guixiaowen commented on PR #1391: URL: https://github.com/apache/auron/pull/1391#issuecomment-3385273388
> i suggest not changing this option because all these aggregate plans are convert to the same NativeAggregateExec. use one unique option is ok here. @richox ri Thank you very much for reviewing the code. In the current implementation, all three types of aggregates are eventually converted to NativeAggregateExec. The reasons for separating them are as follows: From Spark’s design perspective, the three aggregation types (HashAggregate, SortAggregate, and ObjectHashAggregate) are logically independent, each serving different use cases. In the current design, whether a Spark physical execution plan is converted to native is controlled by configuration parameters (for example, FileSourceScanExec). Therefore, HashAggregateExec and SortAggregateExec should also be handled independently. Thank you again for your reply. -- 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: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
