[Task]: grouping on categorical columns should not require Singleton partitioning
Open
dataframe
P2
task
- Dominant language
- Java
- Stars
- 8.7k
- Forks
- 4.7k
- Avg merge
- 1d 20h
- Merged PRs (30d)
- 196
Description
### What needs to happen?
Currently groupby with observed=False (the default) requires aggregating in the Singleton partition since it would otherwise produce results with every index value within every partition.
Originally reported in BEAM-11190, which was erroneously closed at the time of the migration to GitHub Issues.
### Issue Priority
Priority: 2
### Issue Component
Component: dsl-dataframe
Contributor guide
Assessment
This issue has not been assessed yet.