Jeadie commented on code in PR #10236:
URL: https://github.com/apache/arrow-rs/pull/10236#discussion_r3547987431
##########
arrow/benches/filter_kernels.rs:
##########
@@ -295,6 +298,407 @@ fn add_benchmark(c: &mut Criterion) {
c.bench_function("filter run array low selectivity (kept 1/1024)", |b| {
b.iter(|| bench_built_filter(&sparse_filter, &data_array))
});
+
Review Comment:
Good call — most of those were for my own before/after analysis rather than
the permanent suite. I've trimmed to one benchmark per *distinct code path* the
change introduces (in `f8735127a`):
- `filter list i32` — the generic `Slices` child-copy path
- `filter list utf8` — the streamed byte path (`FilterBytes`)
- `filter list utf8view` — the view path
- `filter fixedsizelist` and `filter map` — the two new direct kernels
Dropped the redundant byte variants (`binary`/`largeutf8`/`largebinary`
share the `utf8` path), the per-child-type gate cases (their correctness is
covered by the randomized `take` cross-check tests, so a bench adds little),
and the value-length / row-count investigation sweeps. That takes it from ~17
groups down to 5.
--
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]
