Indeed, this looks like a better idea in most of the cases. Unfortunately, in my case it doesn't seem fit, because there is no common 'characteristic' of /AAA and /BBB. I guess, in my case more logically is to have the special property (for example a multivalue @excludedQueries) against which I can filter out things.
//*[not(@excludedQueries = 'x')] I think that from the pov of logic this would scale better in case I will have to add more 'selective' queries. Am I wrong? ./alex -- .w( the_mindstorm )p. On 5/9/06, Jukka Zitting <[EMAIL PROTECTED]> wrote:
Hi, On 5/9/06, Alexandru Popescu <[EMAIL PROTECTED]> wrote: > Now I have to think about some tricks to be done. For example, let's > say that the nodes under /CCC have a property that is not available on > /AAA and /BBB children. Would it be possible to make use of this in > the XPath? I guess the other way around would work (having in both > /AAA and /BBB children a property that is not available on /CCC). > > Please advise. Any ideas will be highly appreciated, as this > represents a critical stopper for my project. I've quite often used empty mixin types as markers for various things, including queries that cannot be easily expressed otherwise. Is there some characteristic feature of the /AAA and /BBB nodes that make them interesting for the query? Create a marker mixin type for that feature and apply it to whichever nodes you want included in your query. BR, Jukka Zitting -- Yukatan - http://yukatan.fi/ - [EMAIL PROTECTED] Software craftsmanship, JCR consulting, and Java development
