Examples of reduceSinkDesc

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

183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
        plan.setReducer(reducer);
        opTaskMap.put(reducer, currTask);
        if (reducer.getClass() == JoinOperator.class) {
          plan.setNeedsTagging(true);
        }
        ReduceSinkDesc desc = (ReduceSinkDesc) op.getConf();
        plan.setNumReduceTasks(desc.getNumReducers());
      } else {
        opTaskMap.put(op, currTask);
      }

      if (!readInputUnion) {
        GenMRMapJoinCtx mjCtx = opProcCtx.getMapJoinCtx(currMapJoinOp);
        String taskTmpDir;
        TableDesc tt_desc;
        Operator<? extends OperatorDesc> rootOp;

        if (mjCtx.getOldMapJoin() == null || setReducer) {
          taskTmpDir = mjCtx.getTaskTmpDir();
          tt_desc = mjCtx.getTTDesc();
          rootOp = mjCtx.getRootMapJoinOp();
        } else {
          GenMRMapJoinCtx oldMjCtx = opProcCtx.getMapJoinCtx(mjCtx
              .getOldMapJoin());
          taskTmpDir = oldMjCtx.getTaskTmpDir();
          tt_desc = oldMjCtx.getTTDesc();
          rootOp = oldMjCtx.getRootMapJoinOp();
        }

        setTaskPlan(taskTmpDir, taskTmpDir, rootOp, plan, local, tt_desc);
        setupBucketMapJoinInfo(plan, currMapJoinOp, createLocalPlan);
      } else {
        initUnionPlan(opProcCtx, currTask, false);
      }

      opProcCtx.setCurrMapJoinOp(null);
    } else {
      MapJoinDesc desc = (MapJoinDesc) op.getConf();

      // The map is overloaded to keep track of mapjoins also
      opTaskMap.put(op, currTask);

      List<Task<? extends Serializable>> rootTasks = opProcCtx.getRootTasks();
      if (!rootTasks.contains(currTask)) {
        rootTasks.add(currTask);
      }

      assert currTopOp != null;
      List<Operator<? extends OperatorDesc>> seenOps = opProcCtx.getSeenOps();
      String currAliasId = opProcCtx.getCurrAliasId();

      seenOps.add(currTopOp);
      boolean local = (pos == desc.getPosBigTable()) ? false : true;
      setTaskPlan(currAliasId, currTopOp, plan, local, opProcCtx);
      setupBucketMapJoinInfo(plan, (AbstractMapJoinOperator<? extends MapJoinDesc>)op, createLocalPlan);
    }

    opProcCtx.setCurrTask(currTask);
View Full Code Here

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

310
311
312
313
314
315
316
317
318
319
320
321
322
    HashMap<Operator<? extends OperatorDesc>, Task<? extends Serializable>> opTaskMap =
      opProcCtx.getOpTaskMap();

    opTaskMap.put(reducer, unionTask);
    plan.setReducer(reducer);
    ReduceSinkDesc desc = op.getConf();

    plan.setNumReduceTasks(desc.getNumReducers());

    if (reducer.getClass() == JoinOperator.class) {
      plan.setNeedsTagging(true);
    }
View Full Code Here

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

562
563
564
565
566
567
568
569
570
571
572
573
574
        .getConf());
    Operator<? extends OperatorDesc> reducer = op.getChildOperators().get(0);

    // Add the reducer
    cplan.setReducer(reducer);
    ReduceSinkDesc desc = op.getConf();

    cplan.setNumReduceTasks(new Integer(desc.getNumReducers()));

    HashMap<Operator<? extends OperatorDesc>, Task<? extends Serializable>> opTaskMap =
      opProcCtx.getOpTaskMap();
    opTaskMap.put(reducer, redTask);
    Task<? extends Serializable> currTask = opProcCtx.getCurrTask();
View Full Code Here

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

289
290
291
292
293
294
295
296
297
298
299
    ArrayList<String> outputColumns = new ArrayList<String>();
    for (int i = 0; i < valueCols.size(); i++) {
      outputColumns.add(SemanticAnalyzer.getColumnInternalName(i));
    }

    ReduceSinkDesc rsDesc = PlanUtils.getReduceSinkDesc(
        new ArrayList<ExprNodeDesc>(), valueCols, outputColumns, false, -1, -1,
        -1);
    OperatorFactory.getAndMakeChild(rsDesc, inputRS, tsMerge);
    ParseContext parseCtx = ctx.getParseCtx();
    FileSinkDesc fsConf = fsOp.getConf();
View Full Code Here

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

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

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

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

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

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

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

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

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

88
89
90
91
92
93
94
95
96
97
98
99
100
    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

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

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
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
        Operator<? extends Serializable> reducer = op.getChildOperators().get(0);
        plan.setReducer(reducer);
        opTaskMap.put(reducer, currTask);
        if (reducer.getClass() == JoinOperator.class)
          plan.setNeedsTagging(true);
        reduceSinkDesc desc = (reduceSinkDesc)op.getConf();
        plan.setNumReduceTasks(desc.getNumReducers());
      }
      else
        opTaskMap.put(op, currTask);

      if (!readInputUnion) {
        GenMRMapJoinCtx mjCtx = opProcCtx.getMapJoinCtx(currMapJoinOp);
        String taskTmpDir;
        tableDesc tt_desc;
        Operator<? extends Serializable> rootOp;

        if (mjCtx.getOldMapJoin() == null) {
          taskTmpDir = mjCtx.getTaskTmpDir();
          tt_desc = mjCtx.getTTDesc();
          rootOp = mjCtx.getRootMapJoinOp();
        }
        else {
          GenMRMapJoinCtx oldMjCtx = opProcCtx.getMapJoinCtx(mjCtx.getOldMapJoin());
          taskTmpDir = oldMjCtx.getTaskTmpDir();
          tt_desc = oldMjCtx.getTTDesc();
          rootOp = oldMjCtx.getRootMapJoinOp();
        }

        setTaskPlan(taskTmpDir, taskTmpDir, rootOp, plan, local, tt_desc);
      }
      else {
        initUnionPlan(opProcCtx, currTask, false);
      }

      opProcCtx.setCurrMapJoinOp(null);
    }
    else {
      mapJoinDesc desc = (mapJoinDesc)op.getConf();

      // The map is overloaded to keep track of mapjoins also
      opTaskMap.put(op, currTask);

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

      assert currTopOp != null;
      List<Operator<? extends Serializable>> seenOps = opProcCtx.getSeenOps();
      String currAliasId = opProcCtx.getCurrAliasId();

      seenOps.add(currTopOp);
      boolean local = (pos == desc.getPosBigTable()) ? false : true;
      setTaskPlan(currAliasId, currTopOp, plan, local, opProcCtx);
    }

    opProcCtx.setCurrTask(currTask);
    opProcCtx.setCurrTopOp(null);
View Full Code Here
TOP
Copyright © 2018 www.massapi.com. 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.