feat[array]: move over fill_null to scalar_fn array making it lazy #6406

Performance Regression: -54.87%

⚡ 8 improved benchmarks
❌ 14 regressed benchmarks
✅ 1113 untouched benchmarks
⏩ 1268 skipped benchmarks¹

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

	Mode	Benchmark	`BASE`	`HEAD`	Efficiency
❌	Simulation	`chunked_opt_bool_canonical_into[(1000, 10)]`	93.5 µs	148.8 µs	-37.18%
❌	Simulation	`chunked_opt_bool_into_canonical[(100, 100)]`	449.2 µs	890.4 µs	-49.55%
❌	Simulation	`chunked_opt_bool_canonical_into[(100, 100)]`	364.9 µs	808.6 µs	-54.87%
❌	Simulation	`chunked_opt_bool_into_canonical[(1000, 10)]`	114 µs	167 µs	-31.73%
❌	Simulation	`chunked_varbinview_opt_canonical_into[(10, 1000)]`	7.1 ms	11 ms	-35.51%
❌	Simulation	`chunked_varbinview_opt_into_canonical[(10, 1000)]`	7.8 ms	11.7 ms	-33.56%
❌	Simulation	`bench_many_nulls[0.01]`	60.9 µs	81.8 µs	-25.51%
❌	Simulation	`encode_varbin[(1000, 4)]`	205.3 µs	242.6 µs	-15.37%
❌	Simulation	`encode_varbin[(1000, 32)]`	216.8 µs	240.9 µs	-10%
❌	Simulation	`encode_varbin[(1000, 2)]`	203.4 µs	227.1 µs	-10.46%
❌	Simulation	`encode_varbin[(1000, 8)]`	206.7 µs	230.5 µs	-10.33%
❌	Simulation	`take_fsl_nullable_random[16, 100]`	95.4 µs	108.4 µs	-11.98%
⚡	Simulation	`take_map[(0.1, 1.0)]`	4.7 ms	3.5 ms	+34.37%
⚡	Simulation	`take_map[(0.1, 0.1)]`	1,118.8 µs	917.8 µs	+21.9%
⚡	Simulation	`take_map[(0.05, 1.0)]`	4.1 ms	3 ms	+36.26%
⚡	Simulation	`take_map[(0.1, 0.5)]`	2.8 ms	2.1 ms	+35.13%
⚡	Simulation	`take_map[(0.05, 0.1)]`	792 µs	593.8 µs	+33.38%
⚡	Simulation	`patched_take_200k_dispersed`	5.6 ms	4.7 ms	+19.86%
⚡	Simulation	`patched_take_200k_first_chunk_only`	5.4 ms	4.8 ms	+11.95%
❌	Simulation	`compress_alp[f32, (1000, 0.0, 0.95)]`	126.5 µs	141.8 µs	-10.83%
...	...	...	...	...	...

ℹ️ Only the first 20 benchmarks are displayed. Go to the app to view all benchmarks.

_{Comparing ji/fill-null-inline (ae70b43) with develop (aa55054)}

1268 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩