[fix](test) Cast variant subcolumn as json in variant_hirachinal for stable output

[csun5285]

cast(v['c'] as string) on a heterogeneous variant column can produce different whitespace formatting (e.g. "[1, 2, 3]" vs "[1,2,3]") depending on session variables fuzzed by use_fuzzy_session_variable (batch_size, enable_fold_constant_by_be, etc.), causing intermittent regression failures. Casting to json normalizes the serialization path and is stable across fuzzed execution configs (verified 50/50 runs).

What problem does this PR solve?

Issue Number: close #xxx

Related PR: #xxx

Problem Summary:

Release note

None

Check List (For Author)

• Test

  • Regression test
  • Unit Test
  • Manual test (add detailed scripts or steps below)
  • No need to test or manual test. Explain why:
    • This is a refactor/code format and no logic has been changed.
    • Previous test can cover this change.
    • No code files have been changed.
    • Other reason

• Behavior changed:

  • No.
  • Yes.

• Does this need documentation?

  • No.
  • Yes.

Check List (For Reviewer who merge this PR)

• Confirm the release note
• Confirm test cases
• Confirm document
• Add branch pick label

[csun5285]

run buildall

[hello-stephen]

Thank you for your contribution to Apache Doris.
Don't know what should be done next? See How to process your PR.

Please clearly describe your PR:

1. What problem was fixed (it's best to include specific error reporting information). How it was fixed.
2. Which behaviors were modified. What was the previous behavior, what is it now, why was it modified, and what possible impacts might there be.
3. What features were added. Why was this function added?
4. Which code was refactored and why was this part of the code refactored?
5. Which functions were optimized and what is the difference before and after the optimization?

[eldenmoon]

LGTM

[github-actions]

PR approved by at least one committer and no changes requested.

[github-actions]

PR approved by anyone and no changes requested.

[eldenmoon]

/review

[github-actions]

I found one blocking test-stability issue. The PR's goal is to remove flaky checked output caused by cast(v['c'] as string) on heterogeneous variant values, but the same suite still has an immediately preceding checked query using that exact pattern, so the regression can remain flaky under the same fuzzed session variables.

Critical checkpoints:

• Goal/test proof: The goal is to stabilize variant_hirachinal; current change only partially accomplishes it because one checked output still uses unstable string serialization.
• Scope/focus: The code change is small and focused, but incomplete for the stated failure mode.
• Concurrency/lifecycle/config/compatibility/data writes/FE-BE variable passing: Not applicable; this is a regression-test-only change.
• Parallel paths: A parallel checked query in the same suite remains unfixed.
• Test coverage/results: Existing .out values are consistent with the changed json casts, but coverage remains vulnerable to the same flake in the earlier qt_sql.
• Observability/performance/transactions/persistence: Not applicable.

User focus points: No additional user-provided review focus was supplied.

LGTM