Examples with SetHashGroupByMinReduction org.apache.hadoop.hive.ql.optimizer.SetHashGroupByMinReduction used on opensource projects

Example 1 with SetHashGroupByMinReduction

use of org.apache.hadoop.hive.ql.optimizer.SetHashGroupByMinReduction in project hive by apache.

the class TezCompiler method runStatsDependentOptimizations.

private void runStatsDependentOptimizations(OptimizeTezProcContext procCtx) 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<SemanticRule, SemanticNodeProcessor> opRules = new LinkedHashMap<SemanticRule, SemanticNodeProcessor>();
    opRules.put(new RuleRegExp("Set parallelism - ReduceSink", ReduceSinkOperator.getOperatorName() + "%"), new SetReducerParallelism());
    opRules.put(new RuleRegExp("Convert Join to Map-join", JoinOperator.getOperatorName() + "%"), new ConvertJoinMapJoin());
    if (procCtx.conf.getBoolVar(ConfVars.HIVEMAPAGGRHASHMINREDUCTIONSTATSADJUST)) {
        opRules.put(new RuleRegExp("Set min reduction - GBy (Hash)", GroupByOperator.getOperatorName() + "%"), new SetHashGroupByMinReduction());
    }
    // The dispatcher fires the processor corresponding to the closest matching
    // rule and passes the context along
    SemanticDispatcher disp = new DefaultRuleDispatcher(null, opRules, procCtx);
    List<Node> topNodes = new ArrayList<Node>();
    topNodes.addAll(procCtx.parseContext.getTopOps().values());
    SemanticGraphWalker ogw = new ForwardWalker(disp);
    ogw.startWalking(topNodes, null);
}