Java Code Examples for org.apache.parquet.hadoop.util.HadoopInputFile

The following examples show how to use org.apache.parquet.hadoop.util.HadoopInputFile. These examples are extracted from open source projects. You can vote up the ones you like or vote down the ones you don't like, and go to the original project or source file by following the links above each example. You may check out the related API usage on the sidebar.
Example 1
Source Project: parquet-mr   Source File: TransCompressionCommand.java    License: Apache License 2.0 7 votes vote down vote up
@Override
@SuppressWarnings("unchecked")
public int run() throws IOException {
  Preconditions.checkArgument(input != null && output != null,
    "Both input and output parquet file paths are required.");

  Preconditions.checkArgument(codec != null,
    "The codec cannot be null");

  Path inPath = new Path(input);
  Path outPath = new Path(output);
  CompressionCodecName codecName = CompressionCodecName.valueOf(codec);

  ParquetMetadata metaData = ParquetFileReader.readFooter(getConf(), inPath, NO_FILTER);
  MessageType schema = metaData.getFileMetaData().getSchema();
  ParquetFileWriter writer = new ParquetFileWriter(getConf(), schema, outPath, ParquetFileWriter.Mode.CREATE);
  writer.start();

  try (TransParquetFileReader reader = new TransParquetFileReader(HadoopInputFile.fromPath(inPath, getConf()), HadoopReadOptions.builder(getConf()).build())) {
    compressionConverter.processBlocks(reader, writer, metaData, schema, metaData.getFileMetaData().getCreatedBy(), codecName);
  } finally {
    writer.end(metaData.getFileMetaData().getKeyValueMetaData());
  }
  return 0;
}
 
Example 2
Source Project: parquet-mr   Source File: TransCompressionCommand.java    License: Apache License 2.0 6 votes vote down vote up
@Override
public void execute(CommandLine options) throws Exception {
  super.execute(options);
  List<String> args = options.getArgList();
  Path inPath = new Path(args.get(0));
  Path outPath = new Path(args.get(1));
  CompressionCodecName codecName = CompressionCodecName.valueOf(args.get(2));

  ParquetMetadata metaData = ParquetFileReader.readFooter(conf, inPath, NO_FILTER);
  MessageType schema = metaData.getFileMetaData().getSchema();
  ParquetFileWriter writer = new ParquetFileWriter(conf, schema, outPath, ParquetFileWriter.Mode.CREATE);
  writer.start();

  try (TransParquetFileReader reader = new TransParquetFileReader(HadoopInputFile.fromPath(inPath, conf), HadoopReadOptions.builder(conf).build())) {
    compressionConverter.processBlocks(reader, writer, metaData, schema, metaData.getFileMetaData().getCreatedBy(), codecName);
  } finally {
    writer.end(metaData.getFileMetaData().getKeyValueMetaData());
  }
}
 
Example 3
Source Project: parquet-mr   Source File: ParquetReader.java    License: Apache License 2.0 6 votes vote down vote up
public ParquetReader<T> build() throws IOException {
  ParquetReadOptions options = optionsBuilder.build();

  if (path != null) {
    FileSystem fs = path.getFileSystem(conf);
    FileStatus stat = fs.getFileStatus(path);

    if (stat.isFile()) {
      return new ParquetReader<>(
          Collections.singletonList((InputFile) HadoopInputFile.fromStatus(stat, conf)),
          options,
          getReadSupport());

    } else {
      List<InputFile> files = new ArrayList<>();
      for (FileStatus fileStatus : fs.listStatus(path, HiddenFileFilter.INSTANCE)) {
        files.add(HadoopInputFile.fromStatus(fileStatus, conf));
      }
      return new ParquetReader<T>(files, options, getReadSupport());
    }

  } else {
    return new ParquetReader<>(Collections.singletonList(file), options, getReadSupport());
  }
}
 
Example 4
Source Project: parquet-mr   Source File: ParquetFileReader.java    License: Apache License 2.0 6 votes vote down vote up
/**
 * @param configuration the Hadoop conf
 * @param fileMetaData fileMetaData for parquet file
 * @param filePath Path for the parquet file
 * @param blocks the blocks to read
 * @param columns the columns to read (their path)
 * @throws IOException if the file can not be opened
 * @deprecated will be removed in 2.0.0.
 */
@Deprecated
public ParquetFileReader(
    Configuration configuration, FileMetaData fileMetaData,
    Path filePath, List<BlockMetaData> blocks, List<ColumnDescriptor> columns) throws IOException {
  this.converter = new ParquetMetadataConverter(configuration);
  this.file = HadoopInputFile.fromPath(filePath, configuration);
  this.fileMetaData = fileMetaData;
  this.f = file.newStream();
  this.options = HadoopReadOptions.builder(configuration).build();
  this.blocks = filterRowGroups(blocks);
  this.blockIndexStores = listWithNulls(this.blocks.size());
  this.blockRowRanges = listWithNulls(this.blocks.size());
  for (ColumnDescriptor col : columns) {
    paths.put(ColumnPath.get(col.getPath()), col);
  }
  this.crc = options.usePageChecksumVerification() ? new CRC32() : null;
}
 
Example 5
Source Project: parquet-mr   Source File: ParquetFileReader.java    License: Apache License 2.0 6 votes vote down vote up
/**
 * @param conf the Hadoop Configuration
 * @param file Path to a parquet file
 * @param footer a {@link ParquetMetadata} footer already read from the file
 * @throws IOException if the file can not be opened
 * @deprecated will be removed in 2.0.0.
 */
@Deprecated
public ParquetFileReader(Configuration conf, Path file, ParquetMetadata footer) throws IOException {
  this.converter = new ParquetMetadataConverter(conf);
  this.file = HadoopInputFile.fromPath(file, conf);
  this.f = this.file.newStream();
  this.options = HadoopReadOptions.builder(conf).build();
  this.footer = footer;
  this.fileMetaData = footer.getFileMetaData();
  this.blocks = filterRowGroups(footer.getBlocks());
  this.blockIndexStores = listWithNulls(this.blocks.size());
  this.blockRowRanges = listWithNulls(this.blocks.size());
  for (ColumnDescriptor col : footer.getFileMetaData().getSchema().getColumns()) {
    paths.put(ColumnPath.get(col.getPath()), col);
  }
  this.crc = options.usePageChecksumVerification() ? new CRC32() : null;
}
 
Example 6
Source Project: parquet-mr   Source File: TestColumnIndexes.java    License: Apache License 2.0 6 votes vote down vote up
@Test
public void testColumnIndexes() throws IOException {
  LOGGER.info("Starting test with context: {}", context);

  Path file = null;
  try {
    file = context.write(new Path(tmp.getRoot().getAbsolutePath()));
    LOGGER.info("Parquet file \"{}\" is successfully created for the context: {}", file, context);

    List<ContractViolation> violations = ColumnIndexValidator
        .checkContractViolations(HadoopInputFile.fromPath(file, new Configuration()));
    assertTrue(violations.toString(), violations.isEmpty());
  } finally {
    if (file != null) {
      file.getFileSystem(new Configuration()).delete(file, false);
    }
  }
}
 
Example 7
Source Project: parquet-mr   Source File: AvroTestUtil.java    License: Apache License 2.0 6 votes vote down vote up
public static <D> List<D> read(Configuration conf, GenericData model, Schema schema, File file) throws IOException {
  List<D> data = new ArrayList<D>();
  AvroReadSupport.setRequestedProjection(conf, schema);
  AvroReadSupport.setAvroReadSchema(conf, schema);

  try (ParquetReader<D> fileReader = AvroParquetReader
    .<D>builder(HadoopInputFile.fromPath(new Path(file.toString()), conf))
    .withDataModel(model) // reflect disables compatibility
    .build()) {
    D datum;
    while ((datum = fileReader.read()) != null) {
      data.add(datum);
    }
  }

  return data;
}
 
Example 8
private GenericRecord readActualRecord(String parquetPath) throws IOException
{
    try (ParquetReader<GenericRecord> reader = AvroParquetReader
            .<GenericRecord>builder(
                    HadoopInputFile.fromPath(new Path(new File(parquetPath).toURI()), new Configuration()))
            .build())
    {
        return reader.read();
    }
}
 
Example 9
private static <T> List<T> readParquetFile(File file, GenericData dataModel) throws IOException {
	InputFile inFile = HadoopInputFile.fromPath(new org.apache.hadoop.fs.Path(file.toURI()), new Configuration());

	ArrayList<T> results = new ArrayList<>();
	try (ParquetReader<T> reader = AvroParquetReader.<T>builder(inFile).withDataModel(dataModel).build()) {
		T next;
		while ((next = reader.read()) != null) {
			results.add(next);
		}
	}

	return results;
}
 
Example 10
Source Project: flink   Source File: ParquetInputFormat.java    License: Apache License 2.0 5 votes vote down vote up
@Override
public void open(FileInputSplit split) throws IOException {
	// reset the flag when open a new split
	this.skipThisSplit = false;
	org.apache.hadoop.conf.Configuration configuration = new org.apache.hadoop.conf.Configuration();
	InputFile inputFile =
		HadoopInputFile.fromPath(new org.apache.hadoop.fs.Path(split.getPath().toUri()), configuration);
	ParquetReadOptions options = ParquetReadOptions.builder().build();
	ParquetFileReader fileReader = new ParquetFileReader(inputFile, options);
	MessageType fileSchema = fileReader.getFileMetaData().getSchema();
	MessageType readSchema = getReadSchema(fileSchema, split.getPath());
	if (skipThisSplit) {
		LOG.warn(String.format(
			"Escaped the file split [%s] due to mismatch of file schema to expected result schema",
			split.getPath().toString()));
	} else {
		this.parquetRecordReader = new ParquetRecordReader<>(new RowReadSupport(), readSchema,
			filterPredicate == null ? FilterCompat.NOOP : FilterCompat.get(filterPredicate));
		this.parquetRecordReader.initialize(fileReader, configuration);
		this.parquetRecordReader.setSkipCorruptedRecord(this.skipCorruptedRecord);

		if (this.recordConsumed == null) {
			this.recordConsumed = getRuntimeContext().getMetricGroup().counter("parquet-records-consumed");
		}

		LOG.debug(String.format("Open ParquetInputFormat with FileInputSplit [%s]", split.getPath().toString()));
	}
}
 
Example 11
Source Project: flink   Source File: ParquetStreamingFileSinkITCase.java    License: Apache License 2.0 5 votes vote down vote up
private static <T> List<T> readParquetFile(File file, GenericData dataModel) throws IOException {
	InputFile inFile = HadoopInputFile.fromPath(new org.apache.hadoop.fs.Path(file.toURI()), new Configuration());

	ArrayList<T> results = new ArrayList<>();
	try (ParquetReader<T> reader = AvroParquetReader.<T>builder(inFile).withDataModel(dataModel).build()) {
		T next;
		while ((next = reader.read()) != null) {
			results.add(next);
		}
	}

	return results;
}
 
Example 12
Source Project: flink   Source File: ParquetRecordReaderTest.java    License: Apache License 2.0 5 votes vote down vote up
@Test
public void testReadSimpleGroup() throws IOException {
	Long[] array = {1L};
	GenericData.Record record = new GenericRecordBuilder(SIMPLE_SCHEMA)
		.set("bar", "test")
		.set("foo", 32L)
		.set("arr", array).build();

	Path path = createTempParquetFile(tempRoot.getRoot(), SIMPLE_SCHEMA, Collections.singletonList(record));
	MessageType readSchema = (new AvroSchemaConverter()).convert(SIMPLE_SCHEMA);
	ParquetRecordReader<Row> rowReader = new ParquetRecordReader<>(new RowReadSupport(), readSchema);

	InputFile inputFile =
		HadoopInputFile.fromPath(new org.apache.hadoop.fs.Path(path.toUri()), testConfig);
	ParquetReadOptions options = ParquetReadOptions.builder().build();
	ParquetFileReader fileReader = new ParquetFileReader(inputFile, options);

	rowReader.initialize(fileReader, testConfig);
	assertFalse(rowReader.reachEnd());

	Row row = rowReader.nextRecord();
	assertEquals(3, row.getArity());
	assertEquals(32L, row.getField(0));
	assertEquals("test", row.getField(1));
	assertArrayEquals(array, (Long[]) row.getField(2));
	assertTrue(rowReader.reachEnd());
}
 
Example 13
Source Project: flink   Source File: ParquetRecordReaderTest.java    License: Apache License 2.0 5 votes vote down vote up
@Test
public void testReadMultipleSimpleGroup() throws IOException {
	Long[] array = {1L};

	List<IndexedRecord> records = new ArrayList<>();
	for (int i = 0; i < 100; i++) {
		GenericData.Record record = new GenericRecordBuilder(SIMPLE_SCHEMA)
			.set("bar", "test")
			.set("foo", i)
			.set("arr", array).build();
		records.add(record);
	}

	Path path = createTempParquetFile(tempRoot.getRoot(), SIMPLE_SCHEMA, records);
	MessageType readSchema = (new AvroSchemaConverter()).convert(SIMPLE_SCHEMA);
	ParquetRecordReader<Row> rowReader = new ParquetRecordReader<>(new RowReadSupport(), readSchema);

	InputFile inputFile =
		HadoopInputFile.fromPath(new org.apache.hadoop.fs.Path(path.toUri()), testConfig);
	ParquetReadOptions options = ParquetReadOptions.builder().build();
	ParquetFileReader fileReader = new ParquetFileReader(inputFile, options);

	rowReader.initialize(fileReader, testConfig);
	assertTrue(!rowReader.reachEnd());

	for (long i = 0; i < 100; i++) {
		assertFalse(rowReader.reachEnd());
		Row row = rowReader.nextRecord();
		assertEquals(3, row.getArity());
		assertEquals(i, row.getField(0));
		assertEquals("test", row.getField(1));
		assertArrayEquals(array, (Long[]) row.getField(2));
	}

	assertTrue(rowReader.reachEnd());
}
 
Example 14
Source Project: flink   Source File: ParquetRecordReaderTest.java    License: Apache License 2.0 5 votes vote down vote up
@Test
public void testReadNestedGroup() throws IOException {
	Schema schema = unWrapSchema(NESTED_SCHEMA.getField("bar").schema());
	GenericData.Record barRecord = new GenericRecordBuilder(schema)
		.set("spam", 31L).build();

	GenericData.Record record = new GenericRecordBuilder(NESTED_SCHEMA)
		.set("foo", 32L)
		.set("bar", barRecord)
		.build();

	Path path = createTempParquetFile(tempRoot.getRoot(), NESTED_SCHEMA, Collections.singletonList(record));
	MessageType readSchema = (new AvroSchemaConverter()).convert(NESTED_SCHEMA);
	ParquetRecordReader<Row> rowReader = new ParquetRecordReader<>(new RowReadSupport(), readSchema);

	InputFile inputFile =
		HadoopInputFile.fromPath(new org.apache.hadoop.fs.Path(path.toUri()), testConfig);
	ParquetReadOptions options = ParquetReadOptions.builder().build();
	ParquetFileReader fileReader = new ParquetFileReader(inputFile, options);

	rowReader.initialize(fileReader, testConfig);
	assertFalse(rowReader.reachEnd());

	Row row = rowReader.nextRecord();
	assertEquals(7, row.getArity());
	assertEquals(32L, row.getField(0));
	assertEquals(31L, ((Row) row.getField(2)).getField(0));
	assertTrue(rowReader.reachEnd());
}
 
Example 15
Source Project: flink   Source File: ParquetRecordReaderTest.java    License: Apache License 2.0 5 votes vote down vote up
@Test
public void testMapGroup() throws IOException {
	Preconditions.checkState(unWrapSchema(NESTED_SCHEMA.getField("spamMap").schema())
		.getType().equals(Schema.Type.MAP));
	ImmutableMap.Builder<String, String> map = ImmutableMap.builder();
	map.put("testKey", "testValue");

	GenericRecord record = new GenericRecordBuilder(NESTED_SCHEMA)
		.set("foo", 32L)
		.set("spamMap", map.build())
		.build();

	Path path = createTempParquetFile(tempRoot.getRoot(), NESTED_SCHEMA, Collections.singletonList(record));
	MessageType readSchema = (new AvroSchemaConverter()).convert(NESTED_SCHEMA);
	ParquetRecordReader<Row> rowReader = new ParquetRecordReader<>(new RowReadSupport(), readSchema);

	InputFile inputFile =
		HadoopInputFile.fromPath(new org.apache.hadoop.fs.Path(path.toUri()), testConfig);
	ParquetReadOptions options = ParquetReadOptions.builder().build();
	ParquetFileReader fileReader = new ParquetFileReader(inputFile, options);

	rowReader.initialize(fileReader, testConfig);
	assertFalse(rowReader.reachEnd());

	Row row = rowReader.nextRecord();
	assertEquals(7, row.getArity());

	assertEquals(32L, row.getField(0));
	Map<?, ?> result = (Map<?, ?>) row.getField(1);
	assertEquals(result.get("testKey").toString(), "testValue");
	assertTrue(rowReader.reachEnd());
}
 
Example 16
Source Project: flink   Source File: ParquetRecordReaderTest.java    License: Apache License 2.0 5 votes vote down vote up
@Test
public void testNestedMapGroup() throws IOException {
	Schema nestedMapSchema = unWrapSchema(NESTED_SCHEMA.getField("nestedMap").schema());
	Preconditions.checkState(nestedMapSchema.getType().equals(Schema.Type.MAP));

	Schema mapValueSchema = nestedMapSchema.getValueType();
	GenericRecord mapValue = new GenericRecordBuilder(mapValueSchema)
		.set("type", "nested")
		.set("value", "nested_value").build();

	ImmutableMap.Builder<String, GenericRecord> map = ImmutableMap.builder();
	map.put("testKey", mapValue);

	GenericRecord record = new GenericRecordBuilder(NESTED_SCHEMA)
		.set("nestedMap", map.build())
		.set("foo", 34L).build();

	Path path = createTempParquetFile(tempRoot.getRoot(), NESTED_SCHEMA, Collections.singletonList(record));
	MessageType readSchema = (new AvroSchemaConverter()).convert(NESTED_SCHEMA);
	ParquetRecordReader<Row> rowReader = new ParquetRecordReader<>(new RowReadSupport(), readSchema);

	InputFile inputFile =
		HadoopInputFile.fromPath(new org.apache.hadoop.fs.Path(path.toUri()), testConfig);
	ParquetReadOptions options = ParquetReadOptions.builder().build();
	ParquetFileReader fileReader = new ParquetFileReader(inputFile, options);

	rowReader.initialize(fileReader, testConfig);
	assertFalse(rowReader.reachEnd());

	Row row = rowReader.nextRecord();
	assertEquals(7, row.getArity());

	assertEquals(34L, row.getField(0));
	Map result = (Map) row.getField(5);

	Row nestedRow = (Row) result.get("testKey");
	assertEquals("nested", nestedRow.getField(0));
	assertEquals("nested_value", nestedRow.getField(1));
}
 
Example 17
Source Project: garmadon   Source File: ProtoParquetWriterWithOffsetTest.java    License: Apache License 2.0 5 votes vote down vote up
private void checkFileLatestCommittedTimestamp(Path p, long timestamp) throws IOException {
    ParquetFileReader reader = new ParquetFileReader(
        HadoopInputFile.fromPath(p, new Configuration()),
        ParquetReadOptions.builder().build()
    );
    String actualTimestamp = reader.getFooter().getFileMetaData().getKeyValueMetaData().get(ProtoParquetWriterWithOffset.LATEST_TIMESTAMP_META_KEY);
    assertThat(actualTimestamp, is(String.valueOf(timestamp)));
}
 
Example 18
Source Project: pxf   Source File: ParquetFileAccessor.java    License: Apache License 2.0 5 votes vote down vote up
/**
 * Reads the original schema from the parquet file.
 *
 * @param parquetFile the path to the parquet file
 * @param fileSplit   the file split we are accessing
 * @return the original schema from the parquet file
 * @throws IOException when there's an IOException while reading the schema
 */
private MessageType getSchema(Path parquetFile, FileSplit fileSplit) throws IOException {

    final long then = System.nanoTime();
    ParquetMetadataConverter.MetadataFilter filter = ParquetMetadataConverter.range(
            fileSplit.getStart(), fileSplit.getStart() + fileSplit.getLength());
    ParquetReadOptions parquetReadOptions = HadoopReadOptions
            .builder(configuration)
            .withMetadataFilter(filter)
            .build();
    HadoopInputFile inputFile = HadoopInputFile.fromPath(parquetFile, configuration);
    try (ParquetFileReader parquetFileReader =
                 ParquetFileReader.open(inputFile, parquetReadOptions)) {
        FileMetaData metadata = parquetFileReader.getFileMetaData();
        if (LOG.isDebugEnabled()) {
            LOG.debug("{}-{}: Reading file {} with {} records in {} RowGroups",
                    context.getTransactionId(), context.getSegmentId(),
                    parquetFile.getName(), parquetFileReader.getRecordCount(),
                    parquetFileReader.getRowGroups().size());
        }
        final long millis = TimeUnit.NANOSECONDS.toMillis(System.nanoTime() - then);
        LOG.debug("{}-{}: Read schema in {} ms", context.getTransactionId(),
                context.getSegmentId(), millis);
        return metadata.getSchema();
    } catch (Exception e) {
        throw new IOException(e);
    }
}
 
Example 19
Source Project: hugegraph-loader   Source File: ParquetFileLineFetcher.java    License: Apache License 2.0 5 votes vote down vote up
@Override
public void openReader(Readable readable) {
    Path path = new Path(this.source().path());
    try {
        HadoopInputFile file = HadoopInputFile.fromPath(path, this.conf);
        this.reader = ParquetFileReader.open(file);
        this.schema = this.reader.getFooter().getFileMetaData().getSchema();
        this.columnIO = new ColumnIOFactory().getColumnIO(this.schema);
    } catch (IOException e) {
        throw new LoadException("Failed to open parquet reader for '%s'",
                                e, readable);
    }
    this.resetOffset();
}
 
Example 20
Source Project: kafka-connect-fs   Source File: ParquetFileReader.java    License: Apache License 2.0 5 votes vote down vote up
private ParquetReader<GenericRecord> initReader() throws IOException {
    Configuration configuration = getFs().getConf();
    if (this.schema != null) {
        AvroReadSupport.setAvroReadSchema(configuration, this.schema);
    }
    if (this.projection != null) {
        AvroReadSupport.setRequestedProjection(configuration, this.projection);
    }
    return AvroParquetReader
            .<GenericRecord>builder(HadoopInputFile.fromPath(getFilePath(), configuration))
            .build();
}
 
Example 21
Source Project: ignite   Source File: SparkModelParser.java    License: Apache License 2.0 5 votes vote down vote up
/**
 * Load Decision Tree model.
 *
 * @param pathToMdl Path to model.
 * @param learningEnvironment Learning environment.
 */
private static Model loadDecisionTreeModel(String pathToMdl, LearningEnvironment learningEnvironment) {
    try (ParquetFileReader r = ParquetFileReader.open(HadoopInputFile.fromPath(new Path(pathToMdl), new Configuration()))) {
        PageReadStore pages;

        final MessageType schema = r.getFooter().getFileMetaData().getSchema();
        final MessageColumnIO colIO = new ColumnIOFactory().getColumnIO(schema);
        final Map<Integer, NodeData> nodes = new TreeMap<>();

        while (null != (pages = r.readNextRowGroup())) {
            final long rows = pages.getRowCount();
            final RecordReader recordReader = colIO.getRecordReader(pages, new GroupRecordConverter(schema));

            for (int i = 0; i < rows; i++) {
                final SimpleGroup g = (SimpleGroup)recordReader.read();
                NodeData nodeData = extractNodeDataFromParquetRow(g);
                nodes.put(nodeData.id, nodeData);
            }
        }
        return buildDecisionTreeModel(nodes);
    }
    catch (IOException e) {
        String msg = "Error reading parquet file: " + e.getMessage();
        learningEnvironment.logger().log(MLLogger.VerboseLevel.HIGH, msg);
        e.printStackTrace();
    }
    return null;
}
 
Example 22
Source Project: ignite   Source File: SparkModelParser.java    License: Apache License 2.0 5 votes vote down vote up
/**
 * Load SVM model.
 *
 * @param pathToMdl Path to model.
 * @param learningEnvironment Learning environment.
 */
private static Model loadLinearSVMModel(String pathToMdl,
    LearningEnvironment learningEnvironment) {
    Vector coefficients = null;
    double interceptor = 0;

    try (ParquetFileReader r = ParquetFileReader.open(HadoopInputFile.fromPath(new Path(pathToMdl), new Configuration()))) {
        PageReadStore pages;

        final MessageType schema = r.getFooter().getFileMetaData().getSchema();
        final MessageColumnIO colIO = new ColumnIOFactory().getColumnIO(schema);

        while (null != (pages = r.readNextRowGroup())) {
            final long rows = pages.getRowCount();
            final RecordReader recordReader = colIO.getRecordReader(pages, new GroupRecordConverter(schema));
            for (int i = 0; i < rows; i++) {
                final SimpleGroup g = (SimpleGroup)recordReader.read();
                interceptor = readSVMInterceptor(g);
                coefficients = readSVMCoefficients(g);
            }
        }
    }
    catch (IOException e) {
        String msg = "Error reading parquet file: " + e.getMessage();
        learningEnvironment.logger().log(MLLogger.VerboseLevel.HIGH, msg);
        e.printStackTrace();
    }

    return new SVMLinearClassificationModel(coefficients, interceptor);
}
 
Example 23
Source Project: ignite   Source File: SparkModelParser.java    License: Apache License 2.0 5 votes vote down vote up
/**
 * Load linear regression model.
 *
 * @param pathToMdl Path to model.
 * @param learningEnvironment Learning environment.
 */
private static Model loadLinRegModel(String pathToMdl,
    LearningEnvironment learningEnvironment) {
    Vector coefficients = null;
    double interceptor = 0;

    try (ParquetFileReader r = ParquetFileReader.open(HadoopInputFile.fromPath(new Path(pathToMdl), new Configuration()))) {
        PageReadStore pages;

        final MessageType schema = r.getFooter().getFileMetaData().getSchema();
        final MessageColumnIO colIO = new ColumnIOFactory().getColumnIO(schema);

        while (null != (pages = r.readNextRowGroup())) {
            final long rows = pages.getRowCount();
            final RecordReader recordReader = colIO.getRecordReader(pages, new GroupRecordConverter(schema));
            for (int i = 0; i < rows; i++) {
                final SimpleGroup g = (SimpleGroup)recordReader.read();
                interceptor = readLinRegInterceptor(g);
                coefficients = readLinRegCoefficients(g);
            }
        }

    }
    catch (IOException e) {
        String msg = "Error reading parquet file: " + e.getMessage();
        learningEnvironment.logger().log(MLLogger.VerboseLevel.HIGH, msg);
        e.printStackTrace();
    }

    return new LinearRegressionModel(coefficients, interceptor);
}
 
Example 24
Source Project: ignite   Source File: SparkModelParser.java    License: Apache License 2.0 5 votes vote down vote up
/**
 * Load logistic regression model.
 *
 * @param pathToMdl Path to model.
 * @param learningEnvironment Learning environment.
 */
private static Model loadLogRegModel(String pathToMdl,
    LearningEnvironment learningEnvironment) {
    Vector coefficients = null;
    double interceptor = 0;

    try (ParquetFileReader r = ParquetFileReader.open(HadoopInputFile.fromPath(new Path(pathToMdl), new Configuration()))) {
        PageReadStore pages;

        final MessageType schema = r.getFooter().getFileMetaData().getSchema();
        final MessageColumnIO colIO = new ColumnIOFactory().getColumnIO(schema);

        while (null != (pages = r.readNextRowGroup())) {
            final long rows = pages.getRowCount();
            final RecordReader recordReader = colIO.getRecordReader(pages, new GroupRecordConverter(schema));
            for (int i = 0; i < rows; i++) {
                final SimpleGroup g = (SimpleGroup)recordReader.read();
                interceptor = readInterceptor(g);
                coefficients = readCoefficients(g);
            }
        }

    }
    catch (IOException e) {
        String msg = "Error reading parquet file: " + e.getMessage();
        learningEnvironment.logger().log(MLLogger.VerboseLevel.HIGH, msg);
        e.printStackTrace();
    }

    return new LogisticRegressionModel(coefficients, interceptor);
}
 
Example 25
Source Project: flink   Source File: ParquetInputFormat.java    License: Apache License 2.0 5 votes vote down vote up
@Override
public void open(FileInputSplit split) throws IOException {
	// reset the flag when open a new split
	this.skipThisSplit = false;
	org.apache.hadoop.conf.Configuration configuration = new org.apache.hadoop.conf.Configuration();
	InputFile inputFile =
		HadoopInputFile.fromPath(new org.apache.hadoop.fs.Path(split.getPath().toUri()), configuration);
	ParquetReadOptions options = ParquetReadOptions.builder().build();
	ParquetFileReader fileReader = new ParquetFileReader(inputFile, options);
	MessageType fileSchema = fileReader.getFileMetaData().getSchema();
	MessageType readSchema = getReadSchema(fileSchema, split.getPath());
	if (skipThisSplit) {
		LOG.warn(String.format(
			"Escaped the file split [%s] due to mismatch of file schema to expected result schema",
			split.getPath().toString()));
	} else {
		this.parquetRecordReader = new ParquetRecordReader<>(new RowReadSupport(), readSchema,
			filterPredicate == null ? FilterCompat.NOOP : FilterCompat.get(filterPredicate));
		this.parquetRecordReader.initialize(fileReader, configuration);
		this.parquetRecordReader.setSkipCorruptedRecord(this.skipCorruptedRecord);

		if (this.recordConsumed == null) {
			this.recordConsumed = getRuntimeContext().getMetricGroup().counter("parquet-records-consumed");
		}

		LOG.debug(String.format("Open ParquetInputFormat with FileInputSplit [%s]", split.getPath().toString()));
	}
}
 
Example 26
Source Project: flink   Source File: ParquetStreamingFileSinkITCase.java    License: Apache License 2.0 5 votes vote down vote up
private static <T> List<T> readParquetFile(File file, GenericData dataModel) throws IOException {
	InputFile inFile = HadoopInputFile.fromPath(new org.apache.hadoop.fs.Path(file.toURI()), new Configuration());

	ArrayList<T> results = new ArrayList<>();
	try (ParquetReader<T> reader = AvroParquetReader.<T>builder(inFile).withDataModel(dataModel).build()) {
		T next;
		while ((next = reader.read()) != null) {
			results.add(next);
		}
	}

	return results;
}
 
Example 27
Source Project: flink   Source File: ParquetRecordReaderTest.java    License: Apache License 2.0 5 votes vote down vote up
@Test
public void testReadSimpleGroup() throws IOException {
	Long[] array = {1L};
	GenericData.Record record = new GenericRecordBuilder(SIMPLE_SCHEMA)
		.set("bar", "test")
		.set("foo", 32L)
		.set("arr", array).build();

	Path path = createTempParquetFile(tempRoot.getRoot(), SIMPLE_SCHEMA, Collections.singletonList(record));
	MessageType readSchema = (new AvroSchemaConverter()).convert(SIMPLE_SCHEMA);
	ParquetRecordReader<Row> rowReader = new ParquetRecordReader<>(new RowReadSupport(), readSchema);

	InputFile inputFile =
		HadoopInputFile.fromPath(new org.apache.hadoop.fs.Path(path.toUri()), testConfig);
	ParquetReadOptions options = ParquetReadOptions.builder().build();
	ParquetFileReader fileReader = new ParquetFileReader(inputFile, options);

	rowReader.initialize(fileReader, testConfig);
	assertFalse(rowReader.reachEnd());

	Row row = rowReader.nextRecord();
	assertEquals(3, row.getArity());
	assertEquals(32L, row.getField(0));
	assertEquals("test", row.getField(1));
	assertArrayEquals(array, (Long[]) row.getField(2));
	assertTrue(rowReader.reachEnd());
}
 
Example 28
Source Project: flink   Source File: ParquetRecordReaderTest.java    License: Apache License 2.0 5 votes vote down vote up
@Test
public void testReadMultipleSimpleGroup() throws IOException {
	Long[] array = {1L};

	List<IndexedRecord> records = new ArrayList<>();
	for (int i = 0; i < 100; i++) {
		GenericData.Record record = new GenericRecordBuilder(SIMPLE_SCHEMA)
			.set("bar", "test")
			.set("foo", i)
			.set("arr", array).build();
		records.add(record);
	}

	Path path = createTempParquetFile(tempRoot.getRoot(), SIMPLE_SCHEMA, records);
	MessageType readSchema = (new AvroSchemaConverter()).convert(SIMPLE_SCHEMA);
	ParquetRecordReader<Row> rowReader = new ParquetRecordReader<>(new RowReadSupport(), readSchema);

	InputFile inputFile =
		HadoopInputFile.fromPath(new org.apache.hadoop.fs.Path(path.toUri()), testConfig);
	ParquetReadOptions options = ParquetReadOptions.builder().build();
	ParquetFileReader fileReader = new ParquetFileReader(inputFile, options);

	rowReader.initialize(fileReader, testConfig);
	assertTrue(!rowReader.reachEnd());

	for (long i = 0; i < 100; i++) {
		assertFalse(rowReader.reachEnd());
		Row row = rowReader.nextRecord();
		assertEquals(3, row.getArity());
		assertEquals(i, row.getField(0));
		assertEquals("test", row.getField(1));
		assertArrayEquals(array, (Long[]) row.getField(2));
	}

	assertTrue(rowReader.reachEnd());
}
 
Example 29
Source Project: flink   Source File: ParquetRecordReaderTest.java    License: Apache License 2.0 5 votes vote down vote up
@Test
public void testReadNestedGroup() throws IOException {
	Schema schema = unWrapSchema(NESTED_SCHEMA.getField("bar").schema());
	GenericData.Record barRecord = new GenericRecordBuilder(schema)
		.set("spam", 31L).build();

	GenericData.Record record = new GenericRecordBuilder(NESTED_SCHEMA)
		.set("foo", 32L)
		.set("bar", barRecord)
		.build();

	Path path = createTempParquetFile(tempRoot.getRoot(), NESTED_SCHEMA, Collections.singletonList(record));
	MessageType readSchema = (new AvroSchemaConverter()).convert(NESTED_SCHEMA);
	ParquetRecordReader<Row> rowReader = new ParquetRecordReader<>(new RowReadSupport(), readSchema);

	InputFile inputFile =
		HadoopInputFile.fromPath(new org.apache.hadoop.fs.Path(path.toUri()), testConfig);
	ParquetReadOptions options = ParquetReadOptions.builder().build();
	ParquetFileReader fileReader = new ParquetFileReader(inputFile, options);

	rowReader.initialize(fileReader, testConfig);
	assertFalse(rowReader.reachEnd());

	Row row = rowReader.nextRecord();
	assertEquals(7, row.getArity());
	assertEquals(32L, row.getField(0));
	assertEquals(31L, ((Row) row.getField(2)).getField(0));
	assertTrue(rowReader.reachEnd());
}
 
Example 30
Source Project: flink   Source File: ParquetRecordReaderTest.java    License: Apache License 2.0 5 votes vote down vote up
@Test
public void testMapGroup() throws IOException {
	Preconditions.checkState(unWrapSchema(NESTED_SCHEMA.getField("spamMap").schema())
		.getType().equals(Schema.Type.MAP));
	ImmutableMap.Builder<String, String> map = ImmutableMap.builder();
	map.put("testKey", "testValue");

	GenericRecord record = new GenericRecordBuilder(NESTED_SCHEMA)
		.set("foo", 32L)
		.set("spamMap", map.build())
		.build();

	Path path = createTempParquetFile(tempRoot.getRoot(), NESTED_SCHEMA, Collections.singletonList(record));
	MessageType readSchema = (new AvroSchemaConverter()).convert(NESTED_SCHEMA);
	ParquetRecordReader<Row> rowReader = new ParquetRecordReader<>(new RowReadSupport(), readSchema);

	InputFile inputFile =
		HadoopInputFile.fromPath(new org.apache.hadoop.fs.Path(path.toUri()), testConfig);
	ParquetReadOptions options = ParquetReadOptions.builder().build();
	ParquetFileReader fileReader = new ParquetFileReader(inputFile, options);

	rowReader.initialize(fileReader, testConfig);
	assertFalse(rowReader.reachEnd());

	Row row = rowReader.nextRecord();
	assertEquals(7, row.getArity());

	assertEquals(32L, row.getField(0));
	Map<?, ?> result = (Map<?, ?>) row.getField(1);
	assertEquals(result.get("testKey").toString(), "testValue");
	assertTrue(rowReader.reachEnd());
}