gongxun0928 commented on PR #1178: URL: https://github.com/apache/cloudberry/pull/1178#issuecomment-2992227707
> In other words, for the intermediate state data structure NumericAggState in the first stage aggregation, all the int64 fields inside can actually be changed to int32, not just only the above two. This approach is challenging. For instance, both the Finalize HashAggregate and Partial HashAggregate phases rely on the same workflow—advance_aggregates → xxx → do_numeric_accum—to perform numeric accumulation and update the aggregate state. If we were to use different NumericAggState structures for different phases, we would need to refactor a large number of numeric-related functions to support a more compact state data structure. Additionally, in finalize_aggregates, we would need to select different serialization functions depending on the type of aggsplit. All of these changes would result in significant differences between our codebase and upstream PostgreSQL, making future version upgrades more difficult to manage. -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
