[GitHub] [lucene] rmuir commented on pull request #127: LUCENE-9946: Support multi-value fields in range facet counting

[GitBox <gi...@apache.org>]

rmuir commented on pull request #127:
URL: https://github.com/apache/lucene/pull/127#issuecomment-833971467


   @gsmiller do you think we are taking advantage of the sortedness of `SORTED_NUMERIC` here? e.g. do we really need a bitset or could we record hits high-level like this:
   
   ```
   int lastRangeID = -1; // prevents double counting a range within multivalued doc
   for (value : values) {
     rangeID = segmentTree.lookup(value);
     if (rangeID != lastRangeID) {
       hits[rangeID]++;
       lastRangeID = rangeID;
     }
   }
   ```
   
   I'm trying to see if we can accelerate (e.g. avoid bitsets) and maybe reduce overhead between single and multi-valued algorithms, it may simplify code. Of course such an optimization is harder if range can be overlapping.


-- 
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.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org



---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@lucene.apache.org
For additional commands, e-mail: issues-help@lucene.apache.org