Re: ordering and iteration problem

Hi Mark,

> I take it that the generate-id() is needed because there is no other
> way to test for node identity?

Yes - there's no other way of testing for node identity unless there's
something about the node that gives it an identity (such as an ID
attribute). The 'var' attribute looked promising, but it wasn't
present on the nested circuit-breaker-panel elements in Hector's
example.

> Rather than using concat() to generate a compound key, would it also
> work to use the same key as before but then select from the
> resulting node-set only the circuit-breaker with the right parent?

Yes, it would. That way you end up doing more comparisons - for every
circuit-breaker in a particular position you'd have to test whether it
comes from each of the possible circuit-breaker-panels, so it's likely
to be less efficient that including the generated ID in the key. But
then you might not care about that, and even if you did you should
test to see whether it actually made any impact on the transformation
as a whole with your particular processor/environment/source etc.

Cheers,

Jeni

---
Jeni Tennison
http://www.jenitennison.com/


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list