but should show the right direction as far as trying to use parse-xml.