fix: preserve approximate percentile arguments with inline ORDER BY

[haohuaijin]

Which issue does this PR close?

• Closes Inline ORDER BY drops arguments for approximate percentile aggregates #24410.

Rationale for this change

Inline ORDER BY handling incorrectly removes percentile or weight arguments from approximate percentile aggregates.

What changes are included in this PR?

Only remove the first aggregate argument when it matches the inline ordering expression. Add regression coverage for approx_percentile_cont and approx_percentile_cont_with_weight.

Are these changes tested?

Yes. The aggregate.slt regression tests, ./dev/rust_lint.sh, and cargo clippy --all-targets --all-features -- -D warnings all pass.

Are there any user-facing changes?

Valid approximate percentile queries using inline ORDER BY now plan and execute successfully.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 81.23%. Comparing base (bb038a6) to head (014f14d).

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #24411      +/-   ##
==========================================
- Coverage   81.23%   81.23%   -0.01%     
==========================================
  Files        1112     1112              
  Lines      390635   390630       -5     
  Branches   390635   390630       -5     
==========================================
- Hits       317350   317332      -18     
- Misses      54650    54657       +7     
- Partials    18635    18641       +6     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[Jefffrey]

i think this makes sense in keeping existing ordered set aggregates in line with percentile_cont (which doesnt suffer this issue since i think the code was specifically designed for how many arguments it has?)

though @jayzhan211 does have thoughts on how we should handle ORDER BYs with these functions

[haohuaijin]

Thanks @Jefffrey for reviews, i apply suggestion in 32f7528