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

Example 46 with ETLBatchConfig

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

the class PipelineSpecGeneratorTest method testConflictingPipelineProperties.

@Test(expected = IllegalArgumentException.class)
public void testConflictingPipelineProperties() throws ValidationException {
    // populate some mock plugins.
    MockPluginConfigurer pluginConfigurer = new MockPluginConfigurer();
    Set<ArtifactId> artifactIds = ImmutableSet.of(ARTIFACT_ID);
    pluginConfigurer.addMockPlugin(Action.PLUGIN_TYPE, "action1", MockPlugin.builder().putPipelineProperty("prop1", "val1").build(), artifactIds);
    pluginConfigurer.addMockPlugin(Action.PLUGIN_TYPE, "action2", MockPlugin.builder().putPipelineProperty("prop1", "val2").build(), artifactIds);
    Map<String, String> empty = ImmutableMap.of();
    ETLBatchConfig config = ETLBatchConfig.builder().setTimeSchedule("* * * * *").addStage(new ETLStage("a1", new ETLPlugin("action1", Action.PLUGIN_TYPE, empty))).addStage(new ETLStage("a2", new ETLPlugin("action2", Action.PLUGIN_TYPE, empty))).addConnection("a1", "a2").setEngine(Engine.MAPREDUCE).build();
    new BatchPipelineSpecGenerator(NamespaceId.DEFAULT.getNamespace(), pluginConfigurer, null, ImmutableSet.of(BatchSource.PLUGIN_TYPE), ImmutableSet.of(BatchSink.PLUGIN_TYPE), Engine.MAPREDUCE, MOCK_FEATURE_FLAGS_PROVIDER).generateSpec(config);
}

Example 47 with ETLBatchConfig

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

the class PipelineSpecGeneratorTest method testSimpleConditionConnectionWithMultipleTrueBranches.

@Test(expected = IllegalArgumentException.class)
public void testSimpleConditionConnectionWithMultipleTrueBranches() throws ValidationException {
    ETLBatchConfig etlConfig = ETLBatchConfig.builder().setTimeSchedule("* * * * *").addStage(new ETLStage("source", MOCK_SOURCE)).addStage(new ETLStage("condition", MOCK_CONDITION)).addStage(new ETLStage("t1", MOCK_TRANSFORM_A)).addStage(new ETLStage("t2", MOCK_TRANSFORM_B)).addStage(new ETLStage("sink1", MOCK_SINK)).addStage(new ETLStage("sink2", MOCK_SINK)).addConnection("source", "condition").addConnection("condition", "t1", true).addConnection("condition", "t2", true).addConnection("t1", "sink1").addConnection("t2", "sink2").build();
    specGenerator.generateSpec(etlConfig);
}

Example 48 with ETLBatchConfig

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

the class PipelineSpecGeneratorTest method testActionInPipelineMiddle.

@Test
public void testActionInPipelineMiddle() throws ValidationException {
    /*
     * source1 --> sink1 --> action --> source2 --> sink2
     */
    ETLBatchConfig config = ETLBatchConfig.builder().addStage(new ETLStage("source1", MOCK_SOURCE)).addStage(new ETLStage("source2", MOCK_SOURCE)).addStage(new ETLStage("sink1", MOCK_SINK)).addStage(new ETLStage("sink2", MOCK_SINK)).addStage(new ETLStage("action", MOCK_ACTION)).addConnection("source1", "sink1").addConnection("sink1", "action").addConnection("action", "source2").addConnection("source2", "sink2").build();
    try {
        specGenerator.generateSpec(config);
        Assert.fail("Did not fail a pipeline with an action in the middle");
    } catch (IllegalArgumentException e) {
    // expected
    }
    /*
     * action1 --> source --> action2 --> sink --> action3
     */
    config = ETLBatchConfig.builder().addStage(new ETLStage("source", MOCK_SOURCE)).addStage(new ETLStage("sink", MOCK_SINK)).addStage(new ETLStage("action1", MOCK_ACTION)).addStage(new ETLStage("action2", MOCK_ACTION)).addStage(new ETLStage("action3", MOCK_ACTION)).addConnection("action1", "source").addConnection("source", "action2").addConnection("action2", "sink").addConnection("sink", "action3").build();
    try {
        specGenerator.generateSpec(config);
        Assert.fail("Did not fail a pipeline with an action in the middle");
    } catch (IllegalArgumentException e) {
    // expected
    }
}

Example 49 with ETLBatchConfig

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

the class DataPipelineTest method testOuterJoin.

private void testOuterJoin(Engine engine) throws Exception {
    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("c_name", Schema.of(Schema.Type.STRING)));
    String input1Name = "source1OuterJoinInput-" + engine;
    String input2Name = "source2OuterJoinInput-" + engine;
    String input3Name = "source3OuterJoinInput-" + engine;
    String outputName = "outerJoinOutput-" + engine;
    String joinerName = "outerJoiner-" + engine;
    String sinkName = "outerJoinSink-" + engine;
    ETLBatchConfig etlConfig = ETLBatchConfig.builder().addStage(new ETLStage("source1", MockSource.getPlugin(input1Name, inputSchema1))).addStage(new ETLStage("source2", MockSource.getPlugin(input2Name, inputSchema2))).addStage(new ETLStage("source3", MockSource.getPlugin(input3Name, inputSchema3))).addStage(new ETLStage("t1", IdentityTransform.getPlugin())).addStage(new ETLStage("t2", IdentityTransform.getPlugin())).addStage(new ETLStage("t3", IdentityTransform.getPlugin())).addStage(new ETLStage(joinerName, MockJoiner.getPlugin("t1.customer_id=t2.cust_id=t3.c_id&" + "t1.customer_name=t2.cust_name=t3.c_name", "t1", ""))).addStage(new ETLStage(sinkName, MockSink.getPlugin(outputName))).addConnection("source1", "t1").addConnection("source2", "t2").addConnection("source3", "t3").addConnection("t1", joinerName).addConnection("t2", joinerName).addConnection("t3", joinerName).addConnection(joinerName, sinkName).setEngine(engine).build();
    AppRequest<ETLBatchConfig> appRequest = new AppRequest<>(APP_ARTIFACT, etlConfig);
    ApplicationId appId = NamespaceId.DEFAULT.app("OuterJoinApp-" + engine);
    ApplicationManager appManager = deployApplication(appId, appRequest);
    Schema outSchema = Schema.recordOf("join.output", Schema.Field.of("customer_id", Schema.of(Schema.Type.STRING)), Schema.Field.of("customer_name", 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))), 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("c_name", Schema.nullableOf(Schema.of(Schema.Type.STRING))));
    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 recordMartha = StructuredRecord.builder(inputSchema1).set("customer_id", "4").set("customer_name", "martha").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("c_name", "samuel").build();
    StructuredRecord recordTrasPlane = StructuredRecord.builder(inputSchema3).set("t_id", "2").set("c_id", "2").set("c_name", "bob").build();
    StructuredRecord recordTrasBike = StructuredRecord.builder(inputSchema3).set("t_id", "3").set("c_id", "3").set("c_name", "jane").build();
    // write one record to each source
    DataSetManager<Table> inputManager = getDataset(NamespaceId.DEFAULT.dataset(input1Name));
    MockSource.writeInput(inputManager, ImmutableList.of(recordSamuel, recordBob, recordJane, recordMartha));
    inputManager = getDataset(NamespaceId.DEFAULT.dataset(input2Name));
    MockSource.writeInput(inputManager, ImmutableList.of(recordCar, recordBike));
    inputManager = getDataset(NamespaceId.DEFAULT.dataset(input3Name));
    MockSource.writeInput(inputManager, ImmutableList.of(recordTrasCar, recordTrasPlane, recordTrasBike));
    WorkflowManager workflowManager = appManager.getWorkflowManager(SmartWorkflow.NAME);
    workflowManager.start();
    workflowManager.waitForRun(ProgramRunStatus.COMPLETED, 5, TimeUnit.MINUTES);
    StructuredRecord joinRecordSamuel = StructuredRecord.builder(outSchema).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("c_name", "samuel").build();
    StructuredRecord joinRecordBob = StructuredRecord.builder(outSchema).set("customer_id", "2").set("customer_name", "bob").set("t_id", "2").set("c_id", "2").set("c_name", "bob").build();
    StructuredRecord joinRecordJane = StructuredRecord.builder(outSchema).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", "3").set("c_id", "3").set("c_name", "jane").build();
    StructuredRecord joinRecordMartha = StructuredRecord.builder(outSchema).set("customer_id", "4").set("customer_name", "martha").build();
    DataSetManager<Table> sinkManager = getDataset(outputName);
    Set<StructuredRecord> expected = ImmutableSet.of(joinRecordSamuel, joinRecordJane, joinRecordBob, joinRecordMartha);
    Set<StructuredRecord> actual = Sets.newHashSet(MockSink.readOutput(sinkManager));
    Assert.assertEquals(expected, actual);
    validateMetric(4, appId, joinerName + ".records.out");
    validateMetric(4, appId, sinkName + ".records.in");
}

Example 50 with ETLBatchConfig

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

the class DataPipelineTest method runHeadTriggeringPipeline.

private void runHeadTriggeringPipeline(Engine engine, String expectedValue1, String expectedValue2) throws Exception {
    // set runtime arguments
    Map<String, String> runtimeArguments = ImmutableMap.of("head-arg", expectedValue1);
    ETLStage action1 = new ETLStage("action1", MockAction.getPlugin("actionTable", "action1.row", "action1.column", expectedValue2));
    ETLBatchConfig etlConfig = io.cdap.cdap.etl.proto.v2.ETLBatchConfig.builder().addStage(action1).setEngine(engine).build();
    AppRequest<io.cdap.cdap.etl.proto.v2.ETLBatchConfig> appRequest = new AppRequest<>(APP_ARTIFACT, etlConfig);
    ApplicationId appId = NamespaceId.DEFAULT.app("head");
    ApplicationManager appManager = deployApplication(appId, appRequest);
    WorkflowManager manager = appManager.getWorkflowManager(SmartWorkflow.NAME);
    manager.setRuntimeArgs(runtimeArguments);
    manager.start(ImmutableMap.of("logical.start.time", "0"));
    manager.waitForRun(ProgramRunStatus.COMPLETED, 3, TimeUnit.MINUTES);
}