com.pervasive.datarush.ports.record

Class RecordPort

java.lang.Object

com.pervasive.datarush.ports.LogicalPort

com.pervasive.datarush.ports.record.RecordPort

public final class RecordPort
extends LogicalPort

Defines a logical record port.

Nested Class Summary

Nested classes/interfaces inherited from class com.pervasive.datarush.ports.LogicalPort

LogicalPort.Direction

Field Summary

Fields

Modifier and Type	Field and Description
static LogicalPortFactory<RecordPort>	FACTORY The singleton for creating record ports.

Method Summary

Modifier and Type	Method and Description
RecordPort	filtered(LogicalSubgraph subgraph, RecordPortOptions options) Returns a new record port that is a filtered view of this record port.
DataDistribution	getCombinedDataDistribution(MetadataContext ctx) Gets the required data distribution of this port
DataOrdering	getCombinedDataOrdering(MetadataContext ctx) Gets the required data ordering of this port
RecordMetadata	getCombinedMetadata(MetadataContext ctx) Returns the actual metadata that the operator will see given its current requirements and parallelism setting.
LogicalPortFactory<?>	getFactory() Returns the factory that knows how to create ports of this type
RecordInput	getInput(ExecutionContext ctx) Returns the physical input port associated with this logical input port.
RecordInput	getInput(ExecutionContext ctx, int copyNumber) Returns the physical input port associated with the specified copy number of this logical input port.
Class<RecordMetadata>	getMetadataType() Returns the class of metadata that this port uses
RecordOutput	getOutput(ExecutionContext ctx) Returns the physical output port associated with this logical output port.
DataDistribution	getRequiredDataDistribution(MetadataContext ctx) Gets the required data distribution of this port
DataOrdering	getRequiredDataOrdering(MetadataContext ctx) Gets the required data ordering of this port
RecordMetadata	getRequiredMetadata(MetadataContext ctx) Returns the required metadata associated with this port
DataDistribution	getSourceDataDistribution(MetadataContext ctx) Returns source data distribution associated with this input port.
DataOrdering	getSourceDataOrdering(MetadataContext ctx) Returns source data ordering associated with this input port.
RecordMetadata	getSourceMetadata(MetadataContext ctx) Returns the metadata associated with this port
RecordTokenType	getType(MetadataContext ctx) Returns the record type of this port.
void	setIterationStagingOptions(IterativeMetadataContext ctx, DatasetOptions options) Sets the staging options for this input port.
void	setOutputDataDistribution(MetadataCalculationContext ctx, DataDistribution distribution) Sets the data partitioning associated with the given record port.
void	setOutputDataOrdering(MetadataCalculationContext ctx, DataOrdering ordering) Sets the data ordering associated with this output port.
void	setRequiredDataDistribution(MetadataCalculationContext ctx, DataDistribution distribution) Sets the required data distribution of this input port.
void	setRequiredDataOrdering(MetadataCalculationContext ctx, DataOrdering ordering) Sets the required data ordering of the given input.
void	setType(MetadataCalculationContext ctx, RecordTokenType type) Sets the type associated with this port.

Methods inherited from class com.pervasive.datarush.ports.LogicalPort

getDirection, getName, getOwner, isOptional, toString

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, wait, wait, wait

Field Detail

FACTORY

public static final LogicalPortFactory<RecordPort> FACTORY

The singleton for creating record ports.

Method Detail

getSourceMetadata

public RecordMetadata getSourceMetadata(MetadataContext ctx)

Returns the metadata associated with this port

Parameters:

ctx - context used to resolve metadata

Returns:

the metadata associated with this port

getRequiredMetadata

public RecordMetadata getRequiredMetadata(MetadataContext ctx)

Returns the required metadata associated with this port

Parameters:

ctx - context used to resolve metadata

Returns:

the required metadata associated with this port

getCombinedMetadata

public RecordMetadata getCombinedMetadata(MetadataContext ctx)

Returns the actual metadata that the operator will see given its current requirements and parallelism setting.

Parameters:

ctx - context used to resolve metadata

Returns:

the combined metadata associated with this port

getSourceDataOrdering

public DataOrdering getSourceDataOrdering(MetadataContext ctx)

Returns source data ordering associated with this input port. This should be viewed as a hint to operator implementations that may chose a more efficient algorithm depending on the ordering of the input data. Note that in such a case, operators must declare sorting requirements via setRequiredDataOrdering(com.pervasive.datarush.operators.MetadataCalculationContext, com.pervasive.datarush.ports.record.DataOrdering), otherwise the framework is free to provide unsorted data as needed.

Parameters:

ctx - context used to resolve metadata

Returns:

the source data ordering associated with this input port

getSourceDataDistribution

public DataDistribution getSourceDataDistribution(MetadataContext ctx)

Returns source data distribution associated with this input port. This should be viewed as a hint to operator implementations that may chose a more efficient algorithm depending on the distribution of the input data. Note that in such a case, operators must declare distribution requirements via setRequiredDataDistribution(com.pervasive.datarush.operators.MetadataCalculationContext, com.pervasive.datarush.ports.record.DataDistribution), otherwise the framework is free to redistribute as needed.

Parameters:

ctx - context used to resolve metadata

Returns:

the source data distribution associated with this input port

setType

public void setType(MetadataCalculationContext ctx,
                    RecordTokenType type)

Sets the type associated with this port.

Parameters:

ctx - context used to resolve/set metadata

type - the type of this port.

setOutputDataOrdering

public void setOutputDataOrdering(MetadataCalculationContext ctx,
                                  DataOrdering ordering)

Sets the data ordering associated with this output port. Prior to calling this method, type must be set by calling setType(com.pervasive.datarush.operators.MetadataCalculationContext, com.pervasive.datarush.types.RecordTokenType).

Parameters:

ctx - context used to resolve/set metadata

ordering - the data ordering

setOutputDataDistribution

public void setOutputDataDistribution(MetadataCalculationContext ctx,
                                      DataDistribution distribution)

Sets the data partitioning associated with the given record port. Prior to calling this method, type must be set by calling setType(com.pervasive.datarush.operators.MetadataCalculationContext, com.pervasive.datarush.types.RecordTokenType).

Parameters:

ctx - context used to resolve/set metadata

distribution - the output distribution

setRequiredDataOrdering

public void setRequiredDataOrdering(MetadataCalculationContext ctx,
                                    DataOrdering ordering)

Sets the required data ordering of the given input. If unspecified, data ordering will arrive in an unknown order. Specifically, if an operator declares a requiredDataDistribution but does not declare a requiredDataOrdering, data will appear in arbitrary order, regardless of original order, since it is a performance savings to avoid re-sorting data following distribution. In the future there may be additional, similar, optimizations.

Parameters:

ctx - context used to resolve/set metadata

ordering - the required data ordering.

setRequiredDataDistribution

public void setRequiredDataDistribution(MetadataCalculationContext ctx,
                                        DataDistribution distribution)

Sets the required data distribution of this input port. If unspecified, data distribution will currently match the sourceDataDistribution, but this behavior is subject to change in the future. Thus, implementations should always declare their input distribution requirements.

Parameters:

ctx - context used to resolve/set metadata

distribution - the required data distribution of this input port.

filtered

public RecordPort filtered(LogicalSubgraph subgraph,
                           RecordPortOptions options)

Returns a new record port that is a filtered view of this record port. This port must be an input port of an iterative operator. This can provide significant performance boost if only a subset of the iterator's staged columns are needed in each pass.

Parameters:

subgraph - must be a subgraph created from the same iterative operator.

options - specifies selected columns

Returns:

a filtered view of this port.

setIterationStagingOptions

public void setIterationStagingOptions(IterativeMetadataContext ctx,
                                       DatasetOptions options)

Sets the staging options for this input port.

Parameters:

ctx - context used to resolve/set metadata

options - the staging options of this port

getRequiredDataOrdering

public DataOrdering getRequiredDataOrdering(MetadataContext ctx)

Gets the required data ordering of this port

Parameters:

ctx - context used to resolve metadata

Returns:

the required data ordering of this port

getRequiredDataDistribution

public DataDistribution getRequiredDataDistribution(MetadataContext ctx)

Gets the required data distribution of this port

Parameters:

ctx - context used to resolve metadata

Returns:

the required data distribution of this port

getCombinedDataOrdering

public DataOrdering getCombinedDataOrdering(MetadataContext ctx)

Gets the required data ordering of this port

Parameters:

ctx - context used to resolve metadata

Returns:

the required data ordering of this port

getCombinedDataDistribution

public DataDistribution getCombinedDataDistribution(MetadataContext ctx)

Gets the required data distribution of this port

Parameters:

ctx - context used to resolve metadata

Returns:

the required data distribution of this port

getMetadataType

public Class<RecordMetadata> getMetadataType()

Description copied from class: LogicalPort

Returns the class of metadata that this port uses

Specified by:

getMetadataType in class LogicalPort

Returns:

the class of metadata that this port uses

getFactory

public LogicalPortFactory<?> getFactory()

Description copied from class: LogicalPort

Returns the factory that knows how to create ports of this type

Specified by:

getFactory in class LogicalPort

Returns:

the factory that knows how to create ports of this type

getType

public RecordTokenType getType(MetadataContext ctx)

Returns the record type of this port. This port may either be an input port or an output port.

Parameters:

ctx - the context in which to lookup type information

Returns:

the record type of this port

getInput

public RecordInput getInput(ExecutionContext ctx)

Returns the physical input port associated with this logical input port.

Parameters:

ctx - the execution context.

Returns:

the physical input port associated with this logical input port.

getInput

public RecordInput getInput(ExecutionContext ctx,
                            int copyNumber)

Returns the physical input port associated with the specified copy number of this logical input port. This can be used in rare cases when we must examine multiple positions in the same input stream. In order to use this method, operators must also declare an intent to use multiple copies by overriding the method ExecutableOperator#getNumInputCopies.

Parameters:

ctx - the execution context

copyNumber - the copy number, must be in the range [0,numCopies).

Returns:

the physical input port

getOutput

public RecordOutput getOutput(ExecutionContext ctx)

Returns the physical output port associated with this logical output port.

Parameters:

ctx - the execution context.

Returns:

the physical output port associated with this logical output port.