Examples with PostAggregator io.druid.query.aggregation.PostAggregator used on opensource projects

Example 21 with PostAggregator

use of io.druid.query.aggregation.PostAggregator in project druid by druid-io.

the class GroupByQueryRunnerTest method testGroupByWithOrderOnHyperUnique.

@Test
public void testGroupByWithOrderOnHyperUnique() {
    GroupByQuery query = new GroupByQuery.Builder().setDataSource(QueryRunnerTestHelper.dataSource).setGranularity(QueryRunnerTestHelper.allGran).setDimensions(Arrays.<DimensionSpec>asList(new DefaultDimensionSpec(QueryRunnerTestHelper.marketDimension, QueryRunnerTestHelper.marketDimension))).setInterval(QueryRunnerTestHelper.fullOnInterval).setLimitSpec(new DefaultLimitSpec(Lists.newArrayList(new OrderByColumnSpec(QueryRunnerTestHelper.uniqueMetric, OrderByColumnSpec.Direction.DESCENDING)), 3)).setAggregatorSpecs(Lists.<AggregatorFactory>newArrayList(QueryRunnerTestHelper.qualityUniques)).setPostAggregatorSpecs(Lists.<PostAggregator>newArrayList(new HyperUniqueFinalizingPostAggregator(QueryRunnerTestHelper.hyperUniqueFinalizingPostAggMetric, QueryRunnerTestHelper.uniqueMetric))).build();
    List<Row> expectedResults = Arrays.asList(GroupByQueryRunnerTestHelper.createExpectedRow("1970-01-01T00:00:00.000Z", "market", "spot", QueryRunnerTestHelper.uniqueMetric, QueryRunnerTestHelper.UNIQUES_9, QueryRunnerTestHelper.hyperUniqueFinalizingPostAggMetric, QueryRunnerTestHelper.UNIQUES_9), GroupByQueryRunnerTestHelper.createExpectedRow("1970-01-01T00:00:00.000Z", "market", "upfront", QueryRunnerTestHelper.uniqueMetric, QueryRunnerTestHelper.UNIQUES_2, QueryRunnerTestHelper.hyperUniqueFinalizingPostAggMetric, QueryRunnerTestHelper.UNIQUES_2), GroupByQueryRunnerTestHelper.createExpectedRow("1970-01-01T00:00:00.000Z", "market", "total_market", QueryRunnerTestHelper.uniqueMetric, QueryRunnerTestHelper.UNIQUES_2, QueryRunnerTestHelper.hyperUniqueFinalizingPostAggMetric, QueryRunnerTestHelper.UNIQUES_2));
    Iterable<Row> results = GroupByQueryRunnerTestHelper.runQuery(factory, runner, query);
    TestHelper.assertExpectedObjects(expectedResults, results, "order-limit");
}

Example 22 with PostAggregator

use of io.druid.query.aggregation.PostAggregator in project druid by druid-io.

the class GroupByQueryRunnerTest method testGroupByTimeExtractionWithNulls.

@Test
public void testGroupByTimeExtractionWithNulls() {
    final DimExtractionFn nullWednesdays = new DimExtractionFn() {

        @Override
        public String apply(String dimValue) {
            if ("Wednesday".equals(dimValue)) {
                return null;
            } else {
                return dimValue;
            }
        }

        @Override
        public byte[] getCacheKey() {
            throw new UnsupportedOperationException();
        }

        @Override
        public boolean preservesOrdering() {
            return false;
        }

        @Override
        public ExtractionType getExtractionType() {
            return ExtractionType.MANY_TO_ONE;
        }
    };
    GroupByQuery query = GroupByQuery.builder().setDataSource(QueryRunnerTestHelper.dataSource).setQuerySegmentSpec(QueryRunnerTestHelper.fullOnInterval).setDimensions(Lists.newArrayList(new DefaultDimensionSpec("market", "market"), new ExtractionDimensionSpec(Column.TIME_COLUMN_NAME, "dayOfWeek", new CascadeExtractionFn(new ExtractionFn[] { new TimeFormatExtractionFn("EEEE", null, null, null, false), nullWednesdays })))).setAggregatorSpecs(Arrays.asList(QueryRunnerTestHelper.rowsCount, QueryRunnerTestHelper.indexDoubleSum)).setPostAggregatorSpecs(Arrays.<PostAggregator>asList(QueryRunnerTestHelper.addRowsIndexConstant)).setGranularity(QueryRunnerTestHelper.allGran).setDimFilter(new OrDimFilter(Arrays.<DimFilter>asList(new SelectorDimFilter("market", "spot", null), new SelectorDimFilter("market", "upfront", null)))).build();
    List<Row> expectedResults = Arrays.asList(GroupByQueryRunnerTestHelper.createExpectedRow("1970-01-01", "dayOfWeek", null, "market", "spot", "index", 14271.368591308594, "rows", 126L, "addRowsIndexConstant", 14398.368591308594), GroupByQueryRunnerTestHelper.createExpectedRow("1970-01-01", "dayOfWeek", "Friday", "market", "spot", "index", 13219.574157714844, "rows", 117L, "addRowsIndexConstant", 13337.574157714844), GroupByQueryRunnerTestHelper.createExpectedRow("1970-01-01", "dayOfWeek", "Monday", "market", "spot", "index", 13557.738830566406, "rows", 117L, "addRowsIndexConstant", 13675.738830566406), GroupByQueryRunnerTestHelper.createExpectedRow("1970-01-01", "dayOfWeek", "Saturday", "market", "spot", "index", 13493.751281738281, "rows", 117L, "addRowsIndexConstant", 13611.751281738281), GroupByQueryRunnerTestHelper.createExpectedRow("1970-01-01", "dayOfWeek", "Sunday", "market", "spot", "index", 13585.541015625, "rows", 117L, "addRowsIndexConstant", 13703.541015625), GroupByQueryRunnerTestHelper.createExpectedRow("1970-01-01", "dayOfWeek", "Thursday", "market", "spot", "index", 14279.127197265625, "rows", 126L, "addRowsIndexConstant", 14406.127197265625), GroupByQueryRunnerTestHelper.createExpectedRow("1970-01-01", "dayOfWeek", "Tuesday", "market", "spot", "index", 13199.471435546875, "rows", 117L, "addRowsIndexConstant", 13317.471435546875), GroupByQueryRunnerTestHelper.createExpectedRow("1970-01-01", "dayOfWeek", null, "market", "upfront", "index", 28985.5751953125, "rows", 28L, "addRowsIndexConstant", 29014.5751953125), GroupByQueryRunnerTestHelper.createExpectedRow("1970-01-01", "dayOfWeek", "Friday", "market", "upfront", "index", 27297.8623046875, "rows", 26L, "addRowsIndexConstant", 27324.8623046875), GroupByQueryRunnerTestHelper.createExpectedRow("1970-01-01", "dayOfWeek", "Monday", "market", "upfront", "index", 27619.58447265625, "rows", 26L, "addRowsIndexConstant", 27646.58447265625), GroupByQueryRunnerTestHelper.createExpectedRow("1970-01-01", "dayOfWeek", "Saturday", "market", "upfront", "index", 27820.83154296875, "rows", 26L, "addRowsIndexConstant", 27847.83154296875), GroupByQueryRunnerTestHelper.createExpectedRow("1970-01-01", "dayOfWeek", "Sunday", "market", "upfront", "index", 24791.223876953125, "rows", 26L, "addRowsIndexConstant", 24818.223876953125), GroupByQueryRunnerTestHelper.createExpectedRow("1970-01-01", "dayOfWeek", "Thursday", "market", "upfront", "index", 28562.748901367188, "rows", 28L, "addRowsIndexConstant", 28591.748901367188), GroupByQueryRunnerTestHelper.createExpectedRow("1970-01-01", "dayOfWeek", "Tuesday", "market", "upfront", "index", 26968.280639648438, "rows", 26L, "addRowsIndexConstant", 26995.280639648438));
    Iterable<Row> results = GroupByQueryRunnerTestHelper.runQuery(factory, runner, query);
    TestHelper.assertExpectedObjects(expectedResults, results, "");
}

Example 23 with PostAggregator

use of io.druid.query.aggregation.PostAggregator in project druid by druid-io.

the class GroupByQueryRunnerTest method testSubqueryWithMultiColumnAggregators.

@Test
public void testSubqueryWithMultiColumnAggregators() {
    final GroupByQuery subquery = GroupByQuery.builder().setDataSource(QueryRunnerTestHelper.dataSource).setQuerySegmentSpec(QueryRunnerTestHelper.firstToThird).setDimensions(Lists.<DimensionSpec>newArrayList(new DefaultDimensionSpec("quality", "alias"))).setDimFilter(new JavaScriptDimFilter("market", "function(dim){ return true; }", null, JavaScriptConfig.getEnabledInstance())).setAggregatorSpecs(Arrays.asList(QueryRunnerTestHelper.rowsCount, new DoubleSumAggregatorFactory("idx_subagg", "index"), new JavaScriptAggregatorFactory("js_agg", Arrays.asList("index", "market"), "function(current, index, dim){return current + index + dim.length;}", "function(){return 0;}", "function(a,b){return a + b;}", JavaScriptConfig.getEnabledInstance()))).setPostAggregatorSpecs(Arrays.<PostAggregator>asList(new ArithmeticPostAggregator("idx_subpostagg", "+", Arrays.asList(new FieldAccessPostAggregator("the_idx_subagg", "idx_subagg"), new ConstantPostAggregator("thousand", 1000))))).setHavingSpec(new BaseHavingSpec() {

        @Override
        public boolean eval(Row row) {
            return (row.getFloatMetric("idx_subpostagg") < 3800);
        }

        @Override
        public byte[] getCacheKey() {
            return new byte[0];
        }
    }).addOrderByColumn("alias").setGranularity(QueryRunnerTestHelper.dayGran).build();
    final GroupByQuery query = GroupByQuery.builder().setDataSource(subquery).setQuerySegmentSpec(QueryRunnerTestHelper.firstToThird).setDimensions(Lists.<DimensionSpec>newArrayList(new DefaultDimensionSpec("alias", "alias"))).setAggregatorSpecs(Arrays.asList(new LongSumAggregatorFactory("rows", "rows"), new LongSumAggregatorFactory("idx", "idx_subpostagg"), new DoubleSumAggregatorFactory("js_outer_agg", "js_agg"))).setPostAggregatorSpecs(Arrays.<PostAggregator>asList(new ArithmeticPostAggregator("idx_post", "+", Arrays.asList(new FieldAccessPostAggregator("the_idx_agg", "idx"), new ConstantPostAggregator("ten_thousand", 10000))))).setLimitSpec(new DefaultLimitSpec(Arrays.asList(new OrderByColumnSpec("alias", OrderByColumnSpec.Direction.DESCENDING)), 5)).setGranularity(QueryRunnerTestHelper.dayGran).build();
    List<Row> expectedResults = Arrays.asList(GroupByQueryRunnerTestHelper.createExpectedRow("2011-04-01", "alias", "travel", "rows", 1L, "idx_post", 11119.0, "idx", 1119L, "js_outer_agg", 123.92274475097656), GroupByQueryRunnerTestHelper.createExpectedRow("2011-04-01", "alias", "technology", "rows", 1L, "idx_post", 11078.0, "idx", 1078L, "js_outer_agg", 82.62254333496094), GroupByQueryRunnerTestHelper.createExpectedRow("2011-04-01", "alias", "news", "rows", 1L, "idx_post", 11121.0, "idx", 1121L, "js_outer_agg", 125.58358001708984), GroupByQueryRunnerTestHelper.createExpectedRow("2011-04-01", "alias", "health", "rows", 1L, "idx_post", 11120.0, "idx", 1120L, "js_outer_agg", 124.13470458984375), GroupByQueryRunnerTestHelper.createExpectedRow("2011-04-01", "alias", "entertainment", "rows", 1L, "idx_post", 11158.0, "idx", 1158L, "js_outer_agg", 162.74722290039062));
    // Subqueries are handled by the ToolChest
    Iterable<Row> results = GroupByQueryRunnerTestHelper.runQuery(factory, runner, query);
    TestHelper.assertExpectedObjects(expectedResults, results, "");
}

Example 24 with PostAggregator

use of io.druid.query.aggregation.PostAggregator in project druid by druid-io.

the class DefaultLimitSpecTest method testBuildWithExplicitOrder.

@Test
public void testBuildWithExplicitOrder() {
    DefaultLimitSpec limitSpec = new DefaultLimitSpec(ImmutableList.of(new OrderByColumnSpec("k1", OrderByColumnSpec.Direction.ASCENDING)), 2);
    Function<Sequence<Row>, Sequence<Row>> limitFn = limitSpec.build(ImmutableList.<DimensionSpec>of(new DefaultDimensionSpec("k1", "k1")), ImmutableList.<AggregatorFactory>of(new LongSumAggregatorFactory("k2", "k2")), ImmutableList.<PostAggregator>of(new ConstantPostAggregator("k3", 1L)));
    Assert.assertEquals(ImmutableList.of(testRowsList.get(0), testRowsList.get(1)), Sequences.toList(limitFn.apply(testRowsSequence), new ArrayList<Row>()));
    // if there is an aggregator with same name then that is used to build ordering
    limitFn = limitSpec.build(ImmutableList.<DimensionSpec>of(new DefaultDimensionSpec("k1", "k1")), ImmutableList.<AggregatorFactory>of(new LongSumAggregatorFactory("k1", "k1")), ImmutableList.<PostAggregator>of(new ConstantPostAggregator("k3", 1L)));
    Assert.assertEquals(ImmutableList.of(testRowsList.get(2), testRowsList.get(0)), Sequences.toList(limitFn.apply(testRowsSequence), new ArrayList<Row>()));
    // if there is a post-aggregator with same name then that is used to build ordering
    limitFn = limitSpec.build(ImmutableList.<DimensionSpec>of(new DefaultDimensionSpec("k1", "k1")), ImmutableList.<AggregatorFactory>of(new LongSumAggregatorFactory("k2", "k2")), ImmutableList.<PostAggregator>of(new ArithmeticPostAggregator("k1", "+", ImmutableList.<PostAggregator>of(new ConstantPostAggregator("x", 1), new ConstantPostAggregator("y", 1)))));
    Assert.assertEquals((List) ImmutableList.of(testRowsList.get(2), testRowsList.get(0)), (List) Sequences.toList(limitFn.apply(testRowsSequence), new ArrayList<Row>()));
    // makes same result
    limitFn = limitSpec.build(ImmutableList.<DimensionSpec>of(new DefaultDimensionSpec("k1", "k1")), ImmutableList.<AggregatorFactory>of(new LongSumAggregatorFactory("k2", "k2")), ImmutableList.<PostAggregator>of(new ExpressionPostAggregator("k1", "1 + 1")));
    Assert.assertEquals((List) ImmutableList.of(testRowsList.get(2), testRowsList.get(0)), (List) Sequences.toList(limitFn.apply(testRowsSequence), new ArrayList<Row>()));
}

Example 25 with PostAggregator

use of io.druid.query.aggregation.PostAggregator in project druid by druid-io.

the class TopNBinaryFnBenchmark method setUp.

@Override
protected void setUp() throws Exception {
    final ConstantPostAggregator constant = new ConstantPostAggregator("const", 1L);
    final FieldAccessPostAggregator rowsPostAgg = new FieldAccessPostAggregator("rows", "rows");
    final FieldAccessPostAggregator indexPostAgg = new FieldAccessPostAggregator("index", "index");
    final List<AggregatorFactory> aggregatorFactories = new ArrayList<>();
    aggregatorFactories.add(new CountAggregatorFactory("rows"));
    aggregatorFactories.add(new LongSumAggregatorFactory("index", "index"));
    for (int i = 1; i < aggCount; i++) {
        aggregatorFactories.add(new CountAggregatorFactory("rows" + i));
    }
    final List<PostAggregator> postAggregators = new ArrayList<>();
    for (int i = 0; i < postAggCount; i++) {
        postAggregators.add(new ArithmeticPostAggregator("addrowsindexconstant" + i, "+", Lists.newArrayList(constant, rowsPostAgg, indexPostAgg)));
    }
    final DateTime currTime = new DateTime();
    List<Map<String, Object>> list = new ArrayList<>();
    for (int i = 0; i < threshold; i++) {
        Map<String, Object> res = new HashMap<>();
        res.put("testdim", "" + i);
        res.put("rows", 1L);
        for (int j = 0; j < aggCount; j++) {
            res.put("rows" + j, 1L);
        }
        res.put("index", 1L);
        list.add(res);
    }
    result1 = new Result<>(currTime, new TopNResultValue(list));
    List<Map<String, Object>> list2 = new ArrayList<>();
    for (int i = 0; i < threshold; i++) {
        Map<String, Object> res = new HashMap<>();
        res.put("testdim", "" + i);
        res.put("rows", 2L);
        for (int j = 0; j < aggCount; j++) {
            res.put("rows" + j, 2L);
        }
        res.put("index", 2L);
        list2.add(res);
    }
    result2 = new Result<>(currTime, new TopNResultValue(list2));
    fn = new TopNBinaryFn(TopNResultMerger.identity, Granularities.ALL, new DefaultDimensionSpec("testdim", null), new NumericTopNMetricSpec("index"), 100, aggregatorFactories, postAggregators);
}