In XPath I want a expression to select all under div but those nodes that are not a. The tree structure must remain the same - html-agility-pack

Question

How to exclude specific descendants of a node? In this direction, the expression *[not(self::nodetag)] seems just to discriminate at a child level of the node, accepting all other descedants in the returned node set. I want a expression to select all under div but those nodes that are not a, see example below. The tree structure must remain the same.
The approach poste by #Dimitri Novatchev seems to be right but not for HAP implementation:
Using this example document:
<div>
<span>
<a>lala</a>
</span>
</div>
The HAP would return the following structure with his suggested expression /div/descendant::node()[not(self::a)]
<div>
<span>
<a>lala</a>
</span>
</div>
<span>
<a>lala</a>
</span>
If there would be another tag other than a nested on span, it would also return it as a separte tree, any one know about this strange behavior? Is it a HAP bug?
Thanks

Answer 1

I want a expression to select all under div but those nodes that are
not a. The tree structure must remain the same.
Use:
/div/descendant::node()[not(self::a)]
This selects any descendant of the top element div that (the descendant) is not an a.
XSLT - based verification:
<xsl:stylesheet version="1.0"
xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
<xsl:output omit-xml-declaration="yes" indent="yes"/>
<xsl:template match="/">
<xsl:for-each select="/div/descendant::node()[not(self::a)]">
<xsl:value-of select="concat('
', position(), '. "')"/>
<xsl:copy-of select="."/>"
</xsl:for-each>
</xsl:template>
</xsl:stylesheet>
When this transformation is applied on the provided XML document:
<div>
<span>
<a>lala</a>
</span>
</div>
the XPath expression is evaluated and all selected nodes are output with proper formatting to make them well-visible:
1. "
"
2. "<span>
<a>lala</a>
</span>"
3. "
"
4. "lala"
5. "
"
6. "
"
As we can see, 6 nodes are selected -- one span element, four whitespace-only text nodes and one non-whitespace-only text node -- and none of them is an a.
Update:
In a comment the OP has clarified that he actually wants the XML document to be transformed into another, in which any a descendant of a div is omitted.
Here is one such transformation:
<xsl:stylesheet version="1.0"
xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
<xsl:output omit-xml-declaration="yes" indent="yes"/>
<xsl:strip-space elements="*"/>
<xsl:template match="node()|#*">
<xsl:copy>
<xsl:apply-templates select="node()|#*"/>
</xsl:copy>
</xsl:template>
<xsl:template match="div//a"/>
</xsl:stylesheet>
When this transformation is applied on the same XML document (above), the (wwhat I guess is) wanted result is produced:
<div>
<span/>
</div>
If we want to produce only the descendants of any div that has an a descendant, then we need almost the same transformation:
<xsl:stylesheet version="1.0"
xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
<xsl:output omit-xml-declaration="yes" indent="yes"/>
<xsl:strip-space elements="*"/>
<xsl:template match="node()|#*">
<xsl:copy>
<xsl:apply-templates select="node()|#*"/>
</xsl:copy>
</xsl:template>
<xsl:template match="div[.//a]"><xsl:apply-templates/></xsl:template>
<xsl:template match="div//a"/>
</xsl:stylesheet>
The result of this applied to the same XML document as above is:
<span/>

Answer 2

#Devela: you are confusing the set of nodes selected by the XPath expression with the way that they are then displayed by the application that issued the request. It's quite common for an application to display a node by showing the whole subtree rooted at that node. So if your query is //div, and one of the selected div elements contains an <a> node as a descendant, the results will be shown including that <a> element. You can't change that by changing the XPath expression, because the XPath expression didn't select the <a> element; you can only change it by changing the way the results are displayed.
Now, if you want to display a <div> element that is like the <div> element in your source except that the <a> is omitted, then you are outside the scope of what XPath can do. XPath can only choose a subset of the nodes in your input tree, it can't create a modified tree. For that, you need XSLT or XQuery.