BTW, we use Lucene .NET not Java currently, so version is 1.9. Unfortunately we don’t have "setAllowDocsOutOfOrder" but do have "useScorer14" which is almost the same thing for some queries. I did not see much improvement and for other queries it was slower. We are stuck on 1.9 due some stability issues (memory leaks?) in 2.0+ of the .NET port.
What I did finally, was pre-load fields which have a limited set of unique values into memory (using BitArray for >= num_docs/8, and SortedVIntList for < num_docs/8). Then I "optimize" incoming queries by replacing instances of TermQuery for those fields with my own "CachedTermQuery" objects which have custom weight and scorer classes which use the cached field data. -----Original Message----- From: Paul Elschot [mailto:[EMAIL PROTECTED] Sent: Saturday, July 26, 2008 3:48 PM To: java-user@lucene.apache.org Subject: Re: Fastest way to get just the "bits" of matching documents Op Thursday 24 July 2008 23:00:33 schreef Robert Stewart: > Queries are very complex in our case, some have up to 100 or more > clauses (over several fields), including disjunctions and prohibited > clauses. Other than the earlier advice, did you try setAllowDocsOutOfOrder() ? Regards, Paul Elschot --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
