Search in sources :

Example 6 with GroupWriteSupport

use of org.apache.parquet.hadoop.example.GroupWriteSupport in project parquet-mr by apache.

the class ParquetFileTest method createTestParquetFile.

private void createTestParquetFile() throws IOException {
    File file = parquetFile();
    Path fsPath = new Path(file.getPath());
    Configuration conf = new Configuration();
    MessageType schema = createSchema();
    SimpleGroupFactory fact = new SimpleGroupFactory(schema);
    GroupWriteSupport.setSchema(schema, conf);
    try (ParquetWriter<Group> writer = new ParquetWriter<>(fsPath, new GroupWriteSupport(), CompressionCodecName.UNCOMPRESSED, 1024, 1024, 512, true, false, ParquetProperties.WriterVersion.PARQUET_2_0, conf)) {
        for (int i = 0; i < 10; i++) {
            final byte[] bytes = new byte[12];
            ThreadLocalRandom.current().nextBytes(bytes);
            writer.write(fact.newGroup().append(INT32_FIELD, 32 + i).append(INT64_FIELD, 64L + i).append(FLOAT_FIELD, 1.0f + i).append(DOUBLE_FIELD, 2.0d + i).append(BINARY_FIELD, Binary.fromString(COLORS[i % COLORS.length])).append(FIXED_LEN_BYTE_ARRAY_FIELD, Binary.fromConstantByteArray(bytes)));
        }
    }
}
Also used : Path(org.apache.hadoop.fs.Path) Group(org.apache.parquet.example.data.Group) GroupWriteSupport(org.apache.parquet.hadoop.example.GroupWriteSupport) Configuration(org.apache.hadoop.conf.Configuration) ParquetWriter(org.apache.parquet.hadoop.ParquetWriter) SimpleGroupFactory(org.apache.parquet.example.data.simple.SimpleGroupFactory) File(java.io.File) MessageType(org.apache.parquet.schema.MessageType)

Example 7 with GroupWriteSupport

use of org.apache.parquet.hadoop.example.GroupWriteSupport in project parquet-mr by apache.

the class TestParquetWriterNewPage method test.

@Test
public void test() throws Exception {
    Configuration conf = new Configuration();
    Path root = new Path("target/tests/TestParquetWriter/");
    FileSystem fs = root.getFileSystem(conf);
    if (fs.exists(root)) {
        fs.delete(root, true);
    }
    fs.mkdirs(root);
    MessageType schema = parseMessageType("message test { " + "required binary binary_field; " + "required int32 int32_field; " + "required int64 int64_field; " + "required boolean boolean_field; " + "required float float_field; " + "required double double_field; " + "required fixed_len_byte_array(3) flba_field; " + "required int96 int96_field; " + "optional binary null_field; " + "} ");
    GroupWriteSupport.setSchema(schema, conf);
    SimpleGroupFactory f = new SimpleGroupFactory(schema);
    Map<String, Encoding> expected = new HashMap<String, Encoding>();
    expected.put("10-" + PARQUET_1_0, PLAIN_DICTIONARY);
    expected.put("1000-" + PARQUET_1_0, PLAIN);
    expected.put("10-" + PARQUET_2_0, RLE_DICTIONARY);
    expected.put("1000-" + PARQUET_2_0, DELTA_BYTE_ARRAY);
    for (int modulo : asList(10, 1000)) {
        for (WriterVersion version : WriterVersion.values()) {
            Path file = new Path(root, version.name() + "_" + modulo);
            ParquetWriter<Group> writer = new ParquetWriter<Group>(file, new GroupWriteSupport(), UNCOMPRESSED, 1024, 1024, 512, true, false, version, conf);
            for (int i = 0; i < 1000; i++) {
                writer.write(f.newGroup().append("binary_field", "test" + (i % modulo)).append("int32_field", 32).append("int64_field", 64l).append("boolean_field", true).append("float_field", 1.0f).append("double_field", 2.0d).append("flba_field", "foo").append("int96_field", Binary.fromConstantByteArray(new byte[12])));
            }
            writer.close();
            ParquetReader<Group> reader = ParquetReader.builder(new GroupReadSupport(), file).withConf(conf).build();
            for (int i = 0; i < 1000; i++) {
                Group group = reader.read();
                assertEquals("test" + (i % modulo), group.getBinary("binary_field", 0).toStringUsingUTF8());
                assertEquals(32, group.getInteger("int32_field", 0));
                assertEquals(64l, group.getLong("int64_field", 0));
                assertEquals(true, group.getBoolean("boolean_field", 0));
                assertEquals(1.0f, group.getFloat("float_field", 0), 0.001);
                assertEquals(2.0d, group.getDouble("double_field", 0), 0.001);
                assertEquals("foo", group.getBinary("flba_field", 0).toStringUsingUTF8());
                assertEquals(Binary.fromConstantByteArray(new byte[12]), group.getInt96("int96_field", 0));
                assertEquals(0, group.getFieldRepetitionCount("null_field"));
            }
            reader.close();
            ParquetMetadata footer = readFooter(conf, file, NO_FILTER);
            for (BlockMetaData blockMetaData : footer.getBlocks()) {
                for (ColumnChunkMetaData column : blockMetaData.getColumns()) {
                    if (column.getPath().toDotString().equals("binary_field")) {
                        String key = modulo + "-" + version;
                        Encoding expectedEncoding = expected.get(key);
                        assertTrue(key + ":" + column.getEncodings() + " should contain " + expectedEncoding, column.getEncodings().contains(expectedEncoding));
                    }
                }
            }
        }
    }
}
Also used : Path(org.apache.hadoop.fs.Path) Group(org.apache.parquet.example.data.Group) BlockMetaData(org.apache.parquet.hadoop.metadata.BlockMetaData) GroupReadSupport(org.apache.parquet.hadoop.example.GroupReadSupport) Configuration(org.apache.hadoop.conf.Configuration) HashMap(java.util.HashMap) ParquetMetadata(org.apache.parquet.hadoop.metadata.ParquetMetadata) ColumnChunkMetaData(org.apache.parquet.hadoop.metadata.ColumnChunkMetaData) SimpleGroupFactory(org.apache.parquet.example.data.simple.SimpleGroupFactory) Encoding(org.apache.parquet.column.Encoding) WriterVersion(org.apache.parquet.column.ParquetProperties.WriterVersion) GroupWriteSupport(org.apache.parquet.hadoop.example.GroupWriteSupport) FileSystem(org.apache.hadoop.fs.FileSystem) MessageTypeParser.parseMessageType(org.apache.parquet.schema.MessageTypeParser.parseMessageType) MessageType(org.apache.parquet.schema.MessageType) Test(org.junit.Test)

Example 8 with GroupWriteSupport

use of org.apache.parquet.hadoop.example.GroupWriteSupport in project parquet-mr by apache.

the class DataGenerator method generateData.

public void generateData(Path outFile, Configuration configuration, ParquetProperties.WriterVersion version, int blockSize, int pageSize, int fixedLenByteArraySize, CompressionCodecName codec, int nRows) throws IOException {
    if (exists(configuration, outFile)) {
        System.out.println("File already exists " + outFile);
        return;
    }
    System.out.println("Generating data @ " + outFile);
    MessageType schema = parseMessageType("message test { " + "required binary binary_field; " + "required int32 int32_field; " + "required int64 int64_field; " + "required boolean boolean_field; " + "required float float_field; " + "required double double_field; " + "required fixed_len_byte_array(" + fixedLenByteArraySize + ") flba_field; " + "required int96 int96_field; " + "} ");
    GroupWriteSupport.setSchema(schema, configuration);
    SimpleGroupFactory f = new SimpleGroupFactory(schema);
    ParquetWriter<Group> writer = new ParquetWriter<Group>(outFile, new GroupWriteSupport(), codec, blockSize, pageSize, DICT_PAGE_SIZE, true, false, version, configuration);
    // generate some data for the fixed len byte array field
    char[] chars = new char[fixedLenByteArraySize];
    Arrays.fill(chars, '*');
    for (int i = 0; i < nRows; i++) {
        writer.write(f.newGroup().append("binary_field", randomUUID().toString()).append("int32_field", i).append("int64_field", 64l).append("boolean_field", true).append("float_field", 1.0f).append("double_field", 2.0d).append("flba_field", new String(chars)).append("int96_field", Binary.fromConstantByteArray(new byte[12])));
    }
    writer.close();
}
Also used : Group(org.apache.parquet.example.data.Group) GroupWriteSupport(org.apache.parquet.hadoop.example.GroupWriteSupport) ParquetWriter(org.apache.parquet.hadoop.ParquetWriter) SimpleGroupFactory(org.apache.parquet.example.data.simple.SimpleGroupFactory) MessageType(org.apache.parquet.schema.MessageType) MessageTypeParser.parseMessageType(org.apache.parquet.schema.MessageTypeParser.parseMessageType)

Example 9 with GroupWriteSupport

use of org.apache.parquet.hadoop.example.GroupWriteSupport in project parquet-mr by apache.

the class PhoneBookWriter method writeToFile.

public static void writeToFile(File f, List<User> users) throws IOException {
    Configuration conf = new Configuration();
    GroupWriteSupport.setSchema(schema, conf);
    ParquetWriter<Group> writer = new ParquetWriter<Group>(new Path(f.getAbsolutePath()), conf, new GroupWriteSupport());
    for (User u : users) {
        writer.write(groupFromUser(u));
    }
    writer.close();
}
Also used : Path(org.apache.hadoop.fs.Path) Group(org.apache.parquet.example.data.Group) SimpleGroup(org.apache.parquet.example.data.simple.SimpleGroup) GroupWriteSupport(org.apache.parquet.hadoop.example.GroupWriteSupport) Configuration(org.apache.hadoop.conf.Configuration) ParquetWriter(org.apache.parquet.hadoop.ParquetWriter)

Example 10 with GroupWriteSupport

use of org.apache.parquet.hadoop.example.GroupWriteSupport in project zeppelin by apache.

the class SqlInterpreterTest method createParquetFile.

public File createParquetFile(int[] values, ParquetProperties.WriterVersion version) throws IOException {
    File file = File.createTempFile("zeppelin-flink-input", ".par");
    file.delete();
    Path path = new Path(file.getAbsolutePath());
    Configuration conf = new Configuration();
    MessageType schema = MessageTypeParser.parseMessageType("message test { " + "required int32 int32_field; " + "} ");
    GroupWriteSupport.setSchema(schema, conf);
    SimpleGroupFactory f = new SimpleGroupFactory(schema);
    ParquetWriter<Group> writer = new ParquetWriter<Group>(path, new GroupWriteSupport(), CompressionCodecName.UNCOMPRESSED, 1024, 1024, 512, true, false, version, conf);
    for (int i = 0; i < values.length; i++) {
        writer.write(f.newGroup().append("int32_field", values[i]));
    }
    writer.close();
    return file;
}
Also used : Path(org.apache.hadoop.fs.Path) Group(org.apache.parquet.example.data.Group) InterpreterGroup(org.apache.zeppelin.interpreter.InterpreterGroup) GroupWriteSupport(org.apache.parquet.hadoop.example.GroupWriteSupport) Configuration(org.apache.hadoop.conf.Configuration) ParquetWriter(org.apache.parquet.hadoop.ParquetWriter) SimpleGroupFactory(org.apache.parquet.example.data.simple.SimpleGroupFactory) OrcFile(org.apache.orc.OrcFile) File(java.io.File) MessageType(org.apache.parquet.schema.MessageType)

Aggregations

GroupWriteSupport (org.apache.parquet.hadoop.example.GroupWriteSupport)12 Group (org.apache.parquet.example.data.Group)11 Configuration (org.apache.hadoop.conf.Configuration)10 ParquetWriter (org.apache.parquet.hadoop.ParquetWriter)10 MessageType (org.apache.parquet.schema.MessageType)9 Path (org.apache.hadoop.fs.Path)7 SimpleGroupFactory (org.apache.parquet.example.data.simple.SimpleGroupFactory)7 File (java.io.File)6 MessageTypeParser.parseMessageType (org.apache.parquet.schema.MessageTypeParser.parseMessageType)4 AvroParquetWriter (org.apache.parquet.avro.AvroParquetWriter)3 Test (org.junit.Test)3 ParquetProperties (org.apache.parquet.column.ParquetProperties)2 SimpleGroup (org.apache.parquet.example.data.simple.SimpleGroup)2 CompressionCodecName (org.apache.parquet.hadoop.metadata.CompressionCodecName)2 HashMap (java.util.HashMap)1 OperatorTest (org.apache.drill.categories.OperatorTest)1 UnlikelyTest (org.apache.drill.categories.UnlikelyTest)1 ParquetWriterConfiguration (org.apache.gobblin.parquet.writer.ParquetWriterConfiguration)1 ParquetWriterShim (org.apache.gobblin.parquet.writer.ParquetWriterShim)1 FileSystem (org.apache.hadoop.fs.FileSystem)1