> I would like to query a collection of 20,000 documents. But any "collection()" call takes forever (5 minutes) to execute, even when only as a parameter to "ft:query()". I know Lucene quite well and it has no problem handling that quantity of documents. I would just know if there would be a way to call the "ft:query()" method without it requiring loading up the whole collection in memory first? Maybe pass-on only a path string?
20k docs is not that much, so there must be something in the way you
call ft:query. To allow eXist to use the indexes correctly, your call
to ft:query should occur within a predicate, e.g.:
collection(...)//foo[ft:query(., "query string")].