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

Example 6 with FetchWork

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

the class TaskCompiler method genColumnStatsTask.

/**
   * A helper function to generate a column stats task on top of map-red task. The column stats
   * task fetches from the output of the map-red task, constructs the column stats object and
   * persists it to the metastore.
   *
   * This method generates a plan with a column stats task on top of map-red task and sets up the
   * appropriate metadata to be used during execution.
   *
   * @param analyzeRewrite
   * @param loadTableWork
   * @param loadFileWork
   * @param rootTasks
   * @param outerQueryLimit
   */
@SuppressWarnings("unchecked")
protected void genColumnStatsTask(AnalyzeRewriteContext analyzeRewrite, List<LoadFileDesc> loadFileWork, Set<Task<? extends Serializable>> leafTasks, int outerQueryLimit, int numBitVector) {
    ColumnStatsTask cStatsTask = null;
    ColumnStatsWork cStatsWork = null;
    FetchWork fetch = null;
    String tableName = analyzeRewrite.getTableName();
    List<String> colName = analyzeRewrite.getColName();
    List<String> colType = analyzeRewrite.getColType();
    boolean isTblLevel = analyzeRewrite.isTblLvl();
    String cols = loadFileWork.get(0).getColumns();
    String colTypes = loadFileWork.get(0).getColumnTypes();
    String resFileFormat;
    TableDesc resultTab;
    if (SessionState.get().isHiveServerQuery() && conf.getBoolVar(HiveConf.ConfVars.HIVE_SERVER2_THRIFT_RESULTSET_SERIALIZE_IN_TASKS)) {
        resFileFormat = "SequenceFile";
        resultTab = PlanUtils.getDefaultQueryOutputTableDesc(cols, colTypes, resFileFormat, ThriftJDBCBinarySerDe.class);
    } else {
        resFileFormat = HiveConf.getVar(conf, HiveConf.ConfVars.HIVEQUERYRESULTFILEFORMAT);
        resultTab = PlanUtils.getDefaultQueryOutputTableDesc(cols, colTypes, resFileFormat, LazySimpleSerDe.class);
    }
    fetch = new FetchWork(loadFileWork.get(0).getSourcePath(), resultTab, outerQueryLimit);
    ColumnStatsDesc cStatsDesc = new ColumnStatsDesc(tableName, colName, colType, isTblLevel, numBitVector);
    cStatsWork = new ColumnStatsWork(fetch, cStatsDesc);
    cStatsTask = (ColumnStatsTask) TaskFactory.get(cStatsWork, conf);
    for (Task<? extends Serializable> tsk : leafTasks) {
        tsk.addDependentTask(cStatsTask);
    }
}

Example 7 with FetchWork

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

the class TaskCompiler method compile.

@SuppressWarnings({ "nls", "unchecked" })
public void compile(final ParseContext pCtx, final List<Task<? extends Serializable>> rootTasks, final HashSet<ReadEntity> inputs, final HashSet<WriteEntity> outputs) throws SemanticException {
    Context ctx = pCtx.getContext();
    GlobalLimitCtx globalLimitCtx = pCtx.getGlobalLimitCtx();
    List<Task<MoveWork>> mvTask = new ArrayList<Task<MoveWork>>();
    List<LoadTableDesc> loadTableWork = pCtx.getLoadTableWork();
    List<LoadFileDesc> loadFileWork = pCtx.getLoadFileWork();
    boolean isCStats = pCtx.getQueryProperties().isAnalyzeRewrite();
    int outerQueryLimit = pCtx.getQueryProperties().getOuterQueryLimit();
    if (pCtx.getFetchTask() != null) {
        if (pCtx.getFetchTask().getTblDesc() == null) {
            return;
        }
        pCtx.getFetchTask().getWork().setHiveServerQuery(SessionState.get().isHiveServerQuery());
        TableDesc resultTab = pCtx.getFetchTask().getTblDesc();
        // then either the ThriftFormatter or the DefaultFetchFormatter should be used.
        if (!resultTab.getSerdeClassName().equalsIgnoreCase(ThriftJDBCBinarySerDe.class.getName())) {
            if (SessionState.get().isHiveServerQuery()) {
                conf.set(SerDeUtils.LIST_SINK_OUTPUT_FORMATTER, ThriftFormatter.class.getName());
            } else {
                String formatterName = conf.get(SerDeUtils.LIST_SINK_OUTPUT_FORMATTER);
                if (formatterName == null || formatterName.isEmpty()) {
                    conf.set(SerDeUtils.LIST_SINK_OUTPUT_FORMATTER, DefaultFetchFormatter.class.getName());
                }
            }
        }
        return;
    }
    optimizeOperatorPlan(pCtx, inputs, outputs);
    /*
     * In case of a select, use a fetch task instead of a move task.
     * If the select is from analyze table column rewrite, don't create a fetch task. Instead create
     * a column stats task later.
     */
    if (pCtx.getQueryProperties().isQuery() && !isCStats) {
        if ((!loadTableWork.isEmpty()) || (loadFileWork.size() != 1)) {
            throw new SemanticException(ErrorMsg.INVALID_LOAD_TABLE_FILE_WORK.getMsg());
        }
        LoadFileDesc loadFileDesc = loadFileWork.get(0);
        String cols = loadFileDesc.getColumns();
        String colTypes = loadFileDesc.getColumnTypes();
        String resFileFormat;
        TableDesc resultTab = pCtx.getFetchTableDesc();
        if (resultTab == null) {
            resFileFormat = HiveConf.getVar(conf, HiveConf.ConfVars.HIVEQUERYRESULTFILEFORMAT);
            if (SessionState.get().getIsUsingThriftJDBCBinarySerDe() && (resFileFormat.equalsIgnoreCase("SequenceFile"))) {
                resultTab = PlanUtils.getDefaultQueryOutputTableDesc(cols, colTypes, resFileFormat, ThriftJDBCBinarySerDe.class);
                // Set the fetch formatter to be a no-op for the ListSinkOperator, since we'll
                // read formatted thrift objects from the output SequenceFile written by Tasks.
                conf.set(SerDeUtils.LIST_SINK_OUTPUT_FORMATTER, NoOpFetchFormatter.class.getName());
            } else {
                resultTab = PlanUtils.getDefaultQueryOutputTableDesc(cols, colTypes, resFileFormat, LazySimpleSerDe.class);
            }
        } else {
            if (resultTab.getProperties().getProperty(serdeConstants.SERIALIZATION_LIB).equalsIgnoreCase(ThriftJDBCBinarySerDe.class.getName())) {
                // Set the fetch formatter to be a no-op for the ListSinkOperator, since we'll
                // read formatted thrift objects from the output SequenceFile written by Tasks.
                conf.set(SerDeUtils.LIST_SINK_OUTPUT_FORMATTER, NoOpFetchFormatter.class.getName());
            }
        }
        FetchWork fetch = new FetchWork(loadFileDesc.getSourcePath(), resultTab, outerQueryLimit);
        boolean isHiveServerQuery = SessionState.get().isHiveServerQuery();
        fetch.setHiveServerQuery(isHiveServerQuery);
        fetch.setSource(pCtx.getFetchSource());
        fetch.setSink(pCtx.getFetchSink());
        if (isHiveServerQuery && null != resultTab && resultTab.getSerdeClassName().equalsIgnoreCase(ThriftJDBCBinarySerDe.class.getName()) && HiveConf.getBoolVar(conf, HiveConf.ConfVars.HIVE_SERVER2_THRIFT_RESULTSET_SERIALIZE_IN_TASKS)) {
            fetch.setIsUsingThriftJDBCBinarySerDe(true);
        } else {
            fetch.setIsUsingThriftJDBCBinarySerDe(false);
        }
        pCtx.setFetchTask((FetchTask) TaskFactory.get(fetch, conf));
        // For the FetchTask, the limit optimization requires we fetch all the rows
        // in memory and count how many rows we get. It's not practical if the
        // limit factor is too big
        int fetchLimit = HiveConf.getIntVar(conf, HiveConf.ConfVars.HIVELIMITOPTMAXFETCH);
        if (globalLimitCtx.isEnable() && globalLimitCtx.getGlobalLimit() > fetchLimit) {
            LOG.info("For FetchTask, LIMIT " + globalLimitCtx.getGlobalLimit() + " > " + fetchLimit + ". Doesn't qualify limit optimization.");
            globalLimitCtx.disableOpt();
        }
        if (outerQueryLimit == 0) {
            // Believe it or not, some tools do generate queries with limit 0 and than expect
            // query to run quickly. Lets meet their requirement.
            LOG.info("Limit 0. No query execution needed.");
            return;
        }
    } else if (!isCStats) {
        for (LoadTableDesc ltd : loadTableWork) {
            Task<MoveWork> tsk = TaskFactory.get(new MoveWork(null, null, ltd, null, false), conf);
            mvTask.add(tsk);
            // Check to see if we are stale'ing any indexes and auto-update them if we want
            if (HiveConf.getBoolVar(conf, HiveConf.ConfVars.HIVEINDEXAUTOUPDATE)) {
                IndexUpdater indexUpdater = new IndexUpdater(loadTableWork, inputs, conf);
                try {
                    List<Task<? extends Serializable>> indexUpdateTasks = indexUpdater.generateUpdateTasks();
                    for (Task<? extends Serializable> updateTask : indexUpdateTasks) {
                        tsk.addDependentTask(updateTask);
                    }
                } catch (HiveException e) {
                    console.printInfo("WARNING: could not auto-update stale indexes, which are not in sync");
                }
            }
        }
        boolean oneLoadFile = true;
        for (LoadFileDesc lfd : loadFileWork) {
            if (pCtx.getQueryProperties().isCTAS() || pCtx.getQueryProperties().isMaterializedView()) {
                // should not have more than 1 load file for
                assert (oneLoadFile);
                // CTAS
                // make the movetask's destination directory the table's destination.
                Path location;
                String loc = pCtx.getQueryProperties().isCTAS() ? pCtx.getCreateTable().getLocation() : pCtx.getCreateViewDesc().getLocation();
                if (loc == null) {
                    // get the default location
                    Path targetPath;
                    try {
                        String protoName = null;
                        if (pCtx.getQueryProperties().isCTAS()) {
                            protoName = pCtx.getCreateTable().getTableName();
                        } else if (pCtx.getQueryProperties().isMaterializedView()) {
                            protoName = pCtx.getCreateViewDesc().getViewName();
                        }
                        String[] names = Utilities.getDbTableName(protoName);
                        if (!db.databaseExists(names[0])) {
                            throw new SemanticException("ERROR: The database " + names[0] + " does not exist.");
                        }
                        Warehouse wh = new Warehouse(conf);
                        targetPath = wh.getTablePath(db.getDatabase(names[0]), names[1]);
                    } catch (HiveException e) {
                        throw new SemanticException(e);
                    } catch (MetaException e) {
                        throw new SemanticException(e);
                    }
                    location = targetPath;
                } else {
                    location = new Path(loc);
                }
                lfd.setTargetDir(location);
                oneLoadFile = false;
            }
            mvTask.add(TaskFactory.get(new MoveWork(null, null, null, lfd, false), conf));
        }
    }
    generateTaskTree(rootTasks, pCtx, mvTask, inputs, outputs);
    // For each task, set the key descriptor for the reducer
    for (Task<? extends Serializable> rootTask : rootTasks) {
        GenMapRedUtils.setKeyAndValueDescForTaskTree(rootTask);
    }
    // to be used, please do so
    for (Task<? extends Serializable> rootTask : rootTasks) {
        setInputFormat(rootTask);
    }
    optimizeTaskPlan(rootTasks, pCtx, ctx);
    /*
     * If the query was the result of analyze table column compute statistics rewrite, create
     * a column stats task instead of a fetch task to persist stats to the metastore.
     */
    if (isCStats || !pCtx.getColumnStatsAutoGatherContexts().isEmpty()) {
        Set<Task<? extends Serializable>> leafTasks = new LinkedHashSet<Task<? extends Serializable>>();
        getLeafTasks(rootTasks, leafTasks);
        if (isCStats) {
            genColumnStatsTask(pCtx.getAnalyzeRewrite(), loadFileWork, leafTasks, outerQueryLimit, 0);
        } else {
            for (ColumnStatsAutoGatherContext columnStatsAutoGatherContext : pCtx.getColumnStatsAutoGatherContexts()) {
                if (!columnStatsAutoGatherContext.isInsertInto()) {
                    genColumnStatsTask(columnStatsAutoGatherContext.getAnalyzeRewrite(), columnStatsAutoGatherContext.getLoadFileWork(), leafTasks, outerQueryLimit, 0);
                } else {
                    int numBitVector;
                    try {
                        numBitVector = HiveStatsUtils.getNumBitVectorsForNDVEstimation(conf);
                    } catch (Exception e) {
                        throw new SemanticException(e.getMessage());
                    }
                    genColumnStatsTask(columnStatsAutoGatherContext.getAnalyzeRewrite(), columnStatsAutoGatherContext.getLoadFileWork(), leafTasks, outerQueryLimit, numBitVector);
                }
            }
        }
    }
    decideExecMode(rootTasks, ctx, globalLimitCtx);
    if (pCtx.getQueryProperties().isCTAS() && !pCtx.getCreateTable().isMaterialization()) {
        // generate a DDL task and make it a dependent task of the leaf
        CreateTableDesc crtTblDesc = pCtx.getCreateTable();
        crtTblDesc.validate(conf);
        Task<? extends Serializable> crtTblTask = TaskFactory.get(new DDLWork(inputs, outputs, crtTblDesc), conf);
        patchUpAfterCTASorMaterializedView(rootTasks, outputs, crtTblTask);
    } else if (pCtx.getQueryProperties().isMaterializedView()) {
        // generate a DDL task and make it a dependent task of the leaf
        CreateViewDesc viewDesc = pCtx.getCreateViewDesc();
        Task<? extends Serializable> crtViewTask = TaskFactory.get(new DDLWork(inputs, outputs, viewDesc), conf);
        patchUpAfterCTASorMaterializedView(rootTasks, outputs, crtViewTask);
    }
    if (globalLimitCtx.isEnable() && pCtx.getFetchTask() != null) {
        LOG.info("set least row check for FetchTask: " + globalLimitCtx.getGlobalLimit());
        pCtx.getFetchTask().getWork().setLeastNumRows(globalLimitCtx.getGlobalLimit());
    }
    if (globalLimitCtx.isEnable() && globalLimitCtx.getLastReduceLimitDesc() != null) {
        LOG.info("set least row check for LimitDesc: " + globalLimitCtx.getGlobalLimit());
        globalLimitCtx.getLastReduceLimitDesc().setLeastRows(globalLimitCtx.getGlobalLimit());
        List<ExecDriver> mrTasks = Utilities.getMRTasks(rootTasks);
        for (ExecDriver tsk : mrTasks) {
            tsk.setRetryCmdWhenFail(true);
        }
        List<SparkTask> sparkTasks = Utilities.getSparkTasks(rootTasks);
        for (SparkTask sparkTask : sparkTasks) {
            sparkTask.setRetryCmdWhenFail(true);
        }
    }
    Interner<TableDesc> interner = Interners.newStrongInterner();
    for (Task<? extends Serializable> rootTask : rootTasks) {
        GenMapRedUtils.internTableDesc(rootTask, interner);
        GenMapRedUtils.deriveFinalExplainAttributes(rootTask, pCtx.getConf());
    }
}

Example 8 with FetchWork

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

the class GenMRSkewJoinProcessor method processSkewJoin.

/**
   * Create tasks for processing skew joins. The idea is (HIVE-964) to use
   * separated jobs and map-joins to handle skew joins.
   * <p>
   * <ul>
   * <li>
   * Number of mr jobs to handle skew keys is the number of table minus 1 (we
   * can stream the last table, so big keys in the last table will not be a
   * problem).
   * <li>
   * At runtime in Join, we output big keys in one table into one corresponding
   * directories, and all same keys in other tables into different dirs(one for
   * each table). The directories will look like:
   * <ul>
   * <li>
   * dir-T1-bigkeys(containing big keys in T1), dir-T2-keys(containing keys
   * which is big in T1),dir-T3-keys(containing keys which is big in T1), ...
   * <li>
   * dir-T1-keys(containing keys which is big in T2), dir-T2-bigkeys(containing
   * big keys in T2),dir-T3-keys(containing keys which is big in T2), ...
   * <li>
   * dir-T1-keys(containing keys which is big in T3), dir-T2-keys(containing big
   * keys in T3),dir-T3-bigkeys(containing keys which is big in T3), ... .....
   * </ul>
   * </ul>
   * For each table, we launch one mapjoin job, taking the directory containing
   * big keys in this table and corresponding dirs in other tables as input.
   * (Actally one job for one row in the above.)
   *
   * <p>
   * For more discussions, please check
   * https://issues.apache.org/jira/browse/HIVE-964.
   *
   */
@SuppressWarnings("unchecked")
public static void processSkewJoin(JoinOperator joinOp, Task<? extends Serializable> currTask, ParseContext parseCtx) throws SemanticException {
    // now does not work with outer joins
    if (!GenMRSkewJoinProcessor.skewJoinEnabled(parseCtx.getConf(), joinOp)) {
        return;
    }
    List<Task<? extends Serializable>> children = currTask.getChildTasks();
    Path baseTmpDir = parseCtx.getContext().getMRTmpPath();
    JoinDesc joinDescriptor = joinOp.getConf();
    Map<Byte, List<ExprNodeDesc>> joinValues = joinDescriptor.getExprs();
    int numAliases = joinValues.size();
    Map<Byte, Path> bigKeysDirMap = new HashMap<Byte, Path>();
    Map<Byte, Map<Byte, Path>> smallKeysDirMap = new HashMap<Byte, Map<Byte, Path>>();
    Map<Byte, Path> skewJoinJobResultsDir = new HashMap<Byte, Path>();
    Byte[] tags = joinDescriptor.getTagOrder();
    for (int i = 0; i < numAliases; i++) {
        Byte alias = tags[i];
        bigKeysDirMap.put(alias, getBigKeysDir(baseTmpDir, alias));
        Map<Byte, Path> smallKeysMap = new HashMap<Byte, Path>();
        smallKeysDirMap.put(alias, smallKeysMap);
        for (Byte src2 : tags) {
            if (!src2.equals(alias)) {
                smallKeysMap.put(src2, getSmallKeysDir(baseTmpDir, alias, src2));
            }
        }
        skewJoinJobResultsDir.put(alias, getBigKeysSkewJoinResultDir(baseTmpDir, alias));
    }
    joinDescriptor.setHandleSkewJoin(true);
    joinDescriptor.setBigKeysDirMap(bigKeysDirMap);
    joinDescriptor.setSmallKeysDirMap(smallKeysDirMap);
    joinDescriptor.setSkewKeyDefinition(HiveConf.getIntVar(parseCtx.getConf(), HiveConf.ConfVars.HIVESKEWJOINKEY));
    HashMap<Path, Task<? extends Serializable>> bigKeysDirToTaskMap = new HashMap<Path, Task<? extends Serializable>>();
    List<Serializable> listWorks = new ArrayList<Serializable>();
    List<Task<? extends Serializable>> listTasks = new ArrayList<Task<? extends Serializable>>();
    MapredWork currPlan = (MapredWork) currTask.getWork();
    TableDesc keyTblDesc = (TableDesc) currPlan.getReduceWork().getKeyDesc().clone();
    List<String> joinKeys = Utilities.getColumnNames(keyTblDesc.getProperties());
    List<String> joinKeyTypes = Utilities.getColumnTypes(keyTblDesc.getProperties());
    Map<Byte, TableDesc> tableDescList = new HashMap<Byte, TableDesc>();
    Map<Byte, RowSchema> rowSchemaList = new HashMap<Byte, RowSchema>();
    Map<Byte, List<ExprNodeDesc>> newJoinValues = new HashMap<Byte, List<ExprNodeDesc>>();
    Map<Byte, List<ExprNodeDesc>> newJoinKeys = new HashMap<Byte, List<ExprNodeDesc>>();
    // used for create mapJoinDesc, should be in order
    List<TableDesc> newJoinValueTblDesc = new ArrayList<TableDesc>();
    for (Byte tag : tags) {
        newJoinValueTblDesc.add(null);
    }
    for (int i = 0; i < numAliases; i++) {
        Byte alias = tags[i];
        List<ExprNodeDesc> valueCols = joinValues.get(alias);
        String colNames = "";
        String colTypes = "";
        int columnSize = valueCols.size();
        List<ExprNodeDesc> newValueExpr = new ArrayList<ExprNodeDesc>();
        List<ExprNodeDesc> newKeyExpr = new ArrayList<ExprNodeDesc>();
        ArrayList<ColumnInfo> columnInfos = new ArrayList<ColumnInfo>();
        boolean first = true;
        for (int k = 0; k < columnSize; k++) {
            TypeInfo type = valueCols.get(k).getTypeInfo();
            // any name, it does not matter.
            String newColName = i + "_VALUE_" + k;
            ColumnInfo columnInfo = new ColumnInfo(newColName, type, alias.toString(), false);
            columnInfos.add(columnInfo);
            newValueExpr.add(new ExprNodeColumnDesc(columnInfo));
            if (!first) {
                colNames = colNames + ",";
                colTypes = colTypes + ",";
            }
            first = false;
            colNames = colNames + newColName;
            colTypes = colTypes + valueCols.get(k).getTypeString();
        }
        // we are putting join keys at last part of the spilled table
        for (int k = 0; k < joinKeys.size(); k++) {
            if (!first) {
                colNames = colNames + ",";
                colTypes = colTypes + ",";
            }
            first = false;
            colNames = colNames + joinKeys.get(k);
            colTypes = colTypes + joinKeyTypes.get(k);
            ColumnInfo columnInfo = new ColumnInfo(joinKeys.get(k), TypeInfoFactory.getPrimitiveTypeInfo(joinKeyTypes.get(k)), alias.toString(), false);
            columnInfos.add(columnInfo);
            newKeyExpr.add(new ExprNodeColumnDesc(columnInfo));
        }
        newJoinValues.put(alias, newValueExpr);
        newJoinKeys.put(alias, newKeyExpr);
        tableDescList.put(alias, Utilities.getTableDesc(colNames, colTypes));
        rowSchemaList.put(alias, new RowSchema(columnInfos));
        // construct value table Desc
        String valueColNames = "";
        String valueColTypes = "";
        first = true;
        for (int k = 0; k < columnSize; k++) {
            // any name, it does not matter.
            String newColName = i + "_VALUE_" + k;
            if (!first) {
                valueColNames = valueColNames + ",";
                valueColTypes = valueColTypes + ",";
            }
            valueColNames = valueColNames + newColName;
            valueColTypes = valueColTypes + valueCols.get(k).getTypeString();
            first = false;
        }
        newJoinValueTblDesc.set(Byte.valueOf((byte) i), Utilities.getTableDesc(valueColNames, valueColTypes));
    }
    joinDescriptor.setSkewKeysValuesTables(tableDescList);
    joinDescriptor.setKeyTableDesc(keyTblDesc);
    for (int i = 0; i < numAliases - 1; i++) {
        Byte src = tags[i];
        MapWork newPlan = PlanUtils.getMapRedWork().getMapWork();
        // This code has been only added for testing
        boolean mapperCannotSpanPartns = parseCtx.getConf().getBoolVar(HiveConf.ConfVars.HIVE_MAPPER_CANNOT_SPAN_MULTIPLE_PARTITIONS);
        newPlan.setMapperCannotSpanPartns(mapperCannotSpanPartns);
        MapredWork clonePlan = SerializationUtilities.clonePlan(currPlan);
        Operator<? extends OperatorDesc>[] parentOps = new TableScanOperator[tags.length];
        for (int k = 0; k < tags.length; k++) {
            Operator<? extends OperatorDesc> ts = GenMapRedUtils.createTemporaryTableScanOperator(joinOp.getCompilationOpContext(), rowSchemaList.get((byte) k));
            ((TableScanOperator) ts).setTableDesc(tableDescList.get((byte) k));
            parentOps[k] = ts;
        }
        Operator<? extends OperatorDesc> tblScan_op = parentOps[i];
        ArrayList<String> aliases = new ArrayList<String>();
        String alias = src.toString().intern();
        aliases.add(alias);
        Path bigKeyDirPath = bigKeysDirMap.get(src);
        newPlan.addPathToAlias(bigKeyDirPath, aliases);
        newPlan.getAliasToWork().put(alias, tblScan_op);
        PartitionDesc part = new PartitionDesc(tableDescList.get(src), null);
        newPlan.addPathToPartitionInfo(bigKeyDirPath, part);
        newPlan.getAliasToPartnInfo().put(alias, part);
        Operator<? extends OperatorDesc> reducer = clonePlan.getReduceWork().getReducer();
        assert reducer instanceof JoinOperator;
        JoinOperator cloneJoinOp = (JoinOperator) reducer;
        String dumpFilePrefix = "mapfile" + PlanUtils.getCountForMapJoinDumpFilePrefix();
        MapJoinDesc mapJoinDescriptor = new MapJoinDesc(newJoinKeys, keyTblDesc, newJoinValues, newJoinValueTblDesc, newJoinValueTblDesc, joinDescriptor.getOutputColumnNames(), i, joinDescriptor.getConds(), joinDescriptor.getFilters(), joinDescriptor.getNoOuterJoin(), dumpFilePrefix);
        mapJoinDescriptor.setTagOrder(tags);
        mapJoinDescriptor.setHandleSkewJoin(false);
        mapJoinDescriptor.setNullSafes(joinDescriptor.getNullSafes());
        MapredLocalWork localPlan = new MapredLocalWork(new LinkedHashMap<String, Operator<? extends OperatorDesc>>(), new LinkedHashMap<String, FetchWork>());
        Map<Byte, Path> smallTblDirs = smallKeysDirMap.get(src);
        for (int j = 0; j < numAliases; j++) {
            if (j == i) {
                continue;
            }
            Byte small_alias = tags[j];
            Operator<? extends OperatorDesc> tblScan_op2 = parentOps[j];
            localPlan.getAliasToWork().put(small_alias.toString(), tblScan_op2);
            Path tblDir = smallTblDirs.get(small_alias);
            localPlan.getAliasToFetchWork().put(small_alias.toString(), new FetchWork(tblDir, tableDescList.get(small_alias)));
        }
        newPlan.setMapRedLocalWork(localPlan);
        // construct a map join and set it as the child operator of tblScan_op
        MapJoinOperator mapJoinOp = (MapJoinOperator) OperatorFactory.getAndMakeChild(joinOp.getCompilationOpContext(), mapJoinDescriptor, (RowSchema) null, parentOps);
        // change the children of the original join operator to point to the map
        // join operator
        List<Operator<? extends OperatorDesc>> childOps = cloneJoinOp.getChildOperators();
        for (Operator<? extends OperatorDesc> childOp : childOps) {
            childOp.replaceParent(cloneJoinOp, mapJoinOp);
        }
        mapJoinOp.setChildOperators(childOps);
        HiveConf jc = new HiveConf(parseCtx.getConf(), GenMRSkewJoinProcessor.class);
        newPlan.setNumMapTasks(HiveConf.getIntVar(jc, HiveConf.ConfVars.HIVESKEWJOINMAPJOINNUMMAPTASK));
        newPlan.setMinSplitSize(HiveConf.getLongVar(jc, HiveConf.ConfVars.HIVESKEWJOINMAPJOINMINSPLIT));
        newPlan.setInputformat(HiveInputFormat.class.getName());
        MapredWork w = new MapredWork();
        w.setMapWork(newPlan);
        Task<? extends Serializable> skewJoinMapJoinTask = TaskFactory.get(w, jc);
        skewJoinMapJoinTask.setFetchSource(currTask.isFetchSource());
        bigKeysDirToTaskMap.put(bigKeyDirPath, skewJoinMapJoinTask);
        listWorks.add(skewJoinMapJoinTask.getWork());
        listTasks.add(skewJoinMapJoinTask);
    }
    if (children != null) {
        for (Task<? extends Serializable> tsk : listTasks) {
            for (Task<? extends Serializable> oldChild : children) {
                tsk.addDependentTask(oldChild);
            }
        }
        currTask.setChildTasks(new ArrayList<Task<? extends Serializable>>());
        for (Task<? extends Serializable> oldChild : children) {
            oldChild.getParentTasks().remove(currTask);
        }
        listTasks.addAll(children);
    }
    ConditionalResolverSkewJoinCtx context = new ConditionalResolverSkewJoinCtx(bigKeysDirToTaskMap, children);
    ConditionalWork cndWork = new ConditionalWork(listWorks);
    ConditionalTask cndTsk = (ConditionalTask) TaskFactory.get(cndWork, parseCtx.getConf());
    cndTsk.setListTasks(listTasks);
    cndTsk.setResolver(new ConditionalResolverSkewJoin());
    cndTsk.setResolverCtx(context);
    currTask.setChildTasks(new ArrayList<Task<? extends Serializable>>());
    currTask.addDependentTask(cndTsk);
    return;
}

Example 9 with FetchWork

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

the class SortMergeJoinTaskDispatcher method genSMBJoinWork.

// Convert the work in the SMB plan to a regular join
// Note that the operator tree is not fixed, only the path/alias mappings in the
// plan are fixed. The operator tree will still contain the SMBJoinOperator
private void genSMBJoinWork(MapWork currWork, SMBMapJoinOperator smbJoinOp) {
    // Remove the paths which are not part of aliasToPartitionInfo
    Map<String, PartitionDesc> aliasToPartitionInfo = currWork.getAliasToPartnInfo();
    List<Path> removePaths = new ArrayList<>();
    for (Map.Entry<Path, ArrayList<String>> entry : currWork.getPathToAliases().entrySet()) {
        boolean keepPath = false;
        for (String alias : entry.getValue()) {
            if (aliasToPartitionInfo.containsKey(alias)) {
                keepPath = true;
                break;
            }
        }
        // Remove if the path is not present
        if (!keepPath) {
            removePaths.add(entry.getKey());
        }
    }
    List<String> removeAliases = new ArrayList<String>();
    for (Path removePath : removePaths) {
        removeAliases.addAll(currWork.getPathToAliases().get(removePath));
        currWork.removePathToAlias(removePath);
        currWork.removePathToPartitionInfo(removePath);
    }
    for (String alias : removeAliases) {
        currWork.getAliasToPartnInfo().remove(alias);
        currWork.getAliasToWork().remove(alias);
    }
    // Get the MapredLocalWork
    MapredLocalWork localWork = smbJoinOp.getConf().getLocalWork();
    for (Map.Entry<String, Operator<? extends OperatorDesc>> entry : localWork.getAliasToWork().entrySet()) {
        String alias = entry.getKey();
        Operator<? extends OperatorDesc> op = entry.getValue();
        FetchWork fetchWork = localWork.getAliasToFetchWork().get(alias);
        // Add the entry in mapredwork
        currWork.getAliasToWork().put(alias, op);
        PartitionDesc partitionInfo = currWork.getAliasToPartnInfo().get(alias);
        if (fetchWork.getTblDir() != null) {
            currWork.mergeAliasedInput(alias, fetchWork.getTblDir(), partitionInfo);
        } else {
            for (Path pathDir : fetchWork.getPartDir()) {
                currWork.mergeAliasedInput(alias, pathDir, partitionInfo);
            }
        }
    }
    // Remove the dummy store operator from the tree
    for (Operator<? extends OperatorDesc> parentOp : smbJoinOp.getParentOperators()) {
        if (parentOp instanceof DummyStoreOperator) {
            Operator<? extends OperatorDesc> grandParentOp = parentOp.getParentOperators().get(0);
            smbJoinOp.replaceParent(parentOp, grandParentOp);
            grandParentOp.setChildOperators(parentOp.getChildOperators());
            parentOp.setParentOperators(null);
            parentOp.setParentOperators(null);
        }
    }
}

Example 10 with FetchWork

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

the class TestHCatMultiOutputFormat method getTableData.

/**
   * Method to fetch table data
   *
   * @param table table name
   * @param database database
   * @return list of columns in comma seperated way
   * @throws Exception if any error occurs
   */
private List<String> getTableData(String table, String database) throws Exception {
    QueryState queryState = new QueryState(null);
    HiveConf conf = queryState.getConf();
    conf.addResource("hive-site.xml");
    ArrayList<String> results = new ArrayList<String>();
    ArrayList<String> temp = new ArrayList<String>();
    Hive hive = Hive.get(conf);
    org.apache.hadoop.hive.ql.metadata.Table tbl = hive.getTable(database, table);
    FetchWork work;
    if (!tbl.getPartCols().isEmpty()) {
        List<Partition> partitions = hive.getPartitions(tbl);
        List<PartitionDesc> partDesc = new ArrayList<PartitionDesc>();
        List<Path> partLocs = new ArrayList<Path>();
        TableDesc tableDesc = Utilities.getTableDesc(tbl);
        for (Partition part : partitions) {
            partLocs.add(part.getDataLocation());
            partDesc.add(Utilities.getPartitionDescFromTableDesc(tableDesc, part, true));
        }
        work = new FetchWork(partLocs, partDesc, tableDesc);
        work.setLimit(100);
    } else {
        work = new FetchWork(tbl.getDataLocation(), Utilities.getTableDesc(tbl));
    }
    FetchTask task = new FetchTask();
    task.setWork(work);
    task.initialize(queryState, null, null, new CompilationOpContext());
    task.fetch(temp);
    for (String str : temp) {
        results.add(str.replace("\t", ","));
    }
    return results;
}