Skip to content

Review names of configuration settings #2601

Description

@andygrove

What is the problem the feature request solves?

The configuration settings have many inconsistencies.

For example, we have many different prefixes for shuffle-related configs:

  • spark.comet.columnar.shuffle
  • spark.comet.exec.shuffle
  • spark.comet.native.shuffle
  • spark.comet.shuffle

There are also differences in convention for naming within configs:

  • mergeRanges (Camel Case)
  • thread-pool (Hyphenated)
  • read.parallel.io (dot separated)

Renaming any of these configs will be disruptive to users. If we do rename them, I suggest we do it for the 1.0.0 release.

We could also consider introducing a capability to support the old deprecated names for some period of time.

Describe the potential solution

No response

Additional context

No response

Activity

  1. andygrove commented on Sep 28, 2026

    @andygrove
    MemberAuthor

    This was done ahead of 1.0.0 under #4978. #4986 moved every shuffle config under spark.comet.shuffle.*, the hyphenated and dot-separated keys no longer exist, and #4979 added withAlternative so the old names keep working as deprecated aliases. The naming rules are now written down in the Configuration Conventions page of the contributor guide.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    enhancementNew feature or request

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions