Skip to content

CometIcebergWriteDetectionSuite assumes a split Iceberg write plan #6803

Description

@andygrove

Found while enabling spark.comet.exec.localTableScan.enabled by default in draft PR #6780 (Spark 4.1 SQL tests and [scans] job). Part of #4347.

5 tests fail in the Spark 4.1 [scans] job because a LocalTableScan-sourced write is now fully native (CometIcebergWrite over CometLocalTableScan) instead of IcebergWriteExec. Tests: "registration tags a fall-back reason on the write exec", "SparkWrite reflection helpers all resolve...", "no fall-back reason is recorded when the iceberg write feature is disabled", "fall-back: BatchWrite that is not an Iceberg SparkWrite", "Compatible when every partition field is void...". Related to #6664. Once default-on, the explicit localTableScan.enabled in docs and Iceberg diffs becomes redundant.

Activity

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

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions