perf: Use semi-join instead of unique for checking 1:N relationships

[Oliver Borchert (borchero)]

Motivation

The current implementation builds two separate hash tables for no good reason. On some example data, this improves runtime of the filter itself by ~2.5x. End-to-end on the updated benchmark, it yields a ~5-10% runtime speedup.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 100.00%. Comparing base (a523310) to head (3574fd7).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff            @@
##              main      #362   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files           56        56           
  Lines         3475      3475           
=========================================
  Hits          3475      3475           

☔ 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.

[Copilot]

Pull request overview

This PR optimizes the require_relationship_one_to_at_least_one relationship helper by avoiding unnecessary unique() work on the RHS, using a semi-join to filter lhs rows based on key existence in rhs (improving runtime by reducing redundant hashing).

Changes:

• Replace rhs.unique(on) + inner join with how="semi" joins to check key existence more efficiently.
• Preserve the existing optional lhs.unique(..., keep="none") behavior when drop_duplicates=True.