Skip to content

Parquet reader: Add support for reading row index metadata columns #3432

Open

Description

@andygrove

What is the problem the feature request solves?

The native_datafusion scan currently falls back to Spark for queries that read row index metadata columns. It would be nice to support them.

Describe the potential solution

No response

Additional context

No response

Activity

  1. self-assigned this
    on May 4, 2026
  2. mbutrovich commented on May 4, 2026

    @mbutrovich
    Contributor
  3. mbutrovich commented on May 5, 2026

    @mbutrovich
    Contributor

    Opened upstream DataFusion PR that we'll hopefully pick up in 54.0:

    apache/datafusion#22026

  4. added this to the 0.17.0 milestone on May 13, 2026
  5. moved this from Todo to In progress in Comet Developmenton May 13, 2026
  6. modified the milestones: 0.17.0, 1.0.0 on Jun 25, 2026
  7. mbutrovich commented on Jul 22, 2026

    @mbutrovich
    Contributor

    Note that this won't land until DF 55.0.

  8. modified the milestones: 1.0.0, 1.1.0 on Jul 22, 2026
  9. changed the title [-][native_datafusion] Add support for reading row index metadata columns[/-] [+]Parquet reader: Add support for reading row index metadata columns[/+] on Jul 22, 2026
  10. modified the milestones: 1.1.0, 1.2.0 on Sep 11, 2026
  11. modified the milestones: 1.2.0, 1.3.0 on Oct 15, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

enhancementNew feature or requestnative_datafusionSpecific to native_datafusion scan type

Type

No type

Projects

Milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions