Examples with FilteredAggregatorFactory org.apache.druid.query.aggregation.FilteredAggregatorFactory used on opensource projects

Example 6 with FilteredAggregatorFactory

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

the class CalciteQueryTest method testQueryWithSelectProjectAndIdentityProjectDoesNotRename.

@Test
public void testQueryWithSelectProjectAndIdentityProjectDoesNotRename() throws Exception {
    cannotVectorize();
    requireMergeBuffers(3);
    testQuery(PLANNER_CONFIG_NO_HLL.withOverrides(ImmutableMap.of(PlannerConfig.CTX_KEY_USE_GROUPING_SET_FOR_EXACT_DISTINCT, "true")), "SELECT\n" + "(SUM(CASE WHEN (TIMESTAMP '2000-01-04 17:00:00'<=__time AND __time<TIMESTAMP '2022-01-05 17:00:00') THEN 1 ELSE 0 END)*1.0/COUNT(DISTINCT CASE WHEN (TIMESTAMP '2000-01-04 17:00:00'<=__time AND __time<TIMESTAMP '2022-01-05 17:00:00') THEN dim1 END))\n" + "FROM druid.foo\n" + "GROUP BY ()", CalciteTests.REGULAR_USER_AUTH_RESULT, ImmutableList.of(GroupByQuery.builder().setDataSource(new QueryDataSource(GroupByQuery.builder().setDataSource(CalciteTests.DATASOURCE1).setInterval(querySegmentSpec(Filtration.eternity())).setGranularity(Granularities.ALL).setVirtualColumns(expressionVirtualColumn("v0", "case_searched(((947005200000 <= \"__time\") && (\"__time\" < 1641402000000)),\"dim1\",null)", ColumnType.STRING)).setDimensions(dimensions(new DefaultDimensionSpec("v0", "d0", ColumnType.STRING))).setAggregatorSpecs(aggregators(new LongSumAggregatorFactory("a0", null, "case_searched(((947005200000 <= \"__time\") && (\"__time\" < 1641402000000)),1,0)", ExprMacroTable.nil()), new GroupingAggregatorFactory("a1", ImmutableList.of("v0")))).setSubtotalsSpec(ImmutableList.of(ImmutableList.of("d0"), ImmutableList.of())).setContext(QUERY_CONTEXT_DEFAULT).build())).setInterval(querySegmentSpec(Filtration.eternity())).setGranularity(Granularities.ALL).setAggregatorSpecs(aggregators(new FilteredAggregatorFactory(new LongMinAggregatorFactory("_a0", "a0"), selector("a1", "1", null)), new FilteredAggregatorFactory(new CountAggregatorFactory("_a1"), and(not(selector("d0", null, null)), selector("a1", "0", null))))).setPostAggregatorSpecs(Collections.singletonList(new ExpressionPostAggregator("p0", "((\"_a0\" * 1.0) / \"_a1\")", null, ExprMacroTable.nil()))).setContext(QUERY_CONTEXT_DEFAULT).build()), ImmutableList.of(new Object[] { 1.0d }));
}

Example 7 with FilteredAggregatorFactory

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

the class CalciteQueryTest method testTimeseriesEmptyResultsAggregatorDefaultValuesNonVectorized.

@Test
public void testTimeseriesEmptyResultsAggregatorDefaultValuesNonVectorized() throws Exception {
    cannotVectorize();
    // timeseries with all granularity have a single group, so should return default results for given aggregators
    testQuery("SELECT\n" + " ANY_VALUE(dim1, 1024),\n" + " ANY_VALUE(l1),\n" + " EARLIEST(dim1, 1024),\n" + " EARLIEST(l1),\n" + " LATEST(dim1, 1024),\n" + " LATEST(l1),\n" + " ARRAY_AGG(DISTINCT dim3),\n" + " STRING_AGG(DISTINCT dim3, '|'),\n" + " BIT_AND(l1),\n" + " BIT_OR(l1),\n" + " BIT_XOR(l1)\n" + "FROM druid.numfoo WHERE dim2 = 0", ImmutableList.of(Druids.newTimeseriesQueryBuilder().dataSource(CalciteTests.DATASOURCE3).intervals(querySegmentSpec(Filtration.eternity())).filters(bound("dim2", "0", "0", false, false, null, StringComparators.NUMERIC)).granularity(Granularities.ALL).aggregators(aggregators(new StringAnyAggregatorFactory("a0", "dim1", 1024), new LongAnyAggregatorFactory("a1", "l1"), new StringFirstAggregatorFactory("a2", "dim1", null, 1024), new LongFirstAggregatorFactory("a3", "l1", null), new StringLastAggregatorFactory("a4", "dim1", null, 1024), new LongLastAggregatorFactory("a5", "l1", null), new ExpressionLambdaAggregatorFactory("a6", ImmutableSet.of("dim3"), "__acc", "ARRAY<STRING>[]", "ARRAY<STRING>[]", true, true, false, "array_set_add(\"__acc\", \"dim3\")", "array_set_add_all(\"__acc\", \"a6\")", null, null, ExpressionLambdaAggregatorFactory.DEFAULT_MAX_SIZE_BYTES, TestExprMacroTable.INSTANCE), new FilteredAggregatorFactory(new ExpressionLambdaAggregatorFactory("a7", ImmutableSet.of("dim3"), "__acc", "[]", "[]", true, false, false, "array_set_add(\"__acc\", \"dim3\")", "array_set_add_all(\"__acc\", \"a7\")", null, "if(array_length(o) == 0, null, array_to_string(o, '|'))", ExpressionLambdaAggregatorFactory.DEFAULT_MAX_SIZE_BYTES, TestExprMacroTable.INSTANCE), not(selector("dim3", null, null))), new FilteredAggregatorFactory(new ExpressionLambdaAggregatorFactory("a8", ImmutableSet.of("l1"), "__acc", "0", "0", NullHandling.sqlCompatible(), false, false, "bitwiseAnd(\"__acc\", \"l1\")", "bitwiseAnd(\"__acc\", \"a8\")", null, null, ExpressionLambdaAggregatorFactory.DEFAULT_MAX_SIZE_BYTES, TestExprMacroTable.INSTANCE), not(selector("l1", null, null))), new FilteredAggregatorFactory(new ExpressionLambdaAggregatorFactory("a9", ImmutableSet.of("l1"), "__acc", "0", "0", NullHandling.sqlCompatible(), false, false, "bitwiseOr(\"__acc\", \"l1\")", "bitwiseOr(\"__acc\", \"a9\")", null, null, ExpressionLambdaAggregatorFactory.DEFAULT_MAX_SIZE_BYTES, TestExprMacroTable.INSTANCE), not(selector("l1", null, null))), new FilteredAggregatorFactory(new ExpressionLambdaAggregatorFactory("a10", ImmutableSet.of("l1"), "__acc", "0", "0", NullHandling.sqlCompatible(), false, false, "bitwiseXor(\"__acc\", \"l1\")", "bitwiseXor(\"__acc\", \"a10\")", null, null, ExpressionLambdaAggregatorFactory.DEFAULT_MAX_SIZE_BYTES, TestExprMacroTable.INSTANCE), not(selector("l1", null, null))))).context(QUERY_CONTEXT_DEFAULT).build()), ImmutableList.of(useDefault ? new Object[] { "", 0L, "", 0L, "", 0L, null, "", 0L, 0L, 0L } : new Object[] { null, null, null, null, null, null, null, null, null, null, null }));
}

Example 8 with FilteredAggregatorFactory

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

the class CalciteQueryTest method testExpressionCounts.

@Test
public void testExpressionCounts() throws Exception {
    cannotVectorize();
    testQuery("SELECT\n" + " COUNT(reverse(dim2)),\n" + " COUNT(left(dim2, 5)),\n" + " COUNT(strpos(dim2, 'a'))\n" + "FROM druid.numfoo", ImmutableList.of(Druids.newTimeseriesQueryBuilder().dataSource(CalciteTests.DATASOURCE3).intervals(querySegmentSpec(Filtration.eternity())).granularity(Granularities.ALL).virtualColumns(expressionVirtualColumn("v0", "reverse(\"dim2\")", ColumnType.STRING), expressionVirtualColumn("v1", "left(\"dim2\",5)", ColumnType.STRING), expressionVirtualColumn("v2", "(strpos(\"dim2\",'a') + 1)", ColumnType.LONG)).aggregators(aggregators(new FilteredAggregatorFactory(new CountAggregatorFactory("a0"), not(selector("v0", null, null))), new FilteredAggregatorFactory(new CountAggregatorFactory("a1"), not(selector("v1", null, null))), new FilteredAggregatorFactory(new CountAggregatorFactory("a2"), not(selector("v2", null, null))))).context(QUERY_CONTEXT_DEFAULT).build()), ImmutableList.of(useDefault ? // default mode so is 0 + 1 for null rows)
    new Object[] { 3L, 3L, 6L } : new Object[] { 4L, 4L, 4L }));
}

Example 9 with FilteredAggregatorFactory

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

the class CalciteQueryTest method testBitwiseAggregatorsTimeseries.

@Test
public void testBitwiseAggregatorsTimeseries() throws Exception {
    cannotVectorize();
    testQuery("SELECT\n" + " BIT_AND(l1),\n" + " BIT_OR(l1),\n" + " BIT_XOR(l1)\n" + "FROM druid.numfoo", ImmutableList.of(Druids.newTimeseriesQueryBuilder().dataSource(CalciteTests.DATASOURCE3).intervals(querySegmentSpec(Filtration.eternity())).granularity(Granularities.ALL).aggregators(aggregators(new FilteredAggregatorFactory(new ExpressionLambdaAggregatorFactory("a0", ImmutableSet.of("l1"), "__acc", "0", "0", NullHandling.sqlCompatible(), false, false, "bitwiseAnd(\"__acc\", \"l1\")", "bitwiseAnd(\"__acc\", \"a0\")", null, null, ExpressionLambdaAggregatorFactory.DEFAULT_MAX_SIZE_BYTES, TestExprMacroTable.INSTANCE), not(selector("l1", null, null))), new FilteredAggregatorFactory(new ExpressionLambdaAggregatorFactory("a1", ImmutableSet.of("l1"), "__acc", "0", "0", NullHandling.sqlCompatible(), false, false, "bitwiseOr(\"__acc\", \"l1\")", "bitwiseOr(\"__acc\", \"a1\")", null, null, ExpressionLambdaAggregatorFactory.DEFAULT_MAX_SIZE_BYTES, TestExprMacroTable.INSTANCE), not(selector("l1", null, null))), new FilteredAggregatorFactory(new ExpressionLambdaAggregatorFactory("a2", ImmutableSet.of("l1"), "__acc", "0", "0", NullHandling.sqlCompatible(), false, false, "bitwiseXor(\"__acc\", \"l1\")", "bitwiseXor(\"__acc\", \"a2\")", null, null, ExpressionLambdaAggregatorFactory.DEFAULT_MAX_SIZE_BYTES, TestExprMacroTable.INSTANCE), not(selector("l1", null, null))))).context(QUERY_CONTEXT_DEFAULT).build()), ImmutableList.of(useDefault ? new Object[] { 0L, 325327L, 325324L } : new Object[] { 0L, 325327L, 325324L }));
}

Example 10 with FilteredAggregatorFactory

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

the class CalciteJoinQueryTest method testNotInAggregationSubquery.

@Test
@Parameters(source = QueryContextForJoinProvider.class)
public void testNotInAggregationSubquery(Map<String, Object> queryContext) throws Exception {
    // Cannot vectorize JOIN operator.
    cannotVectorize();
    testQuery("SELECT DISTINCT __time FROM druid.foo WHERE __time NOT IN (SELECT MAX(__time) FROM druid.foo)", queryContext, ImmutableList.of(GroupByQuery.builder().setDataSource(join(join(new TableDataSource(CalciteTests.DATASOURCE1), new QueryDataSource(GroupByQuery.builder().setDataSource(Druids.newTimeseriesQueryBuilder().dataSource(CalciteTests.DATASOURCE1).intervals(querySegmentSpec(Filtration.eternity())).granularity(Granularities.ALL).aggregators(new LongMaxAggregatorFactory("a0", "__time")).context(QUERY_CONTEXT_DEFAULT).build()).setInterval(querySegmentSpec(Filtration.eternity())).setGranularity(Granularities.ALL).setAggregatorSpecs(new CountAggregatorFactory("_a0"), NullHandling.sqlCompatible() ? new FilteredAggregatorFactory(new CountAggregatorFactory("_a1"), not(selector("a0", null, null))) : new CountAggregatorFactory("_a1")).setContext(queryContext).build()), "j0.", "1", JoinType.INNER), new QueryDataSource(Druids.newTimeseriesQueryBuilder().dataSource(CalciteTests.DATASOURCE1).intervals(querySegmentSpec(Filtration.eternity())).granularity(Granularities.ALL).aggregators(new LongMaxAggregatorFactory("a0", "__time")).postAggregators(expressionPostAgg("p0", "1")).context(QUERY_CONTEXT_DEFAULT).build()), "_j0.", "(\"__time\" == \"_j0.a0\")", JoinType.LEFT)).setInterval(querySegmentSpec(Filtration.eternity())).setGranularity(Granularities.ALL).setDimFilter(or(selector("j0._a0", "0", null), and(selector("_j0.p0", null, null), expressionFilter("(\"j0._a1\" >= \"j0._a0\")")))).setDimensions(dimensions(new DefaultDimensionSpec("__time", "d0", ColumnType.LONG))).setContext(queryContext).build()), ImmutableList.of(new Object[] { timestamp("2000-01-01") }, new Object[] { timestamp("2000-01-02") }, new Object[] { timestamp("2000-01-03") }, new Object[] { timestamp("2001-01-01") }, new Object[] { timestamp("2001-01-02") }));
}