[FLINK-34543] Support Full Partition Processing On Non-keyed DataStream

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Done
• Affects Version/s: 1.20.0
• Fix Version/s: 1.20.0
• Component/s: API / DataStream
• Labels: pull-request-available

Release Note

New Features:
fullWindowPartition Method: Introduced in the DataStream class to enable full window processing, allowing collection and processing of all records in each subtask.
PartitionWindowedStream Class: Extends DataStream, facilitating full window processing with several new APIs.
APIs for PartitionWindowedStream:
1.mapPartition: Processes records using MapPartitionFunction.
2.sortPartition: Sort records by field or key in the window.
3.aggregate: Aggregate records in the window.
4.reduce: Reduce records in the window.

• Assignee: Wencong Liu
• Reporter: Wencong Liu
• Created: 29/Feb/24 02:15
• Updated: 19/Jul/24 04:16
• Resolved: 14/Mar/24 03:07

Description

Introduce the PartitionWindowedStream and provide multiple full window operations in it.

The related motivation and design can be found in FLIP-380.

Issue Links

• links to: GitHub Pull Request #24398
• links to: GitHub Pull Request #25100
• links to: GitHub Pull Request #25102