Search in sources :

Example 1 with BigQueryTable

use of com.google.cloud.teleport.v2.values.BigQueryTable in project DataflowTemplates by GoogleCloudPlatform.

the class DataplexBigQueryToGcs method buildPipeline.

/**
 * Builds the pipeline with the supplied options.
 *
 * @param options The execution parameters to the pipeline.
 * @return The resulting pipeline.
 */
@VisibleForTesting
static Pipeline buildPipeline(DataplexBigQueryToGcsOptions options, BigQueryMetadataLoader metadataLoader, String targetRootPath, DatasetId datasetId) throws ExecutionException, InterruptedException {
    Pipeline pipeline = Pipeline.create(options);
    List<String> existingTargetFiles = StorageUtils.getFilesInDirectory(targetRootPath);
    LOG.info("Loading BigQuery metadata...");
    List<BigQueryTable> tables = metadataLoader.loadDatasetMetadata(datasetId, new DataplexBigQueryToGcsFilter(options, existingTargetFiles));
    LOG.info("Loaded {} table(s).", tables.size());
    if (!tables.isEmpty()) {
        transformPipeline(pipeline, tables, options, targetRootPath, null, null);
    } else {
        pipeline.apply("Nothing to export", new NoopTransform());
    }
    return pipeline;
}
Also used : NoopTransform(com.google.cloud.teleport.v2.transforms.NoopTransform) BigQueryTable(com.google.cloud.teleport.v2.values.BigQueryTable) DataplexBigQueryToGcsFilter(com.google.cloud.teleport.v2.utils.DataplexBigQueryToGcsFilter) Pipeline(org.apache.beam.sdk.Pipeline) VisibleForTesting(com.google.common.annotations.VisibleForTesting)

Example 2 with BigQueryTable

use of com.google.cloud.teleport.v2.values.BigQueryTable in project DataflowTemplates by GoogleCloudPlatform.

the class DeleteBigQueryDataFn method processElement.

@ProcessElement
public void processElement(@Element KV<BigQueryTable, BigQueryTablePartition> input, PipelineOptions options) {
    BigQueryTable t = input.getKey();
    BigQueryTablePartition p = input.getValue();
    if (t.isPartitioned() && p == null) {
        throw new IllegalStateException(String.format("No partition to delete provided for a partitioned table %s.", t.getTableName()));
    }
    if (!t.isPartitioned() && p != null) {
        throw new IllegalStateException(String.format("Got unexpected partition %s to delete for a non-partitioned table %s.", p.getPartitionName(), t.getTableName()));
    }
    if (!options.as(Options.class).getDeleteSourceData()) {
        if (t.isPartitioned()) {
            LOG.info("Skipping source BigQuery data deletion for partition {}${}.", t.getTableName(), p.getPartitionName());
        } else {
            LOG.info("Skipping source BigQuery data deletion for table {}.", t.getTableName());
        }
        return;
    }
    if (t.isPartitioned()) {
        deletePartition(t, p);
    } else {
        deleteTable(t);
    }
}
Also used : BigQueryTablePartition(com.google.cloud.teleport.v2.values.BigQueryTablePartition) BigQueryTable(com.google.cloud.teleport.v2.values.BigQueryTable)

Example 3 with BigQueryTable

use of com.google.cloud.teleport.v2.values.BigQueryTable in project DataflowTemplates by GoogleCloudPlatform.

the class DataplexBigQueryToGcsTest method setUp.

@Before
public void setUp() throws InterruptedException, IOException {
    options = TestPipeline.testingPipelineOptions().as(DataplexBigQueryToGcsOptions.class);
    options.setProject(PROJECT);
    options.setUpdateDataplexMetadata(true);
    options.setEnforceSamePartitionKey(false);
    // Required when using BigQueryIO.withMethod(EXPORT).
    options.setTempLocation(tmpDir.newFolder("bqTmp").getAbsolutePath());
    outDir = tmpDir.newFolder("out");
    bqSchema = new TableSchema().setFields(ImmutableList.of(new TableFieldSchema().setName("ts").setType("TIMESTAMP"), new TableFieldSchema().setName("s1").setType("STRING"), new TableFieldSchema().setName("d1").setType("DATE"), new TableFieldSchema().setName("i1").setType("INTEGER")));
    avroSchema = new Schema.Parser().parse("{\"type\":\"record\",\"name\":\"__root__\",\"fields\":" + "[{\"name\":\"ts\",\"type\":[\"null\",{\"type\":\"long\",\"logicalType\":\"timestamp-micros\"}]}," + "{\"name\":\"s1\",\"type\":[\"null\",\"string\"]}," + "{\"name\":\"d1\",\"type\":[\"null\",{\"type\":\"int\",\"logicalType\":\"date\"}]}," + "{\"name\":\"i1\",\"type\":[\"null\",\"long\"]}]}");
    long modTime = System.currentTimeMillis() * 1000;
    BigQueryTablePartition p1 = BigQueryTablePartition.builder().setPartitionName("p1").setLastModificationTime(modTime).build();
    BigQueryTablePartition p2 = BigQueryTablePartition.builder().setPartitionName("p2").setLastModificationTime(modTime).build();
    BigQueryTable t1 = BigQueryTable.builder().setTableName("partitioned_table").setProject(PROJECT).setDataset(DATASET).setSchema(avroSchema).setLastModificationTime(modTime).setPartitioningColumn("ts").setPartitions(Arrays.asList(p1, p2)).build();
    BigQueryTable t2 = BigQueryTable.builder().setTableName("unpartitioned_table").setProject(PROJECT).setDataset(DATASET).setSchema(avroSchema).setLastModificationTime(modTime).build();
    tableByName = new HashMap<>();
    tableByName.put(t1.getTableName(), t1);
    tableByName.put(t2.getTableName(), t2);
    defaultRecords = new TableRow[] { new TableRow().set("ts", 1L).set("s1", "1001").set("d1", "1970-01-01").set("i1", 2001L), new TableRow().set("ts", 2L).set("s1", "1002").set("d1", "1970-01-02").set("i1", 2002L), new TableRow().set("ts", 3L).set("s1", "1003").set("d1", "1970-01-03").set("i1", 2003L), new TableRow().set("ts", 4L).set("s1", "1004").set("d1", "1970-01-04").set("i1", null), new TableRow().set("ts", 5L).set("s1", "1005").set("d1", "1970-01-05").set("i1", 2005L) };
    defaultExpectedRecords = new String[] { "{\"ts\": 1, \"s1\": \"1001\", \"d1\": 0, \"i1\": 2001}", "{\"ts\": 2, \"s1\": \"1002\", \"d1\": 1, \"i1\": 2002}", "{\"ts\": 3, \"s1\": \"1003\", \"d1\": 2, \"i1\": 2003}", "{\"ts\": 4, \"s1\": \"1004\", \"d1\": 3, \"i1\": null}", "{\"ts\": 5, \"s1\": \"1005\", \"d1\": 4, \"i1\": 2005}" };
    FakeDatasetService.setUp();
    fakeDatasetService = new FakeDatasetService();
    fakeDatasetService.createDataset(PROJECT, DATASET, "", "", null);
    fakeDatasetService.createTable(new Table().setTableReference(t1.toTableReference()).setSchema(bqSchema).setRequirePartitionFilter(true).setTimePartitioning(new TimePartitioning().setField("ts").setType("DAY")));
    fakeDatasetService.createTable(new Table().setTableReference(t2.toTableReference()).setSchema(bqSchema));
    fakeJobService = new CustomFakeJobService();
    bqFakeServices = new FakeBigQueryServices().withJobService(fakeJobService).withDatasetService(fakeDatasetService);
    when(tableResultMock.iterateAll()).thenReturn(Collections.singleton(fields("unpartitioned_table", "0", null)));
    when(bqMock.query(any())).thenReturn(tableResultMock);
    when(bqMock.delete(any(TableId.class))).thenReturn(true);
    when(bqsMock.createReadSession(any())).thenReturn(ReadSession.newBuilder().setAvroSchema(AvroSchema.newBuilder().setSchema(avroSchema.toString())).build());
    metadataLoader = new BigQueryMetadataLoader(bqMock, bqsMock, MAX_PARALLEL_REQUESTS);
}
Also used : TableId(com.google.cloud.bigquery.TableId) BigQueryTablePartition(com.google.cloud.teleport.v2.values.BigQueryTablePartition) BigQueryTable(com.google.cloud.teleport.v2.values.BigQueryTable) Table(com.google.api.services.bigquery.model.Table) TableSchema(com.google.api.services.bigquery.model.TableSchema) BigQueryMetadataLoader(com.google.cloud.teleport.v2.utils.BigQueryMetadataLoader) TableFieldSchema(com.google.api.services.bigquery.model.TableFieldSchema) TimePartitioning(com.google.api.services.bigquery.model.TimePartitioning) FakeDatasetService(org.apache.beam.sdk.io.gcp.testing.FakeDatasetService) BigQueryTable(com.google.cloud.teleport.v2.values.BigQueryTable) TableRow(com.google.api.services.bigquery.model.TableRow) FakeBigQueryServices(org.apache.beam.sdk.io.gcp.testing.FakeBigQueryServices) DataplexBigQueryToGcsOptions(com.google.cloud.teleport.v2.options.DataplexBigQueryToGcsOptions) Before(org.junit.Before)

Example 4 with BigQueryTable

use of com.google.cloud.teleport.v2.values.BigQueryTable in project DataflowTemplates by GoogleCloudPlatform.

the class DeleteBigQueryDataFnTest method testTransform_withDeleteSourceDataEnabled_doesntTruncateSpecialPartitions.

/**
 * Test that DeleteBigQueryDataFn doesn't attempt to delete special BigQuery partitions even if
 * {@code deleteSourceData = true}.
 *
 * <p>As per <a
 * href="https://cloud.google.com/bigquery/docs/managing-partitioned-tables#delete_a_partition">
 * this documentation</a>, special partitions "__NULL__" and "__UNPARTITIONED__" cannot be
 * deleted.
 */
@Test
@Category(NeedsRunner.class)
public void testTransform_withDeleteSourceDataEnabled_doesntTruncateSpecialPartitions() {
    Options options = TestPipeline.testingPipelineOptions().as(Options.class);
    options.setDeleteSourceData(true);
    BigQueryTablePartition.Builder builder = BigQueryTablePartition.builder().setLastModificationTime(System.currentTimeMillis() * 1000);
    BigQueryTablePartition p1 = builder.setPartitionName("__NULL__").build();
    BigQueryTablePartition p2 = builder.setPartitionName("__UNPARTITIONED__").build();
    BigQueryTablePartition p3 = builder.setPartitionName("NORMAL_PARTITION").build();
    BigQueryTable t1 = table.toBuilder().setPartitions(Arrays.asList(p1, p2, p3)).setPartitioningColumn("column-name-doesnt-matter").build();
    DeleteBigQueryDataFn fn = new DeleteBigQueryDataFn().withTestBqClientFactory(() -> bqMock);
    testPipeline.apply("CreateInput", Create.of(KV.of(t1, p1), KV.of(t1, p2), KV.of(t1, p3)).withCoder(fnCoder)).apply("TestDeleteBigQueryDataFn", ParDo.of(fn));
    testPipeline.run(options);
    verify(bqMock, times(1)).delete(TableId.of("pr1", "d1", "t1$NORMAL_PARTITION"));
    verifyNoMoreInteractions(bqMock);
}
Also used : Options(com.google.cloud.teleport.v2.transforms.DeleteBigQueryDataFn.Options) BigQueryTablePartition(com.google.cloud.teleport.v2.values.BigQueryTablePartition) BigQueryTable(com.google.cloud.teleport.v2.values.BigQueryTable) Category(org.junit.experimental.categories.Category) Test(org.junit.Test)

Example 5 with BigQueryTable

use of com.google.cloud.teleport.v2.values.BigQueryTable in project DataflowTemplates by GoogleCloudPlatform.

the class DataplexBigQueryToGcs method transformPipeline.

@VisibleForTesting
static void transformPipeline(Pipeline pipeline, List<BigQueryTable> tables, DataplexBigQueryToGcsOptions options, String targetRootPath, BigQueryServices testBqServices, BigQueryClientFactory testBqClientFactory) {
    List<PCollection<KV<BigQueryTable, KV<BigQueryTablePartition, String>>>> fileCollections = new ArrayList<>(tables.size());
    tables.forEach(table -> {
        fileCollections.add(pipeline.apply(String.format("ExportTable-%s", table.getTableName()), new BigQueryTableToGcsTransform(table, targetRootPath, options.getFileFormat(), options.getFileCompression(), options.getEnforceSamePartitionKey()).withTestServices(testBqServices)).apply(String.format("AttachTableKeys-%s", table.getTableName()), WithKeys.of(table)));
    });
    PCollection<KV<BigQueryTable, KV<BigQueryTablePartition, String>>> exportFileResults = PCollectionList.of(fileCollections).apply("FlattenTableResults", Flatten.pCollections());
    PCollection<Void> metadataUpdateResults = exportFileResults.apply("UpdateDataplexMetadata", new UpdateDataplexBigQueryToGcsExportMetadataTransform());
    exportFileResults.apply(MapElements.into(TypeDescriptors.kvs(TypeDescriptor.of(BigQueryTable.class), TypeDescriptor.of(BigQueryTablePartition.class))).via((SerializableFunction<KV<BigQueryTable, KV<BigQueryTablePartition, String>>, KV<BigQueryTable, BigQueryTablePartition>>) input -> KV.of(input.getKey(), input.getValue().getKey()))).apply("WaitForMetadataUpdate", Wait.on(metadataUpdateResults)).apply("TruncateBigQueryData", ParDo.of(new DeleteBigQueryDataFn().withTestBqClientFactory(testBqClientFactory)));
}
Also used : BigQueryTablePartition(com.google.cloud.teleport.v2.values.BigQueryTablePartition) SerializableFunction(org.apache.beam.sdk.transforms.SerializableFunction) ArrayList(java.util.ArrayList) DeleteBigQueryDataFn(com.google.cloud.teleport.v2.transforms.DeleteBigQueryDataFn) KV(org.apache.beam.sdk.values.KV) PCollection(org.apache.beam.sdk.values.PCollection) BigQueryTable(com.google.cloud.teleport.v2.values.BigQueryTable) UpdateDataplexBigQueryToGcsExportMetadataTransform(com.google.cloud.teleport.v2.transforms.UpdateDataplexBigQueryToGcsExportMetadataTransform) BigQueryTableToGcsTransform(com.google.cloud.teleport.v2.transforms.BigQueryTableToGcsTransform) VisibleForTesting(com.google.common.annotations.VisibleForTesting)

Aggregations

BigQueryTable (com.google.cloud.teleport.v2.values.BigQueryTable)7 BigQueryTablePartition (com.google.cloud.teleport.v2.values.BigQueryTablePartition)4 Options (com.google.cloud.teleport.v2.transforms.DeleteBigQueryDataFn.Options)2 VisibleForTesting (com.google.common.annotations.VisibleForTesting)2 ArrayList (java.util.ArrayList)2 Test (org.junit.Test)2 Category (org.junit.experimental.categories.Category)2 Table (com.google.api.services.bigquery.model.Table)1 TableFieldSchema (com.google.api.services.bigquery.model.TableFieldSchema)1 TableRow (com.google.api.services.bigquery.model.TableRow)1 TableSchema (com.google.api.services.bigquery.model.TableSchema)1 TimePartitioning (com.google.api.services.bigquery.model.TimePartitioning)1 TableId (com.google.cloud.bigquery.TableId)1 TableResult (com.google.cloud.bigquery.TableResult)1 DataplexBigQueryToGcsOptions (com.google.cloud.teleport.v2.options.DataplexBigQueryToGcsOptions)1 BigQueryTableToGcsTransform (com.google.cloud.teleport.v2.transforms.BigQueryTableToGcsTransform)1 DeleteBigQueryDataFn (com.google.cloud.teleport.v2.transforms.DeleteBigQueryDataFn)1 NoopTransform (com.google.cloud.teleport.v2.transforms.NoopTransform)1 UpdateDataplexBigQueryToGcsExportMetadataTransform (com.google.cloud.teleport.v2.transforms.UpdateDataplexBigQueryToGcsExportMetadataTransform)1 BigQueryMetadataLoader (com.google.cloud.teleport.v2.utils.BigQueryMetadataLoader)1