Examples with HoodieTestDataGenerator org.apache.hudi.common.testutils.HoodieTestDataGenerator used on opensource projects

Example 31 with HoodieTestDataGenerator

use of org.apache.hudi.common.testutils.HoodieTestDataGenerator in project hudi by apache.

the class TestJsonKafkaSource method testJsonKafkaSourceWithConfigurableUpperCap.

@Test
public void testJsonKafkaSourceWithConfigurableUpperCap() {
    // topic setup.
    final String topic = TEST_TOPIC_PREFIX + "testJsonKafkaSourceWithConfigurableUpperCap";
    testUtils.createTopic(topic, 2);
    HoodieTestDataGenerator dataGenerator = new HoodieTestDataGenerator();
    TypedProperties props = createPropsForJsonSource(topic, 500L, "earliest");
    Source jsonSource = new JsonKafkaSource(props, jsc(), spark(), schemaProvider, metrics);
    SourceFormatAdapter kafkaSource = new SourceFormatAdapter(jsonSource);
    // 1. Extract without any checkpoint => get all the data, respecting sourceLimit
    testUtils.sendMessages(topic, jsonifyRecords(dataGenerator.generateInserts("000", 1000)));
    InputBatch<JavaRDD<GenericRecord>> fetch1 = kafkaSource.fetchNewDataInAvroFormat(Option.empty(), 900);
    assertEquals(900, fetch1.getBatch().get().count());
    // 2. Produce new data, extract new data based on upper cap
    testUtils.sendMessages(topic, jsonifyRecords(dataGenerator.generateInserts("001", 1000)));
    InputBatch<Dataset<Row>> fetch2 = kafkaSource.fetchNewDataInRowFormat(Option.of(fetch1.getCheckpointForNextBatch()), Long.MAX_VALUE);
    assertEquals(500, fetch2.getBatch().get().count());
    // fetch data respecting source limit where upper cap > sourceLimit
    InputBatch<JavaRDD<GenericRecord>> fetch3 = kafkaSource.fetchNewDataInAvroFormat(Option.of(fetch1.getCheckpointForNextBatch()), 400);
    assertEquals(400, fetch3.getBatch().get().count());
    // fetch data respecting source limit where upper cap < sourceLimit
    InputBatch<JavaRDD<GenericRecord>> fetch4 = kafkaSource.fetchNewDataInAvroFormat(Option.of(fetch2.getCheckpointForNextBatch()), 600);
    assertEquals(600, fetch4.getBatch().get().count());
    // 3. Extract with previous checkpoint => gives same data back (idempotent)
    InputBatch<JavaRDD<GenericRecord>> fetch5 = kafkaSource.fetchNewDataInAvroFormat(Option.of(fetch1.getCheckpointForNextBatch()), Long.MAX_VALUE);
    assertEquals(fetch2.getBatch().get().count(), fetch5.getBatch().get().count());
    assertEquals(fetch2.getCheckpointForNextBatch(), fetch5.getCheckpointForNextBatch());
    // 4. Extract with latest checkpoint => no new data returned
    InputBatch<JavaRDD<GenericRecord>> fetch6 = kafkaSource.fetchNewDataInAvroFormat(Option.of(fetch4.getCheckpointForNextBatch()), Long.MAX_VALUE);
    assertEquals(Option.empty(), fetch6.getBatch());
}

Example 32 with HoodieTestDataGenerator

use of org.apache.hudi.common.testutils.HoodieTestDataGenerator in project hudi by apache.

the class TestJsonKafkaSourcePostProcessor method testInvalidJsonKafkaSourcePostProcessor.

@Test
public void testInvalidJsonKafkaSourcePostProcessor() {
    // topic setup.
    final String topic = TEST_TOPIC_PREFIX + "testInvalidJsonKafkaSourcePostProcessor";
    testUtils.createTopic(topic, 2);
    HoodieTestDataGenerator dataGenerator = new HoodieTestDataGenerator();
    TypedProperties props = createPropsForJsonSource(topic, null, "earliest");
    // processor class name setup
    props.setProperty(JSON_KAFKA_PROCESSOR_CLASS_OPT.key(), "InvalidJsonKafkaSourcePostProcessor");
    Source jsonSource = new JsonKafkaSource(props, jsc(), spark(), schemaProvider, metrics);
    SourceFormatAdapter kafkaSource = new SourceFormatAdapter(jsonSource);
    testUtils.sendMessages(topic, jsonifyRecords(dataGenerator.generateInserts("000", 1000)));
    Assertions.assertThrows(HoodieSourcePostProcessException.class, () -> kafkaSource.fetchNewDataInAvroFormat(Option.empty(), 900));
}

Example 33 with HoodieTestDataGenerator

use of org.apache.hudi.common.testutils.HoodieTestDataGenerator in project hudi by apache.

the class TestJsonKafkaSourcePostProcessor method testSampleJsonKafkaSourcePostProcessor.

@Test
public void testSampleJsonKafkaSourcePostProcessor() {
    // topic setup.
    final String topic = TEST_TOPIC_PREFIX + "testSampleJsonKafkaSourcePostProcessor";
    testUtils.createTopic(topic, 2);
    HoodieTestDataGenerator dataGenerator = new HoodieTestDataGenerator();
    TypedProperties props = createPropsForJsonSource(topic, null, "earliest");
    // processor class name setup
    props.setProperty(JSON_KAFKA_PROCESSOR_CLASS_OPT.key(), SampleJsonKafkaSourcePostProcessor.class.getName());
    Source jsonSource = new JsonKafkaSource(props, jsc(), spark(), schemaProvider, metrics);
    SourceFormatAdapter kafkaSource = new SourceFormatAdapter(jsonSource);
    testUtils.sendMessages(topic, jsonifyRecords(dataGenerator.generateInserts("000", 1000)));
    InputBatch<JavaRDD<GenericRecord>> fetch1 = kafkaSource.fetchNewDataInAvroFormat(Option.empty(), 900);
    assertNotEquals(900, fetch1.getBatch().get().count());
}

Example 34 with HoodieTestDataGenerator

use of org.apache.hudi.common.testutils.HoodieTestDataGenerator in project hudi by apache.

the class AbstractBaseTestSource method initDataGen.

public static void initDataGen(TypedProperties props, int partition) {
    try {
        boolean useRocksForTestDataGenKeys = props.getBoolean(SourceConfigs.USE_ROCKSDB_FOR_TEST_DATAGEN_KEYS, SourceConfigs.DEFAULT_USE_ROCKSDB_FOR_TEST_DATAGEN_KEYS);
        String baseStoreDir = props.getString(SourceConfigs.ROCKSDB_BASE_DIR_FOR_TEST_DATAGEN_KEYS, File.createTempFile("test_data_gen", ".keys").getParent()) + "/" + partition;
        LOG.info("useRocksForTestDataGenKeys=" + useRocksForTestDataGenKeys + ", BaseStoreDir=" + baseStoreDir);
        dataGeneratorMap.put(partition, new HoodieTestDataGenerator(HoodieTestDataGenerator.DEFAULT_PARTITION_PATHS, useRocksForTestDataGenKeys ? new RocksDBBasedMap<>(baseStoreDir) : new HashMap<>()));
    } catch (IOException e) {
        throw new HoodieIOException(e.getMessage(), e);
    }
}

Example 35 with HoodieTestDataGenerator

use of org.apache.hudi.common.testutils.HoodieTestDataGenerator in project hudi by apache.

the class AbstractBaseTestSource method initDataGen.

public static void initDataGen(SQLContext sqlContext, String globParquetPath, int partition) {
    List<Row> rows = sqlContext.read().format("hudi").load(globParquetPath).select("_hoodie_record_key", "_hoodie_partition_path").collectAsList();
    Map<Integer, HoodieTestDataGenerator.KeyPartition> keyPartitionMap = IntStream.range(0, rows.size()).boxed().collect(Collectors.toMap(Function.identity(), i -> {
        Row r = rows.get(i);
        HoodieTestDataGenerator.KeyPartition kp = new HoodieTestDataGenerator.KeyPartition();
        kp.key = new HoodieKey(r.getString(0), r.getString(1));
        kp.partitionPath = r.getString(1);
        return kp;
    }));
    dataGeneratorMap.put(partition, new HoodieTestDataGenerator(HoodieTestDataGenerator.DEFAULT_PARTITION_PATHS, keyPartitionMap));
}