erikerlandson opened a new pull request #25024: [SPARK-27296][SQL][WIP] User Defined Aggregators that do not ser/de on each input row URL: https://github.com/apache/spark/pull/25024 ## What changes were proposed in this pull request? Defining a new user defined aggregator interface, called `UserDefinedImperativeAggregator`, that does not invoke ser/de on each input row. It is based on the internal `TypedImperativeAggregate` class that spark's predefined aggregators make use of. ## How was this patch tested? Testing is currently very preliminary, and done on the spark-shell REPL. The current PR revision includes a `CountSerDeUDIA` class that runs and verifies that ser/de is only being invoked at partition boundaries and final result compilations. Substantially more testing will be needed.
---------------------------------------------------------------- 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: [email protected] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
