Task (Hive 1.2.2 API)

java.lang.Object
- org.apache.hadoop.hive.ql.exec.Task<T>

All Implemented Interfaces:

Serializable, Node

Direct Known Subclasses:

ColumnStatsTask, ColumnStatsUpdateTask, ColumnTruncateTask, ConditionalTask, CopyTask, DDLTask, DependencyCollectionTask, ExecDriver, ExplainSQRewriteTask, ExplainTask, FetchTask, FunctionTask, IndexMetadataChangeTask, MapredLocalTask, MergeFileTask, MoveTask, PartialScanTask, SparkTask, StatsNoJobTask, StatsTask, TezTask
```
public abstract class Task<T extends Serializable>
extends Object
implements Serializable, Node
```
Task implementation.

See Also:
Serialized Form

Nested Class Summary

Nested Classes
Modifier and Type Class and Description

static class Task.FeedType

Nested Classes
Modifier and Type	Class and Description
`static class`	`Task.FeedType`

Field Summary

Fields
Modifier and Type	Field and Description
`static int`	`BACKUP_COMMON_JOIN`
`protected List<Task<? extends Serializable>>`	`backupChildrenTasks`
`protected Task<? extends Serializable>`	`backupTask`
`protected List<Task<? extends Serializable>>`	`childTasks`
`protected boolean`	`clonedConf`
`static int`	`COMMON_JOIN`
`protected HiveConf`	`conf`
`protected SessionState.LogHelper`	`console`
`static int`	`CONVERTED_MAPJOIN`
`static int`	`CONVERTED_MAPJOIN_LOCAL`
`static int`	`CONVERTED_SORTMERGEJOIN`
`protected Hive`	`db`
`protected DriverContext`	`driverContext`
`protected List<Task<? extends Serializable>>`	`feedSubscribers`
`static int`	`HINTED_MAPJOIN`
`static int`	`HINTED_MAPJOIN_LOCAL`
`protected String`	`id`
`protected boolean`	`initialized`
`protected boolean`	`isdone`
`protected String`	`jobID`
`protected static org.apache.commons.logging.Log`	`LOG`
`static int`	`MAPJOIN_ONLY_NOBACKUP`
`static int`	`NO_TAG`
`protected List<Task<? extends Serializable>>`	`parentTasks`
`protected QueryPlan`	`queryPlan`
`protected boolean`	`queued`
`protected boolean`	`rootTask`
`protected boolean`	`started`
`HashMap<String,Long>`	`taskCounters`
`TaskHandle`	`taskHandle`
`protected int`	`taskTag`
`protected T`	`work`

Constructor Summary

Constructors
Constructor and Description

Task()

Constructors
Constructor and Description
`Task()`

Method Summary

Methods
Modifier and Type	Method and Description
`boolean`	`addDependentTask(Task<? extends Serializable> dependent)` Add a dependent task on the current task.
`protected void`	`cloneConf()`
`boolean`	`done()`
`boolean`	`equals(Object obj)`
`protected abstract int`	`execute(DriverContext driverContext)` This method is overridden in each Task.
`int`	`executeTask()` This method is called in the Driver on every task.
`Task<? extends Serializable>`	`getAndInitBackupTask()`
`List<Task<? extends Serializable>>`	`getBackupChildrenTasks()`
`Task<? extends Serializable>`	`getBackupTask()`
`List<? extends Node>`	`getChildren()` Gets the vector of children nodes.
`List<Task<? extends Serializable>>`	`getChildTasks()`
`HashMap<String,Long>`	`getCounters()`
`List<Task<? extends Serializable>>`	`getDependentTasks()` The default dependent tasks are just child tasks, but different types could implement their own (e.g.
`List<Task<? extends Serializable>>`	`getFeedSubscribers()`
`String`	`getId()`
`boolean`	`getInitialized()`
`String`	`getJobID()`
`Collection<MapWork>`	`getMapWork()`
`int`	`getNumChild()`
`int`	`getNumParent()`
`List<Task<? extends Serializable>>`	`getParentTasks()`
`QueryPlan`	`getQueryPlan()`
`boolean`	`getQueued()`
`Operator<? extends OperatorDesc>`	`getReducer(MapWork work)`
`TaskHandle`	`getTaskHandle()`
`int`	`getTaskTag()`
`Collection<Operator<? extends OperatorDesc>>`	`getTopOperators()`
`abstract StageType`	`getType()` Should be overridden to return the type of the specific task among the types in StageType.
`T`	`getWork()`
`int`	`hashCode()`
`boolean`	`hasReduce()`
`boolean`	`ifRetryCmdWhenFail()`
`void`	`initialize(HiveConf conf, QueryPlan queryPlan, DriverContext driverContext)`
`boolean`	`isFetchSource()`
`boolean`	`isLocalMode()`
`boolean`	`isMapRedLocalTask()`
`boolean`	`isMapRedTask()`
`boolean`	`isRootTask()`
`boolean`	`isRunnable()`
`protected void`	`pushFeed(Task.FeedType feedType, Object feedValue)`
`protected void`	`receiveFeed(Task.FeedType feedType, Object feedValue)`
`void`	`removeDependentTask(Task<? extends Serializable> dependent)` Remove the dependent task.
`void`	`removeFromChildrenTasks()`
`boolean`	`requireLock()`
`void`	`setBackupChildrenTasks(List<Task<? extends Serializable>> backupChildrenTasks)`
`void`	`setBackupTask(Task<? extends Serializable> backupTask)`
`void`	`setChildTasks(List<Task<? extends Serializable>> childTasks)`
`void`	`setConsole(SessionState.LogHelper console)`
`void`	`setDone()`
`protected void`	`setException(Throwable ex)`
`void`	`setFeedSubscribers(List<Task<? extends Serializable>> s)`
`void`	`setFetchSource(boolean fetchSource)`
`void`	`setId(String id)`
`void`	`setInitialized()`
`void`	`setLocalMode(boolean isLocalMode)`
`void`	`setParentTasks(List<Task<? extends Serializable>> parentTasks)`
`void`	`setQueryPlan(QueryPlan queryPlan)`
`void`	`setQueued()`
`void`	`setRetryCmdWhenFail(boolean retryCmdWhenFail)`
`void`	`setRootTask(boolean rootTask)`
`void`	`setStarted()`
`void`	`setTaskTag(int taskTag)`
`void`	`setWork(T work)`
`void`	`shutdown()`
`boolean`	`started()`
`void`	`subscribeFeed(Task<? extends Serializable> publisher)` Subscribe the feed of publisher.
`String`	`toString()`

Methods inherited from class java.lang.Object
clone, finalize, getClass, notify, notifyAll, wait, wait, wait

Methods inherited from interface org.apache.hadoop.hive.ql.lib.Node
getName

Field Detail

taskCounters

public transient HashMap<String,Long> taskCounters

taskHandle
```
public transient TaskHandle taskHandle
```

started
```
protected transient boolean started
```

initialized

protected transient boolean initialized

isdone
```
protected transient boolean isdone
```

queued
```
protected transient boolean queued
```

conf
```
protected transient HiveConf conf
```

db
```
protected transient Hive db
```

console

protected transient SessionState.LogHelper console

queryPlan

protected transient QueryPlan queryPlan

driverContext

protected transient DriverContext driverContext

clonedConf
```
protected transient boolean clonedConf
```

jobID
```
protected transient String jobID
```

backupTask

protected Task<? extends Serializable> backupTask

backupChildrenTasks

protected List<Task<? extends Serializable>> backupChildrenTasks

LOG

protected static transient org.apache.commons.logging.Log LOG

taskTag
```
protected int taskTag
```

NO_TAG
```
public static final int NO_TAG
```
See Also:
Constant Field Values

COMMON_JOIN
```
public static final int COMMON_JOIN
```
See Also:
Constant Field Values

HINTED_MAPJOIN
```
public static final int HINTED_MAPJOIN
```
See Also:
Constant Field Values

HINTED_MAPJOIN_LOCAL
```
public static final int HINTED_MAPJOIN_LOCAL
```
See Also:
Constant Field Values

CONVERTED_MAPJOIN
```
public static final int CONVERTED_MAPJOIN
```
See Also:
Constant Field Values

CONVERTED_MAPJOIN_LOCAL
```
public static final int CONVERTED_MAPJOIN_LOCAL
```
See Also:
Constant Field Values

BACKUP_COMMON_JOIN
```
public static final int BACKUP_COMMON_JOIN
```
See Also:
Constant Field Values

MAPJOIN_ONLY_NOBACKUP
```
public static final int MAPJOIN_ONLY_NOBACKUP
```
See Also:
Constant Field Values

CONVERTED_SORTMERGEJOIN
```
public static final int CONVERTED_SORTMERGEJOIN
```
See Also:
Constant Field Values

feedSubscribers

protected transient List<Task<? extends Serializable>> feedSubscribers

id
```
protected String id
```

work
```
protected T extends Serializable work
```

rootTask
```
protected boolean rootTask
```

childTasks

protected List<Task<? extends Serializable>> childTasks

parentTasks

protected List<Task<? extends Serializable>> parentTasks

Constructor Detail
- Task
```
public Task()
```

Method Detail

getTaskHandle
```
public TaskHandle getTaskHandle()
```

initialize

public void initialize(HiveConf conf,
              QueryPlan queryPlan,
              DriverContext driverContext)

executeTask
```
public int executeTask()
```
This method is called in the Driver on every task. It updates counters and calls execute(), which is overridden in each task

Returns:
return value of execute()

execute
```
protected abstract int execute(DriverContext driverContext)
```
This method is overridden in each Task. TODO execute should return a TaskHandle.

Returns:
status of executing the task

isRootTask
```
public boolean isRootTask()
```

setRootTask

public void setRootTask(boolean rootTask)

setChildTasks

public void setChildTasks(List<Task<? extends Serializable>> childTasks)

getChildren
```
public List<? extends Node> getChildren()
```
Description copied from interface: Node

Gets the vector of children nodes. This is used in the graph walker algorithms.

Specified by:

getChildren in interface Node

Returns:
List

getChildTasks

public List<Task<? extends Serializable>> getChildTasks()

getNumChild
```
public int getNumChild()
```

setParentTasks

public void setParentTasks(List<Task<? extends Serializable>> parentTasks)

getParentTasks

public List<Task<? extends Serializable>> getParentTasks()

getNumParent
```
public int getNumParent()
```

getBackupTask

public Task<? extends Serializable> getBackupTask()

setBackupTask

public void setBackupTask(Task<? extends Serializable> backupTask)

getBackupChildrenTasks

public List<Task<? extends Serializable>> getBackupChildrenTasks()

setBackupChildrenTasks

public void setBackupChildrenTasks(List<Task<? extends Serializable>> backupChildrenTasks)

getAndInitBackupTask

public Task<? extends Serializable> getAndInitBackupTask()

removeFromChildrenTasks
```
public void removeFromChildrenTasks()
```

getDependentTasks
```
public List<Task<? extends Serializable>> getDependentTasks()
```
The default dependent tasks are just child tasks, but different types could implement their own (e.g. ConditionalTask will use the listTasks as dependents).

Returns:
a list of tasks that are dependent on this task.

addDependentTask
```
public boolean addDependentTask(Task<? extends Serializable> dependent)
```
Add a dependent task on the current task. Return if the dependency already existed or is this a new one

Returns:
true if the task got added false if it already existed

removeDependentTask
```
public void removeDependentTask(Task<? extends Serializable> dependent)
```
Remove the dependent task.

Parameters:
dependent - the task to remove

setStarted
```
public void setStarted()
```

started
```
public boolean started()
```

done
```
public boolean done()
```

setDone
```
public void setDone()
```

setQueued
```
public void setQueued()
```

getQueued
```
public boolean getQueued()
```

setInitialized
```
public void setInitialized()
```

getInitialized
```
public boolean getInitialized()
```

isRunnable
```
public boolean isRunnable()
```

setWork
```
public void setWork(T work)
```

getWork
```
public T getWork()
```

getMapWork

public Collection<MapWork> getMapWork()

setId
```
public void setId(String id)
```

getId
```
public String getId()
```

isMapRedTask
```
public boolean isMapRedTask()
```

isMapRedLocalTask
```
public boolean isMapRedLocalTask()
```

getTopOperators

public Collection<Operator<? extends OperatorDesc>> getTopOperators()

hasReduce
```
public boolean hasReduce()
```

getReducer

public Operator<? extends OperatorDesc> getReducer(MapWork work)

getCounters

public HashMap<String,Long> getCounters()

getType
```
public abstract StageType getType()
```
Should be overridden to return the type of the specific task among the types in StageType.

Returns:
StageType.* or null if not overridden

subscribeFeed
```
public void subscribeFeed(Task<? extends Serializable> publisher)
```
Subscribe the feed of publisher. To prevent cycles, a task can only subscribe to its ancestor. Feed is a generic form of execution-time feedback (type, value) pair from one task to another task. Examples include dynamic partitions (which are only available at execution time). The MoveTask may pass the list of dynamic partitions to the StatsTask since after the MoveTask the list of dynamic partitions are lost (MoveTask moves them to the table's destination directory which is mixed with old partitions).

Parameters:
publisher - this feed provider.

getFeedSubscribers

public List<Task<? extends Serializable>> getFeedSubscribers()

setFeedSubscribers

public void setFeedSubscribers(List<Task<? extends Serializable>> s)

pushFeed

protected void pushFeed(Task.FeedType feedType,
            Object feedValue)

receiveFeed

protected void receiveFeed(Task.FeedType feedType,
               Object feedValue)

cloneConf
```
protected void cloneConf()
```

getTaskTag
```
public int getTaskTag()
```

setTaskTag
```
public void setTaskTag(int taskTag)
```

isLocalMode
```
public boolean isLocalMode()
```

setLocalMode

public void setLocalMode(boolean isLocalMode)

requireLock
```
public boolean requireLock()
```

ifRetryCmdWhenFail
```
public boolean ifRetryCmdWhenFail()
```

setRetryCmdWhenFail

public void setRetryCmdWhenFail(boolean retryCmdWhenFail)

getQueryPlan
```
public QueryPlan getQueryPlan()
```

setQueryPlan

public void setQueryPlan(QueryPlan queryPlan)

getJobID
```
public String getJobID()
```

shutdown
```
public void shutdown()
```

setException

protected void setException(Throwable ex)

setConsole

public void setConsole(SessionState.LogHelper console)

isFetchSource
```
public boolean isFetchSource()
```

setFetchSource

public void setFetchSource(boolean fetchSource)

toString
```
public String toString()
```
Overrides:

toString in class Object

hashCode
```
public int hashCode()
```
Overrides:

hashCode in class Object

equals
```
public boolean equals(Object obj)
```
Overrides:

equals in class Object

Class Task<T extends Serializable>

Nested Class Summary

Field Summary

Constructor Summary

Method Summary

Methods inherited from class java.lang.Object

Methods inherited from interface org.apache.hadoop.hive.ql.lib.Node

Field Detail

taskCounters

taskHandle

started

initialized

isdone

queued

conf

db

console

queryPlan

driverContext

clonedConf

jobID

backupTask

backupChildrenTasks

LOG

taskTag

NO_TAG

COMMON_JOIN

HINTED_MAPJOIN

HINTED_MAPJOIN_LOCAL

CONVERTED_MAPJOIN

CONVERTED_MAPJOIN_LOCAL

BACKUP_COMMON_JOIN

MAPJOIN_ONLY_NOBACKUP

CONVERTED_SORTMERGEJOIN

feedSubscribers

id

work

rootTask

childTasks

parentTasks

Constructor Detail

Task

Method Detail

getTaskHandle

initialize

executeTask

execute

isRootTask

setRootTask

setChildTasks

getChildren

getChildTasks

getNumChild

setParentTasks

getParentTasks

getNumParent

getBackupTask

setBackupTask

getBackupChildrenTasks

setBackupChildrenTasks

getAndInitBackupTask

removeFromChildrenTasks

getDependentTasks

addDependentTask

removeDependentTask

setStarted

started

done

setDone

setQueued

getQueued

setInitialized

getInitialized

isRunnable

setWork

getWork

getMapWork

setId

getId

isMapRedTask