org.apache.accumulo.core.client.mapred

Class AccumuloFileOutputFormat

java.lang.Object

org.apache.hadoop.mapred.FileOutputFormat<Key,Value>

org.apache.accumulo.core.client.mapred.AccumuloFileOutputFormat

All Implemented Interfaces:

org.apache.hadoop.mapred.OutputFormat<Key,Value>

public class AccumuloFileOutputFormat
extends org.apache.hadoop.mapred.FileOutputFormat<Key,Value>

This class allows MapReduce jobs to write output in the Accumulo data file format.
Care should be taken to write only sorted data (sorted by Key), as this is an important requirement of Accumulo data files.

The output path to be created must be specified via FileOutputFormat.setOutputPath(JobConf, Path). This is inherited from FileOutputFormat.setOutputPath(JobConf, Path). Other methods from FileOutputFormat are not supported and may be ignored or cause failures. Using other Hadoop configuration options that affect the behavior of the underlying files directly in the Job's configuration may work, but are not directly supported at this time.

Nested Class Summary

Nested classes/interfaces inherited from class org.apache.hadoop.mapred.FileOutputFormat

org.apache.hadoop.mapred.FileOutputFormat.Counter

Field Summary

Fields

Modifier and Type	Field and Description
protected static org.apache.log4j.Logger	log

Constructor Summary

Constructors

Constructor and Description
AccumuloFileOutputFormat()

Method Summary

Methods

Modifier and Type	Method and Description
protected static AccumuloConfiguration	getAccumuloConfiguration(org.apache.hadoop.mapred.JobConf job) This helper method provides an AccumuloConfiguration object constructed from the Accumulo defaults, and overridden with Accumulo properties that have been stored in the Job's configuration.
org.apache.hadoop.mapred.RecordWriter<Key,Value>	getRecordWriter(org.apache.hadoop.fs.FileSystem ignored, org.apache.hadoop.mapred.JobConf job, String name, org.apache.hadoop.util.Progressable progress)
static void	setCompressionType(org.apache.hadoop.mapred.JobConf job, String compressionType) Sets the compression type to use for data blocks.
static void	setDataBlockSize(org.apache.hadoop.mapred.JobConf job, long dataBlockSize) Sets the size for data blocks within each file. Data blocks are a span of key/value pairs stored in the file that are compressed and indexed as a group.
static void	setFileBlockSize(org.apache.hadoop.mapred.JobConf job, long fileBlockSize) Sets the size for file blocks in the file system; file blocks are managed, and replicated, by the underlying file system.
static void	setIndexBlockSize(org.apache.hadoop.mapred.JobConf job, long indexBlockSize) Sets the size for index blocks within each file; smaller blocks means a deeper index hierarchy within the file, while larger blocks mean a more shallow index hierarchy within the file.
static void	setReplication(org.apache.hadoop.mapred.JobConf job, int replication) Sets the file system replication factor for the resulting file, overriding the file system default.

Methods inherited from class org.apache.hadoop.mapred.FileOutputFormat

checkOutputSpecs, getCompressOutput, getOutputCompressorClass, getOutputPath, getPathForCustomFile, getTaskOutputPath, getUniqueName, getWorkOutputPath, setCompressOutput, setOutputCompressorClass, setOutputPath

Methods inherited from class java.lang.Object

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

Field Detail

log

protected static final org.apache.log4j.Logger log

Constructor Detail

AccumuloFileOutputFormat

public AccumuloFileOutputFormat()

Method Detail

getAccumuloConfiguration

protected static AccumuloConfiguration getAccumuloConfiguration(org.apache.hadoop.mapred.JobConf job)

This helper method provides an AccumuloConfiguration object constructed from the Accumulo defaults, and overridden with Accumulo properties that have been stored in the Job's configuration.

Parameters:

job - the Hadoop context for the configured job

Since:

1.5.0

setCompressionType

public static void setCompressionType(org.apache.hadoop.mapred.JobConf job,
                      String compressionType)

Sets the compression type to use for data blocks. Specifying a compression may require additional libraries to be available to your Job.

Parameters:

job - the Hadoop job instance to be configured

compressionType - one of "none", "gz", "lzo", or "snappy"

Since:

1.5.0

setDataBlockSize

public static void setDataBlockSize(org.apache.hadoop.mapred.JobConf job,
                    long dataBlockSize)

Sets the size for data blocks within each file.
Data blocks are a span of key/value pairs stored in the file that are compressed and indexed as a group.

Making this value smaller may increase seek performance, but at the cost of increasing the size of the indexes (which can also affect seek performance).

Parameters:

job - the Hadoop job instance to be configured

dataBlockSize - the block size, in bytes

Since:

1.5.0

setFileBlockSize

public static void setFileBlockSize(org.apache.hadoop.mapred.JobConf job,
                    long fileBlockSize)

Sets the size for file blocks in the file system; file blocks are managed, and replicated, by the underlying file system.

Parameters:

job - the Hadoop job instance to be configured

fileBlockSize - the block size, in bytes

Since:

1.5.0

setIndexBlockSize

public static void setIndexBlockSize(org.apache.hadoop.mapred.JobConf job,
                     long indexBlockSize)

Sets the size for index blocks within each file; smaller blocks means a deeper index hierarchy within the file, while larger blocks mean a more shallow index hierarchy within the file. This can affect the performance of queries.

Parameters:

job - the Hadoop job instance to be configured

indexBlockSize - the block size, in bytes

Since:

1.5.0

setReplication

public static void setReplication(org.apache.hadoop.mapred.JobConf job,
                  int replication)

Sets the file system replication factor for the resulting file, overriding the file system default.

Parameters:

job - the Hadoop job instance to be configured

replication - the number of replicas for produced files

Since:

1.5.0

getRecordWriter

public org.apache.hadoop.mapred.RecordWriter<Key,Value> getRecordWriter(org.apache.hadoop.fs.FileSystem ignored,
                                                               org.apache.hadoop.mapred.JobConf job,
                                                               String name,
                                                               org.apache.hadoop.util.Progressable progress)
                                                                 throws IOException

Specified by:

getRecordWriter in interface org.apache.hadoop.mapred.OutputFormat<Key,Value>

Specified by:

getRecordWriter in class org.apache.hadoop.mapred.FileOutputFormat<Key,Value>

Throws:

IOException