Github user mridulm commented on the issue:
https://github.com/apache/spark/pull/15713
@andrewcraik If we are simply delegating to ArrayBuffer, it makes sense to
remove the class - it is an additional object created for each key being
aggregated.
Alternative would be to implement the array management within CompactBuffer
and specialize array for primitives - reducing overhead if relevant.
Having said that, I do think the original intention for the class continues
to hold - depends on cardination of values per key. For larger cardinality, it
might be as simple changing order of conditions to improve performance:
newIndex > 1 followed by 0, 1.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
