Consider using RangeEndEstimator in a finite range for better Split/Progress
- Dominant language
- Java
- Stars
- 8.7k
- Forks
- 4.7k
- Avg merge
- 2d 2h
- Merged PRs (30d)
- 205
Description
More discussions:
https://github.com/apache/beam/pull/11715#issuecomment-630965824
In PR11715, a RangeEndEstimator is only used when the range is infinite(end \== Long.MAX_VALUE). We may also want to use RangeEndEstimator in a finite range(especially after the first split of infinite range) for a better Split/Progress.
Imported from Jira [BEAM-10032](https://issues.apache.org/jira/browse/BEAM-10032). Original Jira may contain additional context.
Reported by: boyuanz.
Contributor guide
Research direction
Start with the discussion in PR11715, then trace the existing RangeEndEstimator use for infinite ranges and the finite-range Split/Progress behavior. Done means the estimator’s finite-range applicability is resolved and the resulting Split/Progress behavior is covered by the project’s relevant tests.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- java
- Domain
- data-engineering
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 32/100