Re: [xsl] Filtering problem

[Marcus Streets <marcus@xxxxxxxxxxx>]

Wendell Piez wrote:
> Marcus,
> 
> Mukul is correct to point you to the identity transform to build your
> filter.
> 
> In particular, your problem is just a bit trickier than the ordinary
> case since your requirements include things like
> 
>>> there are no attributes starting dim1
>>> or
>>> dim1_a="yes"
> 
> Unfortunately, the proposed match="*[starts-with(name(@*), 'dim1')]
> won't work, since when there are several attributes it will only examine
> the name of the first of them, in document order, which as it happens is
> undefined.

Thanks to both of you.

The part I was really struggling with was priority - I had not got my
head around that.

Once I had that sorted I found I could be much more brutal with the
selection - as I did not need to worry about the corner cases as
priority magically solved those for me.

Marcus