feat(parquet): add sparse-column writer benchmarks (#9654)

# Which issue does this PR close?

- None, but relates to #9652

# Rationale for this change

Measure sparse and all-null cases in benchmarks.

# What changes are included in this PR?

Add three new benchmark cases to the arrow_writer benchmark suite for
evaluating write performance on sparse and all-null data:

- `primitive_sparse_99pct_null`: a flat primitive column with 99% nulls,
exercising long RLE runs in definition levels.
- `list_primitive_sparse_99pct_null`: a list-of-primitive column with
99% nulls, exercising null batching in the list level builder.
- `primitive_all_null`: a flat primitive column with 100% nulls,
exercising the uniform_levels fast path for entirely-null columns.

# Are these changes tested?

N/A

# Are there any user-facing changes?

None.

Signed-off-by: Hippolyte Barraud <hippolyte.barraud@datadoghq.com>

Author: HippoBaro

parquet/benches/arrow_writer.rs

@@ -391,6 +391,15 @@ fn create_batches() -> Vec<(&'static str, RecordBatch)> {
     let batch = create_list_primitive_bench_batch_non_null(BATCH_SIZE, 0.25, 0.75).unwrap();
     batches.push(("list_primitive_non_null", batch));
 
+    let batch = create_primitive_bench_batch(BATCH_SIZE, 0.99, 0.75).unwrap();
+    batches.push(("primitive_sparse_99pct_null", batch));
+
+    let batch = create_list_primitive_bench_batch(BATCH_SIZE, 0.99, 0.75).unwrap();
+    batches.push(("list_primitive_sparse_99pct_null", batch));
+
+    let batch = create_primitive_bench_batch(BATCH_SIZE, 1.0, 0.75).unwrap();
+    batches.push(("primitive_all_null", batch));
+
     batches
 }