gr.uoa.di.madgik.searchlibrary.operatorlibrary.join
Class JoinOp

java.lang.Object
  gr.uoa.di.madgik.searchlibrary.operatorlibrary.join.JoinOp

public class JoinOp
extends Object

Operator class used to perform a hash inner join on its input and produces an output holding the joined results. The operator can also function as part of an intersection operation if its RecordGenerationPolicy is set to RecordGenerationPolicy.KeepLeft (or, less commonly to RecordGenerationPolicy.KeepRight. If not set, the RecordGenerationPolicy is equal to RecordGenerationPolicy.Concatenate which yields standard inner join functionality.

Author:

UoA

Field Summary

static int	BufferCapacityDef The default capacity of the RecordWriters and, if applicable, of all IRecordReaders' buffers
Map<IndexPair,Integer>	producerDefinitionMap A mapping from pairs of definition indices to producer definition indices.
static RecordGenerationPolicy	recordGenerationPolicyDef The default record generation policy.
static long	TimeoutDef The default timeout
static TimeUnit	TimeUnitDef The default time unit

Constructor Summary

JoinOp(URI leftLocator, URI rightLocator, RecordGenerationPolicy recordGenerationPolicy, long timeout, TimeUnit timeUnit, int bufferCapacity, StatsContainer stats)
Creates a new JoinOp with configurable timeout

JoinOp(URI leftLocator, URI rightLocator, RecordGenerationPolicy recordGenerationPolicy, long timeout, TimeUnit timeUnit, StatsContainer stats)
Creates a new JoinOp with configurable timeout

JoinOp(URI leftLocator, URI rightLocator, RecordGenerationPolicy recordGenerationPolicy, StatsContainer stats)

JoinOp(URI leftLocator, URI rightLocator, StatsContainer stats)
Creates a new JoinOp with the default timeout

Method Summary

URI	compute(String leftKeyFieldName, String rightKeyFieldName) Performs the join operation

Methods inherited from class java.lang.Object

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

Field Detail

recordGenerationPolicyDef

public static final RecordGenerationPolicy recordGenerationPolicyDef

The default record generation policy. Currently set to RecordGenerationPolicy.Concatenate which corresponds to the standard inner join policy

TimeoutDef

public static final long TimeoutDef

The default timeout

See Also:

Constant Field Values

TimeUnitDef

public static final TimeUnit TimeUnitDef

The default time unit

BufferCapacityDef

public static final int BufferCapacityDef

The default capacity of the RecordWriters and, if applicable, of all IRecordReaders' buffers

See Also:

Constant Field Values

producerDefinitionMap

public Map<IndexPair,Integer> producerDefinitionMap

A mapping from pairs of definition indices to producer definition indices. Used when record generation policy is RecordGenerationPolicy.Concatenate

Constructor Detail

JoinOp

public JoinOp(URI leftLocator,
              URI rightLocator,
              StatsContainer stats)

Creates a new JoinOp with the default timeout

Parameters:

leftLocator - The locator of the left input

rightLocator - The locator of the right input

stats - statistics

JoinOp

public JoinOp(URI leftLocator,
              URI rightLocator,
              RecordGenerationPolicy recordGenerationPolicy,
              StatsContainer stats)

Parameters:

leftLocator - The locator of the left input

rightLocator - The locator of the right input

recordGenerationPolicy - The record generation policy that should be followed. RecordGenerationPolicy.Concatenate should be used for standard inner join functionality. RecordGenerationPolicy.KeepLeft should be used if the client wishes to use the operator as the first step an intersection operation, the second one being a duplicate elimination step. RecordGenerationPolicy.KeepRight, if used for the same reason as the former, should yield the same results and is included for completeness.

stats - Statistics

JoinOp

public JoinOp(URI leftLocator,
              URI rightLocator,
              RecordGenerationPolicy recordGenerationPolicy,
              long timeout,
              TimeUnit timeUnit,
              StatsContainer stats)

Creates a new JoinOp with configurable timeout

Parameters:

leftLocator - The locator of the left input

rightLocator - The locator of the right input

recordGenerationPolicy - The record generation policy that should be followed. RecordGenerationPolicy.Concatenate should be used for standard inner join functionality. RecordGenerationPolicy.KeepLeft should be used if the client wishes to use the operator as the first step an intersection operation, the second one being a duplicate elimination step. RecordGenerationPolicy.KeepRight, if used for the same reason as the former, should yield the same results and is included for completeness.

timeout - The timeout which will be used by the two RandomReaders and the RecordWriter

timeUnit - The unit of the timeout

stats - Statistics

JoinOp

public JoinOp(URI leftLocator,
              URI rightLocator,
              RecordGenerationPolicy recordGenerationPolicy,
              long timeout,
              TimeUnit timeUnit,
              int bufferCapacity,
              StatsContainer stats)

Creates a new JoinOp with configurable timeout

Parameters:

leftLocator - The locator of the left input

rightLocator - The locator of the right input

recordGenerationPolicy - The record generation policy that should be followed. RecordGenerationPolicy.Concatenate should be used for standard inner join functionality. RecordGenerationPolicy.KeepLeft should be used if the client wishes to use the operator as the first step an intersection operation, the second one being a duplicate elimination step. RecordGenerationPolicy.KeepRight, if used for the same reason as the former, should yield the same results and is included for completeness.

timeout - The timeout which will be used by the two RandomReaders and the RecordWriter

timeUnit - The unit of the timeout

bufferCapacity - The capacity of the buffer which will be used by the RecordWriter and all RandomReaders (if applicable)

stats - Statistics

Method Detail

compute

public URI compute(String leftKeyFieldName,
                   String rightKeyFieldName)
            throws Exception

Performs the join operation

Parameters:

type - The type of resource to create

leftKeyFieldName - The name of the Field of the join key originating from the left locator

rightKeyFieldName - The name of the Field of the join key origiating from the right locator

Returns:

The join result

Throws:

Exception - An unrecoverable for the operation error occured