Support TimestampCombiner in Python streaming mode GroupByKey
Open
bug
core
P3
python
- Dominant language
- Java
- Stars
- 8.7k
- Forks
- 4.7k
- Avg merge
- 2d 5h
- Merged PRs (30d)
- 204
Description
Currently, streaming mode GroupByKey in Python does not respect the specified TimestampCombiner semantics for output elements. We should implement this to better conform to the Beam model.
Imported from Jira [BEAM-2518](https://issues.apache.org/jira/browse/BEAM-2518). Original Jira may contain additional context.
Reported by: ccy.
Contributor guide
Research direction
Start by tracing Python streaming-mode GroupByKey and how TimestampCombiner semantics are handled for output elements. Compare the current behavior with the Beam model and the original BEAM-2518 context; done means streaming GroupByKey respects the specified TimestampCombiner semantics.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- stream-processing
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100