Re: Data Relativity Question

I generally find that if you're doing a join then it help to point a
variable at both the inputs, and avoid relying too heavily on context.

That still leaves the question of which should be your inner loop and which
your outer loop. If you use keys, or if your processor does join
optimization, it shouldn't matter too much.

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

> -----Original Message-----
> From: Bob Portnell [mailto:simply.bobp@xxxxxxxxx]
> Sent: 06 September 2006 21:03
> To: xsl-list@xxxxxxxxxxxxxxxxxxxxxx
> Subject:  Data Relativity Question
>
> Here's a weird notion that just crossed my mind.
>
> My XTVD listings filter is now finished (or at least working
> satisfactorily). The input file is the fairly large file of
> XTVD listings. Most of the execution is in <xsl:for-each>
> loops on the elements of a fairly small outside file of
> titles, people and things that I want to look out for. Those
> loops establish a context outside the input file that I'm
> continually having to step around (thanks to the variable
> trick pointed out by several here). And, as I said, what I
> have works amply well.
>
> Though it wouldn't give me any different visual result, would
> I be better served by using my watch list as the input and
> then go out of that context to get the data in the xtvd file?
> Hmm. Now that I think about it, it probably isn't going to
> make that much difference is it?
> I mean, there's no way around the fact that I'm dealing with
> two files here, and I'm going to run into context issues no
> matter how I slice them.
>
> Wisdom of the List welcome,
>
> Bob P
> simply.bobp@xxxxxxxxx