Prefer header delimiters in auto detection

[quinnj]

Summary

• Prefer delimiter candidates that also appear in the header before falling back to aggregate row consistency.
• Add a regression test for semicolon-delimited data that uses comma decimals.
• Update the comma-decimal docs note to reflect the improved header-aware autodetection behavior.

Root Cause

Delimiter autodetection checked candidate delimiters in a comma-first order and only required aggregate consistency across sampled rows. In files with comma decimal markers, those commas could look like a consistent delimiter even when the header clearly used semicolons.

Fixes #1185.

Validation

• julia --project=. test/runtests.jl -> 1477/1477 passing

Co-authored by Codex

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.79%. Comparing base (2f01e19) to head (1fc8b44).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1186      +/-   ##
==========================================
+ Coverage   90.75%   90.79%   +0.03%     
==========================================
  Files           9        9              
  Lines        2326     2335       +9     
==========================================
+ Hits         2111     2120       +9     
  Misses        215      215              

☔ 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.