RE: Node Set and Node List differentiation

[Kay Michael <Michael.Kay@xxxxxxx>]

> I inferred the following based on the context. Can somebody 
> please confirm or clarify the following.
> 
> Node Set: Nodes that have common parent element type in the 
> hierarchy. This is essential when using UnionExpr since subsequent filters

> would require consistency in the element type of input nodes.

No. A node set is any set of nodes, they do not have to share a parent and
they don't need to be nodes of the same type or elements with the same
element name. For example, "//H1 | //comment()" is a node set containing all
H1 elements and all comments in the document.

> 
> Node List: Could be a 'Node set' but not necessarily. This is 
> a context list of nodes that contain nodes of different types from
arbitrary 
> levels (in the document order).

A node list is a collection of nodes that has a defined order. The order is
not necessarily document order.

Example:

from-ancestors(.) generates a node list containing all the ancestors of the
current node in reverse document order

from-ancestors(.)[position()<4] selects the first three nodes in that
list, returning a new node list containing the parent, grandparent, and
greatgrandparent of
the current node, in that order (reverse document order)

xsl:variable name="elders" expr="from-ancestors(.)[position()<4]"
converts this
node list to a node set (i.e. it loses the ordering information and
potentially eliminates duplicates), so the value of elders is a node set.
 
xsl:apply-templates select="$elders" (in the absence of an xsl:sort
specification) sorts this node set into document order before applying the
relevant template to each node in the sorted list: so the order of
processing is greatgrandparent, grandparent, parent.

Mike Kay


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list