[fast-ut] [reduced-it] [SkipRecovery] Restore Delta deletion vector write coverage [databricks]

[wjxiz1992]

JaCoCo sql-plugin line coverage: not measurable locally — DBR 14.3 shim classfiles do not match the Apache shim350 nightly baseline

Contributes to #12027

Summary

• Remove the [BUG] [DB 14.3] tightBounds stat in Delta Lake tables is set incorrectly #12027-specific expected-failure reason after the fix in Match logic of DV enabling in Gpu Stats Collection [databricks] #12451.
• Restore normal CPU/GPU parity execution for five Delta write cases with deletion vectors enabled.
• Remove the redundant [BUG] [DB 14.3] tightBounds stat in Delta Lake tables is set incorrectly #12027 reason from the managed-table case without counting it as recovered; the remaining Databricks xfails are tracked by [BUG] delta_lake_write_test::test_delta_write_round_trip_managed fails due to mismatched output on Databricks 14.3 #13106 and Delta Lake compaction writes can fallback due to WriteIntoDeltaCommand #11169.

Recovered parameters

• test_delta_write_round_trip_unmanaged[True]
• test_delta_overwrite_round_trip_unmanaged[True]
• test_delta_append_round_trip_unmanaged[True]
• test_delta_write_round_trip_cdf_write_opt[True]
• test_delta_write_aqe_join[True]

Relinked, not recovered

test_delta_write_round_trip_managed[True] is not counted as recovered. This change only removes its redundant #12027 parameter-level reason; the test remains covered by the open Databricks xfails in #13106 and #11169.

Validation

Exact source: 7e457724a96a4367cb74e79e8404d1d35122517f plus this marker-only patch.

• Python syntax: python3 -m py_compile integration_tests/src/main/python/delta_lake_write_test.py
• Reduced-IT selector self-test: python3 integration_tests/src/main/python/reduced_it_selection_test.py (exit 0; all three single-parameter cases retained)
• DBR 14.3 / Spark 3.5.0 / buildver 350db143, focused: 5 passed, 2 warnings in 192.12s
• DBR 17.3 / Spark 4.0.0 / buildver 400db173, focused: 5 passed, 5 warnings in 93.28s
• DBR 14.3 full delta_lake_write_test.py: 150 unique JUnit cases across 8 isolated Spark processes; 106 passed, 23 skipped, 20 xfailed, 1 xpassed; 0 failures/errors
• DBR 17.3 full delta_lake_write_test.py: 150 unique JUnit cases; 123 passed, 2 skipped, 23 xfailed, 2 xpassed, 5 warnings in 901.01s; 0 failures/errors
• JaCoCo diagnostic: the nightly merge showed a raw +107 lines on the matching class subset, but JaCoCo reported 432 runtime/classfile mismatches, so no numeric delta is claimed
• NT local review: 6 reviewers, 0 findings

Documentation

• Updated for new or modified user-facing features or behaviors
• No user-facing change

Testing

• Added or modified tests to cover new code paths
• Covered by existing tests
  (Please provide the names of the existing tests in the PR description.)
• Not required

Performance

• Tests ran and results are added in the PR description
• Issue filed with a link in the PR description
• Not required

[Copilot]

Pull request overview

This PR updates Delta Lake write integration test markers to reflect that the deletion-vector (DV) write coverage gap tracked in #12027 has been fixed (via #12451), restoring normal CPU/GPU parity execution for DV-enabled cases on Databricks runtimes.

Changes:

• Remove the parameter-level xfail reason for DV-enabled writes that referenced #12027, switching back to the standard deletion_vector_values parametrization.
• Keep the managed-table test’s existing Databricks xfails for the still-open issues (#13106, #11169), while dropping the now-redundant #12027 reason.

[greptile-apps]

Greptile Summary

Restores normal CPU/GPU parity coverage for five Delta deletion-vector write cases and removes a redundant parameter-level expected-failure reason from the managed-table case.

• Replaces the issue-specific deletion-vector parameter helper with the standard deletion-vector values.
• Retains the managed-table test’s existing Databricks expected-failure markers.
• Preserves parameter IDs and reduced-IT selection behavior.

Confidence Score: 5/5

The PR appears safe to merge because it only restores execution of previously expected-failing test parameters while preserving collection and parity-check behavior.

The plain deletion-vector parameter list retains the same runtime-dependent values as the former helper, reduced-IT selection retains these single-parameter cases, and the managed test remains covered by its two function-level Databricks xfails.

Important Files Changed

Filename	Overview
integration_tests/src/main/python/delta_lake_write_test.py	Removes the resolved #12027 parameter-level xfail from six Delta write parameterizations without changing their runtime parameter values or CPU/GPU parity assertions.

_{Reviews (1): Last reviewed commit: "test: restore Delta deletion vector writ..." | Re-trigger Greptile}