Examples with ETLConfig io.cdap.cdap.etl.proto.v2.ETLConfig used on opensource projects

Example 66 with ETLConfig

use of io.cdap.cdap.etl.proto.v2.ETLConfig in project cdap by caskdata.

the class DataPipelineTest method testMultipleJoiner.

private void testMultipleJoiner(Engine engine) throws Exception {
    /*
     * source1 ----> t1 ------
     *                        | --> innerjoin ----> t4 ------
     * source2 ----> t2 ------                                 |
     *                                                         | ---> outerjoin --> sink1
     *                                                         |
     * source3 -------------------- t3 ------------------------
     */
    Schema inputSchema1 = Schema.recordOf("customerRecord", Schema.Field.of("customer_id", Schema.of(Schema.Type.STRING)), Schema.Field.of("customer_name", Schema.of(Schema.Type.STRING)));
    Schema inputSchema2 = Schema.recordOf("itemRecord", Schema.Field.of("item_id", Schema.of(Schema.Type.STRING)), Schema.Field.of("item_price", Schema.of(Schema.Type.LONG)), Schema.Field.of("cust_id", Schema.of(Schema.Type.STRING)), Schema.Field.of("cust_name", Schema.of(Schema.Type.STRING)));
    Schema inputSchema3 = Schema.recordOf("transactionRecord", Schema.Field.of("t_id", Schema.of(Schema.Type.STRING)), Schema.Field.of("c_id", Schema.of(Schema.Type.STRING)), Schema.Field.of("i_id", Schema.of(Schema.Type.STRING)));
    Schema outSchema2 = Schema.recordOf("join.output", Schema.Field.of("t_id", Schema.nullableOf(Schema.of(Schema.Type.STRING))), Schema.Field.of("c_id", Schema.nullableOf(Schema.of(Schema.Type.STRING))), Schema.Field.of("i_id", Schema.nullableOf(Schema.of(Schema.Type.STRING))), Schema.Field.of("customer_id", Schema.nullableOf(Schema.of(Schema.Type.STRING))), Schema.Field.of("customer_name", Schema.nullableOf(Schema.of(Schema.Type.STRING))), Schema.Field.of("item_id", Schema.nullableOf(Schema.of(Schema.Type.STRING))), Schema.Field.of("item_price", Schema.nullableOf(Schema.of(Schema.Type.LONG))), Schema.Field.of("cust_id", Schema.nullableOf(Schema.of(Schema.Type.STRING))), Schema.Field.of("cust_name", Schema.nullableOf(Schema.of(Schema.Type.STRING))));
    String source1MulitJoinInput = "multiJoinSource1-" + engine;
    String source2MultiJoinInput = "multiJoinSource2-" + engine;
    String source3MultiJoinInput = "multiJoinSource3-" + engine;
    String outputName = "multiJoinOutput-" + engine;
    String sinkName = "multiJoinOutputSink-" + engine;
    String outerJoinName = "multiJoinOuter-" + engine;
    ETLBatchConfig etlConfig = ETLBatchConfig.builder().addStage(new ETLStage("source1", MockSource.getPlugin(source1MulitJoinInput, inputSchema1))).addStage(new ETLStage("source2", MockSource.getPlugin(source2MultiJoinInput, inputSchema2))).addStage(new ETLStage("source3", MockSource.getPlugin(source3MultiJoinInput, inputSchema3))).addStage(new ETLStage("t1", IdentityTransform.getPlugin())).addStage(new ETLStage("t2", IdentityTransform.getPlugin())).addStage(new ETLStage("t3", IdentityTransform.getPlugin())).addStage(new ETLStage("t4", IdentityTransform.getPlugin())).addStage(new ETLStage("innerjoin", MockJoiner.getPlugin("t1.customer_id=t2.cust_id", "t1,t2", ""))).addStage(new ETLStage(outerJoinName, MockJoiner.getPlugin("t4.item_id=t3.i_id", "", ""))).addStage(new ETLStage(sinkName, MockSink.getPlugin(outputName))).addConnection("source1", "t1").addConnection("source2", "t2").addConnection("source3", "t3").addConnection("t1", "innerjoin").addConnection("t2", "innerjoin").addConnection("innerjoin", "t4").addConnection("t3", outerJoinName).addConnection("t4", outerJoinName).addConnection(outerJoinName, sinkName).setEngine(engine).build();
    AppRequest<ETLBatchConfig> appRequest = new AppRequest<>(APP_ARTIFACT, etlConfig);
    ApplicationId appId = NamespaceId.DEFAULT.app("JoinerApp-" + engine);
    ApplicationManager appManager = deployApplication(appId, appRequest);
    StructuredRecord recordSamuel = StructuredRecord.builder(inputSchema1).set("customer_id", "1").set("customer_name", "samuel").build();
    StructuredRecord recordBob = StructuredRecord.builder(inputSchema1).set("customer_id", "2").set("customer_name", "bob").build();
    StructuredRecord recordJane = StructuredRecord.builder(inputSchema1).set("customer_id", "3").set("customer_name", "jane").build();
    StructuredRecord recordCar = StructuredRecord.builder(inputSchema2).set("item_id", "11").set("item_price", 10000L).set("cust_id", "1").set("cust_name", "samuel").build();
    StructuredRecord recordBike = StructuredRecord.builder(inputSchema2).set("item_id", "22").set("item_price", 100L).set("cust_id", "3").set("cust_name", "jane").build();
    StructuredRecord recordTrasCar = StructuredRecord.builder(inputSchema3).set("t_id", "1").set("c_id", "1").set("i_id", "11").build();
    StructuredRecord recordTrasBike = StructuredRecord.builder(inputSchema3).set("t_id", "2").set("c_id", "3").set("i_id", "22").build();
    StructuredRecord recordTrasPlane = StructuredRecord.builder(inputSchema3).set("t_id", "3").set("c_id", "4").set("i_id", "33").build();
    // write one record to each source
    DataSetManager<Table> inputManager = getDataset(NamespaceId.DEFAULT.dataset(source1MulitJoinInput));
    MockSource.writeInput(inputManager, ImmutableList.of(recordSamuel, recordBob, recordJane));
    inputManager = getDataset(NamespaceId.DEFAULT.dataset(source2MultiJoinInput));
    MockSource.writeInput(inputManager, ImmutableList.of(recordCar, recordBike));
    inputManager = getDataset(NamespaceId.DEFAULT.dataset(source3MultiJoinInput));
    MockSource.writeInput(inputManager, ImmutableList.of(recordTrasCar, recordTrasBike, recordTrasPlane));
    WorkflowManager workflowManager = appManager.getWorkflowManager(SmartWorkflow.NAME);
    workflowManager.start();
    workflowManager.waitForRun(ProgramRunStatus.COMPLETED, 5, TimeUnit.MINUTES);
    StructuredRecord joinRecordSamuel = StructuredRecord.builder(outSchema2).set("customer_id", "1").set("customer_name", "samuel").set("item_id", "11").set("item_price", 10000L).set("cust_id", "1").set("cust_name", "samuel").set("t_id", "1").set("c_id", "1").set("i_id", "11").build();
    StructuredRecord joinRecordJane = StructuredRecord.builder(outSchema2).set("customer_id", "3").set("customer_name", "jane").set("item_id", "22").set("item_price", 100L).set("cust_id", "3").set("cust_name", "jane").set("t_id", "2").set("c_id", "3").set("i_id", "22").build();
    StructuredRecord joinRecordPlane = StructuredRecord.builder(outSchema2).set("t_id", "3").set("c_id", "4").set("i_id", "33").build();
    DataSetManager<Table> sinkManager = getDataset(outputName);
    Set<StructuredRecord> expected = ImmutableSet.of(joinRecordSamuel, joinRecordJane, joinRecordPlane);
    Set<StructuredRecord> actual = Sets.newHashSet(MockSink.readOutput(sinkManager));
    Assert.assertEquals(expected, actual);
    validateMetric(3, appId, outerJoinName + ".records.out");
    validateMetric(3, appId, sinkName + ".records.in");
}

Example 67 with ETLConfig

use of io.cdap.cdap.etl.proto.v2.ETLConfig in project cdap by caskdata.

the class DataPipelineTest method testRuntimeArgs.

private void testRuntimeArgs(Engine engine) throws Exception {
    String sourceName = "runtimeArgInput-" + engine;
    String sinkName = "runtimeArgOutput-" + engine;
    ETLBatchConfig etlConfig = ETLBatchConfig.builder().addStage(new ETLStage("action", MockAction.getPlugin("dumy", "val", "ue", "dwayne"))).addStage(new ETLStage("source", MockSource.getPlugin(sourceName))).addStage(new ETLStage("filter", StringValueFilterTransform.getPlugin("name", "samuel"))).addStage(new ETLStage("sink", MockSink.getPlugin(sinkName))).addConnection("action", "source").addConnection("source", "filter").addConnection("filter", "sink").setEngine(engine).build();
    AppRequest<ETLBatchConfig> appRequest = new AppRequest<>(APP_ARTIFACT, etlConfig);
    ApplicationId appId = NamespaceId.DEFAULT.app("RuntimeArgApp-" + engine);
    ApplicationManager appManager = deployApplication(appId, appRequest);
    // there should be only two programs - one workflow and one mapreduce/spark
    Schema schema = Schema.recordOf("testRecord", Schema.Field.of("name", Schema.of(Schema.Type.STRING)));
    StructuredRecord recordSamuel = StructuredRecord.builder(schema).set("name", "samuel").build();
    StructuredRecord recordDwayne = StructuredRecord.builder(schema).set("name", "dwayne").build();
    // write one record to each source
    DataSetManager<Table> inputManager = getDataset(sourceName);
    MockSource.writeInput(inputManager, ImmutableList.of(recordSamuel, recordDwayne));
    WorkflowManager workflowManager = appManager.getWorkflowManager(SmartWorkflow.NAME);
    workflowManager.start();
    workflowManager.waitForRun(ProgramRunStatus.COMPLETED, 5, TimeUnit.MINUTES);
    // check sink
    DataSetManager<Table> sinkManager = getDataset(sinkName);
    Set<StructuredRecord> expected = ImmutableSet.of(recordSamuel);
    Set<StructuredRecord> actual = Sets.newHashSet(MockSink.readOutput(sinkManager));
    Assert.assertEquals(expected, actual);
}

Example 68 with ETLConfig

use of io.cdap.cdap.etl.proto.v2.ETLConfig in project cdap by caskdata.

the class PreviewDataPipelineTest method testDataPipelinePreviewRun.

private void testDataPipelinePreviewRun(Engine engine) throws Exception {
    PreviewManager previewManager = getPreviewManager();
    String sourceTableName = "singleInput";
    String sinkTableName = "singleOutput";
    Schema schema = Schema.recordOf("testRecord", Schema.Field.of("name", Schema.of(Schema.Type.STRING)));
    /*
     * source --> transform -> sink
     */
    ETLBatchConfig etlConfig = ETLBatchConfig.builder().addStage(new ETLStage("source", MockSource.getPlugin(sourceTableName, schema))).addStage(new ETLStage("transform", IdentityTransform.getPlugin())).addStage(new ETLStage("sink", MockSink.getPlugin(sinkTableName))).addConnection("source", "transform").addConnection("transform", "sink").setEngine(engine).setNumOfRecordsPreview(100).build();
    // Construct the preview config with the program name and program type
    PreviewConfig previewConfig = new PreviewConfig(SmartWorkflow.NAME, ProgramType.WORKFLOW, Collections.<String, String>emptyMap(), 10);
    // Create the table for the mock source
    addDatasetInstance(Table.class.getName(), sourceTableName, DatasetProperties.of(ImmutableMap.of("schema", schema.toString())));
    DataSetManager<Table> inputManager = getDataset(NamespaceId.DEFAULT.dataset(sourceTableName));
    StructuredRecord recordSamuel = StructuredRecord.builder(schema).set("name", "samuel").build();
    StructuredRecord recordBob = StructuredRecord.builder(schema).set("name", "bob").build();
    MockSource.writeInput(inputManager, ImmutableList.of(recordSamuel, recordBob));
    AppRequest<ETLBatchConfig> appRequest = new AppRequest<>(APP_ARTIFACT_RANGE, etlConfig, previewConfig);
    // Start the preview and get the corresponding PreviewRunner.
    ApplicationId previewId = previewManager.start(NamespaceId.DEFAULT, appRequest);
    // Wait for the preview status go into COMPLETED.
    Tasks.waitFor(PreviewStatus.Status.COMPLETED, new Callable<PreviewStatus.Status>() {

        @Override
        public PreviewStatus.Status call() throws Exception {
            PreviewStatus status = previewManager.getStatus(previewId);
            return status == null ? null : status.getStatus();
        }
    }, 5, TimeUnit.MINUTES);
    // Get the data for stage "source" in the PreviewStore, should contain two records.
    checkPreviewStore(previewManager, previewId, "source", 2);
    // Get the data for stage "transform" in the PreviewStore, should contain two records.
    checkPreviewStore(previewManager, previewId, "transform", 2);
    // Get the data for stage "sink" in the PreviewStore, should contain two records.
    checkPreviewStore(previewManager, previewId, "sink", 2);
    // Validate the metrics for preview
    validateMetric(2, previewId, "source.records.in", previewManager);
    validateMetric(2, previewId, "source.records.out", previewManager);
    validateMetric(2, previewId, "transform.records.in", previewManager);
    validateMetric(2, previewId, "transform.records.out", previewManager);
    validateMetric(2, previewId, "sink.records.out", previewManager);
    validateMetric(2, previewId, "sink.records.in", previewManager);
    // Check the sink table is not created in the real space.
    DataSetManager<Table> sinkManager = getDataset(sinkTableName);
    Assert.assertNull(sinkManager.get());
    deleteDatasetInstance(NamespaceId.DEFAULT.dataset(sourceTableName));
    Assert.assertNotNull(previewManager.getRunId(previewId));
}

Example 69 with ETLConfig

use of io.cdap.cdap.etl.proto.v2.ETLConfig in project cdap by caskdata.

the class PreviewDataPipelineTest method testPreviewFailedRun.

private void testPreviewFailedRun(Engine engine) throws Exception {
    PreviewManager previewManager = getPreviewManager();
    String sourceTableName = "singleInput";
    String sinkTableName = "singleOutput";
    Schema schema = Schema.recordOf("testRecord", Schema.Field.of("name", Schema.of(Schema.Type.STRING)));
    /*
     * source --> transform -> sink
     */
    ETLBatchConfig etlConfig = ETLBatchConfig.builder().addStage(new ETLStage("source", MockSource.getPlugin(sourceTableName, schema))).addStage(new ETLStage("transform", ExceptionTransform.getPlugin("name", "samuel"))).addStage(new ETLStage("sink", MockSink.getPlugin(sinkTableName))).addConnection("source", "transform").addConnection("transform", "sink").setNumOfRecordsPreview(100).setEngine(engine).build();
    // Construct the preview config with the program name and program type.
    PreviewConfig previewConfig = new PreviewConfig(SmartWorkflow.NAME, ProgramType.WORKFLOW, Collections.<String, String>emptyMap(), 10);
    // Create the table for the mock source
    addDatasetInstance(Table.class.getName(), sourceTableName, DatasetProperties.of(ImmutableMap.of("schema", schema.toString())));
    DataSetManager<Table> inputManager = getDataset(NamespaceId.DEFAULT.dataset(sourceTableName));
    StructuredRecord recordSamuel = StructuredRecord.builder(schema).set("name", "samuel").build();
    StructuredRecord recordBob = StructuredRecord.builder(schema).set("name", "bob").build();
    MockSource.writeInput(inputManager, "1", recordSamuel);
    MockSource.writeInput(inputManager, "2", recordBob);
    AppRequest<ETLBatchConfig> appRequest = new AppRequest<>(APP_ARTIFACT, etlConfig, previewConfig);
    // Start the preview and get the corresponding PreviewRunner.
    ApplicationId previewId = previewManager.start(NamespaceId.DEFAULT, appRequest);
    // Wait for the preview status go into FAILED.
    Tasks.waitFor(PreviewStatus.Status.RUN_FAILED, new Callable<PreviewStatus.Status>() {

        @Override
        public PreviewStatus.Status call() throws Exception {
            PreviewStatus status = previewManager.getStatus(previewId);
            return status == null ? null : status.getStatus();
        }
    }, 5, TimeUnit.MINUTES);
    // Get the data for stage "source" in the PreviewStore.
    checkPreviewStore(previewManager, previewId, "source", 2);
    // Get the data for stage "transform" in the PreviewStore, should contain one less record than source.
    checkPreviewStore(previewManager, previewId, "transform", 1);
    // Get the data for stage "sink" in the PreviewStore, should contain one less record than source.
    checkPreviewStore(previewManager, previewId, "sink", 1);
    // Validate the metrics for preview
    validateMetric(2, previewId, "source.records.in", previewManager);
    validateMetric(2, previewId, "source.records.out", previewManager);
    validateMetric(2, previewId, "transform.records.in", previewManager);
    validateMetric(1, previewId, "transform.records.out", previewManager);
    validateMetric(1, previewId, "sink.records.out", previewManager);
    validateMetric(1, previewId, "sink.records.in", previewManager);
    // Check the sink table is not created in the real space.
    DataSetManager<Table> sinkManager = getDataset(sinkTableName);
    Assert.assertNull(sinkManager.get());
    deleteDatasetInstance(NamespaceId.DEFAULT.dataset(sourceTableName));
}

Example 70 with ETLConfig

use of io.cdap.cdap.etl.proto.v2.ETLConfig in project cdap by caskdata.

the class PreviewDataPipelineTest method testDataPipelinePreviewStop.

private void testDataPipelinePreviewStop(Engine engine, @Nullable Long sleepInMillis) throws Exception {
    PreviewManager previewManager = getPreviewManager();
    String sourceTableName = "singleInput";
    String sinkTableName = "singleOutput";
    Schema schema = Schema.recordOf("testRecord", Schema.Field.of("name", Schema.of(Schema.Type.STRING)));
    /*
     * source --> transform -> sink
     */
    ETLPlugin sourcePlugin = sleepInMillis == null ? MockSource.getPlugin(sourceTableName, schema) : MockSource.getPlugin(sourceTableName, schema, sleepInMillis);
    ETLBatchConfig etlConfig = ETLBatchConfig.builder().addStage(new ETLStage("source", sourcePlugin)).addStage(new ETLStage("transform", IdentityTransform.getPlugin())).addStage(new ETLStage("sink", MockSink.getPlugin(sinkTableName))).addConnection("source", "transform").addConnection("transform", "sink").setEngine(engine).setNumOfRecordsPreview(100).build();
    // Construct the preview config with the program name and program type
    PreviewConfig previewConfig = new PreviewConfig(SmartWorkflow.NAME, ProgramType.WORKFLOW, Collections.<String, String>emptyMap(), 10);
    // Create the table for the mock source
    addDatasetInstance(Table.class.getName(), sourceTableName, DatasetProperties.of(ImmutableMap.of("schema", schema.toString())));
    DataSetManager<Table> inputManager = getDataset(NamespaceId.DEFAULT.dataset(sourceTableName));
    StructuredRecord recordSamuel = StructuredRecord.builder(schema).set("name", "samuel").build();
    StructuredRecord recordBob = StructuredRecord.builder(schema).set("name", "bob").build();
    MockSource.writeInput(inputManager, ImmutableList.of(recordSamuel, recordBob));
    AppRequest<ETLBatchConfig> appRequest = new AppRequest<>(APP_ARTIFACT_RANGE, etlConfig, previewConfig);
    // Start the preview and get the corresponding PreviewRunner.
    ApplicationId previewId = previewManager.start(NamespaceId.DEFAULT, appRequest);
    if (sleepInMillis != null) {
        // Wait for the preview status go into RUNNING.
        Tasks.waitFor(PreviewStatus.Status.RUNNING, () -> {
            PreviewStatus status = previewManager.getStatus(previewId);
            return status == null ? null : status.getStatus();
        }, 5, TimeUnit.MINUTES);
    }
    previewManager.stopPreview(previewId);
    // Wait for the preview status go into KILLED.
    Tasks.waitFor(PreviewStatus.Status.KILLED, () -> {
        PreviewStatus status = previewManager.getStatus(previewId);
        return status == null ? null : status.getStatus();
    }, 5, TimeUnit.MINUTES);
    // Check the sink table is not created in the real space.
    DataSetManager<Table> sinkManager = getDataset(sinkTableName);
    Assert.assertNull(sinkManager.get());
    deleteDatasetInstance(NamespaceId.DEFAULT.dataset(sourceTableName));
    Assert.assertTrue(previewManager.getRunId(previewId) == null || sleepInMillis != null);
}