Examples with DefaultLimitSpec org.apache.druid.query.groupby.orderby.DefaultLimitSpec used on opensource projects

Example 51 with DefaultLimitSpec

use of org.apache.druid.query.groupby.orderby.DefaultLimitSpec in project druid by druid-io.

the class GroupByQueryRunnerTest method testGroupByLimitPushDownPostAggNotSupported.

@Test
public void testGroupByLimitPushDownPostAggNotSupported() {
    if (!config.getDefaultStrategy().equals(GroupByStrategySelector.STRATEGY_V2)) {
        return;
    }
    expectedException.expect(UnsupportedOperationException.class);
    expectedException.expectMessage("Limit push down when sorting by a post aggregator is not supported.");
    GroupByQuery query = makeQueryBuilder().setDataSource(QueryRunnerTestHelper.DATA_SOURCE).setGranularity(QueryRunnerTestHelper.ALL_GRAN).setDimensions(new DefaultDimensionSpec(QueryRunnerTestHelper.MARKET_DIMENSION, "marketalias")).setInterval(QueryRunnerTestHelper.FULL_ON_INTERVAL_SPEC).setLimitSpec(new DefaultLimitSpec(Collections.singletonList(new OrderByColumnSpec("constant", OrderByColumnSpec.Direction.DESCENDING)), 2)).setAggregatorSpecs(QueryRunnerTestHelper.ROWS_COUNT).setPostAggregatorSpecs(Collections.singletonList(new ConstantPostAggregator("constant", 1))).overrideContext(ImmutableMap.of(GroupByQueryConfig.CTX_KEY_FORCE_LIMIT_PUSH_DOWN, true)).build();
    GroupByQueryRunnerTestHelper.runQuery(factory, runner, query);
}

Example 52 with DefaultLimitSpec

use of org.apache.druid.query.groupby.orderby.DefaultLimitSpec in project druid by druid-io.

the class GroupByQueryRunnerTest method testGroupByWithHavingOnHyperUnique.

@Test
public void testGroupByWithHavingOnHyperUnique() {
    GroupByQuery query = makeQueryBuilder().setDataSource(QueryRunnerTestHelper.DATA_SOURCE).setGranularity(QueryRunnerTestHelper.ALL_GRAN).setDimensions(new DefaultDimensionSpec(QueryRunnerTestHelper.MARKET_DIMENSION, QueryRunnerTestHelper.MARKET_DIMENSION)).setInterval(QueryRunnerTestHelper.FULL_ON_INTERVAL_SPEC).setLimitSpec(new DefaultLimitSpec(Collections.singletonList(new OrderByColumnSpec(QueryRunnerTestHelper.UNIQUE_METRIC, OrderByColumnSpec.Direction.DESCENDING)), 3)).setHavingSpec(new GreaterThanHavingSpec(QueryRunnerTestHelper.UNIQUE_METRIC, 8)).setAggregatorSpecs(QueryRunnerTestHelper.QUALITY_UNIQUES).setPostAggregatorSpecs(Collections.singletonList(new HyperUniqueFinalizingPostAggregator(QueryRunnerTestHelper.HYPER_UNIQUE_FINALIZING_POST_AGG_METRIC, QueryRunnerTestHelper.UNIQUE_METRIC))).build();
    List<ResultRow> expectedResults = Collections.singletonList(makeRow(query, "1970-01-01T00:00:00.000Z", "market", "spot", QueryRunnerTestHelper.UNIQUE_METRIC, QueryRunnerTestHelper.UNIQUES_9, QueryRunnerTestHelper.HYPER_UNIQUE_FINALIZING_POST_AGG_METRIC, QueryRunnerTestHelper.UNIQUES_9));
    Iterable<ResultRow> results = GroupByQueryRunnerTestHelper.runQuery(factory, runner, query);
    TestHelper.assertExpectedObjects(expectedResults, results, "order-limit");
}

Example 53 with DefaultLimitSpec

use of org.apache.druid.query.groupby.orderby.DefaultLimitSpec in project druid by druid-io.

the class GroupByQueryRunnerTest method testMultiValueDimensionAsArrayWithOtherDims.

@Test
public void testMultiValueDimensionAsArrayWithOtherDims() {
    // array types don't work with group by v1
    if (config.getDefaultStrategy().equals(GroupByStrategySelector.STRATEGY_V1)) {
        expectedException.expect(UnsupportedOperationException.class);
        expectedException.expectMessage("GroupBy v1 only supports dimensions with an outputType of STRING");
    }
    // Cannot vectorize due to multi-value dimensions.
    cannotVectorize();
    GroupByQuery query = makeQueryBuilder().setDataSource(QueryRunnerTestHelper.DATA_SOURCE).setQuerySegmentSpec(QueryRunnerTestHelper.FIRST_TO_THIRD).setVirtualColumns(new ExpressionVirtualColumn("v0", "mv_to_array(placementish)", ColumnType.STRING_ARRAY, ExprMacroTable.nil())).setDimensions(new DefaultDimensionSpec("v0", "alias", ColumnType.STRING_ARRAY), new DefaultDimensionSpec("quality", "quality")).setLimitSpec(new DefaultLimitSpec(ImmutableList.of(new OrderByColumnSpec("alias", OrderByColumnSpec.Direction.ASCENDING, StringComparators.LEXICOGRAPHIC), new OrderByColumnSpec("quality", OrderByColumnSpec.Direction.ASCENDING, StringComparators.LEXICOGRAPHIC)), Integer.MAX_VALUE - 1)).setAggregatorSpecs(QueryRunnerTestHelper.ROWS_COUNT, new LongSumAggregatorFactory("idx", "index")).setGranularity(QueryRunnerTestHelper.ALL_GRAN).build();
    List<ResultRow> expectedResults = Arrays.asList(makeRow(query, "2011-04-01", "alias", ComparableStringArray.of("a", "preferred"), "quality", "automotive", "rows", 2L, "idx", 282L), makeRow(query, "2011-04-01", "alias", ComparableStringArray.of("b", "preferred"), "quality", "business", "rows", 2L, "idx", 230L), makeRow(query, "2011-04-01", "alias", ComparableStringArray.of("e", "preferred"), "quality", "entertainment", "rows", 2L, "idx", 324L), makeRow(query, "2011-04-01", "alias", ComparableStringArray.of("h", "preferred"), "quality", "health", "rows", 2L, "idx", 233L), makeRow(query, "2011-04-01", "alias", ComparableStringArray.of("m", "preferred"), "quality", "mezzanine", "rows", 6L, "idx", 5317L), makeRow(query, "2011-04-01", "alias", ComparableStringArray.of("n", "preferred"), "quality", "news", "rows", 2L, "idx", 235L), makeRow(query, "2011-04-01", "alias", ComparableStringArray.of("p", "preferred"), "quality", "premium", "rows", 6L, "idx", 5405L), makeRow(query, "2011-04-01", "alias", ComparableStringArray.of("preferred", "t"), "quality", "technology", "rows", 2L, "idx", 175L), makeRow(query, "2011-04-01", "alias", ComparableStringArray.of("preferred", "t"), "quality", "travel", "rows", 2L, "idx", 245L));
    Iterable<ResultRow> results = GroupByQueryRunnerTestHelper.runQuery(factory, runner, query);
    TestHelper.assertExpectedObjects(expectedResults, results, "multi-value-dims-groupby-arrays");
    query = makeQueryBuilder().setDataSource(QueryRunnerTestHelper.DATA_SOURCE).setQuerySegmentSpec(QueryRunnerTestHelper.FIRST_TO_THIRD).setVirtualColumns(new ExpressionVirtualColumn("v0", "mv_to_array(placementish)", ColumnType.STRING_ARRAY, ExprMacroTable.nil())).setDimensions(new DefaultDimensionSpec("v0", "alias", ColumnType.STRING_ARRAY), new DefaultDimensionSpec("quality", "quality")).setLimitSpec(new DefaultLimitSpec(ImmutableList.of(new OrderByColumnSpec("alias", OrderByColumnSpec.Direction.DESCENDING, StringComparators.LEXICOGRAPHIC), new OrderByColumnSpec("quality", OrderByColumnSpec.Direction.DESCENDING, StringComparators.LEXICOGRAPHIC)), Integer.MAX_VALUE - 1)).setAggregatorSpecs(QueryRunnerTestHelper.ROWS_COUNT, new LongSumAggregatorFactory("idx", "index")).setGranularity(QueryRunnerTestHelper.ALL_GRAN).build();
    Collections.reverse(expectedResults);
    results = GroupByQueryRunnerTestHelper.runQuery(factory, runner, query);
    TestHelper.assertExpectedObjects(expectedResults, results, "multi-value-dims-groupby-arrays-descending");
}

Example 54 with DefaultLimitSpec

use of org.apache.druid.query.groupby.orderby.DefaultLimitSpec in project druid by druid-io.

the class DruidQuery method toGroupByQuery.

/**
 * Return this query as a GroupBy query, or null if this query is not compatible with GroupBy.
 *
 * @return query or null
 */
@Nullable
private GroupByQuery toGroupByQuery(final QueryFeatureInspector queryFeatureInspector) {
    if (grouping == null) {
        return null;
    }
    if (sorting != null && sorting.getOffsetLimit().hasLimit() && sorting.getOffsetLimit().getLimit() <= 0) {
        // Cannot handle zero or negative limits.
        return null;
    }
    final Pair<DataSource, Filtration> dataSourceFiltrationPair = getFiltration(dataSource, filter, virtualColumnRegistry);
    final DataSource newDataSource = dataSourceFiltrationPair.lhs;
    final Filtration filtration = dataSourceFiltrationPair.rhs;
    final DimFilterHavingSpec havingSpec;
    if (grouping.getHavingFilter() != null) {
        havingSpec = new DimFilterHavingSpec(Filtration.create(grouping.getHavingFilter()).optimizeFilterOnly(grouping.getOutputRowSignature()).getDimFilter(), true);
    } else {
        havingSpec = null;
    }
    final List<PostAggregator> postAggregators = new ArrayList<>(grouping.getPostAggregators());
    if (sorting != null && sorting.getProjection() != null) {
        postAggregators.addAll(sorting.getProjection().getPostAggregators());
    }
    GroupByQuery query = new GroupByQuery(newDataSource, filtration.getQuerySegmentSpec(), getVirtualColumns(true), filtration.getDimFilter(), Granularities.ALL, grouping.getDimensionSpecs(), grouping.getAggregatorFactories(), postAggregators, havingSpec, Optional.ofNullable(sorting).orElse(Sorting.none()).limitSpec(), grouping.getSubtotals().toSubtotalsSpec(grouping.getDimensionSpecs()), ImmutableSortedMap.copyOf(plannerContext.getQueryContext()));
    // We don't apply timestamp computation optimization yet when limit is pushed down. Maybe someday.
    if (query.getLimitSpec() instanceof DefaultLimitSpec && query.isApplyLimitPushDown()) {
        return query;
    }
    Map<String, Object> theContext = new HashMap<>();
    Granularity queryGranularity = null;
    // now, part of the query plan logic is handled in GroupByStrategyV2.
    if (!grouping.getDimensions().isEmpty()) {
        for (DimensionExpression dimensionExpression : grouping.getDimensions()) {
            Granularity granularity = Expressions.toQueryGranularity(dimensionExpression.getDruidExpression(), plannerContext.getExprMacroTable());
            if (granularity == null) {
                continue;
            }
            if (queryGranularity != null) {
                // group by more than one timestamp_floor
                // eg: group by timestamp_floor(__time to DAY),timestamp_floor(__time, to HOUR)
                queryGranularity = null;
                break;
            }
            queryGranularity = granularity;
            int timestampDimensionIndexInDimensions = grouping.getDimensions().indexOf(dimensionExpression);
            // these settings will only affect the most inner query sent to the down streaming compute nodes
            theContext.put(GroupByQuery.CTX_TIMESTAMP_RESULT_FIELD, dimensionExpression.getOutputName());
            theContext.put(GroupByQuery.CTX_TIMESTAMP_RESULT_FIELD_INDEX, timestampDimensionIndexInDimensions);
            theContext.put(GroupByQuery.CTX_TIMESTAMP_RESULT_FIELD_GRANULARITY, queryGranularity);
        }
    }
    if (queryGranularity == null) {
        return query;
    }
    return query.withOverriddenContext(theContext);
}

Example 55 with DefaultLimitSpec

use of org.apache.druid.query.groupby.orderby.DefaultLimitSpec in project druid by druid-io.

the class MultiValuedDimensionTest method testGroupByExpressionMultiMultiAutoAutoDupeIdentifier.

@Test
public void testGroupByExpressionMultiMultiAutoAutoDupeIdentifier() {
    GroupByQuery query = GroupByQuery.builder().setDataSource("xx").setQuerySegmentSpec(new LegacySegmentSpec("1970/3000")).setGranularity(Granularities.ALL).setDimensions(new DefaultDimensionSpec("texpr", "texpr")).setVirtualColumns(new ExpressionVirtualColumn("texpr", "concat(tags, tags)", ColumnType.STRING, TestExprMacroTable.INSTANCE)).setLimitSpec(new DefaultLimitSpec(ImmutableList.of(new OrderByColumnSpec("count", OrderByColumnSpec.Direction.DESCENDING)), 5)).setAggregatorSpecs(new CountAggregatorFactory("count")).setContext(context).build();
    Sequence<ResultRow> result = helper.runQueryOnSegmentsObjs(ImmutableList.of(new QueryableIndexSegment(queryableIndex, SegmentId.dummy("sid1")), new IncrementalIndexSegment(incrementalIndex, SegmentId.dummy("sid2"))), query);
    List<ResultRow> expectedResults = Arrays.asList(GroupByQueryRunnerTestHelper.createExpectedRow(query, "1970", "texpr", "t3t3", "count", 4L), GroupByQueryRunnerTestHelper.createExpectedRow(query, "1970", "texpr", "t5t5", "count", 4L), GroupByQueryRunnerTestHelper.createExpectedRow(query, "1970", "texpr", NullHandling.emptyToNullIfNeeded(""), "count", 2L), GroupByQueryRunnerTestHelper.createExpectedRow(query, "1970", "texpr", "t1t1", "count", 2L), GroupByQueryRunnerTestHelper.createExpectedRow(query, "1970", "texpr", "t2t2", "count", 2L));
    TestHelper.assertExpectedObjects(expectedResults, result.toList(), "expr-multi-multi-auto-auto-self");
}