Search in sources :

Example 1 with FunctionSignature

use of com.google.zetasql.FunctionSignature in project beam by apache.

the class AggregateScanConverter method convertAggCall.

private AggregateCall convertAggCall(ResolvedComputedColumn computedColumn, int columnRefOff, int groupCount, RelNode input) {
    ResolvedAggregateFunctionCall aggregateFunctionCall = (ResolvedAggregateFunctionCall) computedColumn.getExpr();
    // Reject AVG(INT64)
    if (aggregateFunctionCall.getFunction().getName().equals("avg")) {
        FunctionSignature signature = aggregateFunctionCall.getSignature();
        if (signature.getFunctionArgumentList().get(0).getType().getKind().equals(TypeKind.TYPE_INT64)) {
            throw new UnsupportedOperationException(AVG_ILLEGAL_LONG_INPUT_TYPE);
        }
    }
    // Reject aggregation DISTINCT
    if (aggregateFunctionCall.getDistinct()) {
        throw new UnsupportedOperationException("Does not support " + aggregateFunctionCall.getFunction().getSqlName() + " DISTINCT. 'SELECT DISTINCT' syntax could be used to deduplicate before" + " aggregation.");
    }
    final SqlAggFunction sqlAggFunction;
    if (aggregateFunctionCall.getFunction().getGroup().equals(BeamZetaSqlCatalog.USER_DEFINED_JAVA_AGGREGATE_FUNCTIONS)) {
        // Create a new operator for user-defined functions.
        SqlReturnTypeInference typeInference = x -> ZetaSqlCalciteTranslationUtils.toCalciteType(aggregateFunctionCall.getFunction().getSignatureList().get(0).getResultType().getType(), // TODO(BEAM-9514) set nullable=true
        false, getCluster().getRexBuilder());
        UdafImpl<?, ?, ?> impl = new UdafImpl<>(getExpressionConverter().userFunctionDefinitions.javaAggregateFunctions().get(aggregateFunctionCall.getFunction().getNamePath()));
        sqlAggFunction = SqlOperators.createUdafOperator(aggregateFunctionCall.getFunction().getName(), typeInference, impl);
    } else {
        // Look up builtin functions in SqlOperatorMappingTable.
        sqlAggFunction = (SqlAggFunction) SqlOperatorMappingTable.create(aggregateFunctionCall);
        if (sqlAggFunction == null) {
            throw new UnsupportedOperationException("Does not support ZetaSQL aggregate function: " + aggregateFunctionCall.getFunction().getName());
        }
    }
    List<Integer> argList = new ArrayList<>();
    ResolvedAggregateFunctionCall expr = ((ResolvedAggregateFunctionCall) computedColumn.getExpr());
    List<ZetaSQLResolvedNodeKind.ResolvedNodeKind> resolvedNodeKinds = Arrays.asList(RESOLVED_CAST, RESOLVED_COLUMN_REF, RESOLVED_GET_STRUCT_FIELD);
    for (int i = 0; i < expr.getArgumentList().size(); i++) {
        // Throw an error if aggregate function's input isn't either a ColumnRef or a cast(ColumnRef).
        // TODO: is there a general way to handle aggregation calls conversion?
        ZetaSQLResolvedNodeKind.ResolvedNodeKind resolvedNodeKind = expr.getArgumentList().get(i).nodeKind();
        if (i == 0 && resolvedNodeKinds.contains(resolvedNodeKind)) {
            argList.add(columnRefOff);
        } else if (i > 0 && resolvedNodeKind == RESOLVED_LITERAL) {
            continue;
        } else {
            throw new UnsupportedOperationException("Aggregate function only accepts Column Reference or CAST(Column Reference) as the first argument and " + "Literals as subsequent arguments as its inputs");
        }
    }
    String aggName = getTrait().resolveAlias(computedColumn.getColumn());
    return AggregateCall.create(sqlAggFunction, false, false, false, argList, -1, null, RelCollations.EMPTY, groupCount, input, // When we pass null as the return type, Calcite infers it for us.
    null, aggName);
}
Also used : ZetaSQLResolvedNodeKind(com.google.zetasql.ZetaSQLResolvedNodeKind) IntStream(java.util.stream.IntStream) AggregateCall(org.apache.beam.vendor.calcite.v1_28_0.org.apache.calcite.rel.core.AggregateCall) ZetaSqlCalciteTranslationUtils(org.apache.beam.sdk.extensions.sql.zetasql.ZetaSqlCalciteTranslationUtils) RelNode(org.apache.beam.vendor.calcite.v1_28_0.org.apache.calcite.rel.RelNode) Arrays(java.util.Arrays) ResolvedNode(com.google.zetasql.resolvedast.ResolvedNode) RexNode(org.apache.beam.vendor.calcite.v1_28_0.org.apache.calcite.rex.RexNode) ImmutableMap(org.apache.beam.vendor.guava.v26_0_jre.com.google.common.collect.ImmutableMap) ResolvedAggregateFunctionCall(com.google.zetasql.resolvedast.ResolvedNodes.ResolvedAggregateFunctionCall) ResolvedAggregateScan(com.google.zetasql.resolvedast.ResolvedNodes.ResolvedAggregateScan) RelCollations(org.apache.beam.vendor.calcite.v1_28_0.org.apache.calcite.rel.RelCollations) ArrayList(java.util.ArrayList) ImmutableBitSet(org.apache.beam.vendor.calcite.v1_28_0.org.apache.calcite.util.ImmutableBitSet) SqlAggFunction(org.apache.beam.vendor.calcite.v1_28_0.org.apache.calcite.sql.SqlAggFunction) RESOLVED_COLUMN_REF(com.google.zetasql.ZetaSQLResolvedNodeKind.ResolvedNodeKind.RESOLVED_COLUMN_REF) LogicalProject(org.apache.beam.vendor.calcite.v1_28_0.org.apache.calcite.rel.logical.LogicalProject) FunctionSignature(com.google.zetasql.FunctionSignature) RESOLVED_CAST(com.google.zetasql.ZetaSQLResolvedNodeKind.ResolvedNodeKind.RESOLVED_CAST) RESOLVED_GET_STRUCT_FIELD(com.google.zetasql.ZetaSQLResolvedNodeKind.ResolvedNodeKind.RESOLVED_GET_STRUCT_FIELD) RESOLVED_LITERAL(com.google.zetasql.ZetaSQLResolvedNodeKind.ResolvedNodeKind.RESOLVED_LITERAL) Collectors(java.util.stream.Collectors) LogicalAggregate(org.apache.beam.vendor.calcite.v1_28_0.org.apache.calcite.rel.logical.LogicalAggregate) List(java.util.List) UdafImpl(org.apache.beam.sdk.extensions.sql.impl.UdafImpl) ResolvedComputedColumn(com.google.zetasql.resolvedast.ResolvedNodes.ResolvedComputedColumn) ResolvedExpr(com.google.zetasql.resolvedast.ResolvedNodes.ResolvedExpr) TypeKind(com.google.zetasql.ZetaSQLType.TypeKind) SqlReturnTypeInference(org.apache.beam.vendor.calcite.v1_28_0.org.apache.calcite.sql.type.SqlReturnTypeInference) ImmutableList(org.apache.beam.vendor.guava.v26_0_jre.com.google.common.collect.ImmutableList) BeamZetaSqlCatalog(org.apache.beam.sdk.extensions.sql.zetasql.BeamZetaSqlCatalog) Collections(java.util.Collections) ArrayList(java.util.ArrayList) SqlAggFunction(org.apache.beam.vendor.calcite.v1_28_0.org.apache.calcite.sql.SqlAggFunction) FunctionSignature(com.google.zetasql.FunctionSignature) UdafImpl(org.apache.beam.sdk.extensions.sql.impl.UdafImpl) ResolvedAggregateFunctionCall(com.google.zetasql.resolvedast.ResolvedNodes.ResolvedAggregateFunctionCall) SqlReturnTypeInference(org.apache.beam.vendor.calcite.v1_28_0.org.apache.calcite.sql.type.SqlReturnTypeInference) ZetaSQLResolvedNodeKind(com.google.zetasql.ZetaSQLResolvedNodeKind) ZetaSQLResolvedNodeKind(com.google.zetasql.ZetaSQLResolvedNodeKind)

Example 2 with FunctionSignature

use of com.google.zetasql.FunctionSignature in project beam by apache.

the class BeamZetaSqlCatalog method addUdfsFromSchema.

private void addUdfsFromSchema() {
    for (String functionName : calciteSchema.getFunctionNames()) {
        Collection<org.apache.beam.vendor.calcite.v1_28_0.org.apache.calcite.schema.Function> functions = calciteSchema.getFunctions(functionName);
        if (functions.size() != 1) {
            throw new IllegalArgumentException(String.format("Expected exactly 1 definition for function '%s', but found %d." + " Beam ZetaSQL supports only a single function definition per function name (BEAM-12073).", functionName, functions.size()));
        }
        for (org.apache.beam.vendor.calcite.v1_28_0.org.apache.calcite.schema.Function function : functions) {
            List<String> path = Arrays.asList(functionName.split("\\."));
            if (function instanceof ScalarFunctionImpl) {
                ScalarFunctionImpl scalarFunction = (ScalarFunctionImpl) function;
                // for unsupported types.
                for (FunctionParameter parameter : scalarFunction.getParameters()) {
                    validateJavaUdfCalciteType(parameter.getType(typeFactory), functionName);
                }
                validateJavaUdfCalciteType(scalarFunction.getReturnType(typeFactory), functionName);
                Method method = scalarFunction.method;
                javaScalarUdfs.put(path, UserFunctionDefinitions.JavaScalarFunction.create(method, ""));
                FunctionArgumentType resultType = new FunctionArgumentType(ZetaSqlCalciteTranslationUtils.toZetaSqlType(scalarFunction.getReturnType(typeFactory)));
                FunctionSignature functionSignature = new FunctionSignature(resultType, getArgumentTypes(scalarFunction), 0L);
                zetaSqlCatalog.addFunction(new Function(path, USER_DEFINED_JAVA_SCALAR_FUNCTIONS, ZetaSQLFunctions.FunctionEnums.Mode.SCALAR, ImmutableList.of(functionSignature)));
            } else if (function instanceof UdafImpl) {
                UdafImpl<?, ?, ?> udaf = (UdafImpl) function;
                javaUdafs.put(path, udaf.getCombineFn());
                FunctionArgumentType resultType = new FunctionArgumentType(ZetaSqlCalciteTranslationUtils.toZetaSqlType(udaf.getReturnType(typeFactory)));
                FunctionSignature functionSignature = new FunctionSignature(resultType, getArgumentTypes(udaf), 0L);
                zetaSqlCatalog.addFunction(new Function(path, USER_DEFINED_JAVA_AGGREGATE_FUNCTIONS, ZetaSQLFunctions.FunctionEnums.Mode.AGGREGATE, ImmutableList.of(functionSignature)));
            } else {
                throw new IllegalArgumentException(String.format("Function %s has unrecognized implementation type %s.", functionName, function.getClass().getName()));
            }
        }
    }
}
Also used : ScalarFunctionImpl(org.apache.beam.sdk.extensions.sql.impl.ScalarFunctionImpl) FunctionArgumentType(com.google.zetasql.FunctionArgumentType) Method(java.lang.reflect.Method) FunctionSignature(com.google.zetasql.FunctionSignature) UdafImpl(org.apache.beam.sdk.extensions.sql.impl.UdafImpl) TableValuedFunction(com.google.zetasql.TableValuedFunction) Function(com.google.zetasql.Function) FunctionParameter(org.apache.beam.vendor.calcite.v1_28_0.org.apache.calcite.schema.FunctionParameter)

Example 3 with FunctionSignature

use of com.google.zetasql.FunctionSignature in project beam by apache.

the class BeamZetaSqlCatalog method addWindowTvfs.

@SuppressWarnings({ // customContext and volatility are in fact nullable, but they are missing the
"nullness" // annotation upstream. TODO Unsuppress when this is fixed in ZetaSQL.
})
private void addWindowTvfs() {
    FunctionArgumentType retType = new FunctionArgumentType(ZetaSQLFunctions.SignatureArgumentKind.ARG_TYPE_RELATION);
    FunctionArgumentType inputTableType = new FunctionArgumentType(ZetaSQLFunctions.SignatureArgumentKind.ARG_TYPE_RELATION);
    FunctionArgumentType descriptorType = new FunctionArgumentType(ZetaSQLFunctions.SignatureArgumentKind.ARG_TYPE_DESCRIPTOR, FunctionArgumentType.FunctionArgumentTypeOptions.builder().setDescriptorResolutionTableOffset(0).build(), 1);
    FunctionArgumentType stringType = new FunctionArgumentType(TypeFactory.createSimpleType(ZetaSQLType.TypeKind.TYPE_STRING));
    // TUMBLE
    zetaSqlCatalog.addTableValuedFunction(new TableValuedFunction.ForwardInputSchemaToOutputSchemaWithAppendedColumnTVF(ImmutableList.of(TVFStreamingUtils.FIXED_WINDOW_TVF), new FunctionSignature(retType, ImmutableList.of(inputTableType, descriptorType, stringType), -1), ImmutableList.of(TVFRelation.Column.create(TVFStreamingUtils.WINDOW_START, TypeFactory.createSimpleType(ZetaSQLType.TypeKind.TYPE_TIMESTAMP)), TVFRelation.Column.create(TVFStreamingUtils.WINDOW_END, TypeFactory.createSimpleType(ZetaSQLType.TypeKind.TYPE_TIMESTAMP))), null, null));
    // HOP
    zetaSqlCatalog.addTableValuedFunction(new TableValuedFunction.ForwardInputSchemaToOutputSchemaWithAppendedColumnTVF(ImmutableList.of(TVFStreamingUtils.SLIDING_WINDOW_TVF), new FunctionSignature(retType, ImmutableList.of(inputTableType, descriptorType, stringType, stringType), -1), ImmutableList.of(TVFRelation.Column.create(TVFStreamingUtils.WINDOW_START, TypeFactory.createSimpleType(ZetaSQLType.TypeKind.TYPE_TIMESTAMP)), TVFRelation.Column.create(TVFStreamingUtils.WINDOW_END, TypeFactory.createSimpleType(ZetaSQLType.TypeKind.TYPE_TIMESTAMP))), null, null));
    // SESSION
    zetaSqlCatalog.addTableValuedFunction(new TableValuedFunction.ForwardInputSchemaToOutputSchemaWithAppendedColumnTVF(ImmutableList.of(TVFStreamingUtils.SESSION_WINDOW_TVF), new FunctionSignature(retType, ImmutableList.of(inputTableType, descriptorType, descriptorType, stringType), -1), ImmutableList.of(TVFRelation.Column.create(TVFStreamingUtils.WINDOW_START, TypeFactory.createSimpleType(ZetaSQLType.TypeKind.TYPE_TIMESTAMP)), TVFRelation.Column.create(TVFStreamingUtils.WINDOW_END, TypeFactory.createSimpleType(ZetaSQLType.TypeKind.TYPE_TIMESTAMP))), null, null));
}
Also used : TableValuedFunction(com.google.zetasql.TableValuedFunction) FunctionArgumentType(com.google.zetasql.FunctionArgumentType) FunctionSignature(com.google.zetasql.FunctionSignature)

Aggregations

FunctionSignature (com.google.zetasql.FunctionSignature)3 FunctionArgumentType (com.google.zetasql.FunctionArgumentType)2 TableValuedFunction (com.google.zetasql.TableValuedFunction)2 UdafImpl (org.apache.beam.sdk.extensions.sql.impl.UdafImpl)2 Function (com.google.zetasql.Function)1 ZetaSQLResolvedNodeKind (com.google.zetasql.ZetaSQLResolvedNodeKind)1 RESOLVED_CAST (com.google.zetasql.ZetaSQLResolvedNodeKind.ResolvedNodeKind.RESOLVED_CAST)1 RESOLVED_COLUMN_REF (com.google.zetasql.ZetaSQLResolvedNodeKind.ResolvedNodeKind.RESOLVED_COLUMN_REF)1 RESOLVED_GET_STRUCT_FIELD (com.google.zetasql.ZetaSQLResolvedNodeKind.ResolvedNodeKind.RESOLVED_GET_STRUCT_FIELD)1 RESOLVED_LITERAL (com.google.zetasql.ZetaSQLResolvedNodeKind.ResolvedNodeKind.RESOLVED_LITERAL)1 TypeKind (com.google.zetasql.ZetaSQLType.TypeKind)1 ResolvedNode (com.google.zetasql.resolvedast.ResolvedNode)1 ResolvedAggregateFunctionCall (com.google.zetasql.resolvedast.ResolvedNodes.ResolvedAggregateFunctionCall)1 ResolvedAggregateScan (com.google.zetasql.resolvedast.ResolvedNodes.ResolvedAggregateScan)1 ResolvedComputedColumn (com.google.zetasql.resolvedast.ResolvedNodes.ResolvedComputedColumn)1 ResolvedExpr (com.google.zetasql.resolvedast.ResolvedNodes.ResolvedExpr)1 Method (java.lang.reflect.Method)1 ArrayList (java.util.ArrayList)1 Arrays (java.util.Arrays)1 Collections (java.util.Collections)1