Re: [xsl] iterate through nodes and determine output by node type

["James Fuller" <james.fuller.2007@xxxxxxxxx>]

if its xsl:document u need, u might try exsl:document

http://www.exslt.org/exsl/elements/document/index.html

which has some support in some XSLT processors.

gl, Jim Fuller

On 10/2/07, Abel Braaksma <abel.online@xxxxxxxxx> wrote:
> Mario Madunic wrote:
> > I know it is a dead end so please stop flogging a dead horse by mentioning it.
> > All I wanted was ideas on how to do something I've not done in the past and not
> > be spoken down to.
> >
> > Also I've used 1.1 successfully in the past and it has given me the results I
> > expected. Also I only use it when I can't use 2.0 but have access to Saxon6. The
> > only reason being I never know when I might need <xsl:document />. This time I'm
> > going with the information given to me by the project lead and was told 1.0 or
> > 1.1 as stated.
>
> Hi Mario,
>
> Since I am the one that started the 1.1 vs 2.0 discussion, I want to
> mention this once more (whether you take the advice or not is up to you
> of course). The xsl:document was introduced to output to multiple
> document streams. XSLT 2.0 has embraced that principle, still, but has
> called it xsl:result-document. You don't have to drop anything, simply
> change one or two statements in your stylesheet (and you already use
> Saxon 8.x, you said).
>
> When you go down that path and want to see how deep and fun the rabbit
> hole is, we are all happy to help you. But when it is uphill with an old
> trainwagon in the backpack, with abandoned technology that is, then both
> you and your project are facing a tough climb. Even then, it might be
> that some people out here still remember some bits and pieces of the
> long-forgotten 1.1 specification, so please don't feel offended, just
> ask (but the response will likely be lower than with 1.0 or 2.0, or a
> new thread is started to point you again to the merits of new technology).
>
> About my original advice, I really don't know if XSLT 1.1 has dropped
> the tedious RTF, but if they did, you can use the pipeline technique I
> presented tp make your overall matching process easier. Otherwise, just
> as with 1.0, you'll have to resort to node-set extension functions
> (which can get heavy).
>
> 2700 files in 40 minutes? I wonder if with some help we can help you
> bring that down to several minutes instead. But that would be with 2.0
> of course ;)
>
> Cheers,
> -- Abel Braaksma