>> > * writing the config file for the GroupingTransformer is less work >> > than doing the same thing in XSLT. >> >> Grouping is a standard XSLT problem. It's not real trivial, it >> usually requires a knowledge of keys, but it's not all that hard >> either. > > though you'll have to rewrite it from scratch each time you need it (AFAIU).
Well yes, but once you've done it it's no different than writing an SQL GROUP BY from scratch each time you need it :-) >> > * calculating things like averages, sums, min and max values is a >> > lot easier than in XSLT (or probably impossible XSLT) >> >> Sum, min, max are usually pretty much trivial. Obviously, if you've >> got sum, average, is not much harder... > > Ah, ok (I was told so by someone else). Though the calculations done by the GroupingTransformer are > based on an expression library, which I think is more powerful than what can be achieved using xslt. Well, as I pointed out XSLT is Turing complete so anything is possible. For some people XSLT Extensions may start to make a lot of sense if you're into math libraries. On the other hand, I should point out there are general XSLT libraries that implement everything from trig functions (log, sin, et al) to pattern generation and functional analysis freely available. For example, join the Mulberry XSLT list and look for stuff from Dimitre Novatchev and his FXSL library... > It would also possible to further extend the GroupingTransformer to do things like sort the groups > based on the calculated summaries (e.g. to place highest sums first). And then there are also plans to > make a crosstabulating transformer some day... Again, I don't really see an issue doing this with XSLT. In a way, sifting through data is one of the things that it's specifically designed for... --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, email: [EMAIL PROTECTED]
