public class GraphDataSink extends BaseCompute implements Collector
Constructor and Description |
---|
GraphDataSink(java.lang.String inputkey) |
Modifier and Type | Method and Description |
---|---|
boolean |
execute(IMessage message)
This method add the received message from the DataObject Source into the data objects.
|
DataPartition<java.util.HashMap<java.lang.String,SsspDefaultVertex>> |
get()
get the collected valued
|
IONames |
getCollectibleNames()
This method should return a set of collectible names, that can be collected from this
collector.
|
void |
prepare(Config cfg,
TaskContext context)
Prepare the task to be executed
|
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
endExecute
public boolean execute(IMessage message)
public void prepare(Config cfg, TaskContext context)
INode
public DataPartition<java.util.HashMap<java.lang.String,SsspDefaultVertex>> get()
Collector
public IONames getCollectibleNames()
Collector
If Collector
C of task graph TG1 collects variable "var1" and C's parallelism is n, If
Receptor
R of task graph TG2 is interested in receiving "var1", R's parallelism should
be equal to n and, R should have the same distribution as C among the workers.
getCollectibleNames
in interface Collector