Optimize GroupIntoBatches for batch Dataflow pipelines
Open
dataflow
improvement
P3
runners
- Dominant language
- Java
- Stars
- 8.7k
- Forks
- 4.7k
- Avg merge
- 1d 20h
- Merged PRs (30d)
- 196
Description
The GroupIntoBatches transform can be significantly optimized on Dataflow since it always ensures that a key K appears in only one bundle after a GroupByKey. This removes the usage of state and timers in the generic GroupIntoBatches transform.
Imported from Jira [BEAM-7912](https://issues.apache.org/jira/browse/BEAM-7912). Original Jira may contain additional context.
Reported by: lcwik.
Contributor guide
Research direction
Start by tracing the GroupIntoBatches transform and the Dataflow batch-pipeline path mentioned in the issue, then locate the relevant implementation and tests. Done means the Dataflow-specific path is optimized without relying on the generic transform's state and timers while preserving batching behavior.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- google-cloud, java
- Domain
- data-engineering, distributed-systems
- Issue type
- Refactor
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100