azecevic000 commented on code in PR #12140:
URL: https://github.com/apache/arrow-rs/pull/12140#discussion_r4581211708


##########
parquet/benches/arrow_writer.rs:
##########
@@ -765,6 +765,17 @@ fn create_string_bench_batch_non_null(
     )?)
 }
 
+fn create_empty_string_bench_batch(size: usize) -> Result<RecordBatch> {
+    let schema = Schema::new(vec![Field::new("_1", DataType::Utf8, false)]);
+    // Like `concat`, `take` and `interleave`, `from_iter_values` gives an
+    // all-empty array a values buffer without an allocation.

Review Comment:
   @Rich-T-kid Sorry for the slightly delayed response. I would say this is a 
grey area regarding meeting the condition of the proper benchmark. Strictly 
speaking, the main value is in guarding against future regressions, which 
historically has happened in this case between glibc version 2.34 and 2.35. I 
guess the value is also in comparing pre/post-merge benchmark results directly 
in CI which is probably one of the main driving factor behind the 
recommendation in `CONTRIBUTING.md`.
   
   In this case its dependent on the underlying system as a whole. It won't 
show any diff on ARM or non AVX-512 capable 86. Considering that it's also 
relatively cheap test I would keep it as some form of regression test, but I 
also understand if you don't see any value in it.



-- 
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