[XML4Lib] XPath: Selecting nodes grouped with parent?

[Conal Tuohy]

David Walker wrote:

> I have a simple set of XML files that include category and 
> database information for various library subscription 
> databases.  Each category has one or more subcategories, and 
> each subcategory has one or more databases, as such:
>  
> <category>
>   <subcategory>
>     <database />
>     <database />
>   </subcategory>
>   <subcategory>
>     <database />
>     <database />
>   </subcategory>
> </category>
>  
> What I want to be able to do is apply a limit to the file 
> based on one or more attributes in the databases.  For 

When you say "apply a limit to the file" what does this mean in
practice? i.e. is this expression used in an XSLT which should show only
a subset of nodes in the input? or in an XQuery? What context is the
XPath expression supposed to work in?

> example, only showing those databases that have a <type> 
> field with a value of 'Bibliography'. But I want to be able 
> to keep them in their subcategory groupings.  In essence, I 
> want to select both children (databases) and parent 
> (subcategory) nodes using an XPath expression with 
> conditional logic based on the children.

> This returns subcategory nodes that have a child database 
> with the type of Bibliography, but it also includes ALL of 
> the child nodes, regardless of whether they have a type of 
> 'Bibliography' or not.
>  
>   //subcategory[database/type = 'Biography']

This expression does return only the subcategory nodes, and does not
include their child nodes.

However, if you were to use the expression in XSLT like so:

<xsl:copy-of select="//subcategory[database/type = 'Biography']"/>

... then you would indeed get all the descendant nodes, but this is an
artifact of xsl:copy-of, not of the XPath expression itself.