[FLINK-8688] Enable distinct aggregation for data stream on Table/SQL API

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Implemented
• Fix Version/s: 1.7.0
• Component/s: Table SQL / API
• Labels: pull-request-available
• Assignee: Rong Rong
• Reporter: Rong Rong
• Created: 18/Feb/18 16:10
• Updated: 19/Sep/19 19:18
• Resolved: 13/Sep/18 21:29

Description

Distinct aggregation is not currently supported on data stream with Table/SQL API. This is an umbrella task for enabling distinct aggregation in various use cases.

Discussion doc can be found here: https://docs.google.com/document/d/1zj6OA-K2hi7ah8Fo-xTQB-mVmYfm6LsN2_NHgTCVmJI/edit?usp=sharing 

Distinct aggregation is a very important feature in SQL processing and there are many JIRAs currently open with various use cases. The goal is to create one solution to both unbounded and bounded distinct aggregation on data stream so that it can easily be extended to support these use cases.

Issue Links

• relates to: FLINK-5315 Support distinct aggregations in table api
• relates to: FLINK-6388 Add support for DISTINCT into Code Generated Aggregations
• relates to: FLINK-6249 Distinct Aggregates for OVER window
• relates to: FLINK-6335 Parse DISTINCT over grouped window in stream SQL
• relates to: FLINK-6373 Add runtime support for distinct aggregation over grouped windows
• links to: Distinct Aggregation Design Doc
• links to: GitHub Pull Request #5502
• links to: GitHub Pull Request #6574