[GitHub] [lucene] easyice opened a new pull request, #12557: Improve refresh speed with softdelete enable

[via GitHub]

easyice opened a new pull request, #12557:
URL: https://github.com/apache/lucene/pull/12557

   I found a flame graph in my production environment, the DocValuesConsumer 
for `___soft_deletes` field accounted for a large proportion
   
   
![image](https://github.com/apache/lucene/assets/23521001/49e4d48a-1ecd-4e92-be0e-7d37fea6c185)
   
   [flame 
(12).html.zip](https://github.com/apache/lucene/files/12619531/flame.12.html.zip)
   
   The docvalue for `___soft_deletes` field is usually sparse, so it will has a 
`IndexedDISI` data structure, the `IndexedDISI` provide operation like 
`advance()`, `index()`, but for  `___soft_deletes` field, the `index()` will 
not used, because the value of the filed is always 1, we can remove the 
calculation of `index()` for this scene.
   
   I ran a benchmark use `IndexWriter#softUpdateDocument` for 
`IndexedDISI.Method.DENSE` case, resulted in  ~16% improvement in performance.
   
   
   


-- 
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