public class StandaloneChunkHandler extends Object implements IChunkHandler
SolutionSetStream. Either way, it is then handed off synchronously using
QueryEngine.acceptChunk(IChunkMessage). That method will queue the chunk for asynchronous processing.
|Modifier and Type||Field and Description|
Instance puts all chunks onto the managed object heap.
Instance puts all chunks onto the native heap.
This instance is explicitly used for several unit tests in the query engine package that are not written to the RDF data model (the native heap version assumes an RDF data model).
|Modifier||Constructor and Description|
|Modifier and Type||Method and Description|
public static final IChunkHandler NATIVE_HEAP_INSTANCE
Vector query engine on native heap.
public static final IChunkHandler MANAGED_HEAP_INSTANCE
public static final IChunkHandler TEST_INSTANCE
public int handleChunk(IRunningQuery query, int bopId, int sinkId, IBindingSet chunk)
IBindingSet chunk generated by some pass over an operator and make it available to the target operator. How this is done depends on whether the query is running against a standalone database or the scale-out database.
Note: The return value is used as part of the termination criteria for
the query which depends on (a) the #of running operator tasks and (b) the
IChunkMessages generated (available) and consumed. The return
value of this method increases the #of
to the query.
query- The query.
bopId- The operator which wrote on the sink.
sinkId- The identifier of the target operator.
chunk- The intermediate results to be passed to that target operator.
IChunkMessagesent. This will always be ONE (1) for scale-up. For scale-out, there will be at least one
IChunkMessageper index partition over which the intermediate results were mapped.
Copyright © 2006–2016 SYSTAP, LLC DBA Blazegraph. All rights reserved.