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

Example 41 with QueryRunner

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

the class SchemalessTestSimpleTest method testTimeBoundary.

@Test
public void testTimeBoundary() {
    TimeBoundaryQuery query = Druids.newTimeBoundaryQueryBuilder().dataSource("testing").build();
    List<Result<TimeBoundaryResultValue>> expectedResults = Collections.singletonList(new Result<TimeBoundaryResultValue>(DateTimes.of("2011-01-12T00:00:00.000Z"), new TimeBoundaryResultValue(ImmutableMap.of(TimeBoundaryQuery.MIN_TIME, DateTimes.of("2011-01-12T00:00:00.000Z"), TimeBoundaryQuery.MAX_TIME, DateTimes.of("2011-01-13T00:00:00.000Z")))));
    QueryRunner runner = TestQueryRunners.makeTimeBoundaryQueryRunner(segment);
    TestHelper.assertExpectedResults(expectedResults, runner.run(QueryPlus.wrap(query)));
}

Example 42 with QueryRunner

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

the class SchemalessTestSimpleTest method testFullOnTimeseries.

@Test
public void testFullOnTimeseries() {
    TimeseriesQuery query = Druids.newTimeseriesQueryBuilder().dataSource(dataSource).granularity(ALL_GRAN).intervals(fullOnInterval).aggregators(Lists.newArrayList(Iterables.concat(commonAggregators, Lists.newArrayList(new DoubleMaxAggregatorFactory("maxIndex", "index"), new DoubleMinAggregatorFactory("minIndex", "index"))))).postAggregators(addRowsIndexConstant).build();
    List<Result<TimeseriesResultValue>> expectedResults = Collections.singletonList(new Result(DateTimes.of("2011-01-12T00:00:00.000Z"), new TimeseriesResultValue(ImmutableMap.<String, Object>builder().put("rows", coalesceAbsentAndEmptyDims ? 10L : 11L).put("index", 900.0).put("addRowsIndexConstant", coalesceAbsentAndEmptyDims ? 911.0 : 912.0).put("uniques", 2.000977198748901D).put("maxIndex", 100.0).put("minIndex", NullHandling.replaceWithDefault() ? 0.0 : 100.0).build())));
    QueryRunner runner = TestQueryRunners.makeTimeSeriesQueryRunner(segment);
    TestHelper.assertExpectedResults(expectedResults, runner.run(QueryPlus.wrap(query)));
}

Example 43 with QueryRunner

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

the class SchemalessTestSimpleTest method testFullOnTopN.

// @Test TODO: Handling of null values is inconsistent right now, need to make it all consistent and re-enable test
// TODO: Complain to Eric when you see this.  It shouldn't be like this...
@Ignore
@SuppressWarnings("unused")
public void testFullOnTopN() {
    TopNQuery query = new TopNQueryBuilder().dataSource(dataSource).granularity(ALL_GRAN).dimension(marketDimension).metric(indexMetric).threshold(3).intervals(fullOnInterval).aggregators(Lists.newArrayList(Iterables.concat(commonAggregators, Lists.newArrayList(new DoubleMaxAggregatorFactory("maxIndex", "index"), new DoubleMinAggregatorFactory("minIndex", "index"))))).postAggregators(addRowsIndexConstant).build();
    List<Result<TopNResultValue>> expectedResults = Collections.singletonList(new Result<>(DateTimes.of("2011-01-12T00:00:00.000Z"), new TopNResultValue(Arrays.asList(new DimensionAndMetricValueExtractor(ImmutableMap.<String, Object>builder().put("market", "spot").put("rows", 4L).put("index", 400.0D).put("addRowsIndexConstant", 405.0D).put("uniques", 1.0002442201269182D).put("maxIndex", 100.0).put("minIndex", 100.0).build()), new DimensionAndMetricValueExtractor(ImmutableMap.<String, Object>builder().put("market", "").put("rows", 2L).put("index", 200.0D).put("addRowsIndexConstant", 203.0D).put("uniques", 0.0).put("maxIndex", 100.0D).put("minIndex", 100.0D).build()), new DimensionAndMetricValueExtractor(ImmutableMap.<String, Object>builder().put("market", "total_market").put("rows", 2L).put("index", 200.0D).put("addRowsIndexConstant", 203.0D).put("uniques", 1.0002442201269182D).put("maxIndex", 100.0D).put("minIndex", 100.0D).build())))));
    try (CloseableStupidPool<ByteBuffer> pool = TestQueryRunners.createDefaultNonBlockingPool()) {
        QueryRunner runner = TestQueryRunners.makeTopNQueryRunner(segment, pool);
        TestHelper.assertExpectedResults(expectedResults, runner.run(QueryPlus.wrap(query)));
    }
}

Example 44 with QueryRunner

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

the class ClientQuerySegmentWalker method inlineIfNecessary.

/**
 * Replace QueryDataSources with InlineDataSources when necessary and possible. "Necessary" is defined as:
 *
 * 1) For outermost subqueries: inlining is necessary if the toolchest cannot handle it.
 * 2) For all other subqueries (e.g. those nested under a join): inlining is always necessary.
 *
 * @param dataSource           datasource to process.
 * @param toolChestIfOutermost if provided, and if the provided datasource is a {@link QueryDataSource}, this method
 *                             will consider whether the toolchest can handle a subquery on the datasource using
 *                             {@link QueryToolChest#canPerformSubquery}. If the toolchest can handle it, then it will
 *                             not be inlined. See {@link org.apache.druid.query.groupby.GroupByQueryQueryToolChest}
 *                             for an example of a toolchest that can handle subqueries.
 * @param subqueryRowLimitAccumulator an accumulator for tracking the number of accumulated rows in all subqueries
 *                                    for a particular master query
 * @param maxSubqueryRows      Max rows that all the subqueries generated by a master query can have, combined
 * @param dryRun               if true, does not actually execute any subqueries, but will inline empty result sets.
 */
// Subquery, toolchest, runner handling all use raw types
@SuppressWarnings({ "rawtypes", "unchecked" })
private DataSource inlineIfNecessary(final DataSource dataSource, @Nullable final QueryToolChest toolChestIfOutermost, final AtomicInteger subqueryRowLimitAccumulator, final int maxSubqueryRows, final boolean dryRun) {
    if (dataSource instanceof QueryDataSource) {
        // This datasource is a subquery.
        final Query subQuery = ((QueryDataSource) dataSource).getQuery();
        final QueryToolChest toolChest = warehouse.getToolChest(subQuery);
        if (toolChestIfOutermost != null && toolChestIfOutermost.canPerformSubquery(subQuery)) {
            // Strip outer queries that are handleable by the toolchest, and inline subqueries that may be underneath
            // them (e.g. subqueries nested under a join).
            final Stack<DataSource> stack = new Stack<>();
            DataSource current = dataSource;
            while (current instanceof QueryDataSource) {
                stack.push(current);
                current = Iterables.getOnlyElement(current.getChildren());
            }
            // lgtm [java/contradictory-type-checks]
            assert !(current instanceof QueryDataSource);
            current = inlineIfNecessary(current, null, subqueryRowLimitAccumulator, maxSubqueryRows, dryRun);
            while (!stack.isEmpty()) {
                current = stack.pop().withChildren(Collections.singletonList(current));
            }
            assert current instanceof QueryDataSource;
            if (toolChest.canPerformSubquery(((QueryDataSource) current).getQuery())) {
                return current;
            } else {
                // We need to consider inlining it.
                return inlineIfNecessary(current, toolChestIfOutermost, subqueryRowLimitAccumulator, maxSubqueryRows, dryRun);
            }
        } else if (canRunQueryUsingLocalWalker(subQuery) || canRunQueryUsingClusterWalker(subQuery)) {
            // Subquery needs to be inlined. Assign it a subquery id and run it.
            final Sequence<?> queryResults;
            if (dryRun) {
                queryResults = Sequences.empty();
            } else {
                final QueryRunner subqueryRunner = subQuery.getRunner(this);
                queryResults = subqueryRunner.run(QueryPlus.wrap(subQuery), DirectDruidClient.makeResponseContextForQuery());
            }
            return toInlineDataSource(subQuery, queryResults, warehouse.getToolChest(subQuery), subqueryRowLimitAccumulator, maxSubqueryRows);
        } else {
            // Cannot inline subquery. Attempt to inline one level deeper, and then try again.
            return inlineIfNecessary(dataSource.withChildren(Collections.singletonList(inlineIfNecessary(Iterables.getOnlyElement(dataSource.getChildren()), null, subqueryRowLimitAccumulator, maxSubqueryRows, dryRun))), toolChestIfOutermost, subqueryRowLimitAccumulator, maxSubqueryRows, dryRun);
        }
    } else {
        // Not a query datasource. Walk children and see if there's anything to inline.
        return dataSource.withChildren(dataSource.getChildren().stream().map(child -> inlineIfNecessary(child, null, subqueryRowLimitAccumulator, maxSubqueryRows, dryRun)).collect(Collectors.toList()));
    }
}

Example 45 with QueryRunner

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

the class ServerManager method buildAndDecorateQueryRunner.

private <T> QueryRunner<T> buildAndDecorateQueryRunner(final QueryRunnerFactory<T, Query<T>> factory, final QueryToolChest<T, Query<T>> toolChest, final SegmentReference segment, final Optional<byte[]> cacheKeyPrefix, final SegmentDescriptor segmentDescriptor, final AtomicLong cpuTimeAccumulator) {
    final SpecificSegmentSpec segmentSpec = new SpecificSegmentSpec(segmentDescriptor);
    final SegmentId segmentId = segment.getId();
    final Interval segmentInterval = segment.getDataInterval();
    // If the segment is closed after this line, ReferenceCountingSegmentQueryRunner will handle and do the right thing.
    if (segmentId == null || segmentInterval == null) {
        return new ReportTimelineMissingSegmentQueryRunner<>(segmentDescriptor);
    }
    String segmentIdString = segmentId.toString();
    MetricsEmittingQueryRunner<T> metricsEmittingQueryRunnerInner = new MetricsEmittingQueryRunner<>(emitter, toolChest, new ReferenceCountingSegmentQueryRunner<>(factory, segment, segmentDescriptor), QueryMetrics::reportSegmentTime, queryMetrics -> queryMetrics.segment(segmentIdString));
    StorageAdapter storageAdapter = segment.asStorageAdapter();
    long segmentMaxTime = storageAdapter.getMaxTime().getMillis();
    long segmentMinTime = storageAdapter.getMinTime().getMillis();
    Interval actualDataInterval = Intervals.utc(segmentMinTime, segmentMaxTime + 1);
    CachingQueryRunner<T> cachingQueryRunner = new CachingQueryRunner<>(segmentIdString, cacheKeyPrefix, segmentDescriptor, actualDataInterval, objectMapper, cache, toolChest, metricsEmittingQueryRunnerInner, cachePopulator, cacheConfig);
    BySegmentQueryRunner<T> bySegmentQueryRunner = new BySegmentQueryRunner<>(segmentId, segmentInterval.getStart(), cachingQueryRunner);
    MetricsEmittingQueryRunner<T> metricsEmittingQueryRunnerOuter = new MetricsEmittingQueryRunner<>(emitter, toolChest, bySegmentQueryRunner, QueryMetrics::reportSegmentAndCacheTime, queryMetrics -> queryMetrics.segment(segmentIdString)).withWaitMeasuredFromNow();
    SpecificSegmentQueryRunner<T> specificSegmentQueryRunner = new SpecificSegmentQueryRunner<>(metricsEmittingQueryRunnerOuter, segmentSpec);
    PerSegmentOptimizingQueryRunner<T> perSegmentOptimizingQueryRunner = new PerSegmentOptimizingQueryRunner<>(specificSegmentQueryRunner, new PerSegmentQueryOptimizationContext(segmentDescriptor));
    return new SetAndVerifyContextQueryRunner<>(serverConfig, CPUTimeMetricQueryRunner.safeBuild(perSegmentOptimizingQueryRunner, toolChest, emitter, cpuTimeAccumulator, false));
}