Reduce timezone integration tests in precommit [fast-ut] [reduced-it] [databricks]

[res-life]

Fixes #15345.

Description

Problem

Premerge currently reuses tz_sensitive_test, the marker intended for comprehensive nightly coverage. Its parameterized tests collect 1,542 cases for each additional timezone, so the two premerge runs take approximately 10 minutes in total.

Using reduced-IT selection alone would still retain 318 cases per timezone. The premerge path needs a smaller, coverage-oriented subset instead of the complete nightly matrix.

Change

Register tz_sensitive_test_for_precommit and apply it to nine representative cases per timezone:

• cast and date-time cases that exercise the core GpuTimeZoneDB conversion paths
• ORC coverage for fixed-offset and recurring-DST timezone pairs
• ORC V1 and V2 readers with one representative reader configuration

Update spark-premerge-build.sh to select the new marker. Existing tz_sensitive_test annotations and nightly selection remain unchanged.

Impact

Premerge collection decreases from 1,542 to 9 cases per timezone, or from 3,084 to 18 cases across both timezone runs—a 99.4% reduction. Nightly CI continues to run the complete timezone-sensitive matrix. There is no user-facing behavior change.

Validation

• Collected 1,542 cases with -m tz_sensitive_test
• Collected 9 cases with -m tz_sensitive_test_for_precommit
• Passed all 9 selected cases with Asia/Shanghai
• Passed all 9 selected cases with America/New_York
• Completed a Spark 3.5.6 clean build with JDK 17

Checklists

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

[res-life]

build

[res-life]

build

[res-life]

build

[res-life]

build