[jira] [Commented] (LUCENE-10396) Automatically create sparse indexes for sort fields

[Ignacio Vera (Jira)]

Title: Message Title

Ignacio Vera commented on LUCENE-10396   Re: Automatically create sparse indexes for sort fields I think you are assuming that we always visit all terms in the index. The iteration might be driven by the result of a query that contains a subset of the documents. In pseudo-code would look like: DocIdSetIterator iterator = executeQuery(); SortedDocValues sortedDocValues = getSortedDocValues(); int doc = iterator.nextDoc(); while (doc != DocIdSetIterator.NO_MORE_DOCS) { if (sortedDocValues.advanceExact(doc)) { BytesRef bytesRef = sortedDocValues.lookupOrd(sortedDocValues.ordValue()); consume(bytesRef); // advance our iterator to the next document with different ordinal doc = iterator.advance(sortedDocValues.advanceOrd()); } else { doc = iterator.nextDoc(); } } I don't see how to do this efficiently with the inverted index only. Add Comment   This message was sent by Atlassian Jira (v8.20.10#820010-sha1:ace47f9)