[GitHub] [lucene] LuXugang commented on pull request #422: LUCENE-10120: Lazy initialize FixedBitSet in LRUQueryCache

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

LuXugang commented on pull request #422:
URL: https://github.com/apache/lucene/pull/422#issuecomment-960382982


   > RoaringDocIdSet will compress dense situations too. It should be used.
   
   As @jpountz said in [LUCENE-10120](https://issues.apache.org/jira/browse/LUCENE-10120), LRUQueryCache always use RoaringDocIdSet for caching no matter dense or sparse until for conjunction optimization, then FixedBitSet was used for caching when the condition is scorer.cost() * 100 >= maxDoc which means very dense, It also means a huge size FixedBitSet will be cached.  more details see  [LUCENE-7339](https://issues.apache.org/jira/browse/LUCENE-7339)  and [LUCENE-7330](https://issues.apache.org/jira/browse/LUCENE-7330).
   
   > Why do we need a range docidset?
   
   So the purpose of range docidset in this RP is trying to only cahce minDoc and maxDoc while no docId gap in  FixedBitSet, and still have a random access ability see then implementation of RangeDocIdSet#bits() in RP.
   
   


-- 
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-unsubscribe@lucene.apache.org

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