Function RecordReaderReadAndSkipRecords

cpp/src/parquet/column_reader_benchmark.cc:206–232

Benchmarks ReadRecords and SkipRecords for RecordReader with the following parameters in order: - repetition: 0 for REQUIRED, 1 for OPTIONAL, 2 for REPEATED. - batch_size: sets how many values to read/skip at each call. - levels_per_page: sets how many levels to read/skip in total.

// - batch_size: sets how many values to read/skip at each call.
// - levels_per_page: sets how many levels to read/skip in total.
static void RecordReaderReadAndSkipRecords(::benchmark::State& state) {
  const auto repetition = static_cast<Repetition::type>(state.range(0));
  const auto batch_size = static_cast<int64_t>(state.range(1));
  const auto levels_per_page = static_cast<int>(state.range(2));

  BenchmarkHelper helper(repetition, /*num_pages=*/16, levels_per_page);

  // Vectors to read the values into.
  for (auto _ : state) {
    state.PauseTiming();
    // read_dense_for_nullable should not matter for skip.
    RecordReader* reader = helper.ResetRecordReader(/*read_dense_for_nullable=*/false);
    int64_t records_read = -1;
    int64_t records_skipped = -1;
    state.ResumeTiming();
    while (records_read != 0 && records_skipped != 0) {
      // ReadRecords may buffer some levels which will be skipped by the following
      // SkipRecords.
      DoNotOptimize(records_read = reader->ReadRecords(batch_size));
      DoNotOptimize(records_skipped = reader->SkipRecords(batch_size));
      reader->Reset();
    }
  }

  state.SetBytesProcessed(state.iterations() * helper.total_size());
  state.SetItemsProcessed(state.iterations() * helper.total_levels());
}

BENCHMARK(ColumnReaderSkipInt32)
    ->ArgNames({"Repetition", "BatchSize"})