feat: preserve ordering through to_timestamp_seconds/millis/micros/nanos and to_unixtime

[haohuaijin]

Which issue does this PR close?

• Closes preserve ordering through to_timestamp_seconds/millis/micros/nanos and to_unixtime #26108

Rationale for this change

Queries over ordered data can perform unnecessary sorts after timestamp conversions. Propagating ordering through these conversions lets the optimizer remove redundant sorts, including the to_timestamp_micros → date_bin → to_unixtime chain.

What changes are included in this PR?

• Add ordering propagation for single-argument to_timestamp_seconds, to_timestamp_millis, to_timestamp_micros, and to_timestamp_nanos calls with integer input.
• Add ordering propagation for single-argument to_unixtime calls with integer, Date32, Date64, or timestamp input.
• Preserve input sort direction and NULL placement. Overflow continues to report an error.
• Keep strict-ordering metadata conservative so precision reduction does not incorrectly eliminate sorting by subsequent keys.

What is the testing strategy for this PR?

integer_timestamp_ordering checks all four timestamp functions across integer types, sort directions, NULL placement, and unsupported inputs. ordering_matches_runtime_conversion checks to_unixtime ordering and temporal runtime conversions across date types, timestamp units, and timezones.

Tests in the existing datetime/timestamps.slt verify plans without redundant SortExec, conversion results and boundaries, the full conversion chain, and a case where precision reduction creates ties and a subsequent sorting key still requires SortExec.

Validation passed:

• Both added unit tests.
• cargo test -p datafusion-sqllogictest --test sqllogictests -- datetime/timestamps.slt
• cargo fmt --all -- --check and git diff --check
• cargo clippy --all-targets --all-features -- -D warnings
• uv run ./dev/rust_lint.sh

Are there any user-facing changes?

Yes. Queries using these conversions on ordered input can avoid redundant sorts. Function results and public APIs are unchanged.

[codecov-commenter]

Codecov Report

❌ Patch coverage is 91.25000% with 14 lines in your changes missing coverage. Please review.
✅ Project coverage is 82.73%. Comparing base (a70dd34) to head (1c88c3d).
⚠️ Report is 3 commits behind head on main.

Files with missing lines	Patch %	Lines
datafusion/functions/src/datetime/to_unixtime.rs	89.01%	0 Missing and 10 partials ⚠️
datafusion/functions/src/datetime/to_timestamp.rs	94.20%	0 Missing and 4 partials ⚠️

Additional details and impacted files

@@           Coverage Diff            @@
##             main   #26111    +/-   ##
========================================
  Coverage   82.72%   82.73%            
========================================
  Files        1147     1147            
  Lines      449087   449373   +286     
  Branches   449087   449373   +286     
========================================
+ Hits       371522   371770   +248     
- Misses      54917    54936    +19     
- Partials    22648    22667    +19     

☔ 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.