[BUG] Fail closed on incomplete aggregate result mapping

[zpuller]

Fixes #15808.

Description

Positional result mapping for partial typed-imperative aggregates previously used an optional lookup that silently retained the original data type when the expected result expression was missing. This could allow GPU conversion to proceed with inconsistent aggregate buffer metadata.

This change records an incomplete positional mapping during metadata construction and marks the aggregate as unsupported during GPU tagging. It adds a unit regression that removes the expected shuffle-facing result expression from a real partial ObjectHashAggregateExec and verifies that GPU conversion fails closed.

Validation:

• sr-test --suite com.nvidia.spark.rapids.HashAggregatesSuite --buildver 357 --no-install
• 321 tests passed with Spark 3.5.7.
• git diff --check

AI assistance disclosure: Codex assisted with implementation, testing, and drafting this description. The author reviewed the complete diff and final description before PR creation.

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

Performance testing is not required because this changes only driver-side plan metadata validation and adds no runtime row or batch processing.

[zpuller]

build