Add NaNCount statistic

[robert3005]

Other part of #3104

[codspeed-hq]

CodSpeed Performance Report

Merging #3120 will degrade performances by 38.89%

_{Comparing rk/nancount (16a8607) with develop (7bed15c)}

Summary

⚡ 3 improvements
❌ 25 regressions
✅ 783 untouched benchmarks

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

Benchmarks breakdown

	Benchmark	BASE	HEAD	Change
❌	decompress_bitpacking_early_filter[i16, 0.03]	201.8 µs	244.2 µs	-17.39%
❌	decompress_bitpacking_early_filter[i16, 0.04]	219.2 µs	261.6 µs	-16.23%
❌	decompress_bitpacking_early_filter[i16, 0.05]	237.8 µs	279.8 µs	-15.01%
❌	decompress_bitpacking_late_filter[i16, 0.0105]	158.2 µs	200.5 µs	-21.08%
❌	decompress_bitpacking_late_filter[i16, 0.01]	157.3 µs	199.6 µs	-21.21%
❌	decompress_bitpacking_late_filter[i16, 0.02]	177.6 µs	219.9 µs	-19.24%
❌	decompress_bitpacking_late_filter[i16, 0.03]	196.3 µs	238.6 µs	-17.73%
❌	decompress_bitpacking_late_filter[i16, 0.04]	214.4 µs	256.7 µs	-16.47%
❌	decompress_bitpacking_late_filter[i16, 0.05]	232.6 µs	274.9 µs	-15.38%
❌	decompress_bitpacking_late_filter[i32, 0.0105]	212.9 µs	300 µs	-29.04%
❌	decompress_bitpacking_late_filter[i32, 0.01]	194.8 µs	299 µs	-34.86%
❌	decompress_bitpacking_late_filter[i32, 0.02]	233.4 µs	320.4 µs	-27.17%
❌	decompress_bitpacking_late_filter[i32, 0.03]	253.1 µs	340.2 µs	-25.59%
❌	decompress_bitpacking_late_filter[i32, 0.04]	272.6 µs	359.6 µs	-24.21%
❌	decompress_bitpacking_late_filter[i32, 0.05]	292 µs	379.2 µs	-23%
❌	decompress_bitpacking_late_filter[i64, 0.0105]	306.4 µs	499.9 µs	-38.7%
❌	decompress_bitpacking_late_filter[i64, 0.01]	304.8 µs	498.8 µs	-38.89%
❌	decompress_bitpacking_late_filter[i64, 0.02]	328.3 µs	522.4 µs	-37.15%
❌	decompress_bitpacking_late_filter[i64, 0.03]	350.7 µs	544.3 µs	-35.57%
❌	decompress_bitpacking_late_filter[i64, 0.04]	388.8 µs	565.7 µs	-31.27%
...	...	...	...	...

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