I don't think anyone is working on it..., because It looks like this is a real hit to performance (the results I posted on LUCENE-1476). We shouldn't do it.
And, in fact, we should consider doing the reverse: taking filters that are now applied via iteration and instead distributing them down to each TermQuery (LUCENE-1536) via random-access API (if the filter can support it); that can give much better performance.
However I need to redo these tests once LUCENE-1345 is in. Mike Jason Rutherglen wrote:
Is anyone working on this? I can't find a patch. I'll start one unless someone has something to post.
--------------------------------------------------------------------- To unsubscribe, e-mail: java-dev-unsubscr...@lucene.apache.org For additional commands, e-mail: java-dev-h...@lucene.apache.org
