apache / apache/hudi

Fix clean planning for very large partitions

Open
#15,479 2 comments 0 reactions 0 assignees View on GitHub
component:cleaning from-jira priority:critical status:pr-available type:improvement
Dominant language
Java
Stars
6.2k
Forks
2.5k
Avg merge
2d 4h
Merged PRs (30d)
112

Description

Within clean planning phase, we do a map() for every partition and then trigger planning for each partition within that. 

 

For very large number of partitions, and if cleaner shuffle parallelism is small, this results in more sequential planning. We can enhance this with mapPartitions call and optimize it 

 

## JIRA info

- Link: https://issues.apache.org/jira/browse/HUDI-5012
- Type: Improvement

Contributor guide

No contributing guide indexed for this repository

Research direction

Start by tracing the clean planning phase and the per-partition map() calls described in the issue, then inspect how cleaner shuffle parallelism affects planning. Compare the existing behavior with the mapPartitions approach; done means very large partition counts no longer cause avoidable sequential planning.

Written by the indexing model from the issue text.

Assessment

Tech stack
java, spark
Domain
data-engineering, distributed-systems
Issue type
Feature
Difficulty
4/5
Estimated time
3-5 days
Activity status
Quiet
Clarity
Mostly clear
Newbie friendliness
52/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.