Examples with NodeProcessor org.apache.hadoop.hive.ql.lib.NodeProcessor used on opensource projects

Example 21 with NodeProcessor

use of org.apache.hadoop.hive.ql.lib.NodeProcessor in project hive by apache.

the class TezCompiler method removeSemijoinOptimizationFromSMBJoins.

private static void removeSemijoinOptimizationFromSMBJoins(OptimizeTezProcContext procCtx) throws SemanticException {
    if (!procCtx.conf.getBoolVar(ConfVars.TEZ_DYNAMIC_SEMIJOIN_REDUCTION) || procCtx.parseContext.getRsToSemiJoinBranchInfo().size() == 0) {
        return;
    }
    Map<Rule, NodeProcessor> opRules = new LinkedHashMap<Rule, NodeProcessor>();
    opRules.put(new RuleRegExp("R1", TableScanOperator.getOperatorName() + "%" + ".*" + TezDummyStoreOperator.getOperatorName() + "%" + CommonMergeJoinOperator.getOperatorName() + "%"), new SMBJoinOpProc());
    SMBJoinOpProcContext ctx = new SMBJoinOpProcContext();
    // The dispatcher finds SMB and if there is semijoin optimization before it, removes it.
    Dispatcher disp = new DefaultRuleDispatcher(null, opRules, ctx);
    List<Node> topNodes = new ArrayList<Node>();
    topNodes.addAll(procCtx.parseContext.getTopOps().values());
    GraphWalker ogw = new PreOrderOnceWalker(disp);
    ogw.startWalking(topNodes, null);
    List<TableScanOperator> tsOps = new ArrayList<>();
    // Iterate over the map and remove semijoin optimizations if needed.
    for (CommonMergeJoinOperator joinOp : ctx.JoinOpToTsOpMap.keySet()) {
        // Get one top level TS Op directly from the stack
        tsOps.add(ctx.JoinOpToTsOpMap.get(joinOp));
        // Get the other one by examining Join Op
        List<Operator<?>> parents = joinOp.getParentOperators();
        for (Operator<?> parent : parents) {
            if (parent instanceof TezDummyStoreOperator) {
                // already accounted for
                continue;
            }
            while (parent != null) {
                if (parent instanceof TableScanOperator) {
                    tsOps.add((TableScanOperator) parent);
                    break;
                }
                parent = parent.getParentOperators().get(0);
            }
        }
    }
    // Now the relevant TableScanOperators are known, find if there exists
    // a semijoin filter on any of them, if so, remove it.
    ParseContext pctx = procCtx.parseContext;
    Set<ReduceSinkOperator> rsSet = new HashSet<>(pctx.getRsToSemiJoinBranchInfo().keySet());
    for (TableScanOperator ts : tsOps) {
        for (ReduceSinkOperator rs : rsSet) {
            SemiJoinBranchInfo sjInfo = pctx.getRsToSemiJoinBranchInfo().get(rs);
            if (sjInfo != null && ts == sjInfo.getTsOp()) {
                // match!
                if (sjInfo.getIsHint()) {
                    throw new SemanticException("Removing hinted semijoin as it is with SMB join " + rs + " : " + ts);
                }
                if (LOG.isDebugEnabled()) {
                    LOG.debug("Semijoin optimization found going to SMB join. Removing semijoin " + OperatorUtils.getOpNamePretty(rs) + " - " + OperatorUtils.getOpNamePretty(ts));
                }
                GenTezUtils.removeBranch(rs);
                GenTezUtils.removeSemiJoinOperator(pctx, rs, ts);
            }
        }
    }
}

Example 22 with NodeProcessor

use of org.apache.hadoop.hive.ql.lib.NodeProcessor in project hive by apache.

the class TezCompiler method runDynamicPartitionPruning.

private void runDynamicPartitionPruning(OptimizeTezProcContext procCtx, Set<ReadEntity> inputs, Set<WriteEntity> outputs) throws SemanticException {
    if (!procCtx.conf.getBoolVar(ConfVars.TEZ_DYNAMIC_PARTITION_PRUNING)) {
        return;
    }
    // Sequence of TableScan operators to be walked
    Deque<Operator<?>> deque = new LinkedList<Operator<?>>();
    deque.addAll(procCtx.parseContext.getTopOps().values());
    Map<Rule, NodeProcessor> opRules = new LinkedHashMap<Rule, NodeProcessor>();
    opRules.put(new RuleRegExp(new String("Dynamic Partition Pruning"), FilterOperator.getOperatorName() + "%"), new DynamicPartitionPruningOptimization());
    // The dispatcher fires the processor corresponding to the closest matching
    // rule and passes the context along
    Dispatcher disp = new DefaultRuleDispatcher(null, opRules, procCtx);
    List<Node> topNodes = new ArrayList<Node>();
    topNodes.addAll(procCtx.parseContext.getTopOps().values());
    GraphWalker ogw = new ForwardWalker(disp);
    ogw.startWalking(topNodes, null);
}

Example 23 with NodeProcessor

use of org.apache.hadoop.hive.ql.lib.NodeProcessor in project hive by apache.

the class TezCompiler method generateTaskTree.

@Override
protected void generateTaskTree(List<Task<? extends Serializable>> rootTasks, ParseContext pCtx, List<Task<MoveWork>> mvTask, Set<ReadEntity> inputs, Set<WriteEntity> outputs) throws SemanticException {
    PerfLogger perfLogger = SessionState.getPerfLogger();
    perfLogger.PerfLogBegin(this.getClass().getName(), PerfLogger.TEZ_COMPILER);
    ParseContext tempParseContext = getParseContext(pCtx, rootTasks);
    GenTezUtils utils = new GenTezUtils();
    GenTezWork genTezWork = new GenTezWork(utils);
    GenTezProcContext procCtx = new GenTezProcContext(conf, tempParseContext, mvTask, rootTasks, inputs, outputs);
    // create a walker which walks the tree in a DFS manner while maintaining
    // the operator stack.
    // The dispatcher generates the plan from the operator tree
    Map<Rule, NodeProcessor> opRules = new LinkedHashMap<Rule, NodeProcessor>();
    opRules.put(new RuleRegExp("Split Work - ReduceSink", ReduceSinkOperator.getOperatorName() + "%"), genTezWork);
    opRules.put(new RuleRegExp("No more walking on ReduceSink-MapJoin", MapJoinOperator.getOperatorName() + "%"), new ReduceSinkMapJoinProc());
    opRules.put(new RuleRegExp("Recognize a Sorted Merge Join operator to setup the right edge and" + " stop traversing the DummyStore-MapJoin", CommonMergeJoinOperator.getOperatorName() + "%"), new MergeJoinProc());
    opRules.put(new RuleRegExp("Split Work + Move/Merge - FileSink", FileSinkOperator.getOperatorName() + "%"), new CompositeProcessor(new FileSinkProcessor(), genTezWork));
    opRules.put(new RuleRegExp("Split work - DummyStore", DummyStoreOperator.getOperatorName() + "%"), genTezWork);
    opRules.put(new RuleRegExp("Handle Potential Analyze Command", TableScanOperator.getOperatorName() + "%"), new ProcessAnalyzeTable(utils));
    opRules.put(new RuleRegExp("Remember union", UnionOperator.getOperatorName() + "%"), new UnionProcessor());
    opRules.put(new RuleRegExp("AppMasterEventOperator", AppMasterEventOperator.getOperatorName() + "%"), new AppMasterEventProcessor());
    // The dispatcher fires the processor corresponding to the closest matching
    // rule and passes the context along
    Dispatcher disp = new DefaultRuleDispatcher(null, opRules, procCtx);
    List<Node> topNodes = new ArrayList<Node>();
    topNodes.addAll(pCtx.getTopOps().values());
    GraphWalker ogw = new GenTezWorkWalker(disp, procCtx);
    ogw.startWalking(topNodes, null);
    // we need to specify the reserved memory for each work that contains Map Join
    for (List<BaseWork> baseWorkList : procCtx.mapJoinWorkMap.values()) {
        for (BaseWork w : baseWorkList) {
            // work should be the smallest unit for memory allocation
            w.setReservedMemoryMB((int) (conf.getLongVar(ConfVars.HIVECONVERTJOINNOCONDITIONALTASKTHRESHOLD) / (1024 * 1024)));
        }
    }
    // we need to clone some operator plans and remove union operators still
    int indexForTezUnion = 0;
    for (BaseWork w : procCtx.workWithUnionOperators) {
        GenTezUtils.removeUnionOperators(procCtx, w, indexForTezUnion++);
    }
    // then we make sure the file sink operators are set up right
    for (FileSinkOperator fileSink : procCtx.fileSinkSet) {
        GenTezUtils.processFileSink(procCtx, fileSink);
    }
    // Connect any edges required for min/max pushdown
    if (pCtx.getRsToRuntimeValuesInfoMap().size() > 0) {
        for (ReduceSinkOperator rs : pCtx.getRsToRuntimeValuesInfoMap().keySet()) {
            // Process min/max
            GenTezUtils.processDynamicSemiJoinPushDownOperator(procCtx, pCtx.getRsToRuntimeValuesInfoMap().get(rs), rs);
        }
    }
    // and finally we hook up any events that need to be sent to the tez AM
    LOG.debug("There are " + procCtx.eventOperatorSet.size() + " app master events.");
    for (AppMasterEventOperator event : procCtx.eventOperatorSet) {
        LOG.debug("Handling AppMasterEventOperator: " + event);
        GenTezUtils.processAppMasterEvent(procCtx, event);
    }
    perfLogger.PerfLogEnd(this.getClass().getName(), PerfLogger.TEZ_COMPILER, "generateTaskTree");
}

Example 24 with NodeProcessor

use of org.apache.hadoop.hive.ql.lib.NodeProcessor in project hive by apache.

the class TezCompiler method runRemoveDynamicPruningOptimization.

private void runRemoveDynamicPruningOptimization(OptimizeTezProcContext procCtx, Set<ReadEntity> inputs, Set<WriteEntity> outputs) throws SemanticException {
    // Sequence of TableScan operators to be walked
    Deque<Operator<?>> deque = new LinkedList<Operator<?>>();
    deque.addAll(procCtx.parseContext.getTopOps().values());
    // create a walker which walks the tree in a DFS manner while maintaining
    // the operator stack.
    Map<Rule, NodeProcessor> opRules = new LinkedHashMap<Rule, NodeProcessor>();
    opRules.put(new RuleRegExp("Remove dynamic pruning by size", AppMasterEventOperator.getOperatorName() + "%"), new RemoveDynamicPruningBySize());
    // The dispatcher fires the processor corresponding to the closest matching
    // rule and passes the context along
    Dispatcher disp = new DefaultRuleDispatcher(null, opRules, procCtx);
    List<Node> topNodes = new ArrayList<Node>();
    topNodes.addAll(procCtx.parseContext.getTopOps().values());
    GraphWalker ogw = new ForwardWalker(disp);
    ogw.startWalking(topNodes, null);
}

Example 25 with NodeProcessor

use of org.apache.hadoop.hive.ql.lib.NodeProcessor in project hive by apache.

the class TypeCheckProcFactory method genExprNode.

protected static Map<ASTNode, ExprNodeDesc> genExprNode(ASTNode expr, TypeCheckCtx tcCtx, TypeCheckProcFactory tf) throws SemanticException {
    // Create the walker, the rules dispatcher and the context.
    // create a walker which walks the tree in a DFS manner while maintaining
    // the operator stack. The dispatcher
    // generates the plan from the operator tree
    Map<Rule, NodeProcessor> opRules = new LinkedHashMap<Rule, NodeProcessor>();
    opRules.put(new RuleRegExp("R1", HiveParser.TOK_NULL + "%"), tf.getNullExprProcessor());
    opRules.put(new RuleRegExp("R2", HiveParser.Number + "%|" + HiveParser.IntegralLiteral + "%|" + HiveParser.NumberLiteral + "%"), tf.getNumExprProcessor());
    opRules.put(new RuleRegExp("R3", HiveParser.Identifier + "%|" + HiveParser.StringLiteral + "%|" + HiveParser.TOK_CHARSETLITERAL + "%|" + HiveParser.TOK_STRINGLITERALSEQUENCE + "%|" + "%|" + HiveParser.KW_IF + "%|" + HiveParser.KW_CASE + "%|" + HiveParser.KW_WHEN + "%|" + HiveParser.KW_IN + "%|" + HiveParser.KW_ARRAY + "%|" + HiveParser.KW_MAP + "%|" + HiveParser.KW_STRUCT + "%|" + HiveParser.KW_EXISTS + "%|" + HiveParser.TOK_SUBQUERY_OP_NOTIN + "%"), tf.getStrExprProcessor());
    opRules.put(new RuleRegExp("R4", HiveParser.KW_TRUE + "%|" + HiveParser.KW_FALSE + "%"), tf.getBoolExprProcessor());
    opRules.put(new RuleRegExp("R5", HiveParser.TOK_DATELITERAL + "%|" + HiveParser.TOK_TIMESTAMPLITERAL + "%|" + HiveParser.TOK_TIMESTAMPLOCALTZLITERAL + "%"), tf.getDateTimeExprProcessor());
    opRules.put(new RuleRegExp("R6", HiveParser.TOK_INTERVAL_YEAR_MONTH_LITERAL + "%|" + HiveParser.TOK_INTERVAL_DAY_TIME_LITERAL + "%|" + HiveParser.TOK_INTERVAL_YEAR_LITERAL + "%|" + HiveParser.TOK_INTERVAL_MONTH_LITERAL + "%|" + HiveParser.TOK_INTERVAL_DAY_LITERAL + "%|" + HiveParser.TOK_INTERVAL_HOUR_LITERAL + "%|" + HiveParser.TOK_INTERVAL_MINUTE_LITERAL + "%|" + HiveParser.TOK_INTERVAL_SECOND_LITERAL + "%"), tf.getIntervalExprProcessor());
    opRules.put(new RuleRegExp("R7", HiveParser.TOK_TABLE_OR_COL + "%"), tf.getColumnExprProcessor());
    opRules.put(new RuleRegExp("R8", HiveParser.TOK_SUBQUERY_EXPR + "%"), tf.getSubQueryExprProcessor());
    // The dispatcher fires the processor corresponding to the closest matching
    // rule and passes the context along
    Dispatcher disp = new DefaultRuleDispatcher(tf.getDefaultExprProcessor(), opRules, tcCtx);
    GraphWalker ogw = new ExpressionWalker(disp);
    // Create a list of top nodes
    ArrayList<Node> topNodes = Lists.<Node>newArrayList(expr);
    HashMap<Node, Object> nodeOutputs = new LinkedHashMap<Node, Object>();
    ogw.startWalking(topNodes, nodeOutputs);
    return convert(nodeOutputs);
}