org.apache.hadoop.hive.ql.exec.tez.tools

Class KeyValueInputMerger

java.lang.Object

org.apache.tez.runtime.api.Reader

org.apache.tez.runtime.library.api.KeyValueReader

org.apache.hadoop.hive.ql.exec.tez.tools.KeyValueInputMerger

public class KeyValueInputMerger
extends org.apache.tez.runtime.library.api.KeyValueReader

A KeyValuesReader implementation that returns a sorted stream of key-values by doing a sorted merge of the key-value in LogicalInputs. Tags are in the last byte of the key, so no special handling for tags is required. Uses a priority queue to pick the KeyValuesReader of the input that is next in sort order.

Field Summary

Fields

Modifier and Type	Field and Description
static org.slf4j.Logger	l4j

Fields inherited from class org.apache.tez.runtime.library.api.KeyValueReader

completedProcessing

Constructor Summary

Constructors

Constructor and Description
KeyValueInputMerger(List<org.apache.tez.runtime.library.api.KeyValueReader> multiMRInputs, Deserializer deserializer, ObjectInspector[] inputObjInspectors, List<String> sortCols)

Method Summary

Modifier and Type	Method and Description
void	clean() Cleanup references
Object	getCurrentKey()
Object	getCurrentValue()
boolean	next()
void	setIOCxt(IOContext ioCxt) Set the IOContext reference so that input path can be changed.

Methods inherited from class org.apache.tez.runtime.library.api.KeyValueReader

hasCompletedProcessing

Methods inherited from class java.lang.Object

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

Field Detail

l4j

public static final org.slf4j.Logger l4j

Constructor Detail

KeyValueInputMerger

public KeyValueInputMerger(List<org.apache.tez.runtime.library.api.KeyValueReader> multiMRInputs,
                           Deserializer deserializer,
                           ObjectInspector[] inputObjInspectors,
                           List<String> sortCols)
                    throws Exception

Throws:

Exception

Method Detail

setIOCxt

public void setIOCxt(IOContext ioCxt)

Set the IOContext reference so that input path can be changed. This is needed because there can be more than one inputs at play at a given time, however, there is one IOContext which needs correct input path. In other joins, even when there are multiple inputs, they are read sequentially which is not the case here.

next

public boolean next()
             throws IOException

Specified by:

next in class org.apache.tez.runtime.library.api.KeyValueReader

Returns:

true if there are more key-values and advances to next key-values

Throws:

IOException

getCurrentKey

public Object getCurrentKey()
                     throws IOException

Specified by:

getCurrentKey in class org.apache.tez.runtime.library.api.KeyValueReader

Throws:

IOException

getCurrentValue

public Object getCurrentValue()
                       throws IOException

Specified by:

getCurrentValue in class org.apache.tez.runtime.library.api.KeyValueReader

Throws:

IOException

clean

public void clean()

Cleanup references