com.pervasive.datarush.operators

Interface MetadataContext

All Known Subinterfaces:

CompositionContext, DeferredCompositionContext, ExecutionContext, IterativeExecutionContext, IterativeMetadataContext, MetadataCalculationContext, StreamingMetadataContext

public interface MetadataContext

Base interface for various contexts in which port metadata can be resolved.

Method Summary

Modifier and Type	Method and Description
PortMetadata	getCombinedMetadata(LogicalPort logicalPort) Returns the source metadata combined with the required metadata to determine the actual metadata that the operator will receive.
EngineConfig	getEngineConfig() Returns the engine settings that were passed to the graph being composed/executed.
FileClient	getFileClient() Returns the file client to be used within the graph
int	getMaxParallelism() Returns the max parallelism of the current operator.
GraphPath	getPath() Returns the path of the operator being composed.
PortMetadata	getRequiredMetadata(LogicalPort logicalPort) Gets the required metadata for the given input.
int	getSourceMaxParallelism(LogicalPort port) Returns whether the max parallelism of the source of this port.
PortMetadata	getSourceMetadata(LogicalPort logicalPort) Returns the source metadata associated with the given port.
boolean	isParallel() Deprecated. Use getMaxParallelism() instead.
boolean	isSourceConnected(LogicalPort inPort) Tests whether the given input port is connected.
boolean	isSourceParallel(LogicalPort port) Deprecated. Use getSourceMaxParallelism(LogicalPort) instead.

Method Detail

getSourceMetadata

PortMetadata getSourceMetadata(LogicalPort logicalPort)

Returns the source metadata associated with the given port. Note that operators should generally not call this method directly. Rather they should use one of the type-safe/convenience methods on the various subclasses of LogicalPort. For example:

    //get the record type for the "input" RecordPort
    input.getType(ctx)
    //get the full record metadata for the "input" RecordPort
    input.getSourceMetadata(ctx)
 

Parameters:

logicalPort - the input port

Returns:

the metadata associated with the given port

getRequiredMetadata

PortMetadata getRequiredMetadata(LogicalPort logicalPort)

Gets the required metadata for the given input. Note that operators should generally not call this method directly. Rather they should use one of the type-safe/convenience methods on the various subclasses of LogicalPort. For example:

    //get the required data ordering for the "input" record port
    input.getRequiredDataOrdering(ctx)
 

Parameters:

logicalPort - must be an input port for a non-composite operator

Returns:

the required metadata for the given input.

getCombinedMetadata

PortMetadata getCombinedMetadata(LogicalPort logicalPort)

Returns the source metadata combined with the required metadata to determine the actual metadata that the operator will receive. This also takes into account the operator's max parallelism setting. This method should always be used when determining output metadata as a function of the input metadata. Note that operators should generally not call this method directly. Rather they should use one of the type-safe/convenience methods on the various subclasses of LogicalPort. For example:

    //get the required data ordering for the "input" record port
    input.getCombinedDataOrdering(ctx)
 

Parameters:

logicalPort - must be an input port for a non-composite operator.

Returns:

the combined metadata

isSourceParallel

boolean isSourceParallel(LogicalPort port)

Deprecated. Use getSourceMaxParallelism(LogicalPort) instead.

Returns whether the source of the given port is parallel. If the given logical port is an input port this will return whether the output to which the input is connected is parallel. This is commonly used by operators that wish to determine their parallelism based on whether their source is parallel.

Parameters:

port - the port

Returns:

whether the source of the given port is parallel

getSourceMaxParallelism

int getSourceMaxParallelism(LogicalPort port)

Returns whether the max parallelism of the source of this port. If the given logical port is an input port this will return whether max parallelism of the output to which the input is connected is parallel. Operators should seldom need this method directly. However, they may use this method indirectly by specifying a ParallelismStrategy. For example, this is used by ParallelismStrategy.NEGOTIATE_BASED_ON_SOURCE to determine parallelism of an operator based on the input parallelism.

Parameters:

port - the port

Returns:

whether the source of the given port is parallel

isParallel

boolean isParallel()

Deprecated. Use getMaxParallelism() instead.

Returns whether the current operator is parallel. Will be true if the operator has declared that it is distributed via StreamingMetadataContext.setParallelizable(boolean) and distribution was not disabled via OperatorSettings#isParallelismDisabled(). For CompositeOperator's and IterativeOperator's this will be true as long as parallelism has not been disabled; if it returns true, the composite node may consist of a mixture of parallel and non-parallel components.

Returns:

whether the current operator is parallel.

getMaxParallelism

int getMaxParallelism()

Returns the max parallelism of the current operator. This will be determined by StreamingMetadataContext.parallelize(ParallelismStrategy). For CompositeOperator's and IterativeOperator's this will be equal to that specified via OperatorSettings.getMaxParallelism(); if it returns a value other than 1, the composite or iterative node may consist of a mixture of parallel and non-parallel components.

Returns:

the max parallelism of the current operator.

isSourceConnected

boolean isSourceConnected(LogicalPort inPort)

Tests whether the given input port is connected. This only applies to ports that were declared as "optional" since graph compilation will fail unless all non-optional ports are connected.

Parameters:

inPort - must be an input port

Returns:

whether the given input port is connected

getPath

GraphPath getPath()

Returns the path of the operator being composed.

Returns:

the path of the operator being composed.

getEngineConfig

EngineConfig getEngineConfig()

Returns the engine settings that were passed to the graph being composed/executed.

Returns:

the engine settings

getFileClient

FileClient getFileClient()

Returns the file client to be used within the graph

Returns:

the file client to be used for all file access