Take a look at previous-sibling axis, contain() function ans
substring-after() function.

On 03/03/06, geoff hopkins <geoffhopkins123@xxxxxxxxx> wrote:
> Take2.
>
> Below is a example of an input xml basically have to
> change the format to that below.  So when START is
> found start reading and stop reading when STOP is
> found.  Have sorted this out ok!
>
> The problem I have is getting the value which is
> outside of the START and STOP tags @doc it is uniquely
> identified in the document as having 'Name:' at the
> start of the text within the node (would prefer it to
> search for first instance of this text)
>
> I can't share the true xml doc for data protection
> issues.
>
> Thanks. Geoff
>
> Example of input doc.
>
> <root>
> <a>sdfsdfsd</a>
> <a>dsfssdf</a>
> <a>Name: DOC_ID1</a>
> <a>iuhs</a>
> <a>sdfsdf</a>
> <a>sdfsd</a>
> <a>START</a>
> <a>A1,A2</a>
> <a>A3,A4</a>
> <a>A5,A6</a>
> <a>A7,A8</a>
> <a>STOP</a>
> <a>sdfsdfds</a>
> <a>sdfsdfsd</a>
> </root>
>
> Example of output:
> <root>
> <doc>
> <relationship parent="A1" child="A2" doc="DOC_ID1"/>
> <relationship parent="A3" child="A4" doc="DOC_ID1"/>
> <relationship parent="A5" child="A6" doc="DOC_ID1"/>
> <relationship parent="A7" child="A8" doc="DOC_ID1"/>
> </doc>
> </root>
>
> __________________________________________________
> Do You Yahoo!?
> Tired of spam?  Yahoo! Mail has the best spam protection around
> http://mail.yahoo.com