Search in sources :

Example 1 with HiveParserQB

use of org.apache.flink.table.planner.delegation.hive.copy.HiveParserQB in project flink by apache.

the class HiveParserCalcitePlanner method genOBLogicalPlan.

private Pair<Sort, RelNode> genOBLogicalPlan(HiveParserQB qb, RelNode srcRel, boolean outermostOB) throws SemanticException {
    Sort sortRel = null;
    RelNode originalOBInput = null;
    HiveParserQBParseInfo qbp = qb.getParseInfo();
    String dest = qbp.getClauseNames().iterator().next();
    HiveParserASTNode obAST = qbp.getOrderByForClause(dest);
    if (obAST != null) {
        // 1. OB Expr sanity test
        // in strict mode, in the presence of order by, limit must be specified
        Integer limit = qb.getParseInfo().getDestLimit(dest);
        if (limit == null) {
            String mapRedMode = semanticAnalyzer.getConf().getVar(HiveConf.ConfVars.HIVEMAPREDMODE);
            boolean banLargeQuery = Boolean.parseBoolean(semanticAnalyzer.getConf().get("hive.strict.checks.large.query", "false"));
            if ("strict".equalsIgnoreCase(mapRedMode) || banLargeQuery) {
                throw new SemanticException(generateErrorMessage(obAST, "Order by-s without limit"));
            }
        }
        // 2. Walk through OB exprs and extract field collations and additional
        // virtual columns needed
        final List<RexNode> virtualCols = new ArrayList<>();
        final List<RelFieldCollation> fieldCollations = new ArrayList<>();
        int fieldIndex;
        List<Node> obASTExprLst = obAST.getChildren();
        HiveParserASTNode obASTExpr;
        HiveParserASTNode nullOrderASTExpr;
        List<Pair<HiveParserASTNode, TypeInfo>> vcASTAndType = new ArrayList<>();
        HiveParserRowResolver inputRR = relToRowResolver.get(srcRel);
        HiveParserRowResolver outputRR = new HiveParserRowResolver();
        HiveParserRexNodeConverter converter = new HiveParserRexNodeConverter(cluster, srcRel.getRowType(), relToHiveColNameCalcitePosMap.get(srcRel), 0, false, funcConverter);
        int numSrcFields = srcRel.getRowType().getFieldCount();
        for (Node node : obASTExprLst) {
            // 2.1 Convert AST Expr to ExprNode
            obASTExpr = (HiveParserASTNode) node;
            nullOrderASTExpr = (HiveParserASTNode) obASTExpr.getChild(0);
            HiveParserASTNode ref = (HiveParserASTNode) nullOrderASTExpr.getChild(0);
            Map<HiveParserASTNode, ExprNodeDesc> astToExprNodeDesc = semanticAnalyzer.genAllExprNodeDesc(ref, inputRR);
            ExprNodeDesc obExprNodeDesc = astToExprNodeDesc.get(ref);
            if (obExprNodeDesc == null) {
                throw new SemanticException("Invalid order by expression: " + obASTExpr.toString());
            }
            // 2.2 Convert ExprNode to RexNode
            RexNode rexNode = converter.convert(obExprNodeDesc).accept(funcConverter);
            // present in the child (& hence we add a child Project Rel)
            if (rexNode instanceof RexInputRef) {
                fieldIndex = ((RexInputRef) rexNode).getIndex();
            } else {
                fieldIndex = numSrcFields + virtualCols.size();
                virtualCols.add(rexNode);
                vcASTAndType.add(new Pair<>(ref, obExprNodeDesc.getTypeInfo()));
            }
            // 2.4 Determine the Direction of order by
            RelFieldCollation.Direction direction = RelFieldCollation.Direction.DESCENDING;
            if (obASTExpr.getType() == HiveASTParser.TOK_TABSORTCOLNAMEASC) {
                direction = RelFieldCollation.Direction.ASCENDING;
            }
            RelFieldCollation.NullDirection nullOrder;
            if (nullOrderASTExpr.getType() == HiveASTParser.TOK_NULLS_FIRST) {
                nullOrder = RelFieldCollation.NullDirection.FIRST;
            } else if (nullOrderASTExpr.getType() == HiveASTParser.TOK_NULLS_LAST) {
                nullOrder = RelFieldCollation.NullDirection.LAST;
            } else {
                throw new SemanticException("Unexpected null ordering option: " + nullOrderASTExpr.getType());
            }
            // 2.5 Add to field collations
            fieldCollations.add(new RelFieldCollation(fieldIndex, direction, nullOrder));
        }
        // 3. Add Child Project Rel if needed, Generate Output RR, input Sel Rel
        // for top constraining Sel
        RelNode obInputRel = srcRel;
        if (!virtualCols.isEmpty()) {
            List<RexNode> originalInputRefs = srcRel.getRowType().getFieldList().stream().map(input -> new RexInputRef(input.getIndex(), input.getType())).collect(Collectors.toList());
            HiveParserRowResolver obSyntheticProjectRR = new HiveParserRowResolver();
            if (!HiveParserRowResolver.add(obSyntheticProjectRR, inputRR)) {
                throw new SemanticException("Duplicates detected when adding columns to RR: see previous message");
            }
            int vcolPos = inputRR.getRowSchema().getSignature().size();
            for (Pair<HiveParserASTNode, TypeInfo> astTypePair : vcASTAndType) {
                obSyntheticProjectRR.putExpression(astTypePair.getKey(), new ColumnInfo(getColumnInternalName(vcolPos), astTypePair.getValue(), null, false));
                vcolPos++;
            }
            obInputRel = genSelectRelNode(CompositeList.of(originalInputRefs, virtualCols), obSyntheticProjectRR, srcRel);
            if (outermostOB) {
                if (!HiveParserRowResolver.add(outputRR, inputRR)) {
                    throw new SemanticException("Duplicates detected when adding columns to RR: see previous message");
                }
            } else {
                if (!HiveParserRowResolver.add(outputRR, obSyntheticProjectRR)) {
                    throw new SemanticException("Duplicates detected when adding columns to RR: see previous message");
                }
            }
            originalOBInput = srcRel;
        } else {
            if (!HiveParserRowResolver.add(outputRR, inputRR)) {
                throw new SemanticException("Duplicates detected when adding columns to RR: see previous message");
            }
        }
        // 4. Construct SortRel
        RelTraitSet traitSet = cluster.traitSet();
        RelCollation canonizedCollation = traitSet.canonize(RelCollationImpl.of(fieldCollations));
        sortRel = LogicalSort.create(obInputRel, canonizedCollation, null, null);
        // 5. Update the maps
        Map<String, Integer> hiveColNameCalcitePosMap = buildHiveToCalciteColumnMap(outputRR);
        relToRowResolver.put(sortRel, outputRR);
        relToHiveColNameCalcitePosMap.put(sortRel, hiveColNameCalcitePosMap);
    }
    return (new Pair<>(sortRel, originalOBInput));
}
Also used : DataType(org.apache.flink.table.types.DataType) Arrays(java.util.Arrays) LogicalSort(org.apache.calcite.rel.logical.LogicalSort) FlinkPlannerImpl(org.apache.flink.table.planner.calcite.FlinkPlannerImpl) JoinType(org.apache.hadoop.hive.ql.parse.JoinType) HiveParserBaseSemanticAnalyzer.getGroupByForClause(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getGroupByForClause) HiveParserContext(org.apache.flink.table.planner.delegation.hive.copy.HiveParserContext) FunctionRegistry(org.apache.hadoop.hive.ql.exec.FunctionRegistry) RelCollationImpl(org.apache.calcite.rel.RelCollationImpl) BigDecimal(java.math.BigDecimal) HiveParserBaseSemanticAnalyzer.unescapeIdentifier(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.unescapeIdentifier) HiveInspectors(org.apache.flink.table.functions.hive.conversion.HiveInspectors) CorrelationId(org.apache.calcite.rel.core.CorrelationId) HiveParserASTNode(org.apache.flink.table.planner.delegation.hive.copy.HiveParserASTNode) Map(java.util.Map) HiveParserASTBuilder(org.apache.flink.table.planner.delegation.hive.copy.HiveParserASTBuilder) RelTraitSet(org.apache.calcite.plan.RelTraitSet) RexWindowBound(org.apache.calcite.rex.RexWindowBound) ImmutableBitSet(org.apache.calcite.util.ImmutableBitSet) ExprNodeDesc(org.apache.hadoop.hive.ql.plan.ExprNodeDesc) HiveParserSubQueryUtils(org.apache.flink.table.planner.delegation.hive.copy.HiveParserSubQueryUtils) HiveParserBaseSemanticAnalyzer.getHiveAggInfo(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getHiveAggInfo) HiveTypeUtil(org.apache.flink.table.catalog.hive.util.HiveTypeUtil) HiveParserBaseSemanticAnalyzer(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer) org.apache.hadoop.hive.serde.serdeConstants(org.apache.hadoop.hive.serde.serdeConstants) Set(java.util.Set) HiveParserUtils.generateErrorMessage(org.apache.flink.table.planner.delegation.hive.HiveParserUtils.generateErrorMessage) RelFieldCollation(org.apache.calcite.rel.RelFieldCollation) HiveASTParseUtils(org.apache.flink.table.planner.delegation.hive.copy.HiveASTParseUtils) HiveParserBaseSemanticAnalyzer.getGroupingSetsForCube(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getGroupingSetsForCube) SqlStdOperatorTable(org.apache.calcite.sql.fun.SqlStdOperatorTable) HiveParserPreCboCtx(org.apache.flink.table.planner.delegation.hive.copy.HiveParserPreCboCtx) RelCollation(org.apache.calcite.rel.RelCollation) HiveParserBaseSemanticAnalyzer.getGroupingSets(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getGroupingSets) HiveParserBaseSemanticAnalyzer.getPartitionKeys(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getPartitionKeys) HiveParserBaseSemanticAnalyzer.removeOBInSubQuery(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.removeOBInSubQuery) HiveASTParser(org.apache.flink.table.planner.delegation.hive.parse.HiveASTParser) ErrorMsg(org.apache.hadoop.hive.ql.ErrorMsg) RexCall(org.apache.calcite.rex.RexCall) StructField(org.apache.hadoop.hive.serde2.objectinspector.StructField) TypeInfoUtils(org.apache.hadoop.hive.serde2.typeinfo.TypeInfoUtils) HiveShim(org.apache.flink.table.catalog.hive.client.HiveShim) HiveParserJoinTypeCheckCtx(org.apache.flink.table.planner.delegation.hive.copy.HiveParserJoinTypeCheckCtx) ArrayList(java.util.ArrayList) LinkedHashMap(java.util.LinkedHashMap) AggInfo(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.AggInfo) HiveParserBaseSemanticAnalyzer.getCorrelationUse(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getCorrelationUse) HiveASTParseDriver(org.apache.flink.table.planner.delegation.hive.copy.HiveASTParseDriver) RexFieldCollation(org.apache.calcite.rex.RexFieldCollation) ViewExpanders(org.apache.calcite.plan.ViewExpanders) LogicalValues(org.apache.calcite.rel.logical.LogicalValues) LogicalCorrelate(org.apache.calcite.rel.logical.LogicalCorrelate) HiveParserBaseSemanticAnalyzer.validateNoHavingReferenceToAlias(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.validateNoHavingReferenceToAlias) HiveParserNamedJoinInfo(org.apache.flink.table.planner.delegation.hive.copy.HiveParserNamedJoinInfo) RelDataType(org.apache.calcite.rel.type.RelDataType) HiveParserUtils.rewriteGroupingFunctionAST(org.apache.flink.table.planner.delegation.hive.HiveParserUtils.rewriteGroupingFunctionAST) LogicalIntersect(org.apache.calcite.rel.logical.LogicalIntersect) HiveParserQBSubQuery(org.apache.flink.table.planner.delegation.hive.copy.HiveParserQBSubQuery) Table(org.apache.hadoop.hive.ql.metadata.Table) HiveParserRowResolver(org.apache.flink.table.planner.delegation.hive.copy.HiveParserRowResolver) Aggregate(org.apache.calcite.rel.core.Aggregate) HiveParserTypeCheckCtx(org.apache.flink.table.planner.delegation.hive.copy.HiveParserTypeCheckCtx) FrameworkConfig(org.apache.calcite.tools.FrameworkConfig) Node(org.apache.hadoop.hive.ql.lib.Node) HiveParserBaseSemanticAnalyzer.getBound(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getBound) HiveParserBaseSemanticAnalyzer.getColumnInternalName(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getColumnInternalName) HiveParserSqlFunctionConverter(org.apache.flink.table.planner.delegation.hive.copy.HiveParserSqlFunctionConverter) LogicalAggregate(org.apache.calcite.rel.logical.LogicalAggregate) JoinRelType(org.apache.calcite.rel.core.JoinRelType) AggregateCall(org.apache.calcite.rel.core.AggregateCall) SqlAggFunction(org.apache.calcite.sql.SqlAggFunction) ArrayDeque(java.util.ArrayDeque) RelDataTypeFactory(org.apache.calcite.rel.type.RelDataTypeFactory) ColumnAccessInfo(org.apache.hadoop.hive.ql.parse.ColumnAccessInfo) HiveParserBaseSemanticAnalyzer.obtainTableType(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.obtainTableType) HiveParserBaseSemanticAnalyzer.convert(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.convert) LogicalFilter(org.apache.calcite.rel.logical.LogicalFilter) RelFactories(org.apache.calcite.rel.core.RelFactories) LoggerFactory(org.slf4j.LoggerFactory) LogicalTableFunctionScan(org.apache.calcite.rel.logical.LogicalTableFunctionScan) SemanticException(org.apache.hadoop.hive.ql.parse.SemanticException) HiveParserBaseSemanticAnalyzer.processPositionAlias(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.processPositionAlias) HiveParserWindowingSpec(org.apache.flink.table.planner.delegation.hive.copy.HiveParserWindowingSpec) RexUtil(org.apache.calcite.rex.RexUtil) LogicalJoin(org.apache.calcite.rel.logical.LogicalJoin) HiveParserErrorMsg(org.apache.flink.table.planner.delegation.hive.parse.HiveParserErrorMsg) RexNode(org.apache.calcite.rex.RexNode) LogicalUnion(org.apache.calcite.rel.logical.LogicalUnion) RelOptCluster(org.apache.calcite.plan.RelOptCluster) LogicalDistribution(org.apache.flink.table.planner.plan.nodes.hive.LogicalDistribution) RexLiteral(org.apache.calcite.rex.RexLiteral) HiveParserQBParseInfo(org.apache.flink.table.planner.delegation.hive.copy.HiveParserQBParseInfo) HiveParserTypeConverter(org.apache.flink.table.planner.delegation.hive.copy.HiveParserTypeConverter) StructObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.StructObjectInspector) Preconditions(org.apache.flink.util.Preconditions) CompositeList(org.apache.calcite.util.CompositeList) Collectors(java.util.stream.Collectors) RexInputRef(org.apache.calcite.rex.RexInputRef) VirtualColumn(org.apache.hadoop.hive.ql.metadata.VirtualColumn) List(java.util.List) Type(java.lang.reflect.Type) Sort(org.apache.calcite.rel.core.Sort) RelDataTypeField(org.apache.calcite.rel.type.RelDataTypeField) GenericUDAFEvaluator(org.apache.hadoop.hive.ql.udf.generic.GenericUDAFEvaluator) HiveParserQueryState(org.apache.flink.table.planner.delegation.hive.copy.HiveParserQueryState) Project(org.apache.calcite.rel.core.Project) CatalogManager(org.apache.flink.table.catalog.CatalogManager) HiveParserBaseSemanticAnalyzer.getGroupingSetsForRollup(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getGroupingSetsForRollup) HiveParserBaseSemanticAnalyzer.getWindowSpecIndx(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getWindowSpecIndx) HashMap(java.util.HashMap) SetOp(org.apache.calcite.rel.core.SetOp) Deque(java.util.Deque) RelOptUtil(org.apache.calcite.plan.RelOptUtil) LogicalMinus(org.apache.calcite.rel.logical.LogicalMinus) DeduplicateCorrelateVariables(org.apache.calcite.sql2rel.DeduplicateCorrelateVariables) HiveParserBaseSemanticAnalyzer.initPhase1Ctx(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.initPhase1Ctx) PlannerContext(org.apache.flink.table.planner.delegation.PlannerContext) HashSet(java.util.HashSet) HiveParserQBExpr(org.apache.flink.table.planner.delegation.hive.copy.HiveParserQBExpr) TableType(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.TableType) Pair(org.apache.calcite.util.Pair) HiveParserBaseSemanticAnalyzer.buildHiveToCalciteColumnMap(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.buildHiveToCalciteColumnMap) SqlOperator(org.apache.calcite.sql.SqlOperator) FlinkCalciteCatalogReader(org.apache.flink.table.planner.plan.FlinkCalciteCatalogReader) RelCollations(org.apache.calcite.rel.RelCollations) ColumnInfo(org.apache.hadoop.hive.ql.exec.ColumnInfo) ExprNodeColumnDesc(org.apache.hadoop.hive.ql.plan.ExprNodeColumnDesc) Logger(org.slf4j.Logger) HiveParserSemanticAnalyzer(org.apache.flink.table.planner.delegation.hive.copy.HiveParserSemanticAnalyzer) LogicalProject(org.apache.calcite.rel.logical.LogicalProject) SqlTypeName(org.apache.calcite.sql.type.SqlTypeName) TypeInfoFactory(org.apache.hadoop.hive.serde2.typeinfo.TypeInfoFactory) Iterator(java.util.Iterator) RexBuilder(org.apache.calcite.rex.RexBuilder) ExprNodeConstantDesc(org.apache.hadoop.hive.ql.plan.ExprNodeConstantDesc) GroupByDesc(org.apache.hadoop.hive.ql.plan.GroupByDesc) HiveConf(org.apache.hadoop.hive.conf.HiveConf) HiveParserQB(org.apache.flink.table.planner.delegation.hive.copy.HiveParserQB) RelNode(org.apache.calcite.rel.RelNode) HiveParserBaseSemanticAnalyzer.genValues(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.genValues) HiveParserUtils.verifyCanHandleAst(org.apache.flink.table.planner.delegation.hive.HiveParserUtils.verifyCanHandleAst) TypeInfo(org.apache.hadoop.hive.serde2.typeinfo.TypeInfo) FieldSchema(org.apache.hadoop.hive.metastore.api.FieldSchema) AbstractMap(java.util.AbstractMap) ObjectPair(org.apache.hadoop.hive.common.ObjectPair) HiveParserBaseSemanticAnalyzer.buildHiveColNameToInputPosMap(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.buildHiveColNameToInputPosMap) HiveParserCreateViewInfo(org.apache.flink.table.planner.delegation.hive.parse.HiveParserCreateViewInfo) SqlUserDefinedTableFunction(org.apache.calcite.sql.validate.SqlUserDefinedTableFunction) HiveParserBaseSemanticAnalyzer.topLevelConjunctCheck(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.topLevelConjunctCheck) Util(org.apache.calcite.util.Util) HiveParserBaseSemanticAnalyzer.addToGBExpr(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.addToGBExpr) Collections(java.util.Collections) HiveParserBaseSemanticAnalyzer.getOrderKeys(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getOrderKeys) HiveParserASTNode(org.apache.flink.table.planner.delegation.hive.copy.HiveParserASTNode) Node(org.apache.hadoop.hive.ql.lib.Node) RexNode(org.apache.calcite.rex.RexNode) RelNode(org.apache.calcite.rel.RelNode) ArrayList(java.util.ArrayList) ColumnInfo(org.apache.hadoop.hive.ql.exec.ColumnInfo) RelTraitSet(org.apache.calcite.plan.RelTraitSet) HiveParserRowResolver(org.apache.flink.table.planner.delegation.hive.copy.HiveParserRowResolver) LogicalSort(org.apache.calcite.rel.logical.LogicalSort) Sort(org.apache.calcite.rel.core.Sort) ExprNodeDesc(org.apache.hadoop.hive.ql.plan.ExprNodeDesc) SemanticException(org.apache.hadoop.hive.ql.parse.SemanticException) Pair(org.apache.calcite.util.Pair) ObjectPair(org.apache.hadoop.hive.common.ObjectPair) HiveParserASTNode(org.apache.flink.table.planner.delegation.hive.copy.HiveParserASTNode) TypeInfo(org.apache.hadoop.hive.serde2.typeinfo.TypeInfo) RelCollation(org.apache.calcite.rel.RelCollation) HiveParserQBParseInfo(org.apache.flink.table.planner.delegation.hive.copy.HiveParserQBParseInfo) RelNode(org.apache.calcite.rel.RelNode) RelFieldCollation(org.apache.calcite.rel.RelFieldCollation) RexInputRef(org.apache.calcite.rex.RexInputRef) RexNode(org.apache.calcite.rex.RexNode)

Example 2 with HiveParserQB

use of org.apache.flink.table.planner.delegation.hive.copy.HiveParserQB in project flink by apache.

the class HiveParserDMLHelper method createInsertOperation.

public Operation createInsertOperation(HiveParserCalcitePlanner analyzer, RelNode queryRelNode) throws SemanticException {
    HiveParserQB topQB = analyzer.getQB();
    QBMetaData qbMetaData = topQB.getMetaData();
    // decide the dest table
    Map<String, Table> nameToDestTable = qbMetaData.getNameToDestTable();
    Map<String, Partition> nameToDestPart = qbMetaData.getNameToDestPartition();
    // for now we only support inserting to a single table
    Preconditions.checkState(nameToDestTable.size() <= 1 && nameToDestPart.size() <= 1, "Only support inserting to 1 table");
    Table destTable;
    String insClauseName;
    if (!nameToDestTable.isEmpty()) {
        insClauseName = nameToDestTable.keySet().iterator().next();
        destTable = nameToDestTable.values().iterator().next();
    } else if (!nameToDestPart.isEmpty()) {
        insClauseName = nameToDestPart.keySet().iterator().next();
        destTable = nameToDestPart.values().iterator().next().getTable();
    } else {
        // happens for INSERT DIRECTORY
        throw new SemanticException("INSERT DIRECTORY is not supported");
    }
    // decide static partition specs
    Map<String, String> staticPartSpec = new LinkedHashMap<>();
    if (destTable.isPartitioned()) {
        List<String> partCols = HiveCatalog.getFieldNames(destTable.getTTable().getPartitionKeys());
        if (!nameToDestPart.isEmpty()) {
            // static partition
            Partition destPart = nameToDestPart.values().iterator().next();
            Preconditions.checkState(partCols.size() == destPart.getValues().size(), "Part cols and static spec doesn't match");
            for (int i = 0; i < partCols.size(); i++) {
                staticPartSpec.put(partCols.get(i), destPart.getValues().get(i));
            }
        } else {
            // dynamic partition
            Map<String, String> spec = qbMetaData.getPartSpecForAlias(insClauseName);
            if (spec != null) {
                for (String partCol : partCols) {
                    String val = spec.get(partCol);
                    if (val != null) {
                        staticPartSpec.put(partCol, val);
                    }
                }
            }
        }
    }
    // decide whether it's overwrite
    boolean overwrite = topQB.getParseInfo().getInsertOverwriteTables().keySet().stream().map(String::toLowerCase).collect(Collectors.toSet()).contains(destTable.getDbName() + "." + destTable.getTableName());
    Tuple4<ObjectIdentifier, QueryOperation, Map<String, String>, Boolean> insertOperationInfo = createInsertOperationInfo(queryRelNode, destTable, staticPartSpec, analyzer.getDestSchemaForClause(insClauseName), overwrite);
    return new SinkModifyOperation(catalogManager.getTableOrError(insertOperationInfo.f0), insertOperationInfo.f1, insertOperationInfo.f2, insertOperationInfo.f3, Collections.emptyMap());
}
Also used : Partition(org.apache.hadoop.hive.ql.metadata.Partition) Table(org.apache.hadoop.hive.ql.metadata.Table) HiveParserQB(org.apache.flink.table.planner.delegation.hive.copy.HiveParserQB) SinkModifyOperation(org.apache.flink.table.operations.SinkModifyOperation) LinkedHashMap(java.util.LinkedHashMap) QBMetaData(org.apache.hadoop.hive.ql.parse.QBMetaData) Map(java.util.Map) LinkedHashMap(java.util.LinkedHashMap) SemanticException(org.apache.hadoop.hive.ql.parse.SemanticException) ObjectIdentifier(org.apache.flink.table.catalog.ObjectIdentifier) PlannerQueryOperation(org.apache.flink.table.planner.operations.PlannerQueryOperation) QueryOperation(org.apache.flink.table.operations.QueryOperation)

Example 3 with HiveParserQB

use of org.apache.flink.table.planner.delegation.hive.copy.HiveParserQB in project flink by apache.

the class HiveParserCalcitePlanner method genDistSortBy.

// Generate plan for sort by, cluster by and distribute by. This is basically same as generating
// order by plan.
// Should refactor to combine them.
private Pair<RelNode, RelNode> genDistSortBy(HiveParserQB qb, RelNode srcRel, boolean outermostOB) throws SemanticException {
    RelNode res = null;
    RelNode originalInput = null;
    HiveParserQBParseInfo qbp = qb.getParseInfo();
    String destClause = qbp.getClauseNames().iterator().next();
    HiveParserASTNode sortAST = qbp.getSortByForClause(destClause);
    HiveParserASTNode distAST = qbp.getDistributeByForClause(destClause);
    HiveParserASTNode clusterAST = qbp.getClusterByForClause(destClause);
    if (sortAST != null || distAST != null || clusterAST != null) {
        List<RexNode> virtualCols = new ArrayList<>();
        List<Pair<HiveParserASTNode, TypeInfo>> vcASTAndType = new ArrayList<>();
        List<RelFieldCollation> fieldCollations = new ArrayList<>();
        List<Integer> distKeys = new ArrayList<>();
        HiveParserRowResolver inputRR = relToRowResolver.get(srcRel);
        HiveParserRexNodeConverter converter = new HiveParserRexNodeConverter(cluster, srcRel.getRowType(), relToHiveColNameCalcitePosMap.get(srcRel), 0, false, funcConverter);
        int numSrcFields = srcRel.getRowType().getFieldCount();
        // handle cluster by
        if (clusterAST != null) {
            if (sortAST != null) {
                throw new SemanticException("Cannot have both CLUSTER BY and SORT BY");
            }
            if (distAST != null) {
                throw new SemanticException("Cannot have both CLUSTER BY and DISTRIBUTE BY");
            }
            for (Node node : clusterAST.getChildren()) {
                HiveParserASTNode childAST = (HiveParserASTNode) node;
                Map<HiveParserASTNode, ExprNodeDesc> astToExprNodeDesc = semanticAnalyzer.genAllExprNodeDesc(childAST, inputRR);
                ExprNodeDesc childNodeDesc = astToExprNodeDesc.get(childAST);
                if (childNodeDesc == null) {
                    throw new SemanticException("Invalid CLUSTER BY expression: " + childAST.toString());
                }
                RexNode childRexNode = converter.convert(childNodeDesc).accept(funcConverter);
                int fieldIndex;
                if (childRexNode instanceof RexInputRef) {
                    fieldIndex = ((RexInputRef) childRexNode).getIndex();
                } else {
                    fieldIndex = numSrcFields + virtualCols.size();
                    virtualCols.add(childRexNode);
                    vcASTAndType.add(new Pair<>(childAST, childNodeDesc.getTypeInfo()));
                }
                // cluster by doesn't support specifying ASC/DESC or NULLS FIRST/LAST, so use
                // default values
                fieldCollations.add(new RelFieldCollation(fieldIndex, RelFieldCollation.Direction.ASCENDING, RelFieldCollation.NullDirection.FIRST));
                distKeys.add(fieldIndex);
            }
        } else {
            // handle sort by
            if (sortAST != null) {
                for (Node node : sortAST.getChildren()) {
                    HiveParserASTNode childAST = (HiveParserASTNode) node;
                    HiveParserASTNode nullOrderAST = (HiveParserASTNode) childAST.getChild(0);
                    HiveParserASTNode fieldAST = (HiveParserASTNode) nullOrderAST.getChild(0);
                    Map<HiveParserASTNode, ExprNodeDesc> astToExprNodeDesc = semanticAnalyzer.genAllExprNodeDesc(fieldAST, inputRR);
                    ExprNodeDesc fieldNodeDesc = astToExprNodeDesc.get(fieldAST);
                    if (fieldNodeDesc == null) {
                        throw new SemanticException("Invalid sort by expression: " + fieldAST.toString());
                    }
                    RexNode childRexNode = converter.convert(fieldNodeDesc).accept(funcConverter);
                    int fieldIndex;
                    if (childRexNode instanceof RexInputRef) {
                        fieldIndex = ((RexInputRef) childRexNode).getIndex();
                    } else {
                        fieldIndex = numSrcFields + virtualCols.size();
                        virtualCols.add(childRexNode);
                        vcASTAndType.add(new Pair<>(childAST, fieldNodeDesc.getTypeInfo()));
                    }
                    RelFieldCollation.Direction direction = RelFieldCollation.Direction.DESCENDING;
                    if (childAST.getType() == HiveASTParser.TOK_TABSORTCOLNAMEASC) {
                        direction = RelFieldCollation.Direction.ASCENDING;
                    }
                    RelFieldCollation.NullDirection nullOrder;
                    if (nullOrderAST.getType() == HiveASTParser.TOK_NULLS_FIRST) {
                        nullOrder = RelFieldCollation.NullDirection.FIRST;
                    } else if (nullOrderAST.getType() == HiveASTParser.TOK_NULLS_LAST) {
                        nullOrder = RelFieldCollation.NullDirection.LAST;
                    } else {
                        throw new SemanticException("Unexpected null ordering option: " + nullOrderAST.getType());
                    }
                    fieldCollations.add(new RelFieldCollation(fieldIndex, direction, nullOrder));
                }
            }
            // handle distribute by
            if (distAST != null) {
                for (Node node : distAST.getChildren()) {
                    HiveParserASTNode childAST = (HiveParserASTNode) node;
                    Map<HiveParserASTNode, ExprNodeDesc> astToExprNodeDesc = semanticAnalyzer.genAllExprNodeDesc(childAST, inputRR);
                    ExprNodeDesc childNodeDesc = astToExprNodeDesc.get(childAST);
                    if (childNodeDesc == null) {
                        throw new SemanticException("Invalid DISTRIBUTE BY expression: " + childAST.toString());
                    }
                    RexNode childRexNode = converter.convert(childNodeDesc).accept(funcConverter);
                    int fieldIndex;
                    if (childRexNode instanceof RexInputRef) {
                        fieldIndex = ((RexInputRef) childRexNode).getIndex();
                    } else {
                        fieldIndex = numSrcFields + virtualCols.size();
                        virtualCols.add(childRexNode);
                        vcASTAndType.add(new Pair<>(childAST, childNodeDesc.getTypeInfo()));
                    }
                    distKeys.add(fieldIndex);
                }
            }
        }
        Preconditions.checkState(!fieldCollations.isEmpty() || !distKeys.isEmpty(), "Both field collations and dist keys are empty");
        // add child SEL if needed
        RelNode realInput = srcRel;
        HiveParserRowResolver outputRR = new HiveParserRowResolver();
        if (!virtualCols.isEmpty()) {
            List<RexNode> originalInputRefs = srcRel.getRowType().getFieldList().stream().map(input -> new RexInputRef(input.getIndex(), input.getType())).collect(Collectors.toList());
            HiveParserRowResolver addedProjectRR = new HiveParserRowResolver();
            if (!HiveParserRowResolver.add(addedProjectRR, inputRR)) {
                throw new SemanticException("Duplicates detected when adding columns to RR: see previous message");
            }
            int vColPos = inputRR.getRowSchema().getSignature().size();
            for (Pair<HiveParserASTNode, TypeInfo> astTypePair : vcASTAndType) {
                addedProjectRR.putExpression(astTypePair.getKey(), new ColumnInfo(getColumnInternalName(vColPos), astTypePair.getValue(), null, false));
                vColPos++;
            }
            realInput = genSelectRelNode(CompositeList.of(originalInputRefs, virtualCols), addedProjectRR, srcRel);
            if (outermostOB) {
                if (!HiveParserRowResolver.add(outputRR, inputRR)) {
                    throw new SemanticException("Duplicates detected when adding columns to RR: see previous message");
                }
            } else {
                if (!HiveParserRowResolver.add(outputRR, addedProjectRR)) {
                    throw new SemanticException("Duplicates detected when adding columns to RR: see previous message");
                }
            }
            originalInput = srcRel;
        } else {
            if (!HiveParserRowResolver.add(outputRR, inputRR)) {
                throw new SemanticException("Duplicates detected when adding columns to RR: see previous message");
            }
        }
        // create rel node
        RelTraitSet traitSet = cluster.traitSet();
        RelCollation canonizedCollation = traitSet.canonize(RelCollationImpl.of(fieldCollations));
        res = LogicalDistribution.create(realInput, canonizedCollation, distKeys);
        Map<String, Integer> hiveColNameCalcitePosMap = buildHiveToCalciteColumnMap(outputRR);
        relToRowResolver.put(res, outputRR);
        relToHiveColNameCalcitePosMap.put(res, hiveColNameCalcitePosMap);
    }
    return (new Pair<>(res, originalInput));
}
Also used : DataType(org.apache.flink.table.types.DataType) Arrays(java.util.Arrays) LogicalSort(org.apache.calcite.rel.logical.LogicalSort) FlinkPlannerImpl(org.apache.flink.table.planner.calcite.FlinkPlannerImpl) JoinType(org.apache.hadoop.hive.ql.parse.JoinType) HiveParserBaseSemanticAnalyzer.getGroupByForClause(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getGroupByForClause) HiveParserContext(org.apache.flink.table.planner.delegation.hive.copy.HiveParserContext) FunctionRegistry(org.apache.hadoop.hive.ql.exec.FunctionRegistry) RelCollationImpl(org.apache.calcite.rel.RelCollationImpl) BigDecimal(java.math.BigDecimal) HiveParserBaseSemanticAnalyzer.unescapeIdentifier(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.unescapeIdentifier) HiveInspectors(org.apache.flink.table.functions.hive.conversion.HiveInspectors) CorrelationId(org.apache.calcite.rel.core.CorrelationId) HiveParserASTNode(org.apache.flink.table.planner.delegation.hive.copy.HiveParserASTNode) Map(java.util.Map) HiveParserASTBuilder(org.apache.flink.table.planner.delegation.hive.copy.HiveParserASTBuilder) RelTraitSet(org.apache.calcite.plan.RelTraitSet) RexWindowBound(org.apache.calcite.rex.RexWindowBound) ImmutableBitSet(org.apache.calcite.util.ImmutableBitSet) ExprNodeDesc(org.apache.hadoop.hive.ql.plan.ExprNodeDesc) HiveParserSubQueryUtils(org.apache.flink.table.planner.delegation.hive.copy.HiveParserSubQueryUtils) HiveParserBaseSemanticAnalyzer.getHiveAggInfo(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getHiveAggInfo) HiveTypeUtil(org.apache.flink.table.catalog.hive.util.HiveTypeUtil) HiveParserBaseSemanticAnalyzer(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer) org.apache.hadoop.hive.serde.serdeConstants(org.apache.hadoop.hive.serde.serdeConstants) Set(java.util.Set) HiveParserUtils.generateErrorMessage(org.apache.flink.table.planner.delegation.hive.HiveParserUtils.generateErrorMessage) RelFieldCollation(org.apache.calcite.rel.RelFieldCollation) HiveASTParseUtils(org.apache.flink.table.planner.delegation.hive.copy.HiveASTParseUtils) HiveParserBaseSemanticAnalyzer.getGroupingSetsForCube(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getGroupingSetsForCube) SqlStdOperatorTable(org.apache.calcite.sql.fun.SqlStdOperatorTable) HiveParserPreCboCtx(org.apache.flink.table.planner.delegation.hive.copy.HiveParserPreCboCtx) RelCollation(org.apache.calcite.rel.RelCollation) HiveParserBaseSemanticAnalyzer.getGroupingSets(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getGroupingSets) HiveParserBaseSemanticAnalyzer.getPartitionKeys(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getPartitionKeys) HiveParserBaseSemanticAnalyzer.removeOBInSubQuery(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.removeOBInSubQuery) HiveASTParser(org.apache.flink.table.planner.delegation.hive.parse.HiveASTParser) ErrorMsg(org.apache.hadoop.hive.ql.ErrorMsg) RexCall(org.apache.calcite.rex.RexCall) StructField(org.apache.hadoop.hive.serde2.objectinspector.StructField) TypeInfoUtils(org.apache.hadoop.hive.serde2.typeinfo.TypeInfoUtils) HiveShim(org.apache.flink.table.catalog.hive.client.HiveShim) HiveParserJoinTypeCheckCtx(org.apache.flink.table.planner.delegation.hive.copy.HiveParserJoinTypeCheckCtx) ArrayList(java.util.ArrayList) LinkedHashMap(java.util.LinkedHashMap) AggInfo(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.AggInfo) HiveParserBaseSemanticAnalyzer.getCorrelationUse(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getCorrelationUse) HiveASTParseDriver(org.apache.flink.table.planner.delegation.hive.copy.HiveASTParseDriver) RexFieldCollation(org.apache.calcite.rex.RexFieldCollation) ViewExpanders(org.apache.calcite.plan.ViewExpanders) LogicalValues(org.apache.calcite.rel.logical.LogicalValues) LogicalCorrelate(org.apache.calcite.rel.logical.LogicalCorrelate) HiveParserBaseSemanticAnalyzer.validateNoHavingReferenceToAlias(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.validateNoHavingReferenceToAlias) HiveParserNamedJoinInfo(org.apache.flink.table.planner.delegation.hive.copy.HiveParserNamedJoinInfo) RelDataType(org.apache.calcite.rel.type.RelDataType) HiveParserUtils.rewriteGroupingFunctionAST(org.apache.flink.table.planner.delegation.hive.HiveParserUtils.rewriteGroupingFunctionAST) LogicalIntersect(org.apache.calcite.rel.logical.LogicalIntersect) HiveParserQBSubQuery(org.apache.flink.table.planner.delegation.hive.copy.HiveParserQBSubQuery) Table(org.apache.hadoop.hive.ql.metadata.Table) HiveParserRowResolver(org.apache.flink.table.planner.delegation.hive.copy.HiveParserRowResolver) Aggregate(org.apache.calcite.rel.core.Aggregate) HiveParserTypeCheckCtx(org.apache.flink.table.planner.delegation.hive.copy.HiveParserTypeCheckCtx) FrameworkConfig(org.apache.calcite.tools.FrameworkConfig) Node(org.apache.hadoop.hive.ql.lib.Node) HiveParserBaseSemanticAnalyzer.getBound(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getBound) HiveParserBaseSemanticAnalyzer.getColumnInternalName(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getColumnInternalName) HiveParserSqlFunctionConverter(org.apache.flink.table.planner.delegation.hive.copy.HiveParserSqlFunctionConverter) LogicalAggregate(org.apache.calcite.rel.logical.LogicalAggregate) JoinRelType(org.apache.calcite.rel.core.JoinRelType) AggregateCall(org.apache.calcite.rel.core.AggregateCall) SqlAggFunction(org.apache.calcite.sql.SqlAggFunction) ArrayDeque(java.util.ArrayDeque) RelDataTypeFactory(org.apache.calcite.rel.type.RelDataTypeFactory) ColumnAccessInfo(org.apache.hadoop.hive.ql.parse.ColumnAccessInfo) HiveParserBaseSemanticAnalyzer.obtainTableType(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.obtainTableType) HiveParserBaseSemanticAnalyzer.convert(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.convert) LogicalFilter(org.apache.calcite.rel.logical.LogicalFilter) RelFactories(org.apache.calcite.rel.core.RelFactories) LoggerFactory(org.slf4j.LoggerFactory) LogicalTableFunctionScan(org.apache.calcite.rel.logical.LogicalTableFunctionScan) SemanticException(org.apache.hadoop.hive.ql.parse.SemanticException) HiveParserBaseSemanticAnalyzer.processPositionAlias(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.processPositionAlias) HiveParserWindowingSpec(org.apache.flink.table.planner.delegation.hive.copy.HiveParserWindowingSpec) RexUtil(org.apache.calcite.rex.RexUtil) LogicalJoin(org.apache.calcite.rel.logical.LogicalJoin) HiveParserErrorMsg(org.apache.flink.table.planner.delegation.hive.parse.HiveParserErrorMsg) RexNode(org.apache.calcite.rex.RexNode) LogicalUnion(org.apache.calcite.rel.logical.LogicalUnion) RelOptCluster(org.apache.calcite.plan.RelOptCluster) LogicalDistribution(org.apache.flink.table.planner.plan.nodes.hive.LogicalDistribution) RexLiteral(org.apache.calcite.rex.RexLiteral) HiveParserQBParseInfo(org.apache.flink.table.planner.delegation.hive.copy.HiveParserQBParseInfo) HiveParserTypeConverter(org.apache.flink.table.planner.delegation.hive.copy.HiveParserTypeConverter) StructObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.StructObjectInspector) Preconditions(org.apache.flink.util.Preconditions) CompositeList(org.apache.calcite.util.CompositeList) Collectors(java.util.stream.Collectors) RexInputRef(org.apache.calcite.rex.RexInputRef) VirtualColumn(org.apache.hadoop.hive.ql.metadata.VirtualColumn) List(java.util.List) Type(java.lang.reflect.Type) Sort(org.apache.calcite.rel.core.Sort) RelDataTypeField(org.apache.calcite.rel.type.RelDataTypeField) GenericUDAFEvaluator(org.apache.hadoop.hive.ql.udf.generic.GenericUDAFEvaluator) HiveParserQueryState(org.apache.flink.table.planner.delegation.hive.copy.HiveParserQueryState) Project(org.apache.calcite.rel.core.Project) CatalogManager(org.apache.flink.table.catalog.CatalogManager) HiveParserBaseSemanticAnalyzer.getGroupingSetsForRollup(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getGroupingSetsForRollup) HiveParserBaseSemanticAnalyzer.getWindowSpecIndx(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getWindowSpecIndx) HashMap(java.util.HashMap) SetOp(org.apache.calcite.rel.core.SetOp) Deque(java.util.Deque) RelOptUtil(org.apache.calcite.plan.RelOptUtil) LogicalMinus(org.apache.calcite.rel.logical.LogicalMinus) DeduplicateCorrelateVariables(org.apache.calcite.sql2rel.DeduplicateCorrelateVariables) HiveParserBaseSemanticAnalyzer.initPhase1Ctx(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.initPhase1Ctx) PlannerContext(org.apache.flink.table.planner.delegation.PlannerContext) HashSet(java.util.HashSet) HiveParserQBExpr(org.apache.flink.table.planner.delegation.hive.copy.HiveParserQBExpr) TableType(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.TableType) Pair(org.apache.calcite.util.Pair) HiveParserBaseSemanticAnalyzer.buildHiveToCalciteColumnMap(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.buildHiveToCalciteColumnMap) SqlOperator(org.apache.calcite.sql.SqlOperator) FlinkCalciteCatalogReader(org.apache.flink.table.planner.plan.FlinkCalciteCatalogReader) RelCollations(org.apache.calcite.rel.RelCollations) ColumnInfo(org.apache.hadoop.hive.ql.exec.ColumnInfo) ExprNodeColumnDesc(org.apache.hadoop.hive.ql.plan.ExprNodeColumnDesc) Logger(org.slf4j.Logger) HiveParserSemanticAnalyzer(org.apache.flink.table.planner.delegation.hive.copy.HiveParserSemanticAnalyzer) LogicalProject(org.apache.calcite.rel.logical.LogicalProject) SqlTypeName(org.apache.calcite.sql.type.SqlTypeName) TypeInfoFactory(org.apache.hadoop.hive.serde2.typeinfo.TypeInfoFactory) Iterator(java.util.Iterator) RexBuilder(org.apache.calcite.rex.RexBuilder) ExprNodeConstantDesc(org.apache.hadoop.hive.ql.plan.ExprNodeConstantDesc) GroupByDesc(org.apache.hadoop.hive.ql.plan.GroupByDesc) HiveConf(org.apache.hadoop.hive.conf.HiveConf) HiveParserQB(org.apache.flink.table.planner.delegation.hive.copy.HiveParserQB) RelNode(org.apache.calcite.rel.RelNode) HiveParserBaseSemanticAnalyzer.genValues(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.genValues) HiveParserUtils.verifyCanHandleAst(org.apache.flink.table.planner.delegation.hive.HiveParserUtils.verifyCanHandleAst) TypeInfo(org.apache.hadoop.hive.serde2.typeinfo.TypeInfo) FieldSchema(org.apache.hadoop.hive.metastore.api.FieldSchema) AbstractMap(java.util.AbstractMap) ObjectPair(org.apache.hadoop.hive.common.ObjectPair) HiveParserBaseSemanticAnalyzer.buildHiveColNameToInputPosMap(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.buildHiveColNameToInputPosMap) HiveParserCreateViewInfo(org.apache.flink.table.planner.delegation.hive.parse.HiveParserCreateViewInfo) SqlUserDefinedTableFunction(org.apache.calcite.sql.validate.SqlUserDefinedTableFunction) HiveParserBaseSemanticAnalyzer.topLevelConjunctCheck(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.topLevelConjunctCheck) Util(org.apache.calcite.util.Util) HiveParserBaseSemanticAnalyzer.addToGBExpr(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.addToGBExpr) Collections(java.util.Collections) HiveParserBaseSemanticAnalyzer.getOrderKeys(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getOrderKeys) HiveParserASTNode(org.apache.flink.table.planner.delegation.hive.copy.HiveParserASTNode) Node(org.apache.hadoop.hive.ql.lib.Node) RexNode(org.apache.calcite.rex.RexNode) RelNode(org.apache.calcite.rel.RelNode) ArrayList(java.util.ArrayList) ColumnInfo(org.apache.hadoop.hive.ql.exec.ColumnInfo) RelTraitSet(org.apache.calcite.plan.RelTraitSet) HiveParserRowResolver(org.apache.flink.table.planner.delegation.hive.copy.HiveParserRowResolver) ExprNodeDesc(org.apache.hadoop.hive.ql.plan.ExprNodeDesc) Pair(org.apache.calcite.util.Pair) ObjectPair(org.apache.hadoop.hive.common.ObjectPair) SemanticException(org.apache.hadoop.hive.ql.parse.SemanticException) HiveParserASTNode(org.apache.flink.table.planner.delegation.hive.copy.HiveParserASTNode) TypeInfo(org.apache.hadoop.hive.serde2.typeinfo.TypeInfo) RelCollation(org.apache.calcite.rel.RelCollation) HiveParserQBParseInfo(org.apache.flink.table.planner.delegation.hive.copy.HiveParserQBParseInfo) RelNode(org.apache.calcite.rel.RelNode) RelFieldCollation(org.apache.calcite.rel.RelFieldCollation) RexInputRef(org.apache.calcite.rex.RexInputRef) RexNode(org.apache.calcite.rex.RexNode)

Example 4 with HiveParserQB

use of org.apache.flink.table.planner.delegation.hive.copy.HiveParserQB in project flink by apache.

the class HiveParserCalcitePlanner method genLogicalPlan.

private RelNode genLogicalPlan(HiveParserQB qb, boolean outerMostQB, Map<String, Integer> outerNameToPosMap, HiveParserRowResolver outerRR) throws SemanticException {
    RelNode res;
    // First generate all the opInfos for the elements in the from clause
    Map<String, RelNode> aliasToRel = new HashMap<>();
    // 0. Check if we can handle the SubQuery;
    // canHandleQbForCbo returns null if the query can be handled.
    String reason = HiveParserUtils.canHandleQbForCbo(semanticAnalyzer.getQueryProperties());
    if (reason != null) {
        String msg = "CBO can not handle Sub Query" + " because it: " + reason;
        throw new SemanticException(msg);
    }
    // 1.1. Recurse over the subqueries to fill the subquery part of the plan
    for (String subqAlias : qb.getSubqAliases()) {
        HiveParserQBExpr qbexpr = qb.getSubqForAlias(subqAlias);
        RelNode relNode = genLogicalPlan(qbexpr);
        aliasToRel.put(subqAlias, relNode);
        if (qb.getViewToTabSchema().containsKey(subqAlias)) {
            if (!(relNode instanceof Project)) {
                throw new SemanticException("View " + subqAlias + " is corresponding to " + relNode.toString() + ", rather than a Project.");
            }
        }
    }
    // 1.2 Recurse over all the source tables
    for (String tableAlias : qb.getTabAliases()) {
        RelNode op = genTableLogicalPlan(tableAlias, qb);
        aliasToRel.put(tableAlias, op);
    }
    if (aliasToRel.isEmpty()) {
        RelNode dummySrc = LogicalValues.createOneRow(cluster);
        aliasToRel.put(HiveParserSemanticAnalyzer.DUMMY_TABLE, dummySrc);
        HiveParserRowResolver dummyRR = new HiveParserRowResolver();
        dummyRR.put(HiveParserSemanticAnalyzer.DUMMY_TABLE, "dummy_col", new ColumnInfo(getColumnInternalName(0), TypeInfoFactory.intTypeInfo, HiveParserSemanticAnalyzer.DUMMY_TABLE, false));
        relToRowResolver.put(dummySrc, dummyRR);
        relToHiveColNameCalcitePosMap.put(dummySrc, buildHiveToCalciteColumnMap(dummyRR));
    }
    if (!qb.getParseInfo().getAliasToLateralViews().isEmpty()) {
        // process lateral views
        res = genLateralViewPlan(qb, aliasToRel);
    } else if (qb.getParseInfo().getJoinExpr() != null) {
        // 1.3 process join
        res = genJoinLogicalPlan(qb.getParseInfo().getJoinExpr(), aliasToRel);
    } else {
        // If no join then there should only be either 1 TS or 1 SubQuery
        res = aliasToRel.values().iterator().next();
    }
    // 2. Build Rel for where Clause
    RelNode filterRel = genFilterLogicalPlan(qb, res, outerNameToPosMap, outerRR);
    res = (filterRel == null) ? res : filterRel;
    RelNode starSrcRel = res;
    // 3. Build Rel for GB Clause
    RelNode gbRel = genGBLogicalPlan(qb, res);
    res = gbRel == null ? res : gbRel;
    // 4. Build Rel for GB Having Clause
    RelNode gbHavingRel = genGBHavingLogicalPlan(qb, res);
    res = gbHavingRel == null ? res : gbHavingRel;
    // 5. Build Rel for Select Clause
    RelNode selectRel = genSelectLogicalPlan(qb, res, starSrcRel, outerNameToPosMap, outerRR);
    res = selectRel == null ? res : selectRel;
    // 6. Build Rel for OB Clause
    Pair<Sort, RelNode> obAndTopProj = genOBLogicalPlan(qb, res, outerMostQB);
    Sort orderRel = obAndTopProj.getKey();
    RelNode topConstrainingProjRel = obAndTopProj.getValue();
    res = orderRel == null ? res : orderRel;
    // Build Rel for SortBy/ClusterBy/DistributeBy. It can happen only if we don't have OrderBy.
    if (orderRel == null) {
        Pair<RelNode, RelNode> distAndTopProj = genDistSortBy(qb, res, outerMostQB);
        RelNode distRel = distAndTopProj.getKey();
        topConstrainingProjRel = distAndTopProj.getValue();
        res = distRel == null ? res : distRel;
    }
    // 7. Build Rel for Limit Clause
    Sort limitRel = genLimitLogicalPlan(qb, res);
    if (limitRel != null) {
        if (orderRel != null) {
            // merge limit into the order-by node
            HiveParserRowResolver orderRR = relToRowResolver.remove(orderRel);
            Map<String, Integer> orderColNameToPos = relToHiveColNameCalcitePosMap.remove(orderRel);
            res = LogicalSort.create(orderRel.getInput(), orderRel.collation, limitRel.offset, limitRel.fetch);
            relToRowResolver.put(res, orderRR);
            relToHiveColNameCalcitePosMap.put(res, orderColNameToPos);
            relToRowResolver.remove(limitRel);
            relToHiveColNameCalcitePosMap.remove(limitRel);
        } else {
            res = limitRel;
        }
    }
    // 8. Introduce top constraining select if needed.
    if (topConstrainingProjRel != null) {
        List<RexNode> originalInputRefs = topConstrainingProjRel.getRowType().getFieldList().stream().map(input -> new RexInputRef(input.getIndex(), input.getType())).collect(Collectors.toList());
        HiveParserRowResolver topConstrainingProjRR = new HiveParserRowResolver();
        if (!HiveParserRowResolver.add(topConstrainingProjRR, relToRowResolver.get(topConstrainingProjRel))) {
            LOG.warn("Duplicates detected when adding columns to RR: see previous message");
        }
        res = genSelectRelNode(originalInputRefs, topConstrainingProjRR, res);
    }
    // TODO: cleanup this
    if (qb.getParseInfo().getAlias() != null) {
        HiveParserRowResolver rr = relToRowResolver.get(res);
        HiveParserRowResolver newRR = new HiveParserRowResolver();
        String alias = qb.getParseInfo().getAlias();
        for (ColumnInfo colInfo : rr.getColumnInfos()) {
            String name = colInfo.getInternalName();
            String[] tmp = rr.reverseLookup(name);
            if ("".equals(tmp[0]) || tmp[1] == null) {
                // ast expression is not a valid column name for table
                tmp[1] = colInfo.getInternalName();
            }
            ColumnInfo newColInfo = new ColumnInfo(colInfo);
            newColInfo.setTabAlias(alias);
            newRR.put(alias, tmp[1], newColInfo);
        }
        relToRowResolver.put(res, newRR);
        relToHiveColNameCalcitePosMap.put(res, buildHiveToCalciteColumnMap(newRR));
    }
    if (LOG.isDebugEnabled()) {
        LOG.debug("Created Plan for Query Block " + qb.getId());
    }
    semanticAnalyzer.setQB(qb);
    return res;
}
Also used : DataType(org.apache.flink.table.types.DataType) Arrays(java.util.Arrays) LogicalSort(org.apache.calcite.rel.logical.LogicalSort) FlinkPlannerImpl(org.apache.flink.table.planner.calcite.FlinkPlannerImpl) JoinType(org.apache.hadoop.hive.ql.parse.JoinType) HiveParserBaseSemanticAnalyzer.getGroupByForClause(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getGroupByForClause) HiveParserContext(org.apache.flink.table.planner.delegation.hive.copy.HiveParserContext) FunctionRegistry(org.apache.hadoop.hive.ql.exec.FunctionRegistry) RelCollationImpl(org.apache.calcite.rel.RelCollationImpl) BigDecimal(java.math.BigDecimal) HiveParserBaseSemanticAnalyzer.unescapeIdentifier(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.unescapeIdentifier) HiveInspectors(org.apache.flink.table.functions.hive.conversion.HiveInspectors) CorrelationId(org.apache.calcite.rel.core.CorrelationId) HiveParserASTNode(org.apache.flink.table.planner.delegation.hive.copy.HiveParserASTNode) Map(java.util.Map) HiveParserASTBuilder(org.apache.flink.table.planner.delegation.hive.copy.HiveParserASTBuilder) RelTraitSet(org.apache.calcite.plan.RelTraitSet) RexWindowBound(org.apache.calcite.rex.RexWindowBound) ImmutableBitSet(org.apache.calcite.util.ImmutableBitSet) ExprNodeDesc(org.apache.hadoop.hive.ql.plan.ExprNodeDesc) HiveParserSubQueryUtils(org.apache.flink.table.planner.delegation.hive.copy.HiveParserSubQueryUtils) HiveParserBaseSemanticAnalyzer.getHiveAggInfo(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getHiveAggInfo) HiveTypeUtil(org.apache.flink.table.catalog.hive.util.HiveTypeUtil) HiveParserBaseSemanticAnalyzer(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer) org.apache.hadoop.hive.serde.serdeConstants(org.apache.hadoop.hive.serde.serdeConstants) Set(java.util.Set) HiveParserUtils.generateErrorMessage(org.apache.flink.table.planner.delegation.hive.HiveParserUtils.generateErrorMessage) RelFieldCollation(org.apache.calcite.rel.RelFieldCollation) HiveASTParseUtils(org.apache.flink.table.planner.delegation.hive.copy.HiveASTParseUtils) HiveParserBaseSemanticAnalyzer.getGroupingSetsForCube(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getGroupingSetsForCube) SqlStdOperatorTable(org.apache.calcite.sql.fun.SqlStdOperatorTable) HiveParserPreCboCtx(org.apache.flink.table.planner.delegation.hive.copy.HiveParserPreCboCtx) RelCollation(org.apache.calcite.rel.RelCollation) HiveParserBaseSemanticAnalyzer.getGroupingSets(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getGroupingSets) HiveParserBaseSemanticAnalyzer.getPartitionKeys(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getPartitionKeys) HiveParserBaseSemanticAnalyzer.removeOBInSubQuery(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.removeOBInSubQuery) HiveASTParser(org.apache.flink.table.planner.delegation.hive.parse.HiveASTParser) ErrorMsg(org.apache.hadoop.hive.ql.ErrorMsg) RexCall(org.apache.calcite.rex.RexCall) StructField(org.apache.hadoop.hive.serde2.objectinspector.StructField) TypeInfoUtils(org.apache.hadoop.hive.serde2.typeinfo.TypeInfoUtils) HiveShim(org.apache.flink.table.catalog.hive.client.HiveShim) HiveParserJoinTypeCheckCtx(org.apache.flink.table.planner.delegation.hive.copy.HiveParserJoinTypeCheckCtx) ArrayList(java.util.ArrayList) LinkedHashMap(java.util.LinkedHashMap) AggInfo(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.AggInfo) HiveParserBaseSemanticAnalyzer.getCorrelationUse(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getCorrelationUse) HiveASTParseDriver(org.apache.flink.table.planner.delegation.hive.copy.HiveASTParseDriver) RexFieldCollation(org.apache.calcite.rex.RexFieldCollation) ViewExpanders(org.apache.calcite.plan.ViewExpanders) LogicalValues(org.apache.calcite.rel.logical.LogicalValues) LogicalCorrelate(org.apache.calcite.rel.logical.LogicalCorrelate) HiveParserBaseSemanticAnalyzer.validateNoHavingReferenceToAlias(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.validateNoHavingReferenceToAlias) HiveParserNamedJoinInfo(org.apache.flink.table.planner.delegation.hive.copy.HiveParserNamedJoinInfo) RelDataType(org.apache.calcite.rel.type.RelDataType) HiveParserUtils.rewriteGroupingFunctionAST(org.apache.flink.table.planner.delegation.hive.HiveParserUtils.rewriteGroupingFunctionAST) LogicalIntersect(org.apache.calcite.rel.logical.LogicalIntersect) HiveParserQBSubQuery(org.apache.flink.table.planner.delegation.hive.copy.HiveParserQBSubQuery) Table(org.apache.hadoop.hive.ql.metadata.Table) HiveParserRowResolver(org.apache.flink.table.planner.delegation.hive.copy.HiveParserRowResolver) Aggregate(org.apache.calcite.rel.core.Aggregate) HiveParserTypeCheckCtx(org.apache.flink.table.planner.delegation.hive.copy.HiveParserTypeCheckCtx) FrameworkConfig(org.apache.calcite.tools.FrameworkConfig) Node(org.apache.hadoop.hive.ql.lib.Node) HiveParserBaseSemanticAnalyzer.getBound(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getBound) HiveParserBaseSemanticAnalyzer.getColumnInternalName(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getColumnInternalName) HiveParserSqlFunctionConverter(org.apache.flink.table.planner.delegation.hive.copy.HiveParserSqlFunctionConverter) LogicalAggregate(org.apache.calcite.rel.logical.LogicalAggregate) JoinRelType(org.apache.calcite.rel.core.JoinRelType) AggregateCall(org.apache.calcite.rel.core.AggregateCall) SqlAggFunction(org.apache.calcite.sql.SqlAggFunction) ArrayDeque(java.util.ArrayDeque) RelDataTypeFactory(org.apache.calcite.rel.type.RelDataTypeFactory) ColumnAccessInfo(org.apache.hadoop.hive.ql.parse.ColumnAccessInfo) HiveParserBaseSemanticAnalyzer.obtainTableType(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.obtainTableType) HiveParserBaseSemanticAnalyzer.convert(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.convert) LogicalFilter(org.apache.calcite.rel.logical.LogicalFilter) RelFactories(org.apache.calcite.rel.core.RelFactories) LoggerFactory(org.slf4j.LoggerFactory) LogicalTableFunctionScan(org.apache.calcite.rel.logical.LogicalTableFunctionScan) SemanticException(org.apache.hadoop.hive.ql.parse.SemanticException) HiveParserBaseSemanticAnalyzer.processPositionAlias(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.processPositionAlias) HiveParserWindowingSpec(org.apache.flink.table.planner.delegation.hive.copy.HiveParserWindowingSpec) RexUtil(org.apache.calcite.rex.RexUtil) LogicalJoin(org.apache.calcite.rel.logical.LogicalJoin) HiveParserErrorMsg(org.apache.flink.table.planner.delegation.hive.parse.HiveParserErrorMsg) RexNode(org.apache.calcite.rex.RexNode) LogicalUnion(org.apache.calcite.rel.logical.LogicalUnion) RelOptCluster(org.apache.calcite.plan.RelOptCluster) LogicalDistribution(org.apache.flink.table.planner.plan.nodes.hive.LogicalDistribution) RexLiteral(org.apache.calcite.rex.RexLiteral) HiveParserQBParseInfo(org.apache.flink.table.planner.delegation.hive.copy.HiveParserQBParseInfo) HiveParserTypeConverter(org.apache.flink.table.planner.delegation.hive.copy.HiveParserTypeConverter) StructObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.StructObjectInspector) Preconditions(org.apache.flink.util.Preconditions) CompositeList(org.apache.calcite.util.CompositeList) Collectors(java.util.stream.Collectors) RexInputRef(org.apache.calcite.rex.RexInputRef) VirtualColumn(org.apache.hadoop.hive.ql.metadata.VirtualColumn) List(java.util.List) Type(java.lang.reflect.Type) Sort(org.apache.calcite.rel.core.Sort) RelDataTypeField(org.apache.calcite.rel.type.RelDataTypeField) GenericUDAFEvaluator(org.apache.hadoop.hive.ql.udf.generic.GenericUDAFEvaluator) HiveParserQueryState(org.apache.flink.table.planner.delegation.hive.copy.HiveParserQueryState) Project(org.apache.calcite.rel.core.Project) CatalogManager(org.apache.flink.table.catalog.CatalogManager) HiveParserBaseSemanticAnalyzer.getGroupingSetsForRollup(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getGroupingSetsForRollup) HiveParserBaseSemanticAnalyzer.getWindowSpecIndx(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getWindowSpecIndx) HashMap(java.util.HashMap) SetOp(org.apache.calcite.rel.core.SetOp) Deque(java.util.Deque) RelOptUtil(org.apache.calcite.plan.RelOptUtil) LogicalMinus(org.apache.calcite.rel.logical.LogicalMinus) DeduplicateCorrelateVariables(org.apache.calcite.sql2rel.DeduplicateCorrelateVariables) HiveParserBaseSemanticAnalyzer.initPhase1Ctx(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.initPhase1Ctx) PlannerContext(org.apache.flink.table.planner.delegation.PlannerContext) HashSet(java.util.HashSet) HiveParserQBExpr(org.apache.flink.table.planner.delegation.hive.copy.HiveParserQBExpr) TableType(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.TableType) Pair(org.apache.calcite.util.Pair) HiveParserBaseSemanticAnalyzer.buildHiveToCalciteColumnMap(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.buildHiveToCalciteColumnMap) SqlOperator(org.apache.calcite.sql.SqlOperator) FlinkCalciteCatalogReader(org.apache.flink.table.planner.plan.FlinkCalciteCatalogReader) RelCollations(org.apache.calcite.rel.RelCollations) ColumnInfo(org.apache.hadoop.hive.ql.exec.ColumnInfo) ExprNodeColumnDesc(org.apache.hadoop.hive.ql.plan.ExprNodeColumnDesc) Logger(org.slf4j.Logger) HiveParserSemanticAnalyzer(org.apache.flink.table.planner.delegation.hive.copy.HiveParserSemanticAnalyzer) LogicalProject(org.apache.calcite.rel.logical.LogicalProject) SqlTypeName(org.apache.calcite.sql.type.SqlTypeName) TypeInfoFactory(org.apache.hadoop.hive.serde2.typeinfo.TypeInfoFactory) Iterator(java.util.Iterator) RexBuilder(org.apache.calcite.rex.RexBuilder) ExprNodeConstantDesc(org.apache.hadoop.hive.ql.plan.ExprNodeConstantDesc) GroupByDesc(org.apache.hadoop.hive.ql.plan.GroupByDesc) HiveConf(org.apache.hadoop.hive.conf.HiveConf) HiveParserQB(org.apache.flink.table.planner.delegation.hive.copy.HiveParserQB) RelNode(org.apache.calcite.rel.RelNode) HiveParserBaseSemanticAnalyzer.genValues(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.genValues) HiveParserUtils.verifyCanHandleAst(org.apache.flink.table.planner.delegation.hive.HiveParserUtils.verifyCanHandleAst) TypeInfo(org.apache.hadoop.hive.serde2.typeinfo.TypeInfo) FieldSchema(org.apache.hadoop.hive.metastore.api.FieldSchema) AbstractMap(java.util.AbstractMap) ObjectPair(org.apache.hadoop.hive.common.ObjectPair) HiveParserBaseSemanticAnalyzer.buildHiveColNameToInputPosMap(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.buildHiveColNameToInputPosMap) HiveParserCreateViewInfo(org.apache.flink.table.planner.delegation.hive.parse.HiveParserCreateViewInfo) SqlUserDefinedTableFunction(org.apache.calcite.sql.validate.SqlUserDefinedTableFunction) HiveParserBaseSemanticAnalyzer.topLevelConjunctCheck(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.topLevelConjunctCheck) Util(org.apache.calcite.util.Util) HiveParserBaseSemanticAnalyzer.addToGBExpr(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.addToGBExpr) Collections(java.util.Collections) HiveParserBaseSemanticAnalyzer.getOrderKeys(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer.getOrderKeys) LinkedHashMap(java.util.LinkedHashMap) HashMap(java.util.HashMap) ColumnInfo(org.apache.hadoop.hive.ql.exec.ColumnInfo) Project(org.apache.calcite.rel.core.Project) LogicalProject(org.apache.calcite.rel.logical.LogicalProject) RelNode(org.apache.calcite.rel.RelNode) HiveParserRowResolver(org.apache.flink.table.planner.delegation.hive.copy.HiveParserRowResolver) LogicalSort(org.apache.calcite.rel.logical.LogicalSort) Sort(org.apache.calcite.rel.core.Sort) RexInputRef(org.apache.calcite.rex.RexInputRef) HiveParserQBExpr(org.apache.flink.table.planner.delegation.hive.copy.HiveParserQBExpr) SemanticException(org.apache.hadoop.hive.ql.parse.SemanticException) RexNode(org.apache.calcite.rex.RexNode)

Example 5 with HiveParserQB

use of org.apache.flink.table.planner.delegation.hive.copy.HiveParserQB in project flink by apache.

the class HiveParserCalcitePlanner method genSubQueryRelNode.

private boolean genSubQueryRelNode(HiveParserQB qb, HiveParserASTNode node, RelNode srcRel, boolean forHavingClause, Map<HiveParserASTNode, RelNode> subQueryToRelNode) throws SemanticException {
    Set<HiveParserASTNode> corrScalarQueriesWithAgg = new HashSet<>();
    // disallow sub-queries which HIVE doesn't currently support
    subqueryRestrictionCheck(qb, node, srcRel, forHavingClause, corrScalarQueriesWithAgg);
    Deque<HiveParserASTNode> stack = new ArrayDeque<>();
    stack.push(node);
    boolean isSubQuery = false;
    while (!stack.isEmpty()) {
        HiveParserASTNode next = stack.pop();
        switch(next.getType()) {
            case HiveASTParser.TOK_SUBQUERY_EXPR:
                // Restriction 2.h Subquery is not allowed in LHS
                if (next.getChildren().size() == 3 && next.getChild(2).getType() == HiveASTParser.TOK_SUBQUERY_EXPR) {
                    throw new SemanticException(HiveParserErrorMsg.getMsg(ErrorMsg.UNSUPPORTED_SUBQUERY_EXPRESSION, next.getChild(2), "SubQuery in LHS expressions are not supported."));
                }
                String sbQueryAlias = "sq_" + qb.incrNumSubQueryPredicates();
                HiveParserQB subQB = new HiveParserQB(qb.getId(), sbQueryAlias, true);
                HiveParserBaseSemanticAnalyzer.Phase1Ctx ctx1 = initPhase1Ctx();
                semanticAnalyzer.doPhase1((HiveParserASTNode) next.getChild(1), subQB, ctx1, null);
                semanticAnalyzer.getMetaData(subQB);
                RelNode subQueryRelNode = genLogicalPlan(subQB, false, relToHiveColNameCalcitePosMap.get(srcRel), relToRowResolver.get(srcRel));
                subQueryToRelNode.put(next, subQueryRelNode);
                isSubQuery = true;
                break;
            default:
                int childCount = next.getChildCount();
                for (int i = childCount - 1; i >= 0; i--) {
                    stack.push((HiveParserASTNode) next.getChild(i));
                }
        }
    }
    return isSubQuery;
}
Also used : HiveParserASTNode(org.apache.flink.table.planner.delegation.hive.copy.HiveParserASTNode) HiveParserQB(org.apache.flink.table.planner.delegation.hive.copy.HiveParserQB) ArrayDeque(java.util.ArrayDeque) RelNode(org.apache.calcite.rel.RelNode) HiveParserBaseSemanticAnalyzer(org.apache.flink.table.planner.delegation.hive.copy.HiveParserBaseSemanticAnalyzer) HashSet(java.util.HashSet) SemanticException(org.apache.hadoop.hive.ql.parse.SemanticException)

Aggregations

ArrayDeque (java.util.ArrayDeque)4 HashSet (java.util.HashSet)4 LinkedHashMap (java.util.LinkedHashMap)4 Map (java.util.Map)4 RelNode (org.apache.calcite.rel.RelNode)4 HiveParserQB (org.apache.flink.table.planner.delegation.hive.copy.HiveParserQB)4 SemanticException (org.apache.hadoop.hive.ql.parse.SemanticException)4 Type (java.lang.reflect.Type)3 BigDecimal (java.math.BigDecimal)3 AbstractMap (java.util.AbstractMap)3 ArrayList (java.util.ArrayList)3 Arrays (java.util.Arrays)3 Collections (java.util.Collections)3 Deque (java.util.Deque)3 HashMap (java.util.HashMap)3 Iterator (java.util.Iterator)3 List (java.util.List)3 Set (java.util.Set)3 Collectors (java.util.stream.Collectors)3 RelOptCluster (org.apache.calcite.plan.RelOptCluster)3