Rich-T-kid opened a new issue, #11927:
URL: https://github.com/apache/arrow-rs/issues/11927

   ### Describe the problem
   
   A number of benchmark files reinvent helpers that already exist (or belong) 
in `arrow/src/util/bench_util.rs`. Examples:
   
   - `arrow/benches/take_kernels.rs` and `arrow/benches/primitive_run_take.rs` 
each define their own `create_random_index`.
   - `arrow/benches/cast_kernels.rs` has `build_array` / 
`build_decimal128_array` / `build_decimal256_array` that mirror the shared 
primitive and decimal builders.
   - `arrow/benches/sort_kernel.rs` has `create_f32_array` / 
`create_bool_array` wrappers around `create_primitive_array` and 
`create_boolean_array`.
   - `arrow/benches/partition_kernels.rs` has a `create_array<T>` wrapper over 
`create_primitive_array`.
   - `arrow/benches/coalesce_kernels.rs` has a local 
`create_fixed_size_binary_array` that duplicates the FSB generator.
   - `parquet/benches/parquet_round_trip.rs` has its own copies of 
`create_fsb_array_with_seed` and a string-view-with-seed helper.
   
   This causes drift (seed strategy, null handling, decimal digit layout) and 
bloats each bench with its own imports and rand plumbing.
   
   ### Describe the proposed solution
   
   Promote the generic builders to `arrow/src/util/bench_util.rs` and have the 
bench files call them directly. Keep only genuinely bench-specific fixtures 
local.
   
   ### Additional context
   
   Picked up while cleaning up the benchmark suite.


-- 
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]

Reply via email to