[fix](regression-test) fix stale nested-type assertions in two_level/three_level_nestedtypes_with_s3data #64706

[hello-stephen]

Cherry-picked from #64706 (master).

Auto cherry-pick failed with a conflict: branch-4.0's copy of these two test files still uses struct_element() for struct field access, while master renamed it to element_at() in an unrelated earlier change. Resolved by taking master's element_at()-based content as-is — this exact content has been run end-to-end on a real branch-4.0 cluster with -forceGenOut and both suites pass (All suites success.), confirming branch-4.0's engine supports element_at() on struct the same way master does.

Summary (from #64706)

• GROUP BY/ORDER BY on plain array/map/struct columns no longer throws, so the old errCode = 2 ... must use with specific function ... group by or order by exception assertions in two_level_nestedtypes_with_s3data.groovy / three_level_nestedtypes_with_s3data.groovy are stale; removed.
• element_at() on a struct returns an untyped result that GROUP BY/ORDER BY and functions like map_keys()/size() can't accept directly — the most-nested-column checks now CAST it to the concrete type at each point of use.
• Struct element_at() out-of-bound access (-1/0/1000) no longer throws the specified field index out of bound either; those stale assertions are removed too.
• Refreshes CPUNetworkUtilization_order.out (brown_p2) to match current output.

Test plan

• Ran both suites end-to-end with -forceGenOut on a real branch-4.0 cluster — All suites success.

Co-Authored-By: Claude Sonnet 4.6 noreply@anthropic.com

https://claude.ai/code/session_01EoF4rikuEzGPKT2ML2WtnM

[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?

[github-actions]

Possible file(s) that should be tracked in LFS detected: 🚨

The following file(s) exceeds the file size limit: 1048576 bytes, as set in the .yml configuration files:

• regression-test/data/datatype_p2/complex_types/three_level_nestedtypes_with_s3data.out
• regression-test/data/datatype_p2/complex_types/two_level_nestedtypes_with_s3data.out

Consider using git-lfs to manage large files.

[hello-stephen]

skip buildall