Good advice, it's true that Jackrabbit is fast for searching by identifier. Thank you Marcel.
Frank Le 2009-11-11 à 4:09 AM, Marcel Reutegger a écrit : > Hi, > > 2009/11/6 François Cassistat <[email protected]>: >> Hi ! >> >> My application plan to use JackRabbit 2.0 (oh yeah!) and needs to do this : >> >> - The user enter a search term (complex or not) >> - We show a filtered list of nodes >> - Then we listen to changes in repository (JCR rocks !) >> - We apply changes to the above list of nodes by applying the same filter, >> so a node should be removed or be added to the list after each event. >> >> Is it possible to easily reexecute a Query only on the node that changed? > > you could add a predicate to the original query that is only true for > the changed node. if your node is referenceable you can use the > jcr:uuid property. hmm... actually, that might even work for any node > because in jackrabbit 2.0 every node exposes the UUID through > getIdentifier(). > > regards > marcel > >> From what I see in the JCR2.0 API, the only option is to recreate a new >> request each time. Maybe JQOM could help me by keeping a part of the query >> tree so it could create query faster? Any suggestion? >> >> >> Frank >> >>
