[GitHub] [druid] abhishekagarwal87 commented on a change in pull request #10316: Optimize InDimFilter hashCode calculation

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

abhishekagarwal87 commented on a change in pull request #10316:
URL: https://github.com/apache/druid/pull/10316#discussion_r476398225



##########
File path: processing/src/main/java/org/apache/druid/query/filter/InDimFilter.java
##########
@@ -382,7 +386,7 @@ public boolean equals(Object o)
   @Override
   public int hashCode()
   {
-    return Objects.hash(values, dimension, extractionFn, filterTuning);
+    return Objects.hash(values.size(), dimension, extractionFn, filterTuning);

Review comment:
       It's interesting though that `values` itself is a HashSet being passed to InDimFilter which would mean hash code is evaluated for all the elements in the set. But that penalty for constructing `values` doesn't show up in the graph. is the full flame graph available to look further? 
   I can see in one place where multiple `InDimFilter` are created with the same `values`.  Maybe that's the part responsible for perf penalty. If there is a `Set` type that remembers its hashCode, using such type for `values` could be more beneficial. 




----------------------------------------------------------------
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: commits-unsubscribe@druid.apache.org
For additional commands, e-mail: commits-help@druid.apache.org