jpountz commented on issue #14961: URL: https://github.com/apache/lucene/issues/14961#issuecomment-3085467207
I'm not a fan of caching these `OrdinalMap` in heap at search time. IMO, there are two long-term directions: either we embrace the use of ordinal maps for sorting, faceting, etc. figure out how to store them in the index, load them off-heap, and rebuild them on every refresh, or we stop using ordinal maps at search time e.g. by aggregating based on segment-level ordinals first, and later merging these per-segment aggregates by values. FWIW Elasticsearch is currently going with the second option for its ES|QL aggregations. -- 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
