com.pervasive.datarush.operators.io

Class WriteSink

java.lang.Object

com.pervasive.datarush.operators.AbstractLogicalOperator

com.pervasive.datarush.operators.StreamingOperator

com.pervasive.datarush.operators.ExecutableOperator

com.pervasive.datarush.operators.io.WriteSink

All Implemented Interfaces:

LogicalOperator, RecordSinkOperator, SinkOperator<RecordPort>

public final class WriteSink
extends ExecutableOperator
implements RecordSinkOperator

Writes a stream of records to a data sink. The sink accepts a sequence of bytes which represent formatted records, identical in logical structure. The mapping between physical and logical structure is encapsulated in a format descriptor which must be provided.

This operator is low-level, providing a generalized model for reading files in a distributed fashion. Typically, WriteSink is not directly used in a graph, instead being indirectly used though a composite operator such as one derived from AbstractWriter providing a more appropriate interface to the end user.

Parallelized writes are supported by creating multiple output sinks from the original target, called fragments, each representing the portion of data on one partition. In the case of a write to as file system, this would result in a directory of files. To write in parallel, the target must support the concept of fragmenting. If a target does not, the write is forced to be non-parallel.

A port is provided which signals completion of the writer. This can be used to express a dependency between operators based on the target having been successfully written.

The writer makes a best-effort attempt to validate the target before execution, but cannot always guarantee correctness, depending on the nature of the target. This is done to try to prevent misconfigured graphs from executing where the writer may not execute until a late phase when a failure may result in the loss of a significant amount of work being lost.

Constructor Summary

Constructors

Constructor and Description
WriteSink() Writes an empty target with default settings.
WriteSink(boolean includeDoneSignal) Writes an empty target with default settings.
WriteSink(ByteSink target, DataFormat format) Reads the specified target using the given format.

Method Summary

Modifier and Type	Method and Description
protected ExecutableOperator	cloneForExecution() Performs a deep copy of the operator for execution.
protected void	computeMetadata(StreamingMetadataContext ctx) Implementations must adhere to the following contracts
protected void	execute(ExecutionContext ctx) Executes the operator.
DataFormat	getFormat() Gets the data format for the configured target.
FormattingOptions	getFormatOptions() Gets the formatting options used by the writer.
boolean	getIncludeDoneSignal() Gets the property value for the include done signal setting.
RecordPort	getInput() Gets the record port providing the records to write to the target sink.
WriteMode	getMode() Gets how existing files should be handled by the writer.
ByteSink	getTarget() Gets the target sink for the reader.
boolean	getWriteOnClient() Indicates whether the writer should write a file on the client.
boolean	getWriteSingleSink() Indicates whether the writer produces a single output file.
boolean	isIgnoreSortOrder() If set to true, the writer will write in any order.
void	setFormat(DataFormat format) Sets the data format for the configured target.
void	setFormatOptions(FormattingOptions options) Sets the formatting options used by the writer.
void	setIgnoreSortOrder(boolean ignoreSortOrder) If set to true, the writer will write in any order.
void	setMode(WriteMode mode) Sets how the writer should handle an existing target.
void	setTarget(ByteSink target) Sets the target sink for the reader.
void	setWriteOnClient(boolean writeOnClient) Set whether the writer should write a file on the local client.
void	setWriteSingleSink(boolean enabled) Set whether the writer should produce a single output file or multiple ones.

Methods inherited from class com.pervasive.datarush.operators.ExecutableOperator

getNumInputCopies, getPortSettings, handleInactiveOutput

Methods inherited from class com.pervasive.datarush.operators.AbstractLogicalOperator

disableParallelism, getInputPorts, getOutputPorts, newInput, newInput, newOutput, newRecordInput, newRecordInput, newRecordOutput, notifyError

Methods inherited from class java.lang.Object

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

Methods inherited from interface com.pervasive.datarush.operators.LogicalOperator

disableParallelism, getInputPorts, getOutputPorts

Constructor Detail

WriteSink

public WriteSink()

Writes an empty target with default settings. Both the target and format must be set before execution or an error will be raised.

See Also:

setTarget(ByteSink), setFormat(DataFormat)

WriteSink

public WriteSink(boolean includeDoneSignal)

Writes an empty target with default settings. Both the target and format must be set before execution or an error will be raised.

See Also:

setTarget(ByteSink), setFormat(DataFormat)

WriteSink

public WriteSink(ByteSink target,
                 DataFormat format)

Reads the specified target using the given format. Default options are used.

Parameters:

target - the target to write

format - the target data format

Method Detail

getInput

public RecordPort getInput()

Gets the record port providing the records to write to the target sink.

Specified by:

getInput in interface RecordSinkOperator

Specified by:

getInput in interface SinkOperator<RecordPort>

Returns:

the input port for records to write

getIncludeDoneSignal

public boolean getIncludeDoneSignal()

Gets the property value for the include done signal setting.

Returns:

true if enabled; false by default

getFormat

public DataFormat getFormat()

Gets the data format for the configured target.

Returns:

the data format

setFormat

public void setFormat(DataFormat format)

Sets the data format for the configured target.

Parameters:

format - the target data format

getTarget

public ByteSink getTarget()

Gets the target sink for the reader.

Returns:

the write target

setTarget

public void setTarget(ByteSink target)

Sets the target sink for the reader.

Parameters:

target - the target to write

getFormatOptions

public FormattingOptions getFormatOptions()

Gets the formatting options used by the writer.

Returns:

the format options

setFormatOptions

public void setFormatOptions(FormattingOptions options)

Sets the formatting options used by the writer.

Parameters:

options - the format options to use

See Also:

DataFormat.DataFormatter

getMode

public WriteMode getMode()

Gets how existing files should be handled by the writer.

Returns:

the configured behavior with respect to target files

setMode

public void setMode(WriteMode mode)

Sets how the writer should handle an existing target.

Parameters:

mode - the behavior to use for existing files

getWriteSingleSink

public boolean getWriteSingleSink()

Indicates whether the writer produces a single output file.

Returns:

true if the writer should write a single output, false otherwise.

setWriteSingleSink

public void setWriteSingleSink(boolean enabled)

Set whether the writer should produce a single output file or multiple ones. By default, an output file will be produced for each partition, if the target sink supports this.

Parameters:

enabled - indicates whether a single output file should be written

getWriteOnClient

public boolean getWriteOnClient()

Indicates whether the writer should write a file on the client. This also disables the parallelism on the sink.

Returns:

true if the writer should a file on the client, false otherwise.

setWriteOnClient

public void setWriteOnClient(boolean writeOnClient)

Set whether the writer should write a file on the local client. This has the effect of disabling the parallelism on the sink.

Parameters:

writeOnClient - indicates whether the writer should write a file on the client.

isIgnoreSortOrder

public boolean isIgnoreSortOrder()

If set to true, the writer will write in any order. If false, the writer will preserve the current sorted state of the input.

Returns:

whether to ignore sort order.

setIgnoreSortOrder

public void setIgnoreSortOrder(boolean ignoreSortOrder)

If set to true, the writer will write in any order. If false, the writer will preserve the current sorted state of the input.

Parameters:

ignoreSortOrder - whether to ignore sort order.

computeMetadata

protected void computeMetadata(StreamingMetadataContext ctx)

Description copied from class: StreamingOperator

Implementations must adhere to the following contracts

General

Regardless of input ports/output port types, all implementations must do the following:

1. Validation. Validation of configuration should always be performed first.
2. Declare parallelizability.. Implementations must declare parallelizability by calling StreamingMetadataContext.parallelize(ParallelismStrategy).

Input record ports

Implementations with input record ports must declare the following:

1. Required data ordering:
Implementations that have data ordering requirements must declare them by calling RecordPort#setRequiredDataOrdering, otherwise data may arrive in any order.
2. Required data distribution (only applies to parallelizable operators):
Implementations that have data distribution requirements must declare them by calling RecordPort#setRequiredDataDistribution, otherwise data will arrive in an unspecified partial distribution.

Note that if the upstream operator's output distribution/ordering is compatible with those required, we avoid a re-sort/re-distribution which is generally a very large savings from a performance standpoint. In addition, some operators may chose to query the upstream output distribution/ordering by calling RecordPort#getSourceDataDistribution and RecordPort#getSourceDataOrdering. These should be viewed as a hints to help chose a more efficient algorithm. In such cases, though, operators must still declare data ordering and data distribution requirements; otherwise there is no guarantee that data will arrive sorted/distributed as required.

Output record ports

Implementations with output record ports must declare the following:

1. Type: Implementations must declare their output type by calling RecordPort#setType.

Implementations with output record ports may declare the following:

1. Output data ordering: Implementations that can make guarantees as to their output ordering may do so by calling RecordPort#setOutputDataOrdering
2. Output data distribution (only applies to parallelizable operators): Implementations that can make guarantees as to their output distribution may do so by calling RecordPort#setOutputDataDistribution

Note that both of these properties are optional; if unspecified, performance may suffer since the framework may unnecessarily re-sort/re-distributed the data.

Input model ports

In general, there is nothing special to declare for input model ports. Models are implicitly duplicated to all partitions when going from non-parallel to parallel operators. The case of a model going from a parallel to a non-parallel node is a special case of a "model reducer" operator. In the case of a model reducer, the downstream operator, must declare the following:

1. Merge handler: Model reducers must declare a merge handler by calling AbstractModelPort#setMergeHandler.

Note that MergeModel is a convenient, re-usable model reducer, parameterized with a merge-handler.

Output model ports

SimpleModelPort's have no associated metadata and therefore there is never any output metadata to declare. PMMLPort's, on the other hand, do have associated metadata. For all PMMLPorts, implementations must declare the following:

1. pmmlModelSpec: Implementations must declare the PMML model spec by calling PMMLPort.setPMMLModelSpec.

Specified by:

computeMetadata in class StreamingOperator

Parameters:

ctx - the context

execute

protected void execute(ExecutionContext ctx)

Description copied from class: ExecutableOperator

Executes the operator. Implementations should adhere to the following contracts:

1. Following execution, all input ports must be at end-of-data.
2. Following execution, all output ports must be at end-of-data.

Specified by:

execute in class ExecutableOperator

Parameters:

ctx - context in which to lookup physical ports bound to logical ports

cloneForExecution

protected ExecutableOperator cloneForExecution()

Description copied from class: ExecutableOperator

Performs a deep copy of the operator for execution. The default implementation is implemented in terms of JSON serialization: we perform a JSON serialization followed by a JSON deserialization. As a best-practice, operator implementations should not override this method. If they must override, though, then they must guarantee that cloneForExecution copies any instance variables that are modified by execute.

Overrides:

cloneForExecution in class ExecutableOperator

Returns:

a deep copy of this operator