Bounded Splittable DoFn Support Status

	Google Cloud Dataflow	Prism Local Runner	Apache Flink	FlareDB	Apache Spark (RDD/DStream based)	Apache Spark Structured Streaming (Dataset based)	Hazelcast Jet	Kafka Streams (experimental, not released)	Twister2	Python Direct FnRunner
Base	Partially : Only Dataflow Runner V2 supports this.	Yes : fully supported	Partially : Only portable Flink Runner supports this.	Yes : fully supported Bounded Splittable DoFns are expanded (pair-with-restriction, split-and-size, truncate, process) and executed with residual handling.	Partially : Only portable Spark Runner in batch mode supports this.	:	:	:	Yes :	No : not implemented
Side Inputs	Partially : Only Dataflow Runner V2 supports this.	Yes : fully supported	Partially : Only portable Flink Runner supports this.	No : not implemented SDF side inputs are modeled in the expansion but not delivered to the harness.	:	:	:	:	:	No : not implemented
Splittable DoFn Initiated Checkpointing	Partially : Only Dataflow Runner v2 supports this.	Yes : fully supported	Partially : Only portable Flink Runner supports this.	Yes : fully supported Residual roots returned by the SDF are processed as new work items with output-watermark holds.	Partially : Only portable Spark Runner in batch mode supports this.	:	:	:	Yes :	No : not implemented
Dynamic Splitting	Partially : Only Dataflow Runner V2 supports this.	Yes : fully supported	No :	No : not implemented Runner-initiated work splitting is not implemented yet.	No :	:	:	:	Yes : Only with Python SDK	No : not implemented
Bundle Finalization	Partially : Only Dataflow Runner V2 supports this.	Yes : fully supported	No :	No : not implemented	No : not implemented	:	:	:	Yes :	No : not implemented

Bounded Splittable DoFn Support Status

Have you found everything you were looking for?