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

Example 86 with FieldAccessPostAggregator

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

the class PvaluefromZscorePostAggregatorTest method setup.

@Before
public void setup() {
    zscore = new FieldAccessPostAggregator("f1", "zscore");
    pvaluefromZscorePostAggregator = new PvaluefromZscorePostAggregator("pvalue", zscore);
}

Example 87 with FieldAccessPostAggregator

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

the class VarianceGroupByQueryTest method testGroupByZtestPostAgg.

@Test
public void testGroupByZtestPostAgg() {
    // test postaggs from 'teststats' package in here since we've already gone to the trouble of setting up the test
    GroupByQuery query = queryBuilder.setDataSource(QueryRunnerTestHelper.DATA_SOURCE).setQuerySegmentSpec(QueryRunnerTestHelper.FIRST_TO_THIRD).setDimensions(new DefaultDimensionSpec("quality", "alias")).setAggregatorSpecs(QueryRunnerTestHelper.ROWS_COUNT, VarianceTestHelper.INDEX_VARIANCE_AGGR, new LongSumAggregatorFactory("idx", "index")).setPostAggregatorSpecs(ImmutableList.of(VarianceTestHelper.STD_DEV_OF_INDEX_POST_AGGR, // these inputs are totally nonsensical, i just want the code path to be executed
    new ZtestPostAggregator("ztest", new FieldAccessPostAggregator("f1", "idx"), new ConstantPostAggregator("f2", 100000L), new FieldAccessPostAggregator("f3", "index_stddev"), new ConstantPostAggregator("f2", 100000L)))).setLimitSpec(new DefaultLimitSpec(OrderByColumnSpec.descending("ztest"), 1)).setGranularity(QueryRunnerTestHelper.DAY_GRAN).build();
    VarianceTestHelper.RowBuilder builder = new VarianceTestHelper.RowBuilder(new String[] { "alias", "rows", "idx", "index_stddev", "index_var", "ztest" });
    List<ResultRow> expectedResults = builder.add("2011-04-01", "premium", 3L, 2900.0, 726.632270328514, 527994.4562827706, 36.54266309285626).build(query);
    Iterable<ResultRow> results = GroupByQueryRunnerTestHelper.runQuery(factory, runner, query);
    TestHelper.assertExpectedObjects(expectedResults, results, "groupBy");
}

Example 88 with FieldAccessPostAggregator

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

the class TopNQueryQueryToolChestTest method doTestCacheStrategyOrderByPost.

private void doTestCacheStrategyOrderByPost(final ColumnType valueType, final Object dimValue) throws IOException {
    CacheStrategy<Result<TopNResultValue>, Object, TopNQuery> strategy = new TopNQueryQueryToolChest(null, null).getCacheStrategy(new TopNQuery(new TableDataSource("dummy"), VirtualColumns.EMPTY, new DefaultDimensionSpec("test", "test", valueType), new NumericTopNMetricSpec("post"), 3, new MultipleIntervalSegmentSpec(ImmutableList.of(Intervals.of("2015-01-01/2015-01-02"))), null, Granularities.ALL, ImmutableList.of(new HyperUniquesAggregatorFactory("metric1", "test", false, false), new CountAggregatorFactory("metric2")), ImmutableList.of(new ArithmeticPostAggregator("post", "+", ImmutableList.of(new FinalizingFieldAccessPostAggregator("metric1", "metric1"), new FieldAccessPostAggregator("metric2", "metric2")))), null));
    HyperLogLogCollector collector = getIntermediateHllCollector(valueType.getType(), dimValue);
    final Result<TopNResultValue> result1 = new Result<>(// test timestamps that result in integer size millis
    DateTimes.utc(123L), new TopNResultValue(Collections.singletonList(ImmutableMap.of("test", dimValue, "metric1", collector, "metric2", 2, "post", collector.estimateCardinality() + 2))));
    Object preparedValue = strategy.prepareForSegmentLevelCache().apply(result1);
    ObjectMapper objectMapper = TestHelper.makeJsonMapper();
    Object fromCacheValue = objectMapper.readValue(objectMapper.writeValueAsBytes(preparedValue), strategy.getCacheObjectClazz());
    Result<TopNResultValue> fromCacheResult = strategy.pullFromSegmentLevelCache().apply(fromCacheValue);
    Assert.assertEquals(result1, fromCacheResult);
    final Result<TopNResultValue> resultLevelCacheResult = new Result<>(// test timestamps that result in integer size millis
    DateTimes.utc(123L), new TopNResultValue(Collections.singletonList(ImmutableMap.of("test", dimValue, "metric1", collector.estimateCardinality(), "metric2", 2, "post", collector.estimateCardinality() + 2))));
    Object preparedResultCacheValue = strategy.prepareForCache(true).apply(resultLevelCacheResult);
    Object fromResultCacheValue = objectMapper.readValue(objectMapper.writeValueAsBytes(preparedResultCacheValue), strategy.getCacheObjectClazz());
    Result<TopNResultValue> fromResultCacheResult = strategy.pullFromCache(true).apply(fromResultCacheValue);
    Assert.assertEquals(resultLevelCacheResult, fromResultCacheResult);
}

Example 89 with FieldAccessPostAggregator

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

the class TopNQueryQueryToolChestTest method testComputeResultLevelCacheKeyWithDifferentPostAgg.

@Test
public void testComputeResultLevelCacheKeyWithDifferentPostAgg() {
    final TopNQuery query1 = new TopNQuery(new TableDataSource("dummy"), VirtualColumns.EMPTY, new DefaultDimensionSpec("test", "test"), new LegacyTopNMetricSpec("metric1"), 3, new MultipleIntervalSegmentSpec(ImmutableList.of(Intervals.of("2015-01-01T18:00:00/2015-01-02T18:00:00"))), null, Granularities.ALL, ImmutableList.of(new LongSumAggregatorFactory("metric1", "metric1"), new LongSumAggregatorFactory("metric2", "metric2")), ImmutableList.of(new ArithmeticPostAggregator("post1", "/", ImmutableList.of(new FieldAccessPostAggregator("metric1", "metric1"), new FieldAccessPostAggregator("metric2", "metric2")))), null);
    final TopNQuery query2 = new TopNQuery(new TableDataSource("dummy"), VirtualColumns.EMPTY, new DefaultDimensionSpec("test", "test"), new LegacyTopNMetricSpec("metric1"), 3, new MultipleIntervalSegmentSpec(ImmutableList.of(Intervals.of("2015-01-01T18:00:00/2015-01-02T18:00:00"))), null, Granularities.ALL, ImmutableList.of(new LongSumAggregatorFactory("metric1", "metric1"), new LongSumAggregatorFactory("metric2", "metric2")), ImmutableList.of(new ArithmeticPostAggregator("post2", "+", ImmutableList.of(new FieldAccessPostAggregator("metric1", "metric1"), new FieldAccessPostAggregator("metric2", "metric2")))), null);
    final CacheStrategy<Result<TopNResultValue>, Object, TopNQuery> strategy1 = new TopNQueryQueryToolChest(null, null).getCacheStrategy(query1);
    final CacheStrategy<Result<TopNResultValue>, Object, TopNQuery> strategy2 = new TopNQueryQueryToolChest(null, null).getCacheStrategy(query2);
    // segment level cache key excludes postaggregates in topn
    Assert.assertTrue(Arrays.equals(strategy1.computeCacheKey(query1), strategy2.computeCacheKey(query2)));
    Assert.assertFalse(Arrays.equals(strategy1.computeCacheKey(query1), strategy1.computeResultLevelCacheKey(query1)));
    Assert.assertFalse(Arrays.equals(strategy1.computeResultLevelCacheKey(query1), strategy2.computeResultLevelCacheKey(query2)));
}

Example 90 with FieldAccessPostAggregator

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

the class TimeseriesQueryRunnerTest method testTimeseriesWithPostAggregatorReferencingTimestampResultField.

@Test
public void testTimeseriesWithPostAggregatorReferencingTimestampResultField() {
    TimeseriesQuery query = Druids.newTimeseriesQueryBuilder().dataSource(QueryRunnerTestHelper.DATA_SOURCE).granularity(QueryRunnerTestHelper.DAY_GRAN).filters(QueryRunnerTestHelper.MARKET_DIMENSION, "spot").intervals(QueryRunnerTestHelper.FIRST_TO_THIRD).postAggregators(new FieldAccessPostAggregator("timestampInPostAgg", "myTimestamp")).descending(descending).context(makeContext(ImmutableMap.of(TimeseriesQuery.CTX_TIMESTAMP_RESULT_FIELD, "myTimestamp"))).build();
    final DateTime aprilFirst = DateTimes.of("2011-04-01");
    final DateTime aprilSecond = DateTimes.of("2011-04-02");
    List<Result<TimeseriesResultValue>> expectedResults = Arrays.asList(new Result<>(aprilFirst, new TimeseriesResultValue(ImmutableMap.of("myTimestamp", aprilFirst.getMillis(), "timestampInPostAgg", aprilFirst.getMillis()))), new Result<>(aprilSecond, new TimeseriesResultValue(ImmutableMap.of("myTimestamp", aprilSecond.getMillis(), "timestampInPostAgg", aprilSecond.getMillis()))));
    Iterable<Result<TimeseriesResultValue>> results = runner.run(QueryPlus.wrap(query)).toList();
    assertExpectedResults(expectedResults, results);
}