Search in sources :

Example 1 with SinkQueryRunners

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

the class SinkQuerySegmentWalker method getQueryRunnerForSegments.

@Override
public <T> QueryRunner<T> getQueryRunnerForSegments(final Query<T> query, final Iterable<SegmentDescriptor> specs) {
    // We only handle one particular dataSource. Make sure that's what we have, then ignore from here on out.
    final DataSourceAnalysis analysis = DataSourceAnalysis.forDataSource(query.getDataSource());
    // Sanity check: make sure the query is based on the table we're meant to handle.
    if (!analysis.getBaseTableDataSource().filter(ds -> dataSource.equals(ds.getName())).isPresent()) {
        throw new ISE("Cannot handle datasource: %s", analysis.getDataSource());
    }
    final QueryRunnerFactory<T, Query<T>> factory = conglomerate.findFactory(query);
    if (factory == null) {
        throw new ISE("Unknown query type[%s].", query.getClass());
    }
    final QueryToolChest<T, Query<T>> toolChest = factory.getToolchest();
    final boolean skipIncrementalSegment = query.getContextValue(CONTEXT_SKIP_INCREMENTAL_SEGMENT, false);
    final AtomicLong cpuTimeAccumulator = new AtomicLong(0L);
    // Make sure this query type can handle the subquery, if present.
    if (analysis.isQuery() && !toolChest.canPerformSubquery(((QueryDataSource) analysis.getDataSource()).getQuery())) {
        throw new ISE("Cannot handle subquery: %s", analysis.getDataSource());
    }
    // segmentMapFn maps each base Segment into a joined Segment if necessary.
    final Function<SegmentReference, SegmentReference> segmentMapFn = joinableFactoryWrapper.createSegmentMapFn(analysis.getJoinBaseTableFilter().map(Filters::toFilter).orElse(null), analysis.getPreJoinableClauses(), cpuTimeAccumulator, analysis.getBaseQuery().orElse(query));
    // We compute the join cache key here itself so it doesn't need to be re-computed for every segment
    final Optional<byte[]> cacheKeyPrefix = analysis.isJoin() ? joinableFactoryWrapper.computeJoinDataSourceCacheKey(analysis) : Optional.of(StringUtils.EMPTY_BYTES);
    Iterable<QueryRunner<T>> perSegmentRunners = Iterables.transform(specs, descriptor -> {
        final PartitionChunk<Sink> chunk = sinkTimeline.findChunk(descriptor.getInterval(), descriptor.getVersion(), descriptor.getPartitionNumber());
        if (chunk == null) {
            return new ReportTimelineMissingSegmentQueryRunner<>(descriptor);
        }
        final Sink theSink = chunk.getObject();
        final SegmentId sinkSegmentId = theSink.getSegment().getId();
        Iterable<QueryRunner<T>> perHydrantRunners = new SinkQueryRunners<>(Iterables.transform(theSink, hydrant -> {
            // Hydrant might swap at any point, but if it's swapped at the start
            // then we know it's *definitely* swapped.
            final boolean hydrantDefinitelySwapped = hydrant.hasSwapped();
            if (skipIncrementalSegment && !hydrantDefinitelySwapped) {
                return new Pair<>(hydrant.getSegmentDataInterval(), new NoopQueryRunner<>());
            }
            // Prevent the underlying segment from swapping when its being iterated
            final Optional<Pair<SegmentReference, Closeable>> maybeSegmentAndCloseable = hydrant.getSegmentForQuery(segmentMapFn);
            // if optional isn't present, we failed to acquire reference to the segment or any joinables
            if (!maybeSegmentAndCloseable.isPresent()) {
                return new Pair<>(hydrant.getSegmentDataInterval(), new ReportTimelineMissingSegmentQueryRunner<>(descriptor));
            }
            final Pair<SegmentReference, Closeable> segmentAndCloseable = maybeSegmentAndCloseable.get();
            try {
                QueryRunner<T> runner = factory.createRunner(segmentAndCloseable.lhs);
                // 2) Hydrants are not the same between replicas, make sure cache is local
                if (hydrantDefinitelySwapped && cache.isLocal()) {
                    StorageAdapter storageAdapter = segmentAndCloseable.lhs.asStorageAdapter();
                    long segmentMinTime = storageAdapter.getMinTime().getMillis();
                    long segmentMaxTime = storageAdapter.getMaxTime().getMillis();
                    Interval actualDataInterval = Intervals.utc(segmentMinTime, segmentMaxTime + 1);
                    runner = new CachingQueryRunner<>(makeHydrantCacheIdentifier(hydrant), cacheKeyPrefix, descriptor, actualDataInterval, objectMapper, cache, toolChest, runner, // Always populate in foreground regardless of config
                    new ForegroundCachePopulator(objectMapper, cachePopulatorStats, cacheConfig.getMaxEntrySize()), cacheConfig);
                }
                // Make it always use Closeable to decrement()
                runner = QueryRunnerHelper.makeClosingQueryRunner(runner, segmentAndCloseable.rhs);
                return new Pair<>(segmentAndCloseable.lhs.getDataInterval(), runner);
            } catch (Throwable e) {
                throw CloseableUtils.closeAndWrapInCatch(e, segmentAndCloseable.rhs);
            }
        }));
        return new SpecificSegmentQueryRunner<>(withPerSinkMetrics(new BySegmentQueryRunner<>(sinkSegmentId, descriptor.getInterval().getStart(), factory.mergeRunners(DirectQueryProcessingPool.INSTANCE, perHydrantRunners)), toolChest, sinkSegmentId, cpuTimeAccumulator), new SpecificSegmentSpec(descriptor));
    });
    final QueryRunner<T> mergedRunner = toolChest.mergeResults(factory.mergeRunners(queryProcessingPool, perSegmentRunners));
    return CPUTimeMetricQueryRunner.safeBuild(new FinalizeResultsQueryRunner<>(mergedRunner, toolChest), toolChest, emitter, cpuTimeAccumulator, true);
}
Also used : DirectQueryProcessingPool(org.apache.druid.query.DirectQueryProcessingPool) QueryRunnerHelper(org.apache.druid.query.QueryRunnerHelper) QueryProcessingPool(org.apache.druid.query.QueryProcessingPool) ForegroundCachePopulator(org.apache.druid.client.cache.ForegroundCachePopulator) StorageAdapter(org.apache.druid.segment.StorageAdapter) Pair(org.apache.druid.java.util.common.Pair) NoopQueryRunner(org.apache.druid.query.NoopQueryRunner) SegmentReference(org.apache.druid.segment.SegmentReference) SpecificSegmentQueryRunner(org.apache.druid.query.spec.SpecificSegmentQueryRunner) QueryRunner(org.apache.druid.query.QueryRunner) FinalizeResultsQueryRunner(org.apache.druid.query.FinalizeResultsQueryRunner) ReportTimelineMissingSegmentQueryRunner(org.apache.druid.query.ReportTimelineMissingSegmentQueryRunner) CacheConfig(org.apache.druid.client.cache.CacheConfig) StringUtils(org.apache.druid.java.util.common.StringUtils) JoinableFactoryWrapper(org.apache.druid.segment.join.JoinableFactoryWrapper) ISE(org.apache.druid.java.util.common.ISE) SpecificSegmentSpec(org.apache.druid.query.spec.SpecificSegmentSpec) BySegmentQueryRunner(org.apache.druid.query.BySegmentQueryRunner) SinkQueryRunners(org.apache.druid.query.SinkQueryRunners) QueryDataSource(org.apache.druid.query.QueryDataSource) ServiceEmitter(org.apache.druid.java.util.emitter.service.ServiceEmitter) Optional(java.util.Optional) FunctionalIterable(org.apache.druid.java.util.common.guava.FunctionalIterable) SegmentId(org.apache.druid.timeline.SegmentId) DataSourceAnalysis(org.apache.druid.query.planning.DataSourceAnalysis) Iterables(com.google.common.collect.Iterables) Intervals(org.apache.druid.java.util.common.Intervals) QueryMetrics(org.apache.druid.query.QueryMetrics) CachingQueryRunner(org.apache.druid.client.CachingQueryRunner) JoinableFactory(org.apache.druid.segment.join.JoinableFactory) Function(java.util.function.Function) PartitionChunk(org.apache.druid.timeline.partition.PartitionChunk) Interval(org.joda.time.Interval) MetricsEmittingQueryRunner(org.apache.druid.query.MetricsEmittingQueryRunner) Query(org.apache.druid.query.Query) CachePopulatorStats(org.apache.druid.client.cache.CachePopulatorStats) Sink(org.apache.druid.segment.realtime.plumber.Sink) QuerySegmentWalker(org.apache.druid.query.QuerySegmentWalker) EmittingLogger(org.apache.druid.java.util.emitter.EmittingLogger) VersionedIntervalTimeline(org.apache.druid.timeline.VersionedIntervalTimeline) ObjectMapper(com.fasterxml.jackson.databind.ObjectMapper) QueryRunnerFactoryConglomerate(org.apache.druid.query.QueryRunnerFactoryConglomerate) QueryToolChest(org.apache.druid.query.QueryToolChest) FireHydrant(org.apache.druid.segment.realtime.FireHydrant) AtomicLong(java.util.concurrent.atomic.AtomicLong) QueryRunnerFactory(org.apache.druid.query.QueryRunnerFactory) Closeable(java.io.Closeable) Preconditions(com.google.common.base.Preconditions) VisibleForTesting(com.google.common.annotations.VisibleForTesting) SegmentDescriptor(org.apache.druid.query.SegmentDescriptor) Cache(org.apache.druid.client.cache.Cache) Filters(org.apache.druid.segment.filter.Filters) CloseableUtils(org.apache.druid.utils.CloseableUtils) CPUTimeMetricQueryRunner(org.apache.druid.query.CPUTimeMetricQueryRunner) Query(org.apache.druid.query.Query) Closeable(java.io.Closeable) StorageAdapter(org.apache.druid.segment.StorageAdapter) DataSourceAnalysis(org.apache.druid.query.planning.DataSourceAnalysis) Filters(org.apache.druid.segment.filter.Filters) ReportTimelineMissingSegmentQueryRunner(org.apache.druid.query.ReportTimelineMissingSegmentQueryRunner) Sink(org.apache.druid.segment.realtime.plumber.Sink) SpecificSegmentQueryRunner(org.apache.druid.query.spec.SpecificSegmentQueryRunner) NoopQueryRunner(org.apache.druid.query.NoopQueryRunner) CachingQueryRunner(org.apache.druid.client.CachingQueryRunner) ISE(org.apache.druid.java.util.common.ISE) Pair(org.apache.druid.java.util.common.Pair) Optional(java.util.Optional) SegmentId(org.apache.druid.timeline.SegmentId) SegmentReference(org.apache.druid.segment.SegmentReference) BySegmentQueryRunner(org.apache.druid.query.BySegmentQueryRunner) NoopQueryRunner(org.apache.druid.query.NoopQueryRunner) SpecificSegmentQueryRunner(org.apache.druid.query.spec.SpecificSegmentQueryRunner) QueryRunner(org.apache.druid.query.QueryRunner) FinalizeResultsQueryRunner(org.apache.druid.query.FinalizeResultsQueryRunner) ReportTimelineMissingSegmentQueryRunner(org.apache.druid.query.ReportTimelineMissingSegmentQueryRunner) BySegmentQueryRunner(org.apache.druid.query.BySegmentQueryRunner) CachingQueryRunner(org.apache.druid.client.CachingQueryRunner) MetricsEmittingQueryRunner(org.apache.druid.query.MetricsEmittingQueryRunner) CPUTimeMetricQueryRunner(org.apache.druid.query.CPUTimeMetricQueryRunner) AtomicLong(java.util.concurrent.atomic.AtomicLong) SinkQueryRunners(org.apache.druid.query.SinkQueryRunners) SpecificSegmentSpec(org.apache.druid.query.spec.SpecificSegmentSpec) ForegroundCachePopulator(org.apache.druid.client.cache.ForegroundCachePopulator) Interval(org.joda.time.Interval)

Aggregations

ObjectMapper (com.fasterxml.jackson.databind.ObjectMapper)1 VisibleForTesting (com.google.common.annotations.VisibleForTesting)1 Preconditions (com.google.common.base.Preconditions)1 Iterables (com.google.common.collect.Iterables)1 Closeable (java.io.Closeable)1 Optional (java.util.Optional)1 AtomicLong (java.util.concurrent.atomic.AtomicLong)1 Function (java.util.function.Function)1 CachingQueryRunner (org.apache.druid.client.CachingQueryRunner)1 Cache (org.apache.druid.client.cache.Cache)1 CacheConfig (org.apache.druid.client.cache.CacheConfig)1 CachePopulatorStats (org.apache.druid.client.cache.CachePopulatorStats)1 ForegroundCachePopulator (org.apache.druid.client.cache.ForegroundCachePopulator)1 ISE (org.apache.druid.java.util.common.ISE)1 Intervals (org.apache.druid.java.util.common.Intervals)1 Pair (org.apache.druid.java.util.common.Pair)1 StringUtils (org.apache.druid.java.util.common.StringUtils)1 FunctionalIterable (org.apache.druid.java.util.common.guava.FunctionalIterable)1 EmittingLogger (org.apache.druid.java.util.emitter.EmittingLogger)1 ServiceEmitter (org.apache.druid.java.util.emitter.service.ServiceEmitter)1