sdf-jkl opened a new pull request, #11773: URL: https://github.com/apache/arrow-rs/pull/11773
# Which issue does this PR close? - Part of #10906 - Supports benchmarking #11402 # Rationale for this change `variant_kernels.rs` benchmarks `json_to_variant` and `variant_get`, but has no coverage for `variant_to_json`. Add benchmarks to track serialization performance and evaluate changes such as avoiding temporary string allocations in #11402. # What changes are included in this PR? Add seven `variant_to_json` benchmarks, each processing 8,192 rows: - Integer scalars and the same values in shredded form. - Small objects containing two strings and an integer. - The same small objects with dictionary and run-end-encoded metadata. - Wide objects using the existing seeded generator with 200 entries per object and mixed scalar values. - Numeric lists with 0–100 elements. Reuse existing fixture generators and the shared row-count constant. Fixture construction and metadata encoding happen outside the timed loops. # Are these changes tested? All seven cases passed a smoke run and completed a release-mode Criterion run with 100 samples each. Compilation, formatting, and spelling checks pass. Clippy with the installed stable toolchain encounters existing lint errors in `arrow-data`. # Are there any significant user-facing changes? No. AI assistance: OpenAI Codex helped implement and refine the benchmarks, run local validation, and draft this description. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
