Examples with ExtractionFn org.apache.druid.query.extraction.ExtractionFn used on opensource projects

Example 11 with ExtractionFn

use of org.apache.druid.query.extraction.ExtractionFn in project druid by druid-io.

the class DimensionHandlerUtils method getEffectiveCapabilities.

/**
 * When determining the capabilities of a column during query processing, this function
 * adjusts the capabilities for columns that cannot be handled as-is to manageable defaults
 * (e.g., treating missing columns as empty String columns)
 */
private static ColumnCapabilities getEffectiveCapabilities(DimensionSpec dimSpec, @Nullable ColumnCapabilities capabilities) {
    if (capabilities == null) {
        capabilities = DEFAULT_STRING_CAPABILITIES;
    }
    // Complex dimension type is not supported
    if (capabilities.is(ValueType.COMPLEX)) {
        capabilities = DEFAULT_STRING_CAPABILITIES;
    }
    // DimensionSelector if an extractionFn is present.
    if (dimSpec.getExtractionFn() != null) {
        ExtractionFn fn = dimSpec.getExtractionFn();
        capabilities = ColumnCapabilitiesImpl.copyOf(capabilities).setType(ColumnType.STRING).setDictionaryValuesUnique(capabilities.isDictionaryEncoded().isTrue() && fn.getExtractionType() == ExtractionFn.ExtractionType.ONE_TO_ONE).setHasMultipleValues(capabilities.hasMultipleValues().isMaybeTrue() || capabilities.isArray()).setDictionaryValuesSorted(capabilities.isDictionaryEncoded().isTrue() && fn.preservesOrdering());
    }
    // we need to wrap selectors on numeric columns with a string casting DimensionSelector.
    if (capabilities.isNumeric()) {
        if (dimSpec.mustDecorate()) {
            capabilities = DEFAULT_STRING_CAPABILITIES;
        }
    }
    return capabilities;
}

Example 12 with ExtractionFn

use of org.apache.druid.query.extraction.ExtractionFn in project druid by druid-io.

the class GroupByQueryRunnerTest method testGroupByTimeExtractionWithNulls.

@Test
public void testGroupByTimeExtractionWithNulls() {
    // Cannot vectorize due to extraction dimension specs.
    cannotVectorize();
    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 = makeQueryBuilder().setDataSource(QueryRunnerTestHelper.DATA_SOURCE).setQuerySegmentSpec(QueryRunnerTestHelper.FULL_ON_INTERVAL_SPEC).setDimensions(new DefaultDimensionSpec("market", "market"), new ExtractionDimensionSpec(ColumnHolder.TIME_COLUMN_NAME, "dayOfWeek", new CascadeExtractionFn(new ExtractionFn[] { new TimeFormatExtractionFn("EEEE", null, null, null, false), nullWednesdays }))).setAggregatorSpecs(QueryRunnerTestHelper.ROWS_COUNT, QueryRunnerTestHelper.INDEX_DOUBLE_SUM).setPostAggregatorSpecs(Collections.singletonList(QueryRunnerTestHelper.ADD_ROWS_INDEX_CONSTANT)).setGranularity(QueryRunnerTestHelper.ALL_GRAN).setDimFilter(new OrDimFilter(Arrays.asList(new SelectorDimFilter("market", "spot", null), new SelectorDimFilter("market", "upfront", null)))).build();
    List<ResultRow> expectedResults = Arrays.asList(makeRow(query, "1970-01-01", "dayOfWeek", null, "market", "spot", "index", 14271.368591308594, "rows", 126L, "addRowsIndexConstant", 14398.368591308594), makeRow(query, "1970-01-01", "dayOfWeek", "Friday", "market", "spot", "index", 13219.574157714844, "rows", 117L, "addRowsIndexConstant", 13337.574157714844), makeRow(query, "1970-01-01", "dayOfWeek", "Monday", "market", "spot", "index", 13557.738830566406, "rows", 117L, "addRowsIndexConstant", 13675.738830566406), makeRow(query, "1970-01-01", "dayOfWeek", "Saturday", "market", "spot", "index", 13493.751281738281, "rows", 117L, "addRowsIndexConstant", 13611.751281738281), makeRow(query, "1970-01-01", "dayOfWeek", "Sunday", "market", "spot", "index", 13585.541015625, "rows", 117L, "addRowsIndexConstant", 13703.541015625), makeRow(query, "1970-01-01", "dayOfWeek", "Thursday", "market", "spot", "index", 14279.127197265625, "rows", 126L, "addRowsIndexConstant", 14406.127197265625), makeRow(query, "1970-01-01", "dayOfWeek", "Tuesday", "market", "spot", "index", 13199.471435546875, "rows", 117L, "addRowsIndexConstant", 13317.471435546875), makeRow(query, "1970-01-01", "dayOfWeek", null, "market", "upfront", "index", 28985.5751953125, "rows", 28L, "addRowsIndexConstant", 29014.5751953125), makeRow(query, "1970-01-01", "dayOfWeek", "Friday", "market", "upfront", "index", 27297.8623046875, "rows", 26L, "addRowsIndexConstant", 27324.8623046875), makeRow(query, "1970-01-01", "dayOfWeek", "Monday", "market", "upfront", "index", 27619.58447265625, "rows", 26L, "addRowsIndexConstant", 27646.58447265625), makeRow(query, "1970-01-01", "dayOfWeek", "Saturday", "market", "upfront", "index", 27820.83154296875, "rows", 26L, "addRowsIndexConstant", 27847.83154296875), makeRow(query, "1970-01-01", "dayOfWeek", "Sunday", "market", "upfront", "index", 24791.223876953125, "rows", 26L, "addRowsIndexConstant", 24818.223876953125), makeRow(query, "1970-01-01", "dayOfWeek", "Thursday", "market", "upfront", "index", 28562.748901367188, "rows", 28L, "addRowsIndexConstant", 28591.748901367188), makeRow(query, "1970-01-01", "dayOfWeek", "Tuesday", "market", "upfront", "index", 26968.280639648438, "rows", 26L, "addRowsIndexConstant", 26995.280639648438));
    Iterable<ResultRow> results = GroupByQueryRunnerTestHelper.runQuery(factory, runner, query);
    TestHelper.assertExpectedObjects(expectedResults, results, "time-extraction");
}

Example 13 with ExtractionFn

use of org.apache.druid.query.extraction.ExtractionFn in project druid by druid-io.

the class GroupByQueryRunnerTest method testBySegmentResultsWithAllFiltersWithExtractionFns.

@Test
public void testBySegmentResultsWithAllFiltersWithExtractionFns() {
    String extractionJsFn = "function(str) { return 'super-' + str; }";
    String jsFn = "function(x) { return(x === 'super-mezzanine') }";
    ExtractionFn extractionFn = new JavaScriptExtractionFn(extractionJsFn, false, JavaScriptConfig.getEnabledInstance());
    List<DimFilter> superFilterList = new ArrayList<>();
    superFilterList.add(new SelectorDimFilter("quality", "super-mezzanine", extractionFn));
    superFilterList.add(new InDimFilter("quality", Arrays.asList("not-super-mezzanine", "FOOBAR", "super-mezzanine"), extractionFn));
    superFilterList.add(new BoundDimFilter("quality", "super-mezzanine", "super-mezzanine", false, false, true, extractionFn, StringComparators.ALPHANUMERIC));
    superFilterList.add(new RegexDimFilter("quality", "super-mezzanine", extractionFn));
    superFilterList.add(new SearchQueryDimFilter("quality", new ContainsSearchQuerySpec("super-mezzanine", true), extractionFn));
    superFilterList.add(new JavaScriptDimFilter("quality", jsFn, extractionFn, JavaScriptConfig.getEnabledInstance()));
    DimFilter superFilter = new AndDimFilter(superFilterList);
    GroupByQuery.Builder builder = makeQueryBuilder().setDataSource(QueryRunnerTestHelper.DATA_SOURCE).setInterval("2011-04-02/2011-04-04").setDimensions(new DefaultDimensionSpec("quality", "alias")).setAggregatorSpecs(QueryRunnerTestHelper.ROWS_COUNT, new LongSumAggregatorFactory("idx", "index")).setGranularity(new PeriodGranularity(new Period("P1M"), null, null)).setDimFilter(superFilter).overrideContext(ImmutableMap.of(QueryContexts.BY_SEGMENT_KEY, true));
    final GroupByQuery fullQuery = builder.build();
    int segmentCount = 32;
    Result<BySegmentResultValue> singleSegmentResult = new Result<>(DateTimes.of("2011-01-12T00:00:00.000Z"), new BySegmentResultValueClass<>(Collections.singletonList(makeRow(fullQuery, "2011-04-01", "alias", "mezzanine", "rows", 6L, "idx", 4420L)), QueryRunnerTestHelper.SEGMENT_ID.toString(), Intervals.of("2011-04-02T00:00:00.000Z/2011-04-04T00:00:00.000Z")));
    List<Result> bySegmentResults = new ArrayList<>();
    for (int i = 0; i < segmentCount; i++) {
        bySegmentResults.add(singleSegmentResult);
    }
    QueryToolChest toolChest = factory.getToolchest();
    List<QueryRunner<ResultRow>> singleSegmentRunners = new ArrayList<>();
    for (int i = 0; i < segmentCount; i++) {
        singleSegmentRunners.add(toolChest.preMergeQueryDecoration(runner));
    }
    ExecutorService exec = Executors.newCachedThreadPool();
    QueryRunner theRunner = toolChest.postMergeQueryDecoration(new FinalizeResultsQueryRunner<>(toolChest.mergeResults(factory.mergeRunners(Executors.newCachedThreadPool(), singleSegmentRunners)), toolChest));
    TestHelper.assertExpectedObjects(bySegmentResults, theRunner.run(QueryPlus.wrap(fullQuery)), "bySegment-filter");
    exec.shutdownNow();
}

Example 14 with ExtractionFn

use of org.apache.druid.query.extraction.ExtractionFn in project druid by druid-io.

the class GroupByQueryRunnerTest method testGroupByCardinalityAggWithExtractionFn.

@Test
public void testGroupByCardinalityAggWithExtractionFn() {
    // Cannot vectorize due to extraction dimension spec.
    cannotVectorize();
    String helloJsFn = "function(str) { return 'hello' }";
    ExtractionFn helloFn = new JavaScriptExtractionFn(helloJsFn, false, JavaScriptConfig.getEnabledInstance());
    GroupByQuery query = makeQueryBuilder().setDataSource(QueryRunnerTestHelper.DATA_SOURCE).setQuerySegmentSpec(QueryRunnerTestHelper.FIRST_TO_THIRD).setDimensions(new DefaultDimensionSpec("market", "alias")).setAggregatorSpecs(QueryRunnerTestHelper.ROWS_COUNT, new CardinalityAggregatorFactory("numVals", ImmutableList.of(new ExtractionDimensionSpec(QueryRunnerTestHelper.QUALITY_DIMENSION, QueryRunnerTestHelper.QUALITY_DIMENSION, helloFn)), false)).setGranularity(QueryRunnerTestHelper.DAY_GRAN).build();
    List<ResultRow> expectedResults = Arrays.asList(makeRow(query, "2011-04-01", "alias", "spot", "rows", 9L, "numVals", 1.0002442201269182d), makeRow(query, "2011-04-01", "alias", "total_market", "rows", 2L, "numVals", 1.0002442201269182d), makeRow(query, "2011-04-01", "alias", "upfront", "rows", 2L, "numVals", 1.0002442201269182d), makeRow(query, "2011-04-02", "alias", "spot", "rows", 9L, "numVals", 1.0002442201269182d), makeRow(query, "2011-04-02", "alias", "total_market", "rows", 2L, "numVals", 1.0002442201269182d), makeRow(query, "2011-04-02", "alias", "upfront", "rows", 2L, "numVals", 1.0002442201269182d));
    Iterable<ResultRow> results = GroupByQueryRunnerTestHelper.runQuery(factory, runner, query);
    TestHelper.assertExpectedObjects(expectedResults, results, "cardinality-agg");
}

Example 15 with ExtractionFn

use of org.apache.druid.query.extraction.ExtractionFn in project druid by druid-io.

the class GroupByQueryRunnerTest method testDimFilterHavingSpecWithExtractionFns.

@Test
public void testDimFilterHavingSpecWithExtractionFns() {
    String extractionJsFn = "function(str) { return 'super-' + str; }";
    ExtractionFn extractionFn = new JavaScriptExtractionFn(extractionJsFn, false, JavaScriptConfig.getEnabledInstance());
    String extractionJsFn2 = "function(num) { return num + 10; }";
    ExtractionFn extractionFn2 = new JavaScriptExtractionFn(extractionJsFn2, false, JavaScriptConfig.getEnabledInstance());
    final DimFilterHavingSpec havingSpec = new DimFilterHavingSpec(new OrDimFilter(ImmutableList.of(new BoundDimFilter("rows", "12", null, true, false, null, extractionFn2, StringComparators.NUMERIC), new SelectorDimFilter("idx", "super-217", extractionFn))), null);
    GroupByQuery.Builder builder = makeQueryBuilder().setDataSource(QueryRunnerTestHelper.DATA_SOURCE).setInterval("2011-04-02/2011-04-04").setDimensions(new DefaultDimensionSpec("quality", "alias")).setAggregatorSpecs(QueryRunnerTestHelper.ROWS_COUNT, new LongSumAggregatorFactory("idx", "index")).setGranularity(new PeriodGranularity(new Period("P1M"), null, null)).setHavingSpec(havingSpec);
    final GroupByQuery fullQuery = builder.build();
    List<ResultRow> expectedResults = Arrays.asList(makeRow(fullQuery, "2011-04-01", "alias", "business", "rows", 2L, "idx", 217L), makeRow(fullQuery, "2011-04-01", "alias", "mezzanine", "rows", 6L, "idx", 4420L), makeRow(fullQuery, "2011-04-01", "alias", "premium", "rows", 6L, "idx", 4416L));
    TestHelper.assertExpectedObjects(expectedResults, GroupByQueryRunnerTestHelper.runQuery(factory, runner, fullQuery), "extractionfn-havingspec");
}