Examples with HashBasedNumberedShardSpec org.apache.druid.timeline.partition.HashBasedNumberedShardSpec used on opensource projects

Example 6 with HashBasedNumberedShardSpec

use of org.apache.druid.timeline.partition.HashBasedNumberedShardSpec in project druid by druid-io.

the class BatchDeltaIngestionTest method makeHadoopDruidIndexerConfig.

private HadoopDruidIndexerConfig makeHadoopDruidIndexerConfig(Map<String, Object> inputSpec, File tmpDir, AggregatorFactory[] aggregators) throws Exception {
    HadoopDruidIndexerConfig config = new HadoopDruidIndexerConfig(new HadoopIngestionSpec(new DataSchema("website", MAPPER.convertValue(new StringInputRowParser(new CSVParseSpec(new TimestampSpec("timestamp", "yyyyMMddHH", null), new DimensionsSpec(DimensionsSpec.getDefaultSchemas(ImmutableList.of("host"))), null, ImmutableList.of("timestamp", "host", "host2", "visited_num"), false, 0), null), Map.class), aggregators != null ? aggregators : new AggregatorFactory[] { new LongSumAggregatorFactory("visited_sum", "visited_num"), new HyperUniquesAggregatorFactory("unique_hosts", "host2") }, new UniformGranularitySpec(Granularities.DAY, Granularities.NONE, ImmutableList.of(INTERVAL_FULL)), null, MAPPER), new HadoopIOConfig(inputSpec, null, tmpDir.getCanonicalPath()), new HadoopTuningConfig(tmpDir.getCanonicalPath(), null, null, null, null, null, null, null, null, false, false, false, false, null, false, false, null, null, false, false, null, null, null, null, null)));
    config.setShardSpecs(ImmutableMap.of(INTERVAL_FULL.getStartMillis(), ImmutableList.of(new HadoopyShardSpec(new HashBasedNumberedShardSpec(0, 1, 0, 1, null, HashPartitionFunction.MURMUR3_32_ABS, HadoopDruidIndexerConfig.JSON_MAPPER), 0))));
    config = HadoopDruidIndexerConfig.fromSpec(config.getSchema());
    return config;
}

Example 7 with HashBasedNumberedShardSpec

use of org.apache.druid.timeline.partition.HashBasedNumberedShardSpec in project druid by druid-io.

the class HashPartitionAdjustingCorePartitionSizeTest method testEqualNumberOfPartitionsToBuckets.

@Test
public void testEqualNumberOfPartitionsToBuckets() throws IOException {
    final File inputDir = temporaryFolder.newFolder();
    for (int i = 0; i < 10; i++) {
        try (final Writer writer = Files.newBufferedWriter(new File(inputDir, "test_" + i).toPath(), StandardCharsets.UTF_8)) {
            writer.write(StringUtils.format("2020-01-01T00:00:00,%s,b1,%d\n", "aa" + (i + 10), 10 * (i + 1)));
        }
    }
    final DimensionBasedPartitionsSpec partitionsSpec = new HashedPartitionsSpec(null, 5, ImmutableList.of("dim1"));
    final Set<DataSegment> segments = runTestTask(TIMESTAMP_SPEC, DIMENSIONS_SPEC, INPUT_FORMAT, null, INTERVAL_TO_INDEX, inputDir, "test_*", partitionsSpec, maxNumConcurrentSubTasks, TaskState.SUCCESS);
    Assert.assertEquals(5, segments.size());
    segments.forEach(segment -> {
        Assert.assertSame(HashBasedNumberedShardSpec.class, segment.getShardSpec().getClass());
        final HashBasedNumberedShardSpec shardSpec = (HashBasedNumberedShardSpec) segment.getShardSpec();
        Assert.assertEquals(5, shardSpec.getNumCorePartitions());
        Assert.assertEquals(5, shardSpec.getNumBuckets());
        Assert.assertEquals(ImmutableList.of("dim1"), shardSpec.getPartitionDimensions());
    });
}

Example 8 with HashBasedNumberedShardSpec

use of org.apache.druid.timeline.partition.HashBasedNumberedShardSpec in project druid by druid-io.

the class HashPartitionMultiPhaseParallelIndexingTest method testAppendLinearlyPartitionedSegmensToHashPartitionedDatasourceSuccessfullyAppend.

@Test
public void testAppendLinearlyPartitionedSegmensToHashPartitionedDatasourceSuccessfullyAppend() {
    final Set<DataSegment> publishedSegments = new HashSet<>();
    publishedSegments.addAll(runTestTask(new HashedPartitionsSpec(null, numShards, ImmutableList.of("dim1", "dim2")), TaskState.SUCCESS, false));
    // Append
    publishedSegments.addAll(runTestTask(new DynamicPartitionsSpec(5, null), TaskState.SUCCESS, true));
    // And append again
    publishedSegments.addAll(runTestTask(new DynamicPartitionsSpec(10, null), TaskState.SUCCESS, true));
    final Map<Interval, List<DataSegment>> intervalToSegments = new HashMap<>();
    publishedSegments.forEach(segment -> intervalToSegments.computeIfAbsent(segment.getInterval(), k -> new ArrayList<>()).add(segment));
    for (Entry<Interval, List<DataSegment>> entry : intervalToSegments.entrySet()) {
        final List<DataSegment> segments = entry.getValue();
        final List<DataSegment> hashedSegments = segments.stream().filter(segment -> segment.getShardSpec().getClass() == HashBasedNumberedShardSpec.class).collect(Collectors.toList());
        final List<DataSegment> linearSegments = segments.stream().filter(segment -> segment.getShardSpec().getClass() == NumberedShardSpec.class).collect(Collectors.toList());
        for (DataSegment hashedSegment : hashedSegments) {
            final HashBasedNumberedShardSpec hashShardSpec = (HashBasedNumberedShardSpec) hashedSegment.getShardSpec();
            for (DataSegment linearSegment : linearSegments) {
                Assert.assertEquals(hashedSegment.getInterval(), linearSegment.getInterval());
                Assert.assertEquals(hashedSegment.getVersion(), linearSegment.getVersion());
                final NumberedShardSpec numberedShardSpec = (NumberedShardSpec) linearSegment.getShardSpec();
                Assert.assertEquals(hashShardSpec.getNumCorePartitions(), numberedShardSpec.getNumCorePartitions());
                Assert.assertTrue(hashShardSpec.getPartitionNum() < numberedShardSpec.getPartitionNum());
            }
        }
    }
}

Example 9 with HashBasedNumberedShardSpec

use of org.apache.druid.timeline.partition.HashBasedNumberedShardSpec in project druid by druid-io.

the class MaterializedViewSupervisorTest method testCreateTask.

/**
 * Verifies that creating HadoopIndexTask compleates without raising exception.
 */
@Test
public void testCreateTask() {
    List<DataSegment> baseSegments = Collections.singletonList(new DataSegment("base", Intervals.of("2015-01-02T00Z/2015-01-03T00Z"), "2015-01-03", ImmutableMap.of(), ImmutableList.of("dim1", "dim2"), ImmutableList.of("m1"), new HashBasedNumberedShardSpec(0, 1, 0, 1, null, null, null), 9, 1024));
    HadoopIndexTask task = spec.createTask(Intervals.of("2015-01-02T00Z/2015-01-03T00Z"), "2015-01-03", baseSegments);
    Assert.assertNotNull(task);
}

Example 10 with HashBasedNumberedShardSpec

use of org.apache.druid.timeline.partition.HashBasedNumberedShardSpec in project druid by druid-io.

the class MaterializedViewSupervisorTest method testCheckSegments.

@Test
public void testCheckSegments() throws IOException {
    Set<DataSegment> baseSegments = Sets.newHashSet(new DataSegment("base", Intervals.of("2015-01-01T00Z/2015-01-02T00Z"), "2015-01-02", ImmutableMap.of(), ImmutableList.of("dim1", "dim2"), ImmutableList.of("m1"), new HashBasedNumberedShardSpec(0, 1, 0, 1, null, null, null), 9, 1024), new DataSegment("base", Intervals.of("2015-01-02T00Z/2015-01-03T00Z"), "2015-01-03", ImmutableMap.of(), ImmutableList.of("dim1", "dim2"), ImmutableList.of("m1"), new HashBasedNumberedShardSpec(0, 1, 0, 1, null, null, null), 9, 1024), new DataSegment("base", Intervals.of("2015-01-03T00Z/2015-01-04T00Z"), "2015-01-04", ImmutableMap.of(), ImmutableList.of("dim1", "dim2"), ImmutableList.of("m1"), new HashBasedNumberedShardSpec(0, 1, 0, 1, null, null, null), 9, 1024));
    Set<DataSegment> derivativeSegments = Sets.newHashSet(new DataSegment(derivativeDatasourceName, Intervals.of("2015-01-01T00Z/2015-01-02T00Z"), "2015-01-02", ImmutableMap.of(), ImmutableList.of("dim1", "dim2"), ImmutableList.of("m1"), new HashBasedNumberedShardSpec(0, 1, 0, 1, null, null, null), 9, 1024), new DataSegment(derivativeDatasourceName, Intervals.of("2015-01-02T00Z/2015-01-03T00Z"), "3015-01-01", ImmutableMap.of(), ImmutableList.of("dim1", "dim2"), ImmutableList.of("m1"), new HashBasedNumberedShardSpec(0, 1, 0, 1, null, null, null), 9, 1024));
    indexerMetadataStorageCoordinator.announceHistoricalSegments(baseSegments);
    indexerMetadataStorageCoordinator.announceHistoricalSegments(derivativeSegments);
    EasyMock.expect(taskMaster.getTaskQueue()).andReturn(Optional.of(taskQueue)).anyTimes();
    EasyMock.expect(taskMaster.getTaskRunner()).andReturn(Optional.absent()).anyTimes();
    EasyMock.expect(taskStorage.getActiveTasks()).andReturn(ImmutableList.of()).anyTimes();
    Pair<SortedMap<Interval, String>, Map<Interval, List<DataSegment>>> toBuildInterval = supervisor.checkSegments();
    Set<Interval> expectedToBuildInterval = Sets.newHashSet(Intervals.of("2015-01-01T00Z/2015-01-02T00Z"));
    Map<Interval, List<DataSegment>> expectedSegments = new HashMap<>();
    expectedSegments.put(Intervals.of("2015-01-01T00Z/2015-01-02T00Z"), Collections.singletonList(new DataSegment("base", Intervals.of("2015-01-01T00Z/2015-01-02T00Z"), "2015-01-02", ImmutableMap.of(), ImmutableList.of("dim1", "dim2"), ImmutableList.of("m1"), new HashBasedNumberedShardSpec(0, 1, 0, 1, null, null, null), 9, 1024)));
    expectedSegments.put(Intervals.of("2015-01-02T00Z/2015-01-03T00Z"), Collections.singletonList(new DataSegment("base", Intervals.of("2015-01-02T00Z/2015-01-03T00Z"), "2015-01-03", ImmutableMap.of(), ImmutableList.of("dim1", "dim2"), ImmutableList.of("m1"), new HashBasedNumberedShardSpec(0, 1, 0, 1, null, null, null), 9, 1024)));
    Assert.assertEquals(expectedToBuildInterval, toBuildInterval.lhs.keySet());
    Assert.assertEquals(expectedSegments, toBuildInterval.rhs);
}