Function BM_ReadIndividualRowGroups

cpp/src/parquet/arrow/reader_writer_benchmark.cc:726–758

//

static void BM_ReadIndividualRowGroups(::benchmark::State& state) {
  std::vector<int64_t> values(BENCHMARK_SIZE, 128);
  std::shared_ptr<Table> table = TableFromVector<Int64Type>(values, true);
  auto output = CreateOutputStream();
  // This writes 10 RowGroups
  EXIT_NOT_OK(
      WriteTable(*table, ::arrow::default_memory_pool(), output, BENCHMARK_SIZE / 10));

  PARQUET_ASSIGN_OR_THROW(auto buffer, output->Finish());

  while (state.KeepRunning()) {
    auto reader =
        ParquetFileReader::Open(std::make_shared<::arrow::io::BufferReader>(buffer));
    auto arrow_reader_result =
        FileReader::Make(::arrow::default_memory_pool(), std::move(reader));
    EXIT_NOT_OK(arrow_reader_result.status());
    auto arrow_reader = std::move(*arrow_reader_result);

    std::vector<std::shared_ptr<Table>> tables;
    for (int i = 0; i < arrow_reader->num_row_groups(); i++) {
      // Only read the even numbered RowGroups
      if ((i % 2) == 0) {
        std::shared_ptr<Table> table;
        EXIT_NOT_OK(arrow_reader->RowGroup(i)->ReadTable(&table));
        tables.push_back(table);
      }
    }

    std::shared_ptr<Table> final_table;
    PARQUET_ASSIGN_OR_THROW(final_table, ConcatenateTables(tables));
  }
  SetBytesProcessed<Int64Type>(state);
}

BENCHMARK(BM_ReadIndividualRowGroups);