| 47 | const char PARQUET_FILENAME[] = "parquet_cpp_example2.parquet"; |
| 48 | |
| 49 | int main(int argc, char** argv) { |
| 50 | /********************************************************************************** |
| 51 | PARQUET WRITER EXAMPLE |
| 52 | **********************************************************************************/ |
| 53 | // parquet::REQUIRED fields do not need definition and repetition level values |
| 54 | // parquet::OPTIONAL fields require only definition level values |
| 55 | // parquet::REPEATED fields require both definition and repetition level values |
| 56 | try { |
| 57 | // Create a local file output stream instance. |
| 58 | using FileClass = ::arrow::io::FileOutputStream; |
| 59 | std::shared_ptr<FileClass> out_file; |
| 60 | PARQUET_ASSIGN_OR_THROW(out_file, FileClass::Open(PARQUET_FILENAME)); |
| 61 | |
| 62 | // Setup the parquet schema |
| 63 | std::shared_ptr<GroupNode> schema = SetupSchema(); |
| 64 | |
| 65 | // Add writer properties |
| 66 | parquet::WriterProperties::Builder builder; |
| 67 | builder.compression(parquet::Compression::SNAPPY); |
| 68 | std::shared_ptr<parquet::WriterProperties> props = builder.build(); |
| 69 | |
| 70 | // Create a ParquetFileWriter instance |
| 71 | std::shared_ptr<parquet::ParquetFileWriter> file_writer = |
| 72 | parquet::ParquetFileWriter::Open(out_file, schema, props); |
| 73 | |
| 74 | // Append a BufferedRowGroup to keep the RowGroup open until a certain size |
| 75 | parquet::RowGroupWriter* rg_writer = file_writer->AppendBufferedRowGroup(); |
| 76 | |
| 77 | int num_columns = file_writer->num_columns(); |
| 78 | std::vector<int64_t> buffered_values_estimate(num_columns, 0); |
| 79 | for (int i = 0; i < NUM_ROWS; i++) { |
| 80 | int64_t estimated_bytes = 0; |
| 81 | // Get the estimated size of the values that are not written to a page yet |
| 82 | for (int n = 0; n < num_columns; n++) { |
| 83 | estimated_bytes += buffered_values_estimate[n]; |
| 84 | } |
| 85 | |
| 86 | // We need to consider the compressed pages |
| 87 | // as well as the values that are not compressed yet |
| 88 | if ((rg_writer->total_bytes_written() + rg_writer->total_compressed_bytes() + |
| 89 | estimated_bytes) > ROW_GROUP_SIZE) { |
| 90 | rg_writer->Close(); |
| 91 | std::fill(buffered_values_estimate.begin(), buffered_values_estimate.end(), 0); |
| 92 | rg_writer = file_writer->AppendBufferedRowGroup(); |
| 93 | } |
| 94 | |
| 95 | int col_id = 0; |
| 96 | // Write the Bool column |
| 97 | parquet::BoolWriter* bool_writer = |
| 98 | static_cast<parquet::BoolWriter*>(rg_writer->column(col_id)); |
| 99 | bool bool_value = ((i % 2) == 0) ? true : false; |
| 100 | bool_writer->WriteBatch(1, nullptr, nullptr, &bool_value); |
| 101 | buffered_values_estimate[col_id] = bool_writer->estimated_buffered_value_bytes(); |
| 102 | |
| 103 | // Write the Int32 column |
| 104 | col_id++; |
| 105 | parquet::Int32Writer* int32_writer = |
| 106 | static_cast<parquet::Int32Writer*>(rg_writer->column(col_id)); |
nothing calls this directly
no test coverage detected