test(databricks): cover catalog metadata text parity

[sd-db]

Resolves #15691

Problem

databricks/dbt-databricks#1428 prevents catalog generation from failing when
metadata from separate schemas receives conflicting agate-inferred types. The
Fusion adapter already represents these fields with Arrow string arrays and
typed Rust strings, but it did not have focused regression coverage for the
same boundary cases.

Solution

Add Databricks metadata adapter tests that verify:

• numeric-looking table, owner, comment, and column values remain strings;
• metadata from separate schema record batches aggregates without a
  value-driven type conflict;
• empty record batches return empty metadata maps without field lookup errors.

No production behavior changes are required because Fusion is already
conformant with the dbt-databricks behavior.

Testing

• cargo test -p dbt-adapter catalog_metadata_ --lib
  • 3 passed; 0 failed; 874 filtered out

Checklist

• I have read the contributing guide and understand what's expected of me.
• I have run this code in development, and it appears to resolve the stated issue.
• This PR includes tests.
• This PR has no interface changes.
• Type annotations are not applicable because this change adds Rust tests only.

[github-actions]

Thank you for your pull request! We could not find a changelog entry for this change. For details on how to document a change, see the contributing guide.

[github-actions]

Thank you for your pull request! We could not find a changelog entry for this change. For details on how to document a change, see the contributing guide.

[sd-db]

This PR adds test-only parity coverage and does not change user-facing behavior or interfaces. Could a maintainer please apply the Skip Changelog label?