Search in sources :

Example 1 with HoodieSparkBootstrapSchemaProvider

use of org.apache.hudi.client.bootstrap.HoodieSparkBootstrapSchemaProvider in project hudi by apache.

the class SparkBootstrapCommitActionExecutor method listAndProcessSourcePartitions.

/**
 * Return Bootstrap Mode selections for partitions listed and figure out bootstrap Schema.
 * @return
 * @throws IOException
 */
private Map<BootstrapMode, List<Pair<String, List<HoodieFileStatus>>>> listAndProcessSourcePartitions() throws IOException {
    List<Pair<String, List<HoodieFileStatus>>> folders = BootstrapUtils.getAllLeafFoldersWithFiles(table.getMetaClient(), bootstrapSourceFileSystem, config.getBootstrapSourceBasePath(), context);
    LOG.info("Fetching Bootstrap Schema !!");
    HoodieBootstrapSchemaProvider sourceSchemaProvider = new HoodieSparkBootstrapSchemaProvider(config);
    bootstrapSchema = sourceSchemaProvider.getBootstrapSchema(context, folders).toString();
    LOG.info("Bootstrap Schema :" + bootstrapSchema);
    BootstrapModeSelector selector = (BootstrapModeSelector) ReflectionUtils.loadClass(config.getBootstrapModeSelectorClass(), config);
    Map<BootstrapMode, List<String>> result = selector.select(folders);
    Map<String, List<HoodieFileStatus>> partitionToFiles = folders.stream().collect(Collectors.toMap(Pair::getKey, Pair::getValue));
    // Ensure all partitions are accounted for
    ValidationUtils.checkArgument(partitionToFiles.keySet().equals(result.values().stream().flatMap(Collection::stream).collect(Collectors.toSet())));
    return result.entrySet().stream().map(e -> Pair.of(e.getKey(), e.getValue().stream().map(p -> Pair.of(p, partitionToFiles.get(p))).collect(Collectors.toList()))).collect(Collectors.toMap(Pair::getKey, Pair::getValue));
}
Also used : HoodieTable(org.apache.hudi.table.HoodieTable) BootstrapMode(org.apache.hudi.client.bootstrap.BootstrapMode) FileSystem(org.apache.hadoop.fs.FileSystem) HoodieInstant(org.apache.hudi.common.table.timeline.HoodieInstant) BaseCommitActionExecutor(org.apache.hudi.table.action.commit.BaseCommitActionExecutor) HoodieJavaRDD(org.apache.hudi.data.HoodieJavaRDD) KeyGeneratorInterface(org.apache.hudi.keygen.KeyGeneratorInterface) Logger(org.apache.log4j.Logger) HoodieFileStatus(org.apache.hudi.avro.model.HoodieFileStatus) Duration(java.time.Duration) Map(java.util.Map) HoodieSparkEngineContext(org.apache.hudi.client.common.HoodieSparkEngineContext) HoodieWriteMetadata(org.apache.hudi.table.action.HoodieWriteMetadata) HoodieActiveTimeline(org.apache.hudi.common.table.timeline.HoodieActiveTimeline) HoodieSparkKeyGeneratorFactory(org.apache.hudi.keygen.factory.HoodieSparkKeyGeneratorFactory) ValidationUtils(org.apache.hudi.common.util.ValidationUtils) BootstrapPartitionPathTranslator(org.apache.hudi.client.bootstrap.translator.BootstrapPartitionPathTranslator) Collection(java.util.Collection) Instant(java.time.Instant) Collectors(java.util.stream.Collectors) StandardCharsets(java.nio.charset.StandardCharsets) List(java.util.List) WRITE_STATUS_STORAGE_LEVEL_VALUE(org.apache.hudi.config.HoodieWriteConfig.WRITE_STATUS_STORAGE_LEVEL_VALUE) HoodieWriteStat(org.apache.hudi.common.model.HoodieWriteStat) BootstrapWriteStatus(org.apache.hudi.client.bootstrap.BootstrapWriteStatus) SparkBulkInsertCommitActionExecutor(org.apache.hudi.table.action.commit.SparkBulkInsertCommitActionExecutor) WriteOperationType(org.apache.hudi.common.model.WriteOperationType) ReflectionUtils(org.apache.hudi.common.util.ReflectionUtils) FullRecordBootstrapDataProvider(org.apache.hudi.client.bootstrap.FullRecordBootstrapDataProvider) MetadataBootstrapHandlerFactory.getMetadataHandler(org.apache.hudi.table.action.bootstrap.MetadataBootstrapHandlerFactory.getMetadataHandler) HoodieBootstrapSchemaProvider(org.apache.hudi.client.bootstrap.HoodieBootstrapSchemaProvider) Option(org.apache.hudi.common.util.Option) HoodieCommitException(org.apache.hudi.exception.HoodieCommitException) HoodieKeyGeneratorException(org.apache.hudi.exception.HoodieKeyGeneratorException) State(org.apache.hudi.common.table.timeline.HoodieInstant.State) BaseSparkCommitActionExecutor(org.apache.hudi.table.action.commit.BaseSparkCommitActionExecutor) HoodieSparkTable(org.apache.hudi.table.HoodieSparkTable) HoodieTableMetaClient(org.apache.hudi.common.table.HoodieTableMetaClient) HoodieTimeline(org.apache.hudi.common.table.timeline.HoodieTimeline) JavaRDD(org.apache.spark.api.java.JavaRDD) BootstrapFileMapping(org.apache.hudi.common.model.BootstrapFileMapping) HoodieRecord(org.apache.hudi.common.model.HoodieRecord) BootstrapModeSelector(org.apache.hudi.client.bootstrap.selector.BootstrapModeSelector) HoodieData(org.apache.hudi.common.data.HoodieData) BootstrapIndex(org.apache.hudi.common.bootstrap.index.BootstrapIndex) TypedProperties(org.apache.hudi.common.config.TypedProperties) HoodieWriteConfig(org.apache.hudi.config.HoodieWriteConfig) WriteMarkersFactory(org.apache.hudi.table.marker.WriteMarkersFactory) Iterator(java.util.Iterator) HoodieCommitMetadata(org.apache.hudi.common.model.HoodieCommitMetadata) IOException(java.io.IOException) HoodieSparkBootstrapSchemaProvider(org.apache.hudi.client.bootstrap.HoodieSparkBootstrapSchemaProvider) WriteStatus(org.apache.hudi.client.WriteStatus) HoodieRecordPayload(org.apache.hudi.common.model.HoodieRecordPayload) SparkValidatorUtils(org.apache.hudi.client.utils.SparkValidatorUtils) HoodieKey(org.apache.hudi.common.model.HoodieKey) HoodieIOException(org.apache.hudi.exception.HoodieIOException) LogManager(org.apache.log4j.LogManager) FSUtils(org.apache.hudi.common.fs.FSUtils) Pair(org.apache.hudi.common.util.collection.Pair) HoodieFileStatus(org.apache.hudi.avro.model.HoodieFileStatus) BootstrapMode(org.apache.hudi.client.bootstrap.BootstrapMode) HoodieSparkBootstrapSchemaProvider(org.apache.hudi.client.bootstrap.HoodieSparkBootstrapSchemaProvider) BootstrapModeSelector(org.apache.hudi.client.bootstrap.selector.BootstrapModeSelector) List(java.util.List) Pair(org.apache.hudi.common.util.collection.Pair) HoodieBootstrapSchemaProvider(org.apache.hudi.client.bootstrap.HoodieBootstrapSchemaProvider)

Aggregations

IOException (java.io.IOException)1 StandardCharsets (java.nio.charset.StandardCharsets)1 Duration (java.time.Duration)1 Instant (java.time.Instant)1 Collection (java.util.Collection)1 Iterator (java.util.Iterator)1 List (java.util.List)1 Map (java.util.Map)1 Collectors (java.util.stream.Collectors)1 FileSystem (org.apache.hadoop.fs.FileSystem)1 HoodieFileStatus (org.apache.hudi.avro.model.HoodieFileStatus)1 WriteStatus (org.apache.hudi.client.WriteStatus)1 BootstrapMode (org.apache.hudi.client.bootstrap.BootstrapMode)1 BootstrapWriteStatus (org.apache.hudi.client.bootstrap.BootstrapWriteStatus)1 FullRecordBootstrapDataProvider (org.apache.hudi.client.bootstrap.FullRecordBootstrapDataProvider)1 HoodieBootstrapSchemaProvider (org.apache.hudi.client.bootstrap.HoodieBootstrapSchemaProvider)1 HoodieSparkBootstrapSchemaProvider (org.apache.hudi.client.bootstrap.HoodieSparkBootstrapSchemaProvider)1 BootstrapModeSelector (org.apache.hudi.client.bootstrap.selector.BootstrapModeSelector)1 BootstrapPartitionPathTranslator (org.apache.hudi.client.bootstrap.translator.BootstrapPartitionPathTranslator)1 HoodieSparkEngineContext (org.apache.hudi.client.common.HoodieSparkEngineContext)1