Compress wide decimals when the writer permits the v2 format

[mhk197]

Enable wide decimal compression through DecimalSchemeV2, with the existing DecimalScheme as its predecessor. The compressor selects the permitted version before matching, estimating, or splitting values. Default editions continue to use the frozen scheme; allowing both byte-parts IDs enables splitting and compression of every part.

The v2 scheme requires both IDs because narrow values still serialize under v1. Tests cover edition fallback, exclusions, nullable i128/i256 inputs, and file round trips. Writer tests verify that pre-encoded lower parts survive when v2 is allowed and become canonical decimals when targeting the frozen core edition.

Stack, in merge order: #9770 → #9808 (math helpers) → #9809 (array behavior) → #9810 (versioned serde) → #9759 (compression/writer integration). Each PR targets the branch immediately before it. This is layer 4 of 4.

Supersedes #9276.

Validation:

• 451 tests passed (1 skipped) across vortex-compressor, vortex-btrblocks, vortex-decimal-byte-parts, vortex-file, and vortex-compat, with all features enabled.
• Workspace cargo clippy --all-targets --all-features passed. The build emitted the existing missing-clang-format warning for a generated DuckDB header.
• Doctests passed for the encoding, compressor, file, and compatibility crates.
• cargo +nightly fmt --all and git diff --check passed.
• The complete stack was validated after the readability refactor and downstream helper updates. CUDA runtime tests were not run.

[codspeed-hq]

Merging this PR will degrade performance by 2.84%

⚠️ Unknown Walltime execution environment detected

Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data.

For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

⚡ 3 improved benchmarks
❌ 2 regressed benchmarks
✅ 2182 untouched benchmarks
🆕 45 new benchmarks
⏩ 224 skipped benchmarks¹

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	WallTime	arrow_checked_add_u32_neon[16384]	13.4 µs	20.3 µs	-34.3%
❌	Simulation	decompress[u64, (4000, 1024)]	70.8 µs	86.1 µs	-17.77%
⚡	WallTime	arrow_checked_add_u32_avx2[16384]	21.3 µs	17.6 µs	+20.79%
⚡	WallTime	filtered_owned_i64_avx512[OneNullInEight]	26.3 µs	22.4 µs	+17.47%
⚡	Simulation	chunked_bool_canonical_into[(1000, 10)]	30.6 µs	27.1 µs	+12.95%
🆕	Simulation	assemble[(I128, 1024)]	N/A	47.3 µs	N/A
🆕	Simulation	assemble[(I128, 65536)]	N/A	2 ms	N/A
🆕	Simulation	assemble[(I128, 8192)]	N/A	265.6 µs	N/A
🆕	Simulation	assemble[(I256, 1024)]	N/A	109 µs	N/A
🆕	Simulation	assemble[(I256, 65536)]	N/A	6 ms	N/A
🆕	Simulation	assemble[(I256, 8192)]	N/A	764.5 µs	N/A
🆕	Simulation	assemble[(I64, 1024)]	N/A	9.3 µs	N/A
🆕	Simulation	assemble[(I64, 65536)]	N/A	7.4 µs	N/A
🆕	Simulation	assemble[(I64, 8192)]	N/A	7.4 µs	N/A
🆕	Simulation	all_null[(I128, 1024)]	N/A	62.7 µs	N/A
🆕	Simulation	all_null[(I128, 65536)]	N/A	2.4 ms	N/A
🆕	Simulation	all_null[(I128, 8192)]	N/A	328.9 µs	N/A
🆕	Simulation	all_null[(I256, 1024)]	N/A	108.3 µs	N/A
🆕	Simulation	all_null[(I256, 65536)]	N/A	4.8 ms	N/A
🆕	Simulation	all_null[(I256, 8192)]	N/A	639.2 µs	N/A
...	...	...	...	...	...

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

Tip

Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing mk/dbp-v2 (52a1cb5) with develop (e3b8eb2)²}

Footnotes

1. 224 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. ↩

2. No successful run was found on mk/compressor-serialized-ids (b28f0b6) during the generation of this report, so develop (e3b8eb2) was used instead as the comparison base. There might be some changes unrelated to this pull request in this report. ↩