jpountz opened a new pull request, #14312: URL: https://github.com/apache/lucene/pull/14312
This is an attempt to resurrect #12194 in a (hopefully) better way. Now that many queries run with `DenseConjunctionBulkScorer`, which scores windows of doc IDs at a time, it becomes natural to skip clauses that have long runs of matches by checking if they match the whole window. This introduces the same `DocIdSetIterator#peekNextNonMatchingDocID()` API that PR #12194 suggested, implements it in `DocIdSetIterator#all`, and uses it in `DenseConjunctionBulkScorer` to skip clauses that match the whole window. For better test coverage, `DenseConjunctionBulkScorer` was refactored to require at least one iterator, which can be a `DocIdSetIterator#all` instance if all docs match. In follow-ups, we should look into supporting other queries that are likely to have long runs of matches, in particular doc-value range queries on fields that are part of the index sort and take advantage of a doc-value skipper. Closes #11915 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For additional commands, e-mail: issues-h...@lucene.apache.org
