ml.shifu.guagua.io

Interface GuaguaRecordReader<KEY extends Bytable,VALUE extends Bytable>

Type Parameters:

KEY - key type for each record

VALUE - value type for each record

public interface GuaguaRecordReader<KEY extends Bytable,VALUE extends Bytable>

GuaguaRecordReader is used for consistent interface to iterate data through FileSplit provided. The typical implementation is HDFS implementation in guagua-mapreduce.

To use it, one should set implementation in AbstractWorkerComputable.

Method Summary

Methods

Modifier and Type	Method and Description
void	close() Close resources at last, for example file descriptors.
KEY	getCurrentKey() Tmp we only support LongWritable key for byte offset in each line, follow LineRecordReader in hadoop.
VALUE	getCurrentValue() Tmp we only support Text value for whole content in each line, follow LineRecordReader in hadoop.
void	initialize(GuaguaFileSplit split) Initialize file split for user to create relative reader instance.
boolean	nextKeyValue() Cursor shift to next and set current key value.

Method Detail

initialize

void initialize(GuaguaFileSplit split)
                throws IOException

Initialize file split for user to create relative reader instance.

Throws:

IOException

nextKeyValue

boolean nextKeyValue()
                     throws IOException

Cursor shift to next and set current key value.

Throws:

IOException

getCurrentKey

KEY getCurrentKey()

Tmp we only support LongWritable key for byte offset in each line, follow LineRecordReader in hadoop.

getCurrentValue

VALUE getCurrentValue()

Tmp we only support Text value for whole content in each line, follow LineRecordReader in hadoop.

close

void close()
           throws IOException

Close resources at last, for example file descriptors. Should be called in finally segment.

Throws:

IOException