Skip to content

Filter manifest lists on pyiceberg-core Arrow output in plan_files #4299

Open

Description

@mbutrovich

Feature Request / Improvement

Scan planning reads the manifest list into ManifestFile objects and then filters them by partition summaries (table/__init__.py#L2805-L2809). #4007 shows that pyiceberg-core parses a manifest list faster than Cython, but building the ManifestFile objects makes the end-to-end path slower.

Apply the approach from #4298 to read_manifest_list: read the list from pyiceberg-core as Arrow, and filter it on the Arrow columns so that only matching manifests become Python objects. This needs apache/iceberg-rust#3649. pyiceberg-core is an optional extra, so the Cython reader stays as the fallback.

Activity

  1. mbutrovich commented on Nov 6, 2026

    @mbutrovich
    ContributorAuthor

    take

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

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions