RE: [xsl] preceding-sibling axis scope

["Michael Kay" <mike@xxxxxxxxxxxx>]

Ah, I think I now see your misunderstanding. It's a common one. When you
have a sequence of (unrelated) nodes, such as the sequence returned by
current-group(), the nodes are not siblings of each other. A node has a
single parent, but can participate in any number of sequences; its siblings
are the children of that parent, which may or may not be members of the same
sequence.

Michael Kay
http://www.saxonica.com/  

> -----Original Message-----
> From: Clint Redwood [mailto:clint@xxxxxxxxxxxxxxx] 
> Sent: 16 April 2009 12:59
> To: xsl-list@xxxxxxxxxxxxxxxxxxxxxx
> Subject: [xsl] preceding-sibling axis scope
> 
> Hi,
> 
> I'm confused by what my XSL is producing, when I'm using the 
> preceding-sibling axis, within a xsl:for-each-group element.
> 
> When I have a reference like this...
> 
> current-group()[preceding-sibling::elemA]
> 
> does the preceding-sibling refer to the current-group() 
> content, or the source document?
> 
> I had expected it to scope to the current-group() but the 
> effect appears to suggest that the scope is the source document.
> 
> Apologies if this is a stupidly obvious question.
> 
> Yours,
> 
> Clint.