Examples with ParquetReaderIterator org.apache.hudi.common.util.ParquetReaderIterator used on opensource projects

Example 1 with ParquetReaderIterator

use of org.apache.hudi.common.util.ParquetReaderIterator in project hudi by apache.

the class ParquetBootstrapMetadataHandler method executeBootstrap.

@Override
void executeBootstrap(HoodieBootstrapHandle<?, ?, ?, ?> bootstrapHandle, Path sourceFilePath, KeyGeneratorInterface keyGenerator, String partitionPath, Schema avroSchema) throws Exception {
    BoundedInMemoryExecutor<GenericRecord, HoodieRecord, Void> wrapper = null;
    try {
        ParquetReader<IndexedRecord> reader = AvroParquetReader.<IndexedRecord>builder(sourceFilePath).withConf(table.getHadoopConf()).build();
        wrapper = new BoundedInMemoryExecutor<GenericRecord, HoodieRecord, Void>(config.getWriteBufferLimitBytes(), new ParquetReaderIterator(reader), new BootstrapRecordConsumer(bootstrapHandle), inp -> {
            String recKey = keyGenerator.getKey(inp).getRecordKey();
            GenericRecord gr = new GenericData.Record(HoodieAvroUtils.RECORD_KEY_SCHEMA);
            gr.put(HoodieRecord.RECORD_KEY_METADATA_FIELD, recKey);
            BootstrapRecordPayload payload = new BootstrapRecordPayload(gr);
            HoodieRecord rec = new HoodieAvroRecord(new HoodieKey(recKey, partitionPath), payload);
            return rec;
        }, table.getPreExecuteRunnable());
        wrapper.execute();
    } catch (Exception e) {
        throw new HoodieException(e);
    } finally {
        bootstrapHandle.close();
        if (null != wrapper) {
            wrapper.shutdownNow();
        }
    }
}

Example 2 with ParquetReaderIterator

use of org.apache.hudi.common.util.ParquetReaderIterator in project hudi by apache.

the class TestBootstrap method generateInputBatch.

private static JavaRDD<HoodieRecord> generateInputBatch(JavaSparkContext jsc, List<Pair<String, List<HoodieFileStatus>>> partitionPaths, Schema writerSchema) {
    List<Pair<String, Path>> fullFilePathsWithPartition = partitionPaths.stream().flatMap(p -> p.getValue().stream().map(x -> Pair.of(p.getKey(), FileStatusUtils.toPath(x.getPath())))).collect(Collectors.toList());
    return jsc.parallelize(fullFilePathsWithPartition.stream().flatMap(p -> {
        try {
            Configuration conf = jsc.hadoopConfiguration();
            AvroReadSupport.setAvroReadSchema(conf, writerSchema);
            Iterator<GenericRecord> recIterator = new ParquetReaderIterator(AvroParquetReader.<GenericRecord>builder(p.getValue()).withConf(conf).build());
            return StreamSupport.stream(Spliterators.spliteratorUnknownSize(recIterator, 0), false).map(gr -> {
                try {
                    String key = gr.get("_row_key").toString();
                    String pPath = p.getKey();
                    return new HoodieAvroRecord<>(new HoodieKey(key, pPath), new RawTripTestPayload(gr.toString(), key, pPath, HoodieTestDataGenerator.TRIP_EXAMPLE_SCHEMA));
                } catch (IOException e) {
                    throw new HoodieIOException(e.getMessage(), e);
                }
            });
        } catch (IOException ioe) {
            throw new HoodieIOException(ioe.getMessage(), ioe);
        }
    }).collect(Collectors.toList()));
}