Examples with ReduceWork org.apache.hadoop.hive.ql.plan.ReduceWork used on opensource projects

Example 1 with ReduceWork

use of org.apache.hadoop.hive.ql.plan.ReduceWork in project hive by apache.

the class TestTezTask method setUp.

@SuppressWarnings("unchecked")
@Before
public void setUp() throws Exception {
    utils = mock(DagUtils.class);
    fs = mock(FileSystem.class);
    path = mock(Path.class);
    when(path.getFileSystem(any(Configuration.class))).thenReturn(fs);
    when(utils.getTezDir(any(Path.class))).thenReturn(path);
    when(utils.createVertex(any(JobConf.class), any(BaseWork.class), any(Path.class), any(LocalResource.class), any(List.class), any(FileSystem.class), any(Context.class), anyBoolean(), any(TezWork.class), any(VertexType.class))).thenAnswer(new Answer<Vertex>() {

        @Override
        public Vertex answer(InvocationOnMock invocation) throws Throwable {
            Object[] args = invocation.getArguments();
            return Vertex.create(((BaseWork) args[1]).getName(), mock(ProcessorDescriptor.class), 0, mock(Resource.class));
        }
    });
    when(utils.createEdge(any(JobConf.class), any(Vertex.class), any(Vertex.class), any(TezEdgeProperty.class), any(VertexType.class))).thenAnswer(new Answer<Edge>() {

        @Override
        public Edge answer(InvocationOnMock invocation) throws Throwable {
            Object[] args = invocation.getArguments();
            return Edge.create((Vertex) args[1], (Vertex) args[2], mock(EdgeProperty.class));
        }
    });
    work = new TezWork("", null);
    mws = new MapWork[] { new MapWork(), new MapWork() };
    rws = new ReduceWork[] { new ReduceWork(), new ReduceWork() };
    work.addAll(mws);
    work.addAll(rws);
    int i = 0;
    for (BaseWork w : work.getAllWork()) {
        w.setName("Work " + (++i));
    }
    op = mock(Operator.class);
    LinkedHashMap<String, Operator<? extends OperatorDesc>> map = new LinkedHashMap<String, Operator<? extends OperatorDesc>>();
    map.put("foo", op);
    mws[0].setAliasToWork(map);
    mws[1].setAliasToWork(map);
    LinkedHashMap<Path, ArrayList<String>> pathMap = new LinkedHashMap<>();
    ArrayList<String> aliasList = new ArrayList<String>();
    aliasList.add("foo");
    pathMap.put(new Path("foo"), aliasList);
    mws[0].setPathToAliases(pathMap);
    mws[1].setPathToAliases(pathMap);
    rws[0].setReducer(op);
    rws[1].setReducer(op);
    TezEdgeProperty edgeProp = new TezEdgeProperty(EdgeType.SIMPLE_EDGE);
    work.connect(mws[0], rws[0], edgeProp);
    work.connect(mws[1], rws[0], edgeProp);
    work.connect(rws[0], rws[1], edgeProp);
    task = new TezTask(utils);
    task.setWork(work);
    task.setConsole(mock(LogHelper.class));
    QueryPlan mockQueryPlan = mock(QueryPlan.class);
    doReturn(UUID.randomUUID().toString()).when(mockQueryPlan).getQueryId();
    task.setQueryPlan(mockQueryPlan);
    conf = new JobConf();
    appLr = mock(LocalResource.class);
    HiveConf hiveConf = new HiveConf();
    hiveConf.setVar(HiveConf.ConfVars.HIVE_AUTHORIZATION_MANAGER, "org.apache.hadoop.hive.ql.security.authorization.plugin.sqlstd.SQLStdHiveAuthorizerFactory");
    SessionState.start(hiveConf);
    session = mock(TezClient.class);
    sessionState = mock(TezSessionState.class);
    when(sessionState.getSession()).thenReturn(session);
    when(session.submitDAG(any(DAG.class))).thenThrow(new SessionNotRunning("")).thenReturn(mock(DAGClient.class));
}

Example 2 with ReduceWork

use of org.apache.hadoop.hive.ql.plan.ReduceWork in project hive by apache.

the class TestExecDriver method populateMapRedPlan1.

@SuppressWarnings("unchecked")
private void populateMapRedPlan1(Table src) throws SemanticException {
    ArrayList<String> outputColumns = new ArrayList<String>();
    for (int i = 0; i < 2; i++) {
        outputColumns.add("_col" + i);
    }
    // map-side work
    Operator<ReduceSinkDesc> op1 = OperatorFactory.get(ctx, PlanUtils.getReduceSinkDesc(Utilities.makeList(getStringColumn("key")), Utilities.makeList(getStringColumn("value")), outputColumns, true, -1, 1, -1, AcidUtils.Operation.NOT_ACID));
    addMapWork(mr, src, "a", op1);
    ReduceWork rWork = new ReduceWork();
    rWork.setNumReduceTasks(Integer.valueOf(1));
    rWork.setKeyDesc(op1.getConf().getKeySerializeInfo());
    rWork.getTagToValueDesc().add(op1.getConf().getValueSerializeInfo());
    mr.setReduceWork(rWork);
    // reduce side work
    Operator<FileSinkDesc> op3 = OperatorFactory.get(ctx, new FileSinkDesc(new Path(tmpdir + File.separator + "mapredplan1.out"), Utilities.defaultTd, false));
    List<ExprNodeDesc> cols = new ArrayList<ExprNodeDesc>();
    cols.add(getStringColumn(Utilities.ReduceField.VALUE.toString() + "." + outputColumns.get(1)));
    List<String> colNames = new ArrayList<String>();
    colNames.add(HiveConf.getColumnInternalName(2));
    Operator<SelectDesc> op2 = OperatorFactory.get(new SelectDesc(cols, colNames), op3);
    rWork.setReducer(op2);
}

Example 3 with ReduceWork

use of org.apache.hadoop.hive.ql.plan.ReduceWork in project hive by apache.

the class TestExecDriver method populateMapRedPlan5.

@SuppressWarnings("unchecked")
private void populateMapRedPlan5(Table src) throws SemanticException {
    // map-side work
    ArrayList<String> outputColumns = new ArrayList<String>();
    for (int i = 0; i < 2; i++) {
        outputColumns.add("_col" + i);
    }
    Operator<ReduceSinkDesc> op0 = OperatorFactory.get(ctx, PlanUtils.getReduceSinkDesc(Utilities.makeList(getStringColumn("0")), Utilities.makeList(getStringColumn("0"), getStringColumn("1")), outputColumns, false, -1, 1, -1, AcidUtils.Operation.NOT_ACID));
    Operator<SelectDesc> op4 = OperatorFactory.get(new SelectDesc(Utilities.makeList(getStringColumn("key"), getStringColumn("value")), outputColumns), op0);
    addMapWork(mr, src, "a", op4);
    ReduceWork rWork = new ReduceWork();
    mr.setReduceWork(rWork);
    rWork.setNumReduceTasks(Integer.valueOf(1));
    rWork.setKeyDesc(op0.getConf().getKeySerializeInfo());
    rWork.getTagToValueDesc().add(op0.getConf().getValueSerializeInfo());
    // reduce side work
    Operator<FileSinkDesc> op3 = OperatorFactory.get(ctx, new FileSinkDesc(new Path(tmpdir + File.separator + "mapredplan5.out"), Utilities.defaultTd, false));
    List<ExprNodeDesc> cols = new ArrayList<ExprNodeDesc>();
    cols.add(getStringColumn(Utilities.ReduceField.KEY + ".reducesinkkey" + 0));
    cols.add(getStringColumn(Utilities.ReduceField.VALUE.toString() + "." + outputColumns.get(1)));
    Operator<SelectDesc> op2 = OperatorFactory.get(new SelectDesc(cols, outputColumns), op3);
    rWork.setReducer(op2);
}

Example 4 with ReduceWork

use of org.apache.hadoop.hive.ql.plan.ReduceWork in project hive by apache.

the class TestExecDriver method populateMapRedPlan2.

@SuppressWarnings("unchecked")
private void populateMapRedPlan2(Table src) throws Exception {
    ArrayList<String> outputColumns = new ArrayList<String>();
    for (int i = 0; i < 2; i++) {
        outputColumns.add("_col" + i);
    }
    // map-side work
    Operator<ReduceSinkDesc> op1 = OperatorFactory.get(ctx, PlanUtils.getReduceSinkDesc(Utilities.makeList(getStringColumn("key")), Utilities.makeList(getStringColumn("key"), getStringColumn("value")), outputColumns, false, -1, 1, -1, AcidUtils.Operation.NOT_ACID));
    addMapWork(mr, src, "a", op1);
    ReduceWork rWork = new ReduceWork();
    rWork.setNumReduceTasks(Integer.valueOf(1));
    rWork.setKeyDesc(op1.getConf().getKeySerializeInfo());
    rWork.getTagToValueDesc().add(op1.getConf().getValueSerializeInfo());
    mr.setReduceWork(rWork);
    // reduce side work
    Operator<FileSinkDesc> op4 = OperatorFactory.get(ctx, new FileSinkDesc(new Path(tmpdir + File.separator + "mapredplan2.out"), Utilities.defaultTd, false));
    Operator<FilterDesc> op3 = OperatorFactory.get(getTestFilterDesc("0"), op4);
    List<ExprNodeDesc> cols = new ArrayList<ExprNodeDesc>();
    cols.add(getStringColumn(Utilities.ReduceField.KEY + ".reducesinkkey" + 0));
    cols.add(getStringColumn(Utilities.ReduceField.VALUE.toString() + "." + outputColumns.get(1)));
    Operator<SelectDesc> op2 = OperatorFactory.get(new SelectDesc(cols, outputColumns), op3);
    rWork.setReducer(op2);
}

Example 5 with ReduceWork

use of org.apache.hadoop.hive.ql.plan.ReduceWork in project hive by apache.

the class TestExecDriver method populateMapRedPlan4.

@SuppressWarnings("unchecked")
private void populateMapRedPlan4(Table src) throws SemanticException {
    // map-side work
    ArrayList<String> outputColumns = new ArrayList<String>();
    for (int i = 0; i < 2; i++) {
        outputColumns.add("_col" + i);
    }
    Operator<ReduceSinkDesc> op1 = OperatorFactory.get(ctx, PlanUtils.getReduceSinkDesc(Utilities.makeList(getStringColumn("tkey")), Utilities.makeList(getStringColumn("tkey"), getStringColumn("tvalue")), outputColumns, false, -1, 1, -1, AcidUtils.Operation.NOT_ACID));
    Operator<ScriptDesc> op0 = OperatorFactory.get(new ScriptDesc("cat", PlanUtils.getDefaultTableDesc("" + Utilities.tabCode, "key,value"), TextRecordWriter.class, PlanUtils.getDefaultTableDesc("" + Utilities.tabCode, "tkey,tvalue"), TextRecordReader.class, TextRecordReader.class, PlanUtils.getDefaultTableDesc("" + Utilities.tabCode, "key")), op1);
    Operator<SelectDesc> op4 = OperatorFactory.get(new SelectDesc(Utilities.makeList(getStringColumn("key"), getStringColumn("value")), outputColumns), op0);
    addMapWork(mr, src, "a", op4);
    ReduceWork rWork = new ReduceWork();
    rWork.setKeyDesc(op1.getConf().getKeySerializeInfo());
    rWork.getTagToValueDesc().add(op1.getConf().getValueSerializeInfo());
    rWork.setNumReduceTasks(Integer.valueOf(1));
    mr.setReduceWork(rWork);
    // reduce side work
    Operator<FileSinkDesc> op3 = OperatorFactory.get(ctx, new FileSinkDesc(new Path(tmpdir + File.separator + "mapredplan4.out"), Utilities.defaultTd, false));
    List<ExprNodeDesc> cols = new ArrayList<ExprNodeDesc>();
    cols.add(getStringColumn(Utilities.ReduceField.KEY + ".reducesinkkey" + 0));
    cols.add(getStringColumn(Utilities.ReduceField.VALUE.toString() + "." + outputColumns.get(1)));
    Operator<SelectDesc> op2 = OperatorFactory.get(new SelectDesc(cols, outputColumns), op3);
    rWork.setReducer(op2);
}