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

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

104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
    GenMRUnionCtx uCtxTask = ctx.getUnionTask(union);
    Task<? extends Serializable> uTask = null;

    Operator<? extends Serializable> parent = union.getParentOperators().get(pos);
    mapredWork uPlan = null;

    // union is encountered for the first time
    if (uCtxTask == null) {
      uCtxTask = new GenMRUnionCtx();
      uPlan = GenMapRedUtils.getMapRedWork();
      uTask = TaskFactory.get(uPlan, parseCtx.getConf());
      uCtxTask.setUTask(uTask);
      ctx.setUnionTask(union, uCtxTask);
    }
    else {
      uTask = uCtxTask.getUTask();
      uPlan = (mapredWork)uTask.getWork();
    }

    // If there is a mapjoin at position 'pos'
    if (uPrsCtx.getMapJoinSubq(pos)) {
      MapJoinOperator mjOp = ctx.getCurrMapJoinOp();
      assert mjOp != null;
      GenMRMapJoinCtx mjCtx = ctx.getMapJoinCtx(mjOp);
      assert mjCtx != null;
      mapredWork plan = (mapredWork) currTask.getWork();

      String taskTmpDir = mjCtx.getTaskTmpDir();
      tableDesc tt_desc = mjCtx.getTTDesc();
      assert plan.getPathToAliases().get(taskTmpDir) == null;
      plan.getPathToAliases().put(taskTmpDir, new ArrayList<String>());
      plan.getPathToAliases().get(taskTmpDir).add(taskTmpDir);
      plan.getPathToPartitionInfo().put(taskTmpDir, new partitionDesc(tt_desc, null));
      plan.getAliasToWork().put(taskTmpDir, mjCtx.getRootMapJoinOp());
    }

    tableDesc tt_desc = PlanUtils.getIntermediateFileTableDesc(
          PlanUtils.getFieldSchemasFromRowSchema(parent.getSchema(), "temporarycol"));
View Full Code Here
835
836
837
838
839
840
841
842
843
844
845
      } catch (Exception e) {
        throw new HiveException(e.getMessage(), e);
      }
    }

    mapredWork plan = Utilities.deserializeMapRedWork(pathData, conf);
    ExecDriver ed = new ExecDriver(plan, conf, isSilent);

    int ret = ed.execute();
    if (ret != 0) {
      System.out.println("Job Failed");
View Full Code Here
906
907
908
909
910
911
912
913
    return true;
  }

  @Override
  public boolean hasReduce() {
    mapredWork w = getWork();
    return w.getReducer() != null;
  }
View Full Code Here
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
179
180
181
182
183
184
185
186
187
188
189
190
      outputColumns.add(SemanticAnalyzer.getColumnInternalName(i));
   
    reduceSinkDesc rsDesc = PlanUtils.getReduceSinkDesc(new ArrayList<exprNodeDesc>(), valueCols,
                                                        outputColumns, false, -1, -1, -1);
    ReduceSinkOperator rsOp = (ReduceSinkOperator)OperatorFactory.getAndMakeChild(rsDesc, fsRS, ts_op);
    mapredWork cplan = GenMapRedUtils.getMapRedWork();
    ParseContext parseCtx = ctx.getParseCtx();

    Task<? extends Serializable> mergeTask = TaskFactory.get(cplan, parseCtx.getConf());
    fileSinkDesc fsConf = fsOp.getConf();
   
    // Add the extract operator to get the value fields
    RowResolver out_rwsch = new RowResolver();
    RowResolver interim_rwsch = ctx.getParseCtx().getOpParseCtx().get(fsOp).getRR();
    Integer pos = Integer.valueOf(0);
    for(ColumnInfo colInfo: interim_rwsch.getColumnInfos()) {
      String [] info = interim_rwsch.reverseLookup(colInfo.getInternalName());
      out_rwsch.put(info[0], info[1],
                    new ColumnInfo(pos.toString(), colInfo.getType(), info[0],
                          colInfo.getIsPartitionCol()));
      pos = Integer.valueOf(pos.intValue() + 1);
    }

    Operator extract =
      OperatorFactory.getAndMakeChild(
        new extractDesc(new exprNodeColumnDesc(TypeInfoFactory.stringTypeInfo,
                                Utilities.ReduceField.VALUE.toString(), "", false)),
        new RowSchema(out_rwsch.getColumnInfos()));
   
    tableDesc ts = (tableDesc)fsConf.getTableInfo().clone();
    fsConf.getTableInfo().getProperties().remove(org.apache.hadoop.hive.metastore.api.Constants.META_TABLE_PARTITION_COLUMNS);
    FileSinkOperator newOutput =
      (FileSinkOperator)OperatorFactory.getAndMakeChild(
         new fileSinkDesc(finalName, ts,
                          parseCtx.getConf().getBoolVar(HiveConf.ConfVars.COMPRESSRESULT)),
         fsRS, extract);

    cplan.setReducer(extract);
    ArrayList<String> aliases = new ArrayList<String>();
    aliases.add(fsConf.getDirName());
    cplan.getPathToAliases().put(fsConf.getDirName(), aliases);
    cplan.getAliasToWork().put(fsConf.getDirName(), ts_op);   
    cplan.getPathToPartitionInfo().put(fsConf.getDirName(), new partitionDesc(fsConf.getTableInfo(), null));
    cplan.setNumReduceTasks(-1);
   
    moveWork dummyMv = new moveWork(null, null, null, new loadFileDesc(fsOp.getConf().getDirName(), finalName, true, null, null), false);
    Task<? extends Serializable> dummyMergeTask = TaskFactory.get(dummyMv, ctx.getConf());
    List<Serializable> listWorks = new ArrayList<Serializable>();
    listWorks.add(dummyMv);
View Full Code Here
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
    MapJoinOperator currMapJoinOp = ctx.getCurrMapJoinOp();
   
    if  (currMapJoinOp != null) {
      opTaskMap.put(null, currTask);
      GenMRMapJoinCtx mjCtx = ctx.getMapJoinCtx(currMapJoinOp);
      mapredWork plan = (mapredWork) currTask.getWork();

      String taskTmpDir = mjCtx.getTaskTmpDir();
      tableDesc tt_desc = mjCtx.getTTDesc();
      assert plan.getPathToAliases().get(taskTmpDir) == null;
      plan.getPathToAliases().put(taskTmpDir, new ArrayList<String>());
      plan.getPathToAliases().get(taskTmpDir).add(taskTmpDir);
      plan.getPathToPartitionInfo().put(taskTmpDir, new partitionDesc(tt_desc, null));
      plan.getAliasToWork().put(taskTmpDir, mjCtx.getRootMapJoinOp());
      return dest;
    }
   
    return dest;
  }
View Full Code Here
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
  public static void initPlan(ReduceSinkOperator op, GenMRProcContext opProcCtx) throws SemanticException {
    Operator<? extends Serializable> reducer = op.getChildOperators().get(0);
    Map<Operator<? extends Serializable>, GenMapRedCtx> mapCurrCtx = opProcCtx.getMapCurrCtx();
    GenMapRedCtx mapredCtx = mapCurrCtx.get(op.getParentOperators().get(0));
    Task<? extends Serializable> currTask    = mapredCtx.getCurrTask();
    mapredWork plan = (mapredWork) currTask.getWork();
    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)
      plan.setNeedsTagging(true);

    assert currTopOp != null;
    List<Operator<? extends Serializable>> seenOps = opProcCtx.getSeenOps();
    String currAliasId = opProcCtx.getCurrAliasId();
View Full Code Here
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
    Map<Operator<? extends Serializable>, GenMapRedCtx> mapCurrCtx = opProcCtx.getMapCurrCtx();
    assert (((pos == -1) && (readInputMapJoin)) || (pos != -1));
    int parentPos = (pos == -1) ? 0 : pos;
    GenMapRedCtx mapredCtx = mapCurrCtx.get(op.getParentOperators().get(parentPos));
    Task<? extends Serializable> currTask    = mapredCtx.getCurrTask();
    mapredWork plan = (mapredWork) currTask.getWork();
    HashMap<Operator<? extends Serializable>, Task<? extends Serializable>> opTaskMap = opProcCtx.getOpTaskMap();
    Operator<? extends Serializable> currTopOp = opProcCtx.getCurrTopOp();

    // The mapjoin has already been encountered. Some context must be stored about that
    if (readInputMapJoin) {
      MapJoinOperator currMapJoinOp = opProcCtx.getCurrMapJoinOp();
      assert currMapJoinOp != null;
      boolean local = ((pos == -1) || (pos == ((mapJoinDesc)currMapJoinOp.getConf()).getPosBigTable())) ? false : true;

      if (setReducer) {
        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) {
View Full Code Here
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
  public static void initUnionPlan(ReduceSinkOperator op, GenMRProcContext opProcCtx) throws SemanticException {
    Operator<? extends Serializable> reducer = op.getChildOperators().get(0);
    Map<Operator<? extends Serializable>, GenMapRedCtx> mapCurrCtx = opProcCtx.getMapCurrCtx();
    GenMapRedCtx mapredCtx = mapCurrCtx.get(op.getParentOperators().get(0));
    Task<? extends Serializable> currTask    = mapredCtx.getCurrTask();
    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());

    if (reducer.getClass() == JoinOperator.class)
      plan.setNeedsTagging(true);

    initUnionPlan(opProcCtx, currTask, false);
  }
View Full Code Here
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
  /*
   * It is a idempotent function to add various intermediate files as the source for the
   * union. The plan has already been created.
   */
  public static void initUnionPlan(GenMRProcContext opProcCtx, Task<? extends Serializable> currTask, boolean local) {
    mapredWork plan = (mapredWork) currTask.getWork();
    UnionOperator currUnionOp = opProcCtx.getCurrUnionOp();
    assert currUnionOp != null;
    GenMRUnionCtx uCtx = opProcCtx.getUnionTask(currUnionOp);
    assert uCtx != null;

    List<String>    taskTmpDirLst = uCtx.getTaskTmpDir();
    List<tableDesc> tt_descLst    = uCtx.getTTDesc();
    assert !taskTmpDirLst.isEmpty() && !tt_descLst.isEmpty();
    assert taskTmpDirLst.size() == tt_descLst.size();
    int size = taskTmpDirLst.size();
    assert local == false;

    for (int pos = 0; pos < size; pos++) {
      String taskTmpDir = taskTmpDirLst.get(pos);
      tableDesc tt_desc = tt_descLst.get(pos);
      if (plan.getPathToAliases().get(taskTmpDir) == null) {
        plan.getPathToAliases().put(taskTmpDir, new ArrayList<String>());
        plan.getPathToAliases().get(taskTmpDir).add(taskTmpDir);
        plan.getPathToPartitionInfo().put(taskTmpDir, new partitionDesc(tt_desc, null));
        plan.getAliasToWork().put(taskTmpDir, currUnionOp);
      }
    }
  }
View Full Code Here
270
271
272
273
274
275
276
277
278
279
280
                              GenMRProcContext opProcCtx,
                              int pos, boolean split,
                              boolean readMapJoinData,
                              boolean readUnionData) throws SemanticException {
    Task<? extends Serializable> currTask = task;
    mapredWork plan = (mapredWork) currTask.getWork();
    Operator<? extends Serializable> currTopOp = opProcCtx.getCurrTopOp();
    List<Task<? extends Serializable>> parTasks = null;

    // terminate the old task and make current task dependent on it
    if (split) {
View Full Code Here
TOP

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

  • com.sap.hadoop.windowing.runtime2.mr.MRExecutor
  • com.twitter.ambrose.hive.HiveDAGTransformer
  • org.apache.hadoop.hive.ql.exec.ExecDriver
  • org.apache.hadoop.hive.ql.exec.ExecMapper
  • org.apache.hadoop.hive.ql.exec.ExecReducer
  • org.apache.hadoop.hive.ql.exec.MapRedTask
  • org.apache.hadoop.hive.ql.exec.MoveTask
  • org.apache.hadoop.hive.ql.exec.mr.ExecDriver
  • org.apache.hadoop.hive.ql.exec.mr.MapRedTask
  • org.apache.hadoop.hive.ql.exec.TestOperators
    • 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.