Skip to content

Convert an Arrow batch's schema once instead of once per row #9521

Closed

Description

@aglinxinyuan

Task Summary

ArrowUtils.getTexeraTuple(rowIndex, vectorSchemaRoot) converts the root's Arrow schema into a Texera Schema on every call, so PythonProxyServer rebuilds it for each row of every Data batch a Python worker sends, and ArrowSourceOpExec for each row it reads from an .arrow file. Convert it once per VectorSchemaRoot and pass it to a getTexeraTuple overload, without changing the tuples either caller produces. Raised by kunwp1 in the review of #9515.

Task Type

  • Refactor / Cleanup
  • Performance

Activity

  1. aglinxinyuan commented on Nov 17, 2026

    @aglinxinyuan
    ContributorAuthor

    /take

  2. added a commit that references this issue on Nov 17, 2026
    55e179a
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

No labels
No labels

Type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions