Examples with SegmentDescriptor org.apache.druid.query.SegmentDescriptor used on opensource projects

Example 21 with SegmentDescriptor

use of org.apache.druid.query.SegmentDescriptor in project druid by druid-io.

the class AppenderatorDriverRealtimeIndexTaskTest method testLateData.

@Test(timeout = 60_000L)
public void testLateData() throws Exception {
    expectPublishedSegments(1);
    final AppenderatorDriverRealtimeIndexTask task = makeRealtimeTask(null);
    final ListenableFuture<TaskStatus> statusFuture = runTask(task);
    // Wait for firehose to show up, it starts off null.
    while (task.getFirehose() == null) {
        Thread.sleep(50);
    }
    final TestFirehose firehose = (TestFirehose) task.getFirehose();
    firehose.addRows(ImmutableList.of(ImmutableMap.of("t", now.getMillis(), "dim1", "foo", "met1", "1"), // Data is from 2 days ago, should still be processed
    ImmutableMap.of("t", now.minus(new Period("P2D")).getMillis(), "dim2", "bar", "met1", 2.0)));
    // Stop the firehose, this will drain out existing events.
    firehose.close();
    // Wait for publish.
    Collection<DataSegment> publishedSegments = awaitSegments();
    // Check metrics.
    Assert.assertEquals(2, task.getRowIngestionMeters().getProcessed());
    Assert.assertEquals(0, task.getRowIngestionMeters().getThrownAway());
    Assert.assertEquals(0, task.getRowIngestionMeters().getUnparseable());
    // Do some queries.
    Assert.assertEquals(2, sumMetric(task, null, "rows").longValue());
    Assert.assertEquals(3, sumMetric(task, null, "met1").longValue());
    awaitHandoffs();
    for (DataSegment publishedSegment : publishedSegments) {
        Pair<Executor, Runnable> executorRunnablePair = handOffCallbacks.get(new SegmentDescriptor(publishedSegment.getInterval(), publishedSegment.getVersion(), publishedSegment.getShardSpec().getPartitionNum()));
        Assert.assertNotNull(publishedSegment + " missing from handoff callbacks: " + handOffCallbacks, executorRunnablePair);
        // Simulate handoff.
        executorRunnablePair.lhs.execute(executorRunnablePair.rhs);
    }
    handOffCallbacks.clear();
    // Wait for the task to finish.
    final TaskStatus taskStatus = statusFuture.get();
    Assert.assertEquals(TaskState.SUCCESS, taskStatus.getStatusCode());
}

Example 22 with SegmentDescriptor

use of org.apache.druid.query.SegmentDescriptor in project druid by druid-io.

the class AppenderatorDriverRealtimeIndexTaskTest method makeToolboxFactory.

private void makeToolboxFactory(final File directory) {
    taskStorage = new HeapMemoryTaskStorage(new TaskStorageConfig(null));
    publishedSegments = new CopyOnWriteArrayList<>();
    ObjectMapper mapper = new DefaultObjectMapper();
    mapper.registerSubtypes(LinearShardSpec.class);
    mapper.registerSubtypes(NumberedShardSpec.class);
    IndexerSQLMetadataStorageCoordinator mdc = new IndexerSQLMetadataStorageCoordinator(mapper, derbyConnectorRule.metadataTablesConfigSupplier().get(), derbyConnectorRule.getConnector()) {

        @Override
        public Set<DataSegment> announceHistoricalSegments(Set<DataSegment> segments) throws IOException {
            Set<DataSegment> result = super.announceHistoricalSegments(segments);
            Assert.assertFalse("Segment latch not initialized, did you forget to call expectPublishSegments?", segmentLatch == null);
            publishedSegments.addAll(result);
            segments.forEach(s -> segmentLatch.countDown());
            return result;
        }

        @Override
        public SegmentPublishResult announceHistoricalSegments(Set<DataSegment> segments, Set<DataSegment> segmentsToDrop, DataSourceMetadata startMetadata, DataSourceMetadata endMetadata) throws IOException {
            SegmentPublishResult result = super.announceHistoricalSegments(segments, segmentsToDrop, startMetadata, endMetadata);
            Assert.assertFalse("Segment latch not initialized, did you forget to call expectPublishSegments?", segmentLatch == null);
            publishedSegments.addAll(result.getSegments());
            result.getSegments().forEach(s -> segmentLatch.countDown());
            return result;
        }
    };
    taskLockbox = new TaskLockbox(taskStorage, mdc);
    final TaskConfig taskConfig = new TaskConfig(directory.getPath(), null, null, 50000, null, true, null, null, null, false, false, TaskConfig.BATCH_PROCESSING_MODE_DEFAULT.name());
    final TaskActionToolbox taskActionToolbox = new TaskActionToolbox(taskLockbox, taskStorage, mdc, EMITTER, EasyMock.createMock(SupervisorManager.class));
    final TaskActionClientFactory taskActionClientFactory = new LocalTaskActionClientFactory(taskStorage, taskActionToolbox, new TaskAuditLogConfig(false));
    final QueryRunnerFactoryConglomerate conglomerate = new DefaultQueryRunnerFactoryConglomerate(ImmutableMap.of(TimeseriesQuery.class, new TimeseriesQueryRunnerFactory(new TimeseriesQueryQueryToolChest(), new TimeseriesQueryEngine(), (query, future) -> {
    // do nothing
    })));
    handOffCallbacks = new ConcurrentHashMap<>();
    final SegmentHandoffNotifierFactory handoffNotifierFactory = dataSource -> new SegmentHandoffNotifier() {

        @Override
        public boolean registerSegmentHandoffCallback(SegmentDescriptor descriptor, Executor exec, Runnable handOffRunnable) {
            handOffCallbacks.put(descriptor, new Pair<>(exec, handOffRunnable));
            handoffLatch.countDown();
            return true;
        }

        @Override
        public void start() {
        // Noop
        }

        @Override
        public void close() {
        // Noop
        }
    };
    final TestUtils testUtils = new TestUtils();
    taskToolboxFactory = new TaskToolboxFactory(taskConfig, new DruidNode("druid/middlemanager", "localhost", false, 8091, null, true, false), taskActionClientFactory, EMITTER, new TestDataSegmentPusher(), new TestDataSegmentKiller(), // DataSegmentMover
    null, // DataSegmentArchiver
    null, new TestDataSegmentAnnouncer(), EasyMock.createNiceMock(DataSegmentServerAnnouncer.class), handoffNotifierFactory, () -> conglomerate, // queryExecutorService
    DirectQueryProcessingPool.INSTANCE, NoopJoinableFactory.INSTANCE, () -> EasyMock.createMock(MonitorScheduler.class), new SegmentCacheManagerFactory(testUtils.getTestObjectMapper()), testUtils.getTestObjectMapper(), testUtils.getTestIndexIO(), MapCache.create(1024), new CacheConfig(), new CachePopulatorStats(), testUtils.getTestIndexMergerV9(), EasyMock.createNiceMock(DruidNodeAnnouncer.class), EasyMock.createNiceMock(DruidNode.class), new LookupNodeService("tier"), new DataNodeService("tier", 1000, ServerType.INDEXER_EXECUTOR, 0), new SingleFileTaskReportFileWriter(reportsFile), null, AuthTestUtils.TEST_AUTHORIZER_MAPPER, new NoopChatHandlerProvider(), testUtils.getRowIngestionMetersFactory(), new TestAppenderatorsManager(), new NoopIndexingServiceClient(), null, null, null);
}

Example 23 with SegmentDescriptor

use of org.apache.druid.query.SegmentDescriptor in project druid by druid-io.

the class AppenderatorDriverRealtimeIndexTaskTest method testMaxRowsPerSegment.

@Test(timeout = 60_000L)
public void testMaxRowsPerSegment() throws Exception {
    // Expect 2 segments as we will hit maxRowsPerSegment
    expectPublishedSegments(2);
    final AppenderatorDriverRealtimeIndexTask task = makeRealtimeTask(null);
    final ListenableFuture<TaskStatus> statusFuture = runTask(task);
    // Wait for firehose to show up, it starts off null.
    while (task.getFirehose() == null) {
        Thread.sleep(50);
    }
    final TestFirehose firehose = (TestFirehose) task.getFirehose();
    // maxRowsPerSegment is 1000 as configured in #makeRealtimeTask
    for (int i = 0; i < 2000; i++) {
        firehose.addRows(ImmutableList.of(ImmutableMap.of("t", now.getMillis(), "dim1", "foo-" + i, "met1", "1")));
    }
    // Stop the firehose, this will drain out existing events.
    firehose.close();
    // Wait for publish.
    Collection<DataSegment> publishedSegments = awaitSegments();
    // Check metrics.
    Assert.assertEquals(2000, task.getRowIngestionMeters().getProcessed());
    Assert.assertEquals(0, task.getRowIngestionMeters().getThrownAway());
    Assert.assertEquals(0, task.getRowIngestionMeters().getUnparseable());
    // Do some queries.
    Assert.assertEquals(2000, sumMetric(task, null, "rows").longValue());
    Assert.assertEquals(2000, sumMetric(task, null, "met1").longValue());
    awaitHandoffs();
    for (DataSegment publishedSegment : publishedSegments) {
        Pair<Executor, Runnable> executorRunnablePair = handOffCallbacks.get(new SegmentDescriptor(publishedSegment.getInterval(), publishedSegment.getVersion(), publishedSegment.getShardSpec().getPartitionNum()));
        Assert.assertNotNull(publishedSegment + " missing from handoff callbacks: " + handOffCallbacks, executorRunnablePair);
        // Simulate handoff.
        executorRunnablePair.lhs.execute(executorRunnablePair.rhs);
    }
    handOffCallbacks.clear();
    // Wait for the task to finish.
    final TaskStatus taskStatus = statusFuture.get();
    Assert.assertEquals(TaskState.SUCCESS, taskStatus.getStatusCode());
}

Example 24 with SegmentDescriptor

use of org.apache.druid.query.SegmentDescriptor in project druid by druid-io.

the class AppenderatorDriverRealtimeIndexTaskTest method testBasics.

@Test(timeout = 60_000L)
public void testBasics() throws Exception {
    expectPublishedSegments(1);
    final AppenderatorDriverRealtimeIndexTask task = makeRealtimeTask(null);
    Assert.assertTrue(task.supportsQueries());
    final ListenableFuture<TaskStatus> statusFuture = runTask(task);
    // Wait for firehose to show up, it starts off null.
    while (task.getFirehose() == null) {
        Thread.sleep(50);
    }
    final TestFirehose firehose = (TestFirehose) task.getFirehose();
    firehose.addRows(ImmutableList.of(ImmutableMap.of("t", now.getMillis(), "dim1", "foo", "met1", "1"), ImmutableMap.of("t", now.getMillis(), "dim2", "bar", "met1", 2.0)));
    // Stop the firehose, this will drain out existing events.
    firehose.close();
    // Wait for publish.
    Collection<DataSegment> publishedSegments = awaitSegments();
    // Check metrics.
    Assert.assertEquals(2, task.getRowIngestionMeters().getProcessed());
    Assert.assertEquals(0, task.getRowIngestionMeters().getThrownAway());
    Assert.assertEquals(0, task.getRowIngestionMeters().getUnparseable());
    // Do some queries.
    Assert.assertEquals(2, sumMetric(task, null, "rows").longValue());
    Assert.assertEquals(3, sumMetric(task, null, "met1").longValue());
    awaitHandoffs();
    for (DataSegment publishedSegment : publishedSegments) {
        Pair<Executor, Runnable> executorRunnablePair = handOffCallbacks.get(new SegmentDescriptor(publishedSegment.getInterval(), publishedSegment.getVersion(), publishedSegment.getShardSpec().getPartitionNum()));
        Assert.assertNotNull(publishedSegment + " missing from handoff callbacks: " + handOffCallbacks, executorRunnablePair);
        // Simulate handoff.
        executorRunnablePair.lhs.execute(executorRunnablePair.rhs);
    }
    handOffCallbacks.clear();
    // Wait for the task to finish.
    final TaskStatus taskStatus = statusFuture.get();
    Assert.assertEquals(TaskState.SUCCESS, taskStatus.getStatusCode());
}

Example 25 with SegmentDescriptor

use of org.apache.druid.query.SegmentDescriptor in project druid by druid-io.

the class AppenderatorDriverRealtimeIndexTaskTest method testTransformSpec.

@Test(timeout = 60_000L)
public void testTransformSpec() throws Exception {
    expectPublishedSegments(2);
    final TransformSpec transformSpec = new TransformSpec(new SelectorDimFilter("dim1", "foo", null), ImmutableList.of(new ExpressionTransform("dim1t", "concat(dim1,dim1)", ExprMacroTable.nil())));
    final AppenderatorDriverRealtimeIndexTask task = makeRealtimeTask(null, transformSpec, true, 0, true, 0, 1);
    final ListenableFuture<TaskStatus> statusFuture = runTask(task);
    // Wait for firehose to show up, it starts off null.
    while (task.getFirehose() == null) {
        Thread.sleep(50);
    }
    final TestFirehose firehose = (TestFirehose) task.getFirehose();
    firehose.addRows(ImmutableList.of(ImmutableMap.of("t", now.getMillis(), "dim1", "foo", "met1", "1"), ImmutableMap.of("t", now.minus(new Period("P1D")).getMillis(), "dim1", "foo", "met1", 2.0), ImmutableMap.of("t", now.getMillis(), "dim2", "bar", "met1", 2.0)));
    // Stop the firehose, this will drain out existing events.
    firehose.close();
    Collection<DataSegment> publishedSegments = awaitSegments();
    // Check metrics.
    Assert.assertEquals(2, task.getRowIngestionMeters().getProcessed());
    Assert.assertEquals(1, task.getRowIngestionMeters().getThrownAway());
    Assert.assertEquals(0, task.getRowIngestionMeters().getUnparseable());
    // Do some queries.
    Assert.assertEquals(2, sumMetric(task, null, "rows").longValue());
    Assert.assertEquals(2, sumMetric(task, new SelectorDimFilter("dim1t", "foofoo", null), "rows").longValue());
    if (NullHandling.replaceWithDefault()) {
        Assert.assertEquals(0, sumMetric(task, new SelectorDimFilter("dim1t", "barbar", null), "metric1").longValue());
    } else {
        Assert.assertNull(sumMetric(task, new SelectorDimFilter("dim1t", "barbar", null), "metric1"));
    }
    Assert.assertEquals(3, sumMetric(task, null, "met1").longValue());
    awaitHandoffs();
    for (DataSegment publishedSegment : publishedSegments) {
        Pair<Executor, Runnable> executorRunnablePair = handOffCallbacks.get(new SegmentDescriptor(publishedSegment.getInterval(), publishedSegment.getVersion(), publishedSegment.getShardSpec().getPartitionNum()));
        Assert.assertNotNull(publishedSegment + " missing from handoff callbacks: " + handOffCallbacks, executorRunnablePair);
        // Simulate handoff.
        executorRunnablePair.lhs.execute(executorRunnablePair.rhs);
    }
    handOffCallbacks.clear();
    // Wait for the task to finish.
    final TaskStatus taskStatus = statusFuture.get();
    Assert.assertEquals(TaskState.SUCCESS, taskStatus.getStatusCode());
}