[APEXMALHAR-2472] Implement Kudu Input Operator

• Type: New Feature
• Status: Closed
• Priority: Major
• Resolution: Implemented
• Fix Version/s: 3.8.0
• Component/s: adapters database
• Assignee: Ananth
• Reporter: Ananth
• Created: 01/Apr/17 22:13
• Updated: 11/Nov/17 11:25
• Resolved: 27/Sep/17 14:59

Description

This operator would allow Kudu to be used as an Input store. This has multiple advantages like :

• Ability to solve the dedup problem from entire data set perspective. The dedupe operators we have today are primarily window based and this might not meet all of the use cases in real world.
• Ability to selectively stream data based on a SQL expression. Since Kudu is a structural store, we could effectively allow a SQL expression based "input" definition that would allow for selective streaming for all downstream operators. This could potentially be an alternative streaming store pattern as compared to Kafka as Kafka does not allow for selective streaming of tuples.

Issue Links

• relates to: APEXMALHAR-2546 Mocking dependencies should ideally be in the maven parent instead of individual modules
• links to: GitHub Pull Request #667