org.apache.hawq.pxf.plugins.hdfs

Class AvroFileAccessor

java.lang.Object

org.apache.hawq.pxf.api.utilities.Plugin

org.apache.hawq.pxf.plugins.hdfs.HdfsSplittableDataAccessor

org.apache.hawq.pxf.plugins.hdfs.AvroFileAccessor

All Implemented Interfaces:

ReadAccessor

public class AvroFileAccessor
extends HdfsSplittableDataAccessor

A PXF Accessor for reading Avro File records

Field Summary

Fields inherited from class org.apache.hawq.pxf.plugins.hdfs.HdfsSplittableDataAccessor

conf, data, inputFormat, iter, jobConf, key, reader

Fields inherited from class org.apache.hawq.pxf.api.utilities.Plugin

inputData

Constructor Summary

Constructors

Constructor and Description
AvroFileAccessor(InputData input) Constructs a AvroFileAccessor that creates the job configuration and accesses the avro file to fetch the avro schema

Method Summary

Modifier and Type	Method and Description
protected java.lang.Object	getReader(org.apache.hadoop.mapred.JobConf jobConf, org.apache.hadoop.mapred.InputSplit split) Specialized accessors will override this method and implement their own recordReader.
OneRow	readNextObject() readNextObject The AVRO accessor is currently the only specialized accessor that overrides this method.

Methods inherited from class org.apache.hawq.pxf.plugins.hdfs.HdfsSplittableDataAccessor

closeForRead, getNextSplit, isThreadSafe, openForRead

Methods inherited from class java.lang.Object

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

Constructor Detail

AvroFileAccessor

public AvroFileAccessor(InputData input)
                 throws java.lang.Exception

Constructs a AvroFileAccessor that creates the job configuration and accesses the avro file to fetch the avro schema

Parameters:

input - all input parameters coming from the client

Throws:

java.lang.Exception - if getting the avro schema fails

Method Detail

getReader

protected java.lang.Object getReader(org.apache.hadoop.mapred.JobConf jobConf,
                                     org.apache.hadoop.mapred.InputSplit split)
                              throws java.io.IOException

Description copied from class: HdfsSplittableDataAccessor

Specialized accessors will override this method and implement their own recordReader. For example, a plain delimited text accessor may want to return a LineRecordReader.

Specified by:

getReader in class HdfsSplittableDataAccessor

Parameters:

jobConf - the hadoop jobconf to use for the selected InputFormat

split - the input split to be read by the accessor

Returns:

a recordreader to be used for reading the data records of the split

Throws:

java.io.IOException - if recordreader could not be created

readNextObject

public OneRow readNextObject()
                      throws java.io.IOException

readNextObject The AVRO accessor is currently the only specialized accessor that overrides this method. This happens, because the special AvroRecordReader.next() semantics (use of the AvroWrapper), so it cannot use the RecordReader's default implementation in SplittableFileAccessor

Specified by:

readNextObject in interface ReadAccessor

Overrides:

readNextObject in class HdfsSplittableDataAccessor

Returns:

the object which was read

Throws:

java.io.IOException