Search in sources :

Example 1 with VarianceAggregatorCollector

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

the class VarianceSqlAggregatorTest method testStdDevWithVirtualColumns.

@Test
public void testStdDevWithVirtualColumns() throws Exception {
    VarianceAggregatorCollector holder1 = new VarianceAggregatorCollector();
    VarianceAggregatorCollector holder2 = new VarianceAggregatorCollector();
    VarianceAggregatorCollector holder3 = new VarianceAggregatorCollector();
    for (InputRow row : CalciteTests.ROWS1_WITH_NUMERIC_DIMS) {
        Object raw1 = row.getRaw("d1");
        Object raw2 = row.getRaw("f1");
        Object raw3 = row.getRaw("l1");
        addToHolder(holder1, raw1, 7);
        addToHolder(holder2, raw2, 7);
        addToHolder(holder3, raw3, 7);
    }
    final List<Object[]> expectedResults = ImmutableList.of(new Object[] { Math.sqrt(holder1.getVariance(false)), (float) Math.sqrt(holder2.getVariance(false)), (long) Math.sqrt(holder3.getVariance(false)) });
    testQuery("SELECT\n" + "STDDEV(d1*7),\n" + "STDDEV(f1*7),\n" + "STDDEV(l1*7)\n" + "FROM numfoo", ImmutableList.of(Druids.newTimeseriesQueryBuilder().dataSource(CalciteTests.DATASOURCE3).intervals(new MultipleIntervalSegmentSpec(ImmutableList.of(Filtration.eternity()))).granularity(Granularities.ALL).virtualColumns(BaseCalciteQueryTest.expressionVirtualColumn("v0", "(\"d1\" * 7)", ColumnType.DOUBLE), BaseCalciteQueryTest.expressionVirtualColumn("v1", "(\"f1\" * 7)", ColumnType.FLOAT), BaseCalciteQueryTest.expressionVirtualColumn("v2", "(\"l1\" * 7)", ColumnType.LONG)).aggregators(ImmutableList.of(new VarianceAggregatorFactory("a0:agg", "v0", "sample", "double"), new VarianceAggregatorFactory("a1:agg", "v1", "sample", "float"), new VarianceAggregatorFactory("a2:agg", "v2", "sample", "long"))).postAggregators(new StandardDeviationPostAggregator("a0", "a0:agg", "sample"), new StandardDeviationPostAggregator("a1", "a1:agg", "sample"), new StandardDeviationPostAggregator("a2", "a2:agg", "sample")).context(BaseCalciteQueryTest.QUERY_CONTEXT_DEFAULT).build()), expectedResults);
}
Also used : VarianceAggregatorCollector(org.apache.druid.query.aggregation.variance.VarianceAggregatorCollector) StandardDeviationPostAggregator(org.apache.druid.query.aggregation.variance.StandardDeviationPostAggregator) InputRow(org.apache.druid.data.input.InputRow) MultipleIntervalSegmentSpec(org.apache.druid.query.spec.MultipleIntervalSegmentSpec) VarianceAggregatorFactory(org.apache.druid.query.aggregation.variance.VarianceAggregatorFactory) BaseCalciteQueryTest(org.apache.druid.sql.calcite.BaseCalciteQueryTest) Test(org.junit.Test)

Example 2 with VarianceAggregatorCollector

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

the class VarianceSqlAggregatorTest method testVarSamp.

@Test
public void testVarSamp() throws Exception {
    VarianceAggregatorCollector holder1 = new VarianceAggregatorCollector();
    VarianceAggregatorCollector holder2 = new VarianceAggregatorCollector();
    VarianceAggregatorCollector holder3 = new VarianceAggregatorCollector();
    for (InputRow row : CalciteTests.ROWS1_WITH_NUMERIC_DIMS) {
        Object raw1 = row.getRaw("d1");
        Object raw2 = row.getRaw("f1");
        Object raw3 = row.getRaw("l1");
        addToHolder(holder1, raw1);
        addToHolder(holder2, raw2);
        addToHolder(holder3, raw3);
    }
    final List<Object[]> expectedResults = ImmutableList.of(new Object[] { holder1.getVariance(false), holder2.getVariance(false).floatValue(), holder3.getVariance(false).longValue() });
    testQuery("SELECT\n" + "VAR_SAMP(d1),\n" + "VAR_SAMP(f1),\n" + "VAR_SAMP(l1)\n" + "FROM numfoo", ImmutableList.of(Druids.newTimeseriesQueryBuilder().dataSource(CalciteTests.DATASOURCE3).intervals(new MultipleIntervalSegmentSpec(ImmutableList.of(Filtration.eternity()))).granularity(Granularities.ALL).aggregators(ImmutableList.of(new VarianceAggregatorFactory("a0:agg", "d1", "sample", "double"), new VarianceAggregatorFactory("a1:agg", "f1", "sample", "float"), new VarianceAggregatorFactory("a2:agg", "l1", "sample", "long"))).context(BaseCalciteQueryTest.QUERY_CONTEXT_DEFAULT).build()), expectedResults);
}
Also used : VarianceAggregatorCollector(org.apache.druid.query.aggregation.variance.VarianceAggregatorCollector) InputRow(org.apache.druid.data.input.InputRow) MultipleIntervalSegmentSpec(org.apache.druid.query.spec.MultipleIntervalSegmentSpec) VarianceAggregatorFactory(org.apache.druid.query.aggregation.variance.VarianceAggregatorFactory) BaseCalciteQueryTest(org.apache.druid.sql.calcite.BaseCalciteQueryTest) Test(org.junit.Test)

Example 3 with VarianceAggregatorCollector

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

the class VarianceSqlAggregatorTest method testVarPop.

@Test
public void testVarPop() throws Exception {
    VarianceAggregatorCollector holder1 = new VarianceAggregatorCollector();
    VarianceAggregatorCollector holder2 = new VarianceAggregatorCollector();
    VarianceAggregatorCollector holder3 = new VarianceAggregatorCollector();
    for (InputRow row : CalciteTests.ROWS1_WITH_NUMERIC_DIMS) {
        Object raw1 = row.getRaw("d1");
        Object raw2 = row.getRaw("f1");
        Object raw3 = row.getRaw("l1");
        addToHolder(holder1, raw1);
        addToHolder(holder2, raw2);
        addToHolder(holder3, raw3);
    }
    final List<Object[]> expectedResults = ImmutableList.of(new Object[] { holder1.getVariance(true), holder2.getVariance(true).floatValue(), holder3.getVariance(true).longValue() });
    testQuery("SELECT\n" + "VAR_POP(d1),\n" + "VAR_POP(f1),\n" + "VAR_POP(l1)\n" + "FROM numfoo", ImmutableList.of(Druids.newTimeseriesQueryBuilder().dataSource(CalciteTests.DATASOURCE3).intervals(new MultipleIntervalSegmentSpec(ImmutableList.of(Filtration.eternity()))).granularity(Granularities.ALL).aggregators(ImmutableList.of(new VarianceAggregatorFactory("a0:agg", "d1", "population", "double"), new VarianceAggregatorFactory("a1:agg", "f1", "population", "float"), new VarianceAggregatorFactory("a2:agg", "l1", "population", "long"))).context(BaseCalciteQueryTest.QUERY_CONTEXT_DEFAULT).build()), expectedResults);
}
Also used : VarianceAggregatorCollector(org.apache.druid.query.aggregation.variance.VarianceAggregatorCollector) InputRow(org.apache.druid.data.input.InputRow) MultipleIntervalSegmentSpec(org.apache.druid.query.spec.MultipleIntervalSegmentSpec) VarianceAggregatorFactory(org.apache.druid.query.aggregation.variance.VarianceAggregatorFactory) BaseCalciteQueryTest(org.apache.druid.sql.calcite.BaseCalciteQueryTest) Test(org.junit.Test)

Example 4 with VarianceAggregatorCollector

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

the class VarianceBenchmark method collectVarianceInBatch.

@Benchmark
@BenchmarkMode(Mode.AverageTime)
@OutputTimeUnit(TimeUnit.NANOSECONDS)
public void collectVarianceInBatch(Blackhole blackhole) {
    double sum = 0, nvariance = 0;
    for (float v : randomValues) {
        sum += v;
    }
    double mean = sum / randomValues.length;
    for (float v : randomValues) {
        nvariance += (v - mean) * (v - mean);
    }
    VarianceAggregatorCollector collector = new VarianceAggregatorCollector(randomValues.length, sum, nvariance);
    blackhole.consume(collector);
}
Also used : VarianceAggregatorCollector(org.apache.druid.query.aggregation.variance.VarianceAggregatorCollector) BenchmarkMode(org.openjdk.jmh.annotations.BenchmarkMode) Benchmark(org.openjdk.jmh.annotations.Benchmark) OutputTimeUnit(org.openjdk.jmh.annotations.OutputTimeUnit)

Example 5 with VarianceAggregatorCollector

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

the class VarianceBenchmark method collectVarianceOneByOne.

@Benchmark
@BenchmarkMode(Mode.AverageTime)
@OutputTimeUnit(TimeUnit.NANOSECONDS)
public void collectVarianceOneByOne(Blackhole blackhole) {
    VarianceAggregatorCollector collector = new VarianceAggregatorCollector();
    for (float v : randomValues) {
        collector.add(v);
    }
    blackhole.consume(collector);
}
Also used : VarianceAggregatorCollector(org.apache.druid.query.aggregation.variance.VarianceAggregatorCollector) BenchmarkMode(org.openjdk.jmh.annotations.BenchmarkMode) Benchmark(org.openjdk.jmh.annotations.Benchmark) OutputTimeUnit(org.openjdk.jmh.annotations.OutputTimeUnit)

Aggregations

VarianceAggregatorCollector (org.apache.druid.query.aggregation.variance.VarianceAggregatorCollector)7 InputRow (org.apache.druid.data.input.InputRow)5 VarianceAggregatorFactory (org.apache.druid.query.aggregation.variance.VarianceAggregatorFactory)5 MultipleIntervalSegmentSpec (org.apache.druid.query.spec.MultipleIntervalSegmentSpec)5 BaseCalciteQueryTest (org.apache.druid.sql.calcite.BaseCalciteQueryTest)5 Test (org.junit.Test)5 StandardDeviationPostAggregator (org.apache.druid.query.aggregation.variance.StandardDeviationPostAggregator)3 Benchmark (org.openjdk.jmh.annotations.Benchmark)2 BenchmarkMode (org.openjdk.jmh.annotations.BenchmarkMode)2 OutputTimeUnit (org.openjdk.jmh.annotations.OutputTimeUnit)2