Examples with DataFormat com.google.cloud.bigquery.storage.v1.DataFormat used on opensource projects

Example 1 with DataFormat

use of com.google.cloud.bigquery.storage.v1.DataFormat in project beam by apache.

the class BigQueryStorageSourceBase method split.

@Override
public List<BigQueryStorageStreamSource<T>> split(long desiredBundleSizeBytes, PipelineOptions options) throws Exception {
    BigQueryOptions bqOptions = options.as(BigQueryOptions.class);
    Table targetTable = getTargetTable(bqOptions);
    ReadSession.Builder readSessionBuilder = ReadSession.newBuilder();
    if (targetTable != null) {
        readSessionBuilder.setTable(BigQueryHelpers.toTableResourceName(targetTable.getTableReference()));
    } else {
        // If the table does not exist targetTable will be null.
        // Construct the table id if we can generate it. For error recording/logging.
        @Nullable String tableReferenceId = getTargetTableId(bqOptions);
        if (tableReferenceId != null) {
            readSessionBuilder.setTable(tableReferenceId);
        }
    }
    if (selectedFieldsProvider != null || rowRestrictionProvider != null) {
        ReadSession.TableReadOptions.Builder tableReadOptionsBuilder = ReadSession.TableReadOptions.newBuilder();
        if (selectedFieldsProvider != null) {
            tableReadOptionsBuilder.addAllSelectedFields(selectedFieldsProvider.get());
        }
        if (rowRestrictionProvider != null) {
            tableReadOptionsBuilder.setRowRestriction(rowRestrictionProvider.get());
        }
        readSessionBuilder.setReadOptions(tableReadOptionsBuilder);
    }
    if (format != null) {
        readSessionBuilder.setDataFormat(format);
    }
    int streamCount = 0;
    if (desiredBundleSizeBytes > 0) {
        long tableSizeBytes = (targetTable != null) ? targetTable.getNumBytes() : 0;
        streamCount = (int) Math.min(tableSizeBytes / desiredBundleSizeBytes, MAX_SPLIT_COUNT);
    }
    streamCount = Math.max(streamCount, MIN_SPLIT_COUNT);
    CreateReadSessionRequest createReadSessionRequest = CreateReadSessionRequest.newBuilder().setParent(BigQueryHelpers.toProjectResourceName(bqOptions.getBigQueryProject() == null ? bqOptions.getProject() : bqOptions.getBigQueryProject())).setReadSession(readSessionBuilder).setMaxStreamCount(streamCount).build();
    ReadSession readSession;
    try (StorageClient client = bqServices.getStorageClient(bqOptions)) {
        readSession = client.createReadSession(createReadSessionRequest);
        LOG.info("Sent BigQuery Storage API CreateReadSession request '{}'; received response '{}'.", createReadSessionRequest, readSession);
    }
    if (readSession.getStreamsList().isEmpty()) {
        // The underlying table is empty or all rows have been pruned.
        return ImmutableList.of();
    }
    Schema sessionSchema;
    if (readSession.getDataFormat() == DataFormat.ARROW) {
        org.apache.arrow.vector.types.pojo.Schema schema = ArrowConversion.arrowSchemaFromInput(readSession.getArrowSchema().getSerializedSchema().newInput());
        org.apache.beam.sdk.schemas.Schema beamSchema = ArrowConversion.ArrowSchemaTranslator.toBeamSchema(schema);
        sessionSchema = AvroUtils.toAvroSchema(beamSchema);
    } else if (readSession.getDataFormat() == DataFormat.AVRO) {
        sessionSchema = new Schema.Parser().parse(readSession.getAvroSchema().getSchema());
    } else {
        throw new IllegalArgumentException("data is not in a supported dataFormat: " + readSession.getDataFormat());
    }
    TableSchema trimmedSchema = BigQueryAvroUtils.trimBigQueryTableSchema(targetTable.getSchema(), sessionSchema);
    List<BigQueryStorageStreamSource<T>> sources = Lists.newArrayList();
    for (ReadStream readStream : readSession.getStreamsList()) {
        sources.add(BigQueryStorageStreamSource.create(readSession, readStream, trimmedSchema, parseFn, outputCoder, bqServices));
    }
    return ImmutableList.copyOf(sources);
}