Details
-
Sub-task
-
Status: Closed
-
Major
-
Resolution: Won't Do
-
None
-
None
-
None
-
None
Description
Currently HoodieWriteClient does something like `hoodieRecordRDD.map().sort()` internally.. if we want to support Flink DataStream as the object, then we need to somehow define an abstraction like HoodieExecutionContext<T> which will have a common set of map(T) -> T, filter(), repartition() methods. There will be subclass like HoodieFlinkDataStreamExecutionContext<DataStream> which will implement it in Flink specific ways and hand back the transformed T object.