feat(core): Implement null-safe equality operator

[f4t4nt]

Adds support for null-safe equality comparisons (<=>) in Daft, similar to MySQL's implementation. This operator treats NULL values specially:

• NULL <=> NULL returns TRUE
• NULL <=> non-NULL returns FALSE
• non-NULL <=> non-NULL behaves like regular equals

Includes comprehensive test coverage for both expression-based and SQL-based usage.

[codspeed-hq]

CodSpeed Performance Report

Merging #3663 will degrade performances by 40.33%

_{Comparing nishant-null-safe-equality (631412c) with main (feab49a)}

Summary

❌ 1 regressions
✅ 26 untouched benchmarks

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Benchmarks breakdown

	Benchmark	main	nishant-null-safe-equality	Change
❌	test_iter_rows_first_row[100 Small Files]	148.5 ms	248.9 ms	-40.33%

[codecov]

Codecov Report

Attention: Patch coverage is 64.10256% with 154 lines in your changes missing coverage. Please review.

Project coverage is 77.76%. Comparing base (c932ec9) to head (631412c).
Report is 8 commits behind head on main.

Files with missing lines	Patch %	Lines
src/daft-core/src/array/ops/comparison.rs	64.32%	147 Missing ⚠️
src/daft-dsl/src/expr/mod.rs	20.00%	4 Missing ⚠️
src/daft-stats/src/column_stats/comparison.rs	0.00%	3 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3663      +/-   ##
==========================================
- Coverage   78.06%   77.76%   -0.31%     
==========================================
  Files         728      728              
  Lines       89967    90346     +379     
==========================================
+ Hits        70236    70259      +23     
- Misses      19731    20087     +356     

Files with missing lines	Coverage Δ
daft/expressions/expressions.py	93.54% <100.00%> (+0.02%)	⬆️
...onnect/src/translation/expr/unresolved_function.rs	74.46% <100.00%> (+0.27%)	⬆️
src/daft-core/src/series/ops/comparison.rs	95.00% <ø> (ø)
src/daft-dsl/src/python.rs	91.16% <100.00%> (+0.06%)	⬆️
src/daft-functions/src/list/list_fill.rs	96.40% <ø> (-0.06%)	⬇️
src/daft-sql/src/planner.rs	75.18% <100.00%> (+1.29%)	⬆️
src/daft-table/src/lib.rs	83.74% <100.00%> (+0.02%)	⬆️
src/daft-stats/src/column_stats/comparison.rs	72.41% <0.00%> (-1.93%)	⬇️
src/daft-dsl/src/expr/mod.rs	75.97% <20.00%> (-0.35%)	⬇️
src/daft-core/src/array/ops/comparison.rs	73.87% <64.32%> (-2.72%)	⬇️

... and 18 files with indirect coverage changes

[f4t4nt]

The test coverage is low bc I'm not able to test FixedSizeBinary properly (see src/daft-core/src/array/ops/comparison.rs:2117, that print statement is never hit). Also, I am not able to test lit <=> col hence the

match (self.len(), rhs.len()) {
...
(1, r_size) => {

clauses are also not tested but this issue persists for regular equality so I have not seriously tried fixing it yet. All current tests are successful.

[desmondcheongzx]

Looks mostly good to me! Some comments

[desmondcheongzx]

Looks great! Thanks for implementing this!