edu.yale.cs.hadoopdb.connector
Class DBInputSplit

java.lang.Object
  edu.yale.cs.hadoopdb.connector.DBInputSplit

All Implemented Interfaces:

org.apache.hadoop.io.Writable, org.apache.hadoop.mapred.InputSplit

Direct Known Subclasses:

SMSInputSplit

public class DBInputSplit
extends java.lang.Object
implements org.apache.hadoop.mapred.InputSplit

DBInputSplit links each Map to a DB Chunk. Splits serialize their DB connection information so that when instantiated at Map nodes, they can connect to their respective DB Chunks.

Field Summary

protected DBChunk	chunk
protected java.lang.String[]	locations
static org.apache.commons.logging.Log	LOG
protected java.lang.String	relation

Constructor Summary

DBInputSplit()

Method Summary

private DBChunk	deserializeChunk(java.io.DataInput in) Deserializes DBChunk
DBChunk	getChunk()
long	getLength() Returns 1 now...
java.lang.String[]	getLocations() Returns locations (host addresses) of the chunk's hosts.
java.lang.String	getRelation()
void	readFields(java.io.DataInput in) Deserializes relation and DBChunk object.
private void	serializeChunk(DBChunk chunk, java.io.DataOutput out) Serializes DBChunk
void	setChunk(DBChunk chunk) Sets a DBChunk and updates split locations
private void	setLocations() This method is called by readFields or setChunk on split instantiation or creation.
void	setRelation(java.lang.String relation)
void	write(java.io.DataOutput out) Serializes the relation and Chunk object.

Methods inherited from class java.lang.Object

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

Field Detail

chunk

protected DBChunk chunk

locations

protected java.lang.String[] locations

LOG

public static final org.apache.commons.logging.Log LOG

relation

protected java.lang.String relation

Constructor Detail

DBInputSplit

public DBInputSplit()

Method Detail

deserializeChunk

private DBChunk deserializeChunk(java.io.DataInput in)
                          throws java.io.IOException

Deserializes DBChunk

Throws:

java.io.IOException

getChunk

public DBChunk getChunk()

getLength

public long getLength()
               throws java.io.IOException

Returns 1 now... could in the future return information about the size of the Chunk (e.g. number of rows)

Specified by:

getLength in interface org.apache.hadoop.mapred.InputSplit

Throws:

java.io.IOException

getLocations

public java.lang.String[] getLocations()
                                throws java.io.IOException

Returns locations (host addresses) of the chunk's hosts.

Specified by:

getLocations in interface org.apache.hadoop.mapred.InputSplit

Throws:

java.io.IOException

getRelation

public java.lang.String getRelation()

readFields

public void readFields(java.io.DataInput in)
                throws java.io.IOException

Deserializes relation and DBChunk object. Then creates the list of locations from the DBChunk object.

Specified by:

readFields in interface org.apache.hadoop.io.Writable

Throws:

java.io.IOException

serializeChunk

private void serializeChunk(DBChunk chunk,
                            java.io.DataOutput out)
                     throws java.io.IOException

Serializes DBChunk

Throws:

java.io.IOException

setChunk

public void setChunk(DBChunk chunk)

Sets a DBChunk and updates split locations

setLocations

private void setLocations()

This method is called by readFields or setChunk on split instantiation or creation. A chunk could be stored on one or more locations. The locations array is therefore populated with the different host locations of the split's chunk.

setRelation

public void setRelation(java.lang.String relation)

write

public void write(java.io.DataOutput out)
           throws java.io.IOException

Serializes the relation and Chunk object.

Specified by:

write in interface org.apache.hadoop.io.Writable

Throws:

java.io.IOException