[FLINK-12094] Introduce sort merge join operator to blink batch

• Type: New Feature
• Status: Closed
• Priority: Major
• Resolution: Implemented
• Fix Version/s: 1.9.0
• Component/s: Table SQL / Runtime
• Labels: pull-request-available
• Assignee: Jingsong Lee
• Reporter: Jingsong Lee
• Created: 03/Apr/19 07:30
• Updated: 08/Apr/19 02:17
• Resolved: 08/Apr/19 02:17

Description

Introduce SortMergeJoinOperator: An implementation that realizes the joining through a sort-merge join strategy.

Support all sql types: INNER, LEFT, RIGHT, FULL, SEMI, ANTI

1.In most cases, its performance is weaker than HashJoin.
2.It is more stable than HashJoin, and most of the data can be sorted stably.
3.SortMergeJoin should be the best choice if sort can be omitted in the case of multi-level join cascade with the same key.

Issue Links

• is a child of: FLINK-11490 Add an initial Blink SQL batch runtime
• links to: GitHub Pull Request #8107