Examples with ExtractionDimensionSpec org.apache.druid.query.dimension.ExtractionDimensionSpec used on opensource projects

Example 41 with ExtractionDimensionSpec

use of org.apache.druid.query.dimension.ExtractionDimensionSpec in project druid by druid-io.

the class ExpressionSelectorBenchmark method timeFloorUsingExtractionFn.

@Benchmark
public void timeFloorUsingExtractionFn(Blackhole blackhole) {
    final Sequence<Cursor> cursors = new QueryableIndexStorageAdapter(index).makeCursors(null, index.getDataInterval(), VirtualColumns.EMPTY, Granularities.ALL, false, null);
    final List<?> results = cursors.map(cursor -> {
        final DimensionSelector selector = cursor.getColumnSelectorFactory().makeDimensionSelector(new ExtractionDimensionSpec(ColumnHolder.TIME_COLUMN_NAME, "v", new TimeFormatExtractionFn(null, null, null, Granularities.HOUR, true)));
        consumeDimension(cursor, selector, blackhole);
        return null;
    }).toList();
    blackhole.consume(results);
}

Example 42 with ExtractionDimensionSpec

use of org.apache.druid.query.dimension.ExtractionDimensionSpec in project druid by druid-io.

the class ExpressionSelectorBenchmark method strlenUsingExtractionFn.

@Benchmark
public void strlenUsingExtractionFn(Blackhole blackhole) {
    final Sequence<Cursor> cursors = new QueryableIndexStorageAdapter(index).makeCursors(null, index.getDataInterval(), VirtualColumns.EMPTY, Granularities.ALL, false, null);
    final List<?> results = cursors.map(cursor -> {
        final DimensionSelector selector = cursor.getColumnSelectorFactory().makeDimensionSelector(new ExtractionDimensionSpec("x", "v", StrlenExtractionFn.instance()));
        consumeDimension(cursor, selector, blackhole);
        return null;
    }).toList();
    blackhole.consume(results);
}

Example 43 with ExtractionDimensionSpec

use of org.apache.druid.query.dimension.ExtractionDimensionSpec in project druid by druid-io.

the class ExpressionSelectorBenchmark method timeFormatUsingExtractionFn.

@Benchmark
public void timeFormatUsingExtractionFn(Blackhole blackhole) {
    final Sequence<Cursor> cursors = new QueryableIndexStorageAdapter(index).makeCursors(null, index.getDataInterval(), VirtualColumns.EMPTY, Granularities.ALL, false, null);
    final List<?> results = cursors.map(cursor -> {
        final DimensionSelector selector = cursor.getColumnSelectorFactory().makeDimensionSelector(new ExtractionDimensionSpec(ColumnHolder.TIME_COLUMN_NAME, "v", new TimeFormatExtractionFn("yyyy-MM-dd", null, null, null, false)));
        consumeDimension(cursor, selector, blackhole);
        return null;
    }).toList();
    blackhole.consume(results);
}

Example 44 with ExtractionDimensionSpec

use of org.apache.druid.query.dimension.ExtractionDimensionSpec in project druid by druid-io.

the class BloomFilterSqlAggregatorTest method testBloomFilterAggExtractionFn.

@Test
public void testBloomFilterAggExtractionFn() throws Exception {
    cannotVectorize();
    BloomKFilter expected1 = new BloomKFilter(TEST_NUM_ENTRIES);
    for (InputRow row : CalciteTests.ROWS1_WITH_NUMERIC_DIMS) {
        String raw = NullHandling.emptyToNullIfNeeded((String) row.getRaw("dim1"));
        // empty string extractionFn produces null
        if (raw == null || "".equals(raw)) {
            expected1.addBytes(null, 0, 0);
        } else {
            expected1.addString(raw.substring(0, 1));
        }
    }
    testQuery("SELECT\n" + "BLOOM_FILTER(SUBSTRING(dim1, 1, 1), 1000)\n" + "FROM numfoo", ImmutableList.of(Druids.newTimeseriesQueryBuilder().dataSource(CalciteTests.DATASOURCE3).intervals(new MultipleIntervalSegmentSpec(ImmutableList.of(Filtration.eternity()))).granularity(Granularities.ALL).aggregators(ImmutableList.of(new BloomFilterAggregatorFactory("a0:agg", new ExtractionDimensionSpec("dim1", "a0:dim1", new SubstringDimExtractionFn(0, 1)), TEST_NUM_ENTRIES))).context(BaseCalciteQueryTest.QUERY_CONTEXT_DEFAULT).build()), ImmutableList.of(new Object[] { CalciteTests.getJsonMapper().writeValueAsString(expected1) }));
}

Example 45 with ExtractionDimensionSpec

use of org.apache.druid.query.dimension.ExtractionDimensionSpec in project druid by druid-io.

the class BloomFilterSqlAggregator method toDruidAggregation.

@Nullable
@Override
public Aggregation toDruidAggregation(PlannerContext plannerContext, RowSignature rowSignature, VirtualColumnRegistry virtualColumnRegistry, RexBuilder rexBuilder, String name, AggregateCall aggregateCall, Project project, List<Aggregation> existingAggregations, boolean finalizeAggregations) {
    final RexNode inputOperand = Expressions.fromFieldAccess(rowSignature, project, aggregateCall.getArgList().get(0));
    final DruidExpression input = Expressions.toDruidExpression(plannerContext, rowSignature, inputOperand);
    if (input == null) {
        return null;
    }
    final AggregatorFactory aggregatorFactory;
    final String aggName = StringUtils.format("%s:agg", name);
    final RexNode maxNumEntriesOperand = Expressions.fromFieldAccess(rowSignature, project, aggregateCall.getArgList().get(1));
    if (!maxNumEntriesOperand.isA(SqlKind.LITERAL)) {
        // maxNumEntriesOperand must be a literal in order to plan.
        return null;
    }
    final int maxNumEntries = ((Number) RexLiteral.value(maxNumEntriesOperand)).intValue();
    // Look for existing matching aggregatorFactory.
    for (final Aggregation existing : existingAggregations) {
        for (AggregatorFactory factory : existing.getAggregatorFactories()) {
            if (factory instanceof BloomFilterAggregatorFactory) {
                final BloomFilterAggregatorFactory theFactory = (BloomFilterAggregatorFactory) factory;
                // Check input for equivalence.
                final boolean inputMatches;
                final DruidExpression virtualInput = virtualColumnRegistry.findVirtualColumnExpressions(theFactory.requiredFields()).stream().findFirst().orElse(null);
                if (virtualInput == null) {
                    if (input.isDirectColumnAccess()) {
                        inputMatches = input.getDirectColumn().equals(theFactory.getField().getDimension());
                    } else {
                        inputMatches = input.getSimpleExtraction().getColumn().equals(theFactory.getField().getDimension()) && input.getSimpleExtraction().getExtractionFn().equals(theFactory.getField().getExtractionFn());
                    }
                } else {
                    inputMatches = virtualInput.equals(input);
                }
                final boolean matches = inputMatches && theFactory.getMaxNumEntries() == maxNumEntries;
                if (matches) {
                    // Found existing one. Use this.
                    return Aggregation.create(theFactory);
                }
            }
        }
    }
    // No existing match found. Create a new one.
    ColumnType valueType = Calcites.getColumnTypeForRelDataType(inputOperand.getType());
    final DimensionSpec spec;
    if (input.isDirectColumnAccess()) {
        spec = new DefaultDimensionSpec(input.getSimpleExtraction().getColumn(), StringUtils.format("%s:%s", name, input.getSimpleExtraction().getColumn()), valueType);
    } else if (input.isSimpleExtraction()) {
        spec = new ExtractionDimensionSpec(input.getSimpleExtraction().getColumn(), StringUtils.format("%s:%s", name, input.getSimpleExtraction().getColumn()), valueType, input.getSimpleExtraction().getExtractionFn());
    } else {
        String virtualColumnName = virtualColumnRegistry.getOrCreateVirtualColumnForExpression(input, inputOperand.getType());
        spec = new DefaultDimensionSpec(virtualColumnName, StringUtils.format("%s:%s", name, virtualColumnName));
    }
    aggregatorFactory = new BloomFilterAggregatorFactory(aggName, spec, maxNumEntries);
    return Aggregation.create(aggregatorFactory);
}