use of org.apache.druid.query.QueryDataSource in project druid by druid-io.
the class CalciteJoinQueryTest method testJoinWithLimitBeforeJoining.
@Test
public void testJoinWithLimitBeforeJoining() throws Exception {
// Cannot vectorize JOIN operator.
cannotVectorize();
testQuery("SELECT t1.dim2, AVG(t1.m2) FROM (SELECT * FROM foo LIMIT 10) AS t1 INNER JOIN foo AS t2 ON t1.m1 = t2.m1 GROUP BY t1.dim2", ImmutableList.of(GroupByQuery.builder().setDataSource(join(new QueryDataSource(newScanQueryBuilder().dataSource(CalciteTests.DATASOURCE1).intervals(querySegmentSpec(Filtration.eternity())).columns("dim2", "m1", "m2").context(QUERY_CONTEXT_DEFAULT).limit(10).build()), new QueryDataSource(newScanQueryBuilder().dataSource(CalciteTests.DATASOURCE1).intervals(querySegmentSpec(Filtration.eternity())).columns(ImmutableList.of("m1")).resultFormat(ScanQuery.ResultFormat.RESULT_FORMAT_COMPACTED_LIST).context(QUERY_CONTEXT_DEFAULT).build()), "j0.", equalsCondition(DruidExpression.ofColumn(ColumnType.FLOAT, "m1"), DruidExpression.ofColumn(ColumnType.FLOAT, "j0.m1")), JoinType.INNER)).setInterval(querySegmentSpec(Filtration.eternity())).setDimensions(new DefaultDimensionSpec("dim2", "d0", ColumnType.STRING)).setGranularity(Granularities.ALL).setAggregatorSpecs(useDefault ? aggregators(new DoubleSumAggregatorFactory("a0:sum", "m2"), new CountAggregatorFactory("a0:count")) : aggregators(new DoubleSumAggregatorFactory("a0:sum", "m2"), new FilteredAggregatorFactory(new CountAggregatorFactory("a0:count"), not(selector("m2", null, null))))).setPostAggregatorSpecs(ImmutableList.of(new ArithmeticPostAggregator("a0", "quotient", ImmutableList.of(new FieldAccessPostAggregator(null, "a0:sum"), new FieldAccessPostAggregator(null, "a0:count"))))).setContext(QUERY_CONTEXT_DEFAULT).build()), NullHandling.sqlCompatible() ? ImmutableList.of(new Object[] { null, 4.0 }, new Object[] { "", 3.0 }, new Object[] { "a", 2.5 }, new Object[] { "abc", 5.0 }) : ImmutableList.of(new Object[] { "", 3.6666666666666665 }, new Object[] { "a", 2.5 }, new Object[] { "abc", 5.0 }));
}
use of org.apache.druid.query.QueryDataSource in project druid by druid-io.
the class CalciteCorrelatedQueryTest method testCorrelatedSubqueryWithCorrelatedQueryFilter.
@Test
@Parameters(source = QueryContextForJoinProvider.class)
public void testCorrelatedSubqueryWithCorrelatedQueryFilter(Map<String, Object> queryContext) throws Exception {
cannotVectorize();
queryContext = withLeftDirectAccessEnabled(queryContext);
testQuery("select country, ANY_VALUE(\n" + " select max(\"users\") from (\n" + " select floor(__time to day), count(user) \"users\" from visits f where f.country = visits.country and f.city = 'A' group by 1\n" + " )\n" + " ) as \"dailyVisits\"\n" + "from visits \n" + " where city = 'B'" + " group by 1", queryContext, ImmutableList.of(GroupByQuery.builder().setDataSource(join(new TableDataSource(CalciteTests.USERVISITDATASOURCE), new QueryDataSource(GroupByQuery.builder().setDataSource(GroupByQuery.builder().setDataSource(CalciteTests.USERVISITDATASOURCE).setQuerySegmentSpec(querySegmentSpec(Intervals.ETERNITY)).setVirtualColumns(new ExpressionVirtualColumn("v0", "timestamp_floor(\"__time\",'P1D',null,'UTC')", ColumnType.LONG, TestExprMacroTable.INSTANCE)).setDimensions(new DefaultDimensionSpec("v0", "d0", ColumnType.LONG), new DefaultDimensionSpec("country", "d1")).setAggregatorSpecs(new FilteredAggregatorFactory(new CountAggregatorFactory("a0"), not(selector("user", null, null)))).setDimFilter(and(selector("city", "A", null), not(selector("country", null, null)))).setContext(withTimestampResultContext(queryContext, "d0", Granularities.DAY)).setGranularity(new AllGranularity()).build()).setQuerySegmentSpec(querySegmentSpec(Intervals.ETERNITY)).setDimensions(new DefaultDimensionSpec("d1", "_d0")).setAggregatorSpecs(new LongMaxAggregatorFactory("_a0", "a0")).setGranularity(new AllGranularity()).setContext(queryContext).build()), "j0.", equalsCondition(makeColumnExpression("country"), makeColumnExpression("j0._d0")), JoinType.LEFT, selector("city", "B", null))).setQuerySegmentSpec(querySegmentSpec(Intervals.ETERNITY)).setDimensions(new DefaultDimensionSpec("country", "d0")).setAggregatorSpecs(new LongAnyAggregatorFactory("a0", "j0._a0")).setGranularity(new AllGranularity()).setContext(queryContext).build()), ImmutableList.of(new Object[] { "canada", 2L }));
}
use of org.apache.druid.query.QueryDataSource in project druid by druid-io.
the class CalciteCorrelatedQueryTest method testCorrelatedSubquery.
@Test
@Parameters(source = QueryContextForJoinProvider.class)
public void testCorrelatedSubquery(Map<String, Object> queryContext) throws Exception {
cannotVectorize();
queryContext = withLeftDirectAccessEnabled(queryContext);
testQuery("select country, ANY_VALUE(\n" + " select avg(\"users\") from (\n" + " select floor(__time to day), count(distinct user) \"users\" from visits f where f.country = visits.country group by 1\n" + " )\n" + " ) as \"DAU\"\n" + "from visits \n" + "group by 1", queryContext, ImmutableList.of(GroupByQuery.builder().setDataSource(join(new TableDataSource(CalciteTests.USERVISITDATASOURCE), new QueryDataSource(GroupByQuery.builder().setDataSource(GroupByQuery.builder().setDataSource(CalciteTests.USERVISITDATASOURCE).setQuerySegmentSpec(querySegmentSpec(Intervals.ETERNITY)).setVirtualColumns(new ExpressionVirtualColumn("v0", "timestamp_floor(\"__time\",'P1D',null,'UTC')", ColumnType.LONG, TestExprMacroTable.INSTANCE)).setDimFilter(not(selector("country", null, null))).setDimensions(new DefaultDimensionSpec("v0", "d0", ColumnType.LONG), new DefaultDimensionSpec("country", "d1")).setAggregatorSpecs(new CardinalityAggregatorFactory("a0:a", null, Collections.singletonList(new DefaultDimensionSpec("user", "user")), false, true)).setPostAggregatorSpecs(Collections.singletonList(new HyperUniqueFinalizingPostAggregator("a0", "a0:a"))).setContext(withTimestampResultContext(queryContext, "d0", Granularities.DAY)).setGranularity(new AllGranularity()).build()).setQuerySegmentSpec(querySegmentSpec(Intervals.ETERNITY)).setDimensions(new DefaultDimensionSpec("d1", "_d0")).setAggregatorSpecs(new LongSumAggregatorFactory("_a0:sum", "a0"), useDefault ? new CountAggregatorFactory("_a0:count") : new FilteredAggregatorFactory(new CountAggregatorFactory("_a0:count"), not(selector("a0", null, null)))).setPostAggregatorSpecs(Collections.singletonList(new ArithmeticPostAggregator("_a0", "quotient", Arrays.asList(new FieldAccessPostAggregator(null, "_a0:sum"), new FieldAccessPostAggregator(null, "_a0:count"))))).setGranularity(new AllGranularity()).setContext(queryContext).build()), "j0.", equalsCondition(makeColumnExpression("country"), makeColumnExpression("j0._d0")), JoinType.LEFT)).setQuerySegmentSpec(querySegmentSpec(Intervals.ETERNITY)).setDimensions(new DefaultDimensionSpec("country", "d0")).setAggregatorSpecs(new LongAnyAggregatorFactory("a0", "j0._a0")).setGranularity(new AllGranularity()).setContext(queryContext).build()), ImmutableList.of(new Object[] { "India", 2L }, new Object[] { "USA", 1L }, new Object[] { "canada", 3L }));
}
use of org.apache.druid.query.QueryDataSource in project druid by druid-io.
the class CalciteJoinQueryTest method testJoinOnTimeseriesWithFloorOnTime.
@Test
public void testJoinOnTimeseriesWithFloorOnTime() throws Exception {
// Cannot vectorize JOIN operator.
cannotVectorize();
testQuery("SELECT CAST(__time AS BIGINT), m1, ANY_VALUE(dim3, 100) FROM foo WHERE (TIME_FLOOR(__time, 'PT1H'), m1) IN\n" + " (\n" + " SELECT TIME_FLOOR(__time, 'PT1H') AS t1, MIN(m1) AS t2 FROM foo WHERE dim3 = 'b'\n" + " AND __time BETWEEN '1994-04-29 00:00:00' AND '2020-01-11 00:00:00' GROUP BY 1\n" + " )\n" + "GROUP BY 1, 2\n", ImmutableList.of(GroupByQuery.builder().setDataSource(join(new TableDataSource(CalciteTests.DATASOURCE1), new QueryDataSource(Druids.newTimeseriesQueryBuilder().dataSource(CalciteTests.DATASOURCE1).intervals(querySegmentSpec(Intervals.of("1994-04-29/2020-01-11T00:00:00.001Z"))).filters(selector("dim3", "b", null)).granularity(new PeriodGranularity(Period.hours(1), null, DateTimeZone.UTC)).aggregators(aggregators(new FloatMinAggregatorFactory("a0", "m1"))).context(getTimeseriesContextWithFloorTime(TIMESERIES_CONTEXT_BY_GRAN, "d0")).build()), "j0.", "((timestamp_floor(\"__time\",'PT1H',null,'UTC') == \"j0.d0\") && (\"m1\" == \"j0.a0\"))", JoinType.INNER)).setInterval(querySegmentSpec(Filtration.eternity())).setDimensions(new DefaultDimensionSpec("__time", "d0", ColumnType.LONG), new DefaultDimensionSpec("m1", "d1", ColumnType.FLOAT)).setGranularity(Granularities.ALL).setAggregatorSpecs(aggregators(new StringAnyAggregatorFactory("a0", "dim3", 100))).setContext(QUERY_CONTEXT_DEFAULT).build()), ImmutableList.of(new Object[] { 946684800000L, 1.0f, "[a, b]" }, new Object[] { 946771200000L, 2.0f, "[b, c]" }));
}
use of org.apache.druid.query.QueryDataSource in project druid by druid-io.
the class CalciteJoinQueryTest method testInnerJoinTwoLookupsToTableUsingNumericColumnInReverse.
@Test
@Parameters(source = QueryContextForJoinProvider.class)
public void testInnerJoinTwoLookupsToTableUsingNumericColumnInReverse(Map<String, Object> queryContext) throws Exception {
// Like "testInnerJoinTwoLookupsToTableUsingNumericColumn", but the tables are specified backwards.
cannotVectorize();
testQuery("SELECT COUNT(*)\n" + "FROM lookup.lookyloo l1\n" + "INNER JOIN lookup.lookyloo l2 ON l1.k = l2.k\n" + "INNER JOIN foo on l2.k = foo.m1", queryContext, ImmutableList.of(Druids.newTimeseriesQueryBuilder().dataSource(join(join(new LookupDataSource("lookyloo"), new LookupDataSource("lookyloo"), "j0.", equalsCondition(makeColumnExpression("k"), makeColumnExpression("j0.k")), JoinType.INNER), new QueryDataSource(newScanQueryBuilder().dataSource(CalciteTests.DATASOURCE1).intervals(querySegmentSpec(Filtration.eternity())).columns("m1").context(QUERY_CONTEXT_DEFAULT).build()), "_j0.", equalsCondition(makeExpression(ColumnType.DOUBLE, "CAST(\"j0.k\", 'DOUBLE')"), DruidExpression.ofColumn(ColumnType.DOUBLE, "_j0.m1")), JoinType.INNER)).intervals(querySegmentSpec(Filtration.eternity())).granularity(Granularities.ALL).aggregators(new CountAggregatorFactory("a0")).context(QUERY_CONTEXT_DEFAULT).build()), ImmutableList.of(new Object[] { 1L }));
}
Aggregations