datafu.hourglass.mapreduce
Class PartitioningReducer

java.lang.Object
  datafu.hourglass.mapreduce.ObjectProcessor
      datafu.hourglass.mapreduce.ObjectReducer
          datafu.hourglass.mapreduce.PartitioningReducer

All Implemented Interfaces:

java.io.Serializable

public class PartitioningReducer
extends ObjectReducer
implements java.io.Serializable

The reducer used by AbstractPartitionPreservingIncrementalJob and its derived classes.

An implementation of Accumulator is used to perform aggregation and produce the output value.

The input key is assumed to have time and value fields. The value here is the true key, and the time represents the input partition the data was derived from. The true key is used as the key in the reducer output and the time is dropped. This reducer uses multiple outputs; the time is used to determine which output to write to, where the named outputs have the form yyyyMMdd derived from the time.

Author:

"Matthew Hayes"

See Also:

Serialized Form

Constructor Summary

PartitioningReducer()

Method Summary

void	close()
Accumulator<org.apache.avro.generic.GenericRecord,org.apache.avro.generic.GenericRecord>	getAccumulator() Gets the accumulator used to perform aggregation.
PartitionPreservingSchemas	getSchemas() Gets the Avro schemas
void	reduce(java.lang.Object keyObj, java.lang.Iterable<java.lang.Object> values, org.apache.hadoop.mapreduce.ReduceContext<java.lang.Object,java.lang.Object,java.lang.Object,java.lang.Object> context)
void	setAccumulator(Accumulator<org.apache.avro.generic.GenericRecord,org.apache.avro.generic.GenericRecord> acc) Sets the accumulator used to perform aggregation.
void	setContext(org.apache.hadoop.mapreduce.TaskInputOutputContext<java.lang.Object,java.lang.Object,java.lang.Object,java.lang.Object> context)
void	setSchemas(PartitionPreservingSchemas schemas) Sets the Avro schemas.

Methods inherited from class datafu.hourglass.mapreduce.ObjectProcessor

getContext

Methods inherited from class java.lang.Object

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

Constructor Detail

PartitioningReducer

public PartitioningReducer()

Method Detail

reduce

public void reduce(java.lang.Object keyObj,
                   java.lang.Iterable<java.lang.Object> values,
                   org.apache.hadoop.mapreduce.ReduceContext<java.lang.Object,java.lang.Object,java.lang.Object,java.lang.Object> context)
            throws java.io.IOException,
                   java.lang.InterruptedException

Specified by:

reduce in class ObjectReducer

Throws:

java.io.IOException

java.lang.InterruptedException

setContext

public void setContext(org.apache.hadoop.mapreduce.TaskInputOutputContext<java.lang.Object,java.lang.Object,java.lang.Object,java.lang.Object> context)

Overrides:

setContext in class ObjectProcessor

setAccumulator

public void setAccumulator(Accumulator<org.apache.avro.generic.GenericRecord,org.apache.avro.generic.GenericRecord> acc)

Sets the accumulator used to perform aggregation.

Parameters:

acc - The accumulator

getAccumulator

public Accumulator<org.apache.avro.generic.GenericRecord,org.apache.avro.generic.GenericRecord> getAccumulator()

Gets the accumulator used to perform aggregation.

Returns:

The accumulator

setSchemas

public void setSchemas(PartitionPreservingSchemas schemas)

Sets the Avro schemas.

Parameters:

schemas -

getSchemas

public PartitionPreservingSchemas getSchemas()

Gets the Avro schemas

Returns:

schemas

close

public void close()
           throws java.io.IOException,
                  java.lang.InterruptedException

Overrides:

close in class ObjectProcessor

Throws:

java.io.IOException

java.lang.InterruptedException