/**
* A {@link SpoutWrapper} wraps an {@link IRichSpout} in order to execute it within a Flink Streaming program. It
* takes the spout's output tuples and transforms them into Flink tuples of type {@code OUT} (see
* {@link SpoutCollector} for supported types).
*
* Per default, {@link SpoutWrapper} calls the wrapped spout's {@link IRichSpout#nextTuple() nextTuple()} method in
* an infinite loop.
* Alternatively, {@link SpoutWrapper} can call {@link IRichSpout#nextTuple() nextTuple()} for a finite number of
* times and terminate automatically afterwards (for finite input streams). The number of {@code nextTuple()} calls can
* be specified as a certain number of invocations or can be undefined. In the undefined case, {@link SpoutWrapper}
* terminates if no record was emitted to the output collector for the first time during a call to
* {@link IRichSpout#nextTuple() nextTuple()}.
* If the given spout implements {@link FiniteSpout} interface and {@link #numberOfInvocations} is not provided or
* is {@code null}, {@link SpoutWrapper} calls {@link IRichSpout#nextTuple() nextTuple()} method until
* {@link FiniteSpout#reachedEnd()} returns true.
*/
public final class SpoutWrapper extends RichParallelSourceFunction implements StoppableFunction {
//......
/** The number of {@link IRichSpout#nextTuple()} calls. */
private Integer numberOfInvocations; // do not use int -> null indicates an infinite loop
/**
* Instantiates a new {@link SpoutWrapper} that calls the {@link IRichSpout#nextTuple() nextTuple()} method of
* the given {@link IRichSpout spout} a finite number of times. The output type will be one of {@link Tuple0} to
* {@link Tuple25} depending on the spout's declared number of attributes.
*
* @param spout
* The {@link IRichSpout spout} to be used.
* @param numberOfInvocations
* The number of calls to {@link IRichSpout#nextTuple()}. If value is negative, {@link SpoutWrapper}
* terminates if no tuple was emitted for the first time. If value is {@code null}, finite invocation is
* disabled.
* @throws IllegalArgumentException
* If the number of declared output attributes is not with range [0;25].
*/
public SpoutWrapper(final IRichSpout spout, final Integer numberOfInvocations)
throws IllegalArgumentException {
this(spout, (Collection) null, numberOfInvocations);
}
/**
* Instantiates a new {@link SpoutWrapper} that calls the {@link IRichSpout#nextTuple() nextTuple()} method of
* the given {@link IRichSpout spout} in an infinite loop. The output type will be one of {@link Tuple0} to
* {@link Tuple25} depending on the spout's declared number of attributes.
*
* @param spout
* The {@link IRichSpout spout} to be used.
* @throws IllegalArgumentException
* If the number of declared output attributes is not with range [0;25].
*/
public SpoutWrapper(final IRichSpout spout) throws IllegalArgumentException {
this(spout, (Collection) null, null);
}
@Override
public final void run(final SourceContext ctx) throws Exception {
final GlobalJobParameters config = super.getRuntimeContext().getExecutionConfig()
.getGlobalJobParameters();
StormConfig stormConfig = new StormConfig();
if (config != null) {
if (config instanceof StormConfig) {
stormConfig = (StormConfig) config;
} else {
stormConfig.putAll(config.toMap());
}
}
final TopologyContext stormTopologyContext = WrapperSetupHelper.createTopologyContext(
(StreamingRuntimeContext) super.getRuntimeContext(), this.spout, this.name,
this.stormTopology, stormConfig);
SpoutCollector collector = new SpoutCollector(this.numberOfAttributes,
stormTopologyContext.getThisTaskId(), ctx);
this.spout.open(stormConfig, stormTopologyContext, new SpoutOutputCollector(collector));
this.spout.activate();
if (numberOfInvocations == null) {
if (this.spout instanceof FiniteSpout) {
final FiniteSpout finiteSpout = (FiniteSpout) this.spout;
while (this.isRunning && !finiteSpout.reachedEnd()) {
finiteSpout.nextTuple();
}
} else {
while (this.isRunning) {
this.spout.nextTuple();
}
}
} else {
int counter = this.numberOfInvocations;
if (counter >= 0) {
while ((--counter >= 0) && this.isRunning) {
this.spout.nextTuple();
}
} else {
do {
collector.tupleEmitted = false;
this.spout.nextTuple();
} while (collector.tupleEmitted && this.isRunning);
}
}
}
/**
* {@inheritDoc}
*
*
Sets the {@link #isRunning} flag to {@code false}.
*/
@Override
public void cancel() {
this.isRunning = false;
}
/**
* {@inheritDoc}
*
*
Sets the {@link #isRunning} flag to {@code false}.
*/
@Override
public void stop() {
this.isRunning = false;
}
@Override
public void close() throws Exception {
this.spout.close();
}
}
/**
* A {@link SpoutCollector} is used by {@link SpoutWrapper} to provided an Storm
* compatible output collector to the wrapped spout. It transforms the emitted Storm tuples into
* Flink tuples and emits them via the provide {@link SourceContext} object.
*/
class SpoutCollector extends AbstractStormCollector implements ISpoutOutputCollector {
/** The Flink source context object. */
private final SourceContext flinkContext;
/**
* Instantiates a new {@link SpoutCollector} that emits Flink tuples to the given Flink source context. If the
* number of attributes is specified as zero, any output type is supported. If the number of attributes is between 0
* to 25, the output type is {@link Tuple0} to {@link Tuple25}, respectively.
*
* @param numberOfAttributes
* The number of attributes of the emitted tuples.
* @param taskId
* The ID of the producer task (negative value for unknown).
* @param flinkContext
* The Flink source context to be used.
* @throws UnsupportedOperationException
* if the specified number of attributes is greater than 25
*/
SpoutCollector(final HashMap numberOfAttributes, final int taskId,
final SourceContext flinkContext) throws UnsupportedOperationException {
super(numberOfAttributes, taskId);
assert (flinkContext != null);
this.flinkContext = flinkContext;
}
@Override
protected List doEmit(final OUT flinkTuple) {
this.flinkContext.collect(flinkTuple);
// TODO
return null;
}
@Override
public void reportError(final Throwable error) {
// not sure, if Flink can support this
}
@Override
public List emit(final String streamId, final List
/**
* The Task represents one execution of a parallel subtask on a TaskManager.
* A Task wraps a Flink operator (which may be a user function) and
* runs it, providing all services necessary for example to consume input data,
* produce its results (intermediate result partitions) and communicate
* with the JobManager.
*
*
The Flink operators (implemented as subclasses of
* {@link AbstractInvokable} have only data readers, -writers, and certain event callbacks.
* The task connects those to the network stack and actor messages, and tracks the state
* of the execution and handles exceptions.
*
*
Tasks have no knowledge about how they relate to other tasks, or whether they
* are the first attempt to execute the task, or a repeated attempt. All of that
* is only known to the JobManager. All the task knows are its own runnable code,
* the task's configuration, and the IDs of the intermediate results to consume and
* produce (if any).
*
*
Each Task is run by one dedicated thread.
*/
public class Task implements Runnable, TaskActions, CheckpointListener {
//......
/**
* The core work method that bootstraps the task and executes its code.
*/
@Override
public void run() {
//......
// now load and instantiate the task's invokable code
invokable = loadAndInstantiateInvokable(userCodeClassLoader, nameOfInvokableClass, env);
// ----------------------------------------------------------------
// actual task core work
// ----------------------------------------------------------------
// we must make strictly sure that the invokable is accessible to the cancel() call
// by the time we switched to running.
this.invokable = invokable;
// switch to the RUNNING state, if that fails, we have been canceled/failed in the meantime
if (!transitionState(ExecutionState.DEPLOYING, ExecutionState.RUNNING)) {
throw new CancelTaskException();
}
// notify everyone that we switched to running
notifyObservers(ExecutionState.RUNNING, null);
taskManagerActions.updateTaskExecutionState(new TaskExecutionState(jobId, executionId, ExecutionState.RUNNING));
// make sure the user code classloader is accessible thread-locally
executingThread.setContextClassLoader(userCodeClassLoader);
// run the invokable
invokable.invoke();
//......
}
}
/**
* Base class for all streaming tasks. A task is the unit of local processing that is deployed
* and executed by the TaskManagers. Each task runs one or more {@link StreamOperator}s which form
* the Task's operator chain. Operators that are chained together execute synchronously in the
* same thread and hence on the same stream partition. A common case for these chains
* are successive map/flatmap/filter tasks.
*
*
The task chain contains one "head" operator and multiple chained operators.
* The StreamTask is specialized for the type of the head operator: one-input and two-input tasks,
* as well as for sources, iteration heads and iteration tails.
*
*
The Task class deals with the setup of the streams read by the head operator, and the streams
* produced by the operators at the ends of the operator chain. Note that the chain may fork and
* thus have multiple ends.
*
*
The life cycle of the task is set up as follows:
*
{@code
* -- setInitialState -> provides state of all operators in the chain
*
* -- invoke()
* |
* +----> Create basic utils (config, etc) and load the chain of operators
* +----> operators.setup()
* +----> task specific init()
* +----> initialize-operator-states()
* +----> open-operators()
* +----> run()
* +----> close-operators()
* +----> dispose-operators()
* +----> common cleanup
* +----> task specific cleanup()
* }
*
*
The {@code StreamTask} has a lock object called {@code lock}. All calls to methods on a
* {@code StreamOperator} must be synchronized on this lock object to ensure that no methods
* are called concurrently.
*
* @param
* @param
*/
@Internal
public abstract class StreamTask>
extends AbstractInvokable
implements AsyncExceptionHandler {
//......
@Override
public final void invoke() throws Exception {
boolean disposed = false;
try {
//......
// let the task do its work
isRunning = true;
run();
// if this left the run() method cleanly despite the fact that this was canceled,
// make sure the "clean shutdown" is not attempted
if (canceled) {
throw new CancelTaskException();
}
LOG.debug("Finished task {}", getName());
//......
}
finally {
// clean up everything we initialized
isRunning = false;
//......
}
}
}
/**
* {@link StreamTask} for executing a {@link StreamSource}.
*
*
One important aspect of this is that the checkpointing and the emission of elements must never
* occur at the same time. The execution must be serial. This is achieved by having the contract
* with the StreamFunction that it must only modify its state or emit elements in
* a synchronized block that locks on the lock Object. Also, the modification of the state
* and the emission of elements must happen in the same block of code that is protected by the
* synchronized block.
*
* @param Type of the output elements of this source.
* @param Type of the source function for the stream source operator
* @param Type of the stream source operator
*/
@Internal
public class SourceStreamTask, OP extends StreamSource>
extends StreamTask {
//......
@Override
protected void run() throws Exception {
headOperator.run(getCheckpointLock(), getStreamStatusMaintainer());
}
}
/**
* {@link StreamOperator} for streaming sources.
*
* @param Type of the output elements
* @param Type of the source function of this stream source operator
*/
@Internal
public class StreamSource>
extends AbstractUdfStreamOperator implements StreamOperator {
//......
public void run(final Object lockingObject, final StreamStatusMaintainer streamStatusMaintainer) throws Exception {
run(lockingObject, streamStatusMaintainer, output);
}
public void run(final Object lockingObject,
final StreamStatusMaintainer streamStatusMaintainer,
final Output> collector) throws Exception {
final TimeCharacteristic timeCharacteristic = getOperatorConfig().getTimeCharacteristic();
final Configuration configuration = this.getContainingTask().getEnvironment().getTaskManagerInfo().getConfiguration();
final long latencyTrackingInterval = getExecutionConfig().isLatencyTrackingConfigured()
? getExecutionConfig().getLatencyTrackingInterval()
: configuration.getLong(MetricOptions.LATENCY_INTERVAL);
LatencyMarksEmitter latencyEmitter = null;
if (latencyTrackingInterval > 0) {
latencyEmitter = new LatencyMarksEmitter<>(
getProcessingTimeService(),
collector,
latencyTrackingInterval,
this.getOperatorID(),
getRuntimeContext().getIndexOfThisSubtask());
}
final long watermarkInterval = getRuntimeContext().getExecutionConfig().getAutoWatermarkInterval();
this.ctx = StreamSourceContexts.getSourceContext(
timeCharacteristic,
getProcessingTimeService(),
lockingObject,
streamStatusMaintainer,
collector,
watermarkInterval,
-1);
try {
userFunction.run(ctx);
// if we get here, then the user function either exited after being done (finite source)
// or the function was canceled or stopped. For the finite source case, we should emit
// a final watermark that indicates that we reached the end of event-time
if (!isCanceledOrStopped()) {
ctx.emitWatermark(Watermark.MAX_WATERMARK);
}
} finally {
// make sure that the context is closed in any case
ctx.close();
if (latencyEmitter != null) {
latencyEmitter.close();
}
}
}
1. aggregateByKey的运行机制
/**
* Aggregate the values of each key, using given combine functions and a neutral "zero value".
* This function can return a different result type
spark-sql是Spark bin目录下的一个可执行脚本,它的目的是通过这个脚本执行Hive的命令,即原来通过
hive>输入的指令可以通过spark-sql>输入的指令来完成。
spark-sql可以使用内置的Hive metadata-store,也可以使用已经独立安装的Hive的metadata store
关于Hive build into Spark
// Max value in Array
var arr = [1,2,3,5,3,2];Math.max.apply(null, arr); // 5
// Max value in Jaon Array
var arr = [{"x":"8/11/2009","y":0.026572007},{"x"
在使用XMlhttpRequest对象发送请求和响应之前,必须首先使用javaScript对象创建一个XMLHttpRquest对象。
var xmlhttp;
function getXMLHttpRequest(){
if(window.ActiveXObject){
xmlhttp:new ActiveXObject("Microsoft.XMLHTTP