Package org.apache.hadoop.hive.ql.plan

Examples of org.apache.hadoop.hive.ql.plan.ReduceSinkDesc

188
189
190
191
192
193
194
195
196
197
198
199
200
    opTaskMap.put(reducer, unionTask);

    plan.setReduceWork(new ReduceWork());
    plan.getReduceWork().setReducer(reducer);
    plan.getReduceWork().setReducer(reducer);
    ReduceSinkDesc desc = op.getConf();

    plan.getReduceWork().setNumReduceTasks(desc.getNumReducers());

    if (needsTagging(plan.getReduceWork())) {
      plan.getReduceWork().setNeedsTagging(true);
    }
View Full Code Here
400
401
402
403
404
405
406
407
408
409
410
411
    // Add the reducer
    ReduceWork rWork = new ReduceWork();
    childPlan.setReduceWork(rWork);
    rWork.setReducer(reducer);
    ReduceSinkDesc desc = cRS.getConf();
    childPlan.getReduceWork().setNumReduceTasks(new Integer(desc.getNumReducers()));

    opProcCtx.getOpTaskMap().put(reducer, childTask);

    splitTasks(cRS, parentTask, childTask, opProcCtx);
  }
View Full Code Here
91
92
93
94
95
96
97
98
99
100
101
    });

    fs = new FileSinkOperator();
    fs.setConf(new FileSinkDesc());
    rs = new ReduceSinkOperator();
    rs.setConf(new ReduceSinkDesc());
    ts = new TableScanOperator();
    ts.setConf(new TableScanDesc());
    ts.getChildOperators().add(rs);
    rs.getParentOperators().add(ts);
    rs.getChildOperators().add(fs);
View Full Code Here
214
215
216
217
218
219
220
221
222
223
224
      Map<String, ExprNodeDesc> colExprMap = Maps.newHashMap();
      for (ColumnInfo ci : valColInfo) {
        newValueCols.add(new ExprNodeColumnDesc(ci));
        colExprMap.put(ci.getInternalName(), newValueCols.get(newValueCols.size() - 1));
      }
      ReduceSinkDesc rsConf = getReduceSinkDesc(partitionPositions, sortPositions, sortOrder,
          newValueCols, bucketColumns, numBuckets, fsParent, fsOp.getConf().getWriteType());

      if (!bucketColumns.isEmpty()) {
        String tableAlias = outRR.getColumnInfos().get(0).getTabAlias();
        ColumnInfo ci = new ColumnInfo(BUCKET_NUMBER_COL_NAME, TypeInfoFactory.stringTypeInfo,
View Full Code Here
448
449
450
451
452
453
454
455
456
457
458
459
460
          outValColNames, 0, "");
      TableDesc valueTable = PlanUtils.getReduceValueTableDesc(valFields);
      List<List<Integer>> distinctColumnIndices = Lists.newArrayList();

      // Number of reducers is set to default (-1)
      ReduceSinkDesc rsConf = new ReduceSinkDesc(newKeyCols, newKeyCols.size(), newValueCols,
          outputKeyCols, distinctColumnIndices, outValColNames, -1, newPartCols, -1, keyTable,
          valueTable, writeType);
      rsConf.setBucketCols(bucketColumns);
      rsConf.setNumBuckets(numBuckets);

      return rsConf;
    }
View Full Code Here
487
488
489
490
491
492
493
494
495
496
497
498
499
    Map<Byte, List<ExprNodeDesc>> keyExprMap  = new HashMap<Byte, List<ExprNodeDesc>>();
    List<Operator<? extends OperatorDesc>> parentOps = joinOp.getParentOperators();
    // get the join keys from parent ReduceSink operators
    for (Operator<? extends OperatorDesc> parentOp : parentOps) {
      ReduceSinkDesc rsconf = ((ReduceSinkOperator)parentOp).getConf();
      Byte tag = (byte) rsconf.getTag();
      List<ExprNodeDesc> keys = rsconf.getKeyCols();
      keyExprMap.put(tag, keys);
    }

    context.setKeyExprMap(keyExprMap);
    // Make a deep copy of the aliases so that they are not changed in the context
View Full Code Here
80
81
82
83
84
85
86
87
88
89
90
91
92
    HashMap<Operator<? extends Serializable>, Task<? extends Serializable>> opTaskMap = opProcCtx.getOpTaskMap();
    Operator<? extends Serializable> currTopOp = opProcCtx.getCurrTopOp();

    opTaskMap.put(reducer, currTask);
    plan.setReducer(reducer);
    reduceSinkDesc desc = (reduceSinkDesc)op.getConf();
   
    plan.setNumReduceTasks(desc.getNumReducers());

    List<Task<? extends Serializable>> rootTasks = opProcCtx.getRootTasks();

    rootTasks.add(currTask);
    if (reducer.getClass() == JoinOperator.class)
View Full Code Here
121
122
123
124
125
126
127
128
129
130
131
132
133
    mapredWork plan = (mapredWork) currTask.getWork();
    HashMap<Operator<? extends Serializable>, Task<? extends Serializable>> opTaskMap = opProcCtx.getOpTaskMap();

    opTaskMap.put(reducer, currTask);
    plan.setReducer(reducer);
    reduceSinkDesc desc = (reduceSinkDesc)op.getConf();
   
    plan.setNumReduceTasks(desc.getNumReducers());

    List<Task<? extends Serializable>> rootTasks = opProcCtx.getRootTasks();

    //    rootTasks.add(currTask);
    if (reducer.getClass() == JoinOperator.class)
View Full Code Here
211
212
213
214
215
216
217
218
219
220
221
222
223
    Task<? extends Serializable> redTask = TaskFactory.get(cplan, parseCtx.getConf());
    Operator<? extends Serializable> reducer = op.getChildOperators().get(0);

    // Add the reducer
    cplan.setReducer(reducer);
    reduceSinkDesc desc = (reduceSinkDesc)op.getConf();
   
    cplan.setNumReduceTasks(new Integer(desc.getNumReducers()));

    HashMap<Operator<? extends Serializable>, Task<? extends Serializable>> opTaskMap = opProcCtx.getOpTaskMap();
    opTaskMap.put(reducer, redTask);
    Task<? extends Serializable> currTask    = opProcCtx.getCurrTask();
View Full Code Here
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
      ReduceSinkOperator op = (ReduceSinkOperator)nd;
      ColumnPrunerProcCtx cppCtx = (ColumnPrunerProcCtx)ctx;
      HashMap<Operator<? extends Serializable>, OpParseContext> opToParseCtxMap =
          cppCtx.getOpToParseCtxMap();
      RowResolver redSinkRR = opToParseCtxMap.get(op).getRR();
      reduceSinkDesc conf = op.getConf();
      List<Operator<? extends Serializable>> childOperators = op.getChildOperators();
      List<Operator<? extends Serializable>> parentOperators = op.getParentOperators();
      List<String> childColLists = new ArrayList<String>();

      for(Operator<? extends Serializable> child: childOperators)
        childColLists = Utilities.mergeUniqElems(childColLists, cppCtx.getPrunedColLists().get(child));

      List<String> colLists = new ArrayList<String>();
      ArrayList<exprNodeDesc> keys = conf.getKeyCols();
      for (exprNodeDesc key : keys)
        colLists = Utilities.mergeUniqElems(colLists, key.getCols());

      if ((childOperators.size() == 1) && (childOperators.get(0) instanceof JoinOperator)) {
        assert parentOperators.size() == 1;
        Operator<? extends Serializable> par = parentOperators.get(0);
        RowResolver parRR = opToParseCtxMap.get(par).getRR();
        RowResolver childRR = opToParseCtxMap.get(childOperators.get(0)).getRR();

        for (String childCol : childColLists) {
          String [] nm = childRR.reverseLookup(childCol);
          ColumnInfo cInfo = redSinkRR.get(nm[0],nm[1]);
          if (cInfo != null) {
            cInfo = parRR.get(nm[0], nm[1]);
            if (!colLists.contains(cInfo.getInternalName()))
              colLists.add(cInfo.getInternalName());
          }
        }
      }
      else {
        // Reduce Sink contains the columns needed - no need to aggregate from children
        ArrayList<exprNodeDesc> vals = conf.getValueCols();
        for (exprNodeDesc val : vals)
          colLists = Utilities.mergeUniqElems(colLists, val.getCols());
      }

      cppCtx.getPrunedColLists().put(op, colLists);
View Full Code Here
TOP

Related Classes of org.apache.hadoop.hive.ql.plan.ReduceSinkDesc

  • org.apache.hadoop.hive.ql.exec.vector.VectorReduceSinkOperator
  • org.apache.hadoop.hive.ql.optimizer.AbstractSMBJoinProc
  • org.apache.hadoop.hive.ql.optimizer.ColumnPrunerProcFactory
  • org.apache.hadoop.hive.ql.optimizer.ColumnPrunerProcFactory$ColumnPrunerReduceSinkProc
  • org.apache.hadoop.hive.ql.optimizer.ConstantPropagateProcFactory$ConstantPropagateReduceSinkProc
  • org.apache.hadoop.hive.ql.optimizer.correlation.ReduceSinkDeDuplication$AbsctractReducerReducerProc
  • org.apache.hadoop.hive.ql.optimizer.GenMapRedUtils
  • org.apache.hadoop.hive.ql.optimizer.GenMRFileSink1
  • org.apache.hadoop.hive.ql.optimizer.lineage.OpProcFactory$JoinLineage
  • org.apache.hadoop.hive.ql.optimizer.lineage.OpProcFactory$ReduceSinkLineage
    • Copyright © 2018 www.massapicom. All rights reserved.
    All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.