com.fasterxml.sort
Class Sorter<T>

java.lang.Object
  com.fasterxml.sort.Sorter<T>

All Implemented Interfaces:

SortingState

Direct Known Subclasses:

TextFileSorter

public class Sorter<T>
extends Object
implements SortingState

Main entry point for sorting functionality; object that drives the sorting process from pre-sort to final output. Instances are not thread-safe, although they are reusable. Since the cost of creating new instances is trivial, there is usally no benefit from reusing instances, other than possible convenience.

Nested Class Summary

Nested classes/interfaces inherited from interface com.fasterxml.sort.SortingState

SortingState.Phase

Field Summary

protected Exception	_cancelForException
protected AtomicBoolean	_cancelRequest
protected Comparator<T>	_comparator
protected SortConfig	_config
protected int	_currentSortRound
protected SortingState.Phase	_phase
protected int	_presortFileCount
protected DataReaderFactory<T>	_readerFactory Factory used for reading intermediate sorted files.
protected int	_sortRoundCount
protected DataWriterFactory<T>	_writerFactory Factory used for writing intermediate sorted files.

Constructor Summary

Sorter(SortConfig config, DataReaderFactory<T> readerFactory, DataWriterFactory<T> writerFactory, Comparator<T> comparator)

Method Summary

protected static int	_calculateRoundCount(int files, int mergeFactor)
protected boolean	_checkForCancel()
protected boolean	_checkForCancel(Collection<File> tmpFilesToDelete)
protected File	_merge(List<File> inputs)
protected void	_merge(List<File> inputs, DataWriter<T> writer)
protected Comparator<Object>	_rawComparator()
protected void	_writeAll(DataWriter<T> resultWriter, Object[] items)
protected File	_writePresorted(Object[] items)
void	cancel() Method that can be used to try to cancel executing sort operation.
void	cancel(IOException e) Method that can be used to try to cancel executing sort operation.
void	cancel(RuntimeException e) Method that can be used to try to cancel executing sort operation.
int	getNumberOfPreSortFiles() Accessor for checking how many pre-sort files were created during pre-sort phase.
int	getNumberOfSortRounds() Accessor for figuring out how many regular sorting rounds need to be taken to complete sorting, if known.
SortingState.Phase	getPhase()
int	getSortRound() Accessor for checking which sorting round sorter is doing: for pre-sort it basically means number of segment (0-based) that is being processed in-memory, for regular sort it is number of (0-based) sorting round.
boolean	isCompleted() Accessor for determining whether sorting has been succesfully completed or not.
boolean	isPreSorting() Accessor for determining whether sorter is in its in-memory pre-sorting phase.
boolean	isSorting() Accessor for determining whether sorter is in regular merge-sort phase or not.
protected void	merge(List<File> presorted, DataWriter<T> resultWriter) Main-level merge method called during once during sorting.
protected List<File>	presort(DataReader<T> inputReader, SegmentedBuffer buffer, Object[] firstSortedBatch, T nextValue)
boolean	sort(DataReader<T> inputReader, DataWriter<T> resultWriter) Method that will perform full sort on input data read using given DataReader, and written out using specified DataWriter.
void	sort(InputStream source, OutputStream destination) Method that will perform full sort on specified input, writing results into specified destination.

Methods inherited from class java.lang.Object

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

Field Detail

_config

protected final SortConfig _config

_readerFactory

protected final DataReaderFactory<T> _readerFactory

Factory used for reading intermediate sorted files.

_writerFactory

protected final DataWriterFactory<T> _writerFactory

Factory used for writing intermediate sorted files.

_comparator

protected final Comparator<T> _comparator

_phase

protected SortingState.Phase _phase

_presortFileCount

protected int _presortFileCount

_sortRoundCount

protected int _sortRoundCount

_currentSortRound

protected int _currentSortRound

_cancelRequest

protected final AtomicBoolean _cancelRequest

_cancelForException

protected Exception _cancelForException

Constructor Detail

Sorter

public Sorter(SortConfig config,
              DataReaderFactory<T> readerFactory,
              DataWriterFactory<T> writerFactory,
              Comparator<T> comparator)

Parameters:

config - Configuration for the sorter

readerFactory - Factory used for creating readers for pre-sorted data; as well as for input if an InputStream is passed as source

writerFactory - Factory used for creating writers for storing pre-sorted data; as well as for results if an OutputStream is passed as destination.

Method Detail

cancel

public void cancel()

Description copied from interface: SortingState

Method that can be used to try to cancel executing sort operation. No exception will be thrown; sorting will just be stopped as soon as sorting thread notices request.

Specified by:

cancel in interface SortingState

cancel

public void cancel(RuntimeException e)

Description copied from interface: SortingState

Method that can be used to try to cancel executing sort operation. Exception object can be specified; if non-null instance is given, it will be thrown to indicate erroneous result, otherwise sorting is just interrupted but execution returns normally.

Specified by:

cancel in interface SortingState

cancel

public void cancel(IOException e)

Description copied from interface: SortingState

Method that can be used to try to cancel executing sort operation. Exception object can be specified; if non-null instance is given, it will be thrown to indicate erroneous result, otherwise sorting is just interrupted but execution returns normally.

Specified by:

cancel in interface SortingState

getPhase

public SortingState.Phase getPhase()

Specified by:

getPhase in interface SortingState

getNumberOfSortRounds

public int getNumberOfSortRounds()

Description copied from interface: SortingState

Accessor for figuring out how many regular sorting rounds need to be taken to complete sorting, if known. If information is not known, will return -1. This information generally becomes available after pre-sorting round.

Specified by:

getNumberOfSortRounds in interface SortingState

getNumberOfPreSortFiles

public int getNumberOfPreSortFiles()

Description copied from interface: SortingState

Accessor for checking how many pre-sort files were created during pre-sort phase. Can be zero if the whole data fit in memory during pre-sorting.

Specified by:

getNumberOfPreSortFiles in interface SortingState

getSortRound

public int getSortRound()

Description copied from interface: SortingState

Accessor for checking which sorting round sorter is doing: for pre-sort it basically means number of segment (0-based) that is being processed in-memory, for regular sort it is number of (0-based) sorting round.

Specified by:

getSortRound in interface SortingState

isCompleted

public boolean isCompleted()

Description copied from interface: SortingState

Accessor for determining whether sorting has been succesfully completed or not.

Specified by:

isCompleted in interface SortingState

isPreSorting

public boolean isPreSorting()

Description copied from interface: SortingState

Accessor for determining whether sorter is in its in-memory pre-sorting phase.

Specified by:

isPreSorting in interface SortingState

isSorting

public boolean isSorting()

Description copied from interface: SortingState

Accessor for determining whether sorter is in regular merge-sort phase or not.

Specified by:

isSorting in interface SortingState

sort

public void sort(InputStream source,
                 OutputStream destination)
          throws IOException

Method that will perform full sort on specified input, writing results into specified destination. Data conversions needed are done using DataReaderFactory and DataWriterFactory configured for this sorter.

Throws:

IOException

sort

public boolean sort(DataReader<T> inputReader,
                    DataWriter<T> resultWriter)
             throws IOException

Method that will perform full sort on input data read using given DataReader, and written out using specified DataWriter. Conversions to and from intermediate sort files is done using DataReaderFactory and DataWriterFactory configured for this sorter.

Returns:

true if sorting completed succesfully; false if it was cancelled

Throws:

IOException

presort

protected List<File> presort(DataReader<T> inputReader,
                             SegmentedBuffer buffer,
                             Object[] firstSortedBatch,
                             T nextValue)
                      throws IOException

Throws:

IOException

_writePresorted

protected File _writePresorted(Object[] items)
                        throws IOException

Throws:

IOException

merge

protected void merge(List<File> presorted,
                     DataWriter<T> resultWriter)
              throws IOException

Main-level merge method called during once during sorting.

Throws:

IOException

_writeAll

protected void _writeAll(DataWriter<T> resultWriter,
                         Object[] items)
                  throws IOException

Throws:

IOException

_merge

protected File _merge(List<File> inputs)
               throws IOException

Throws:

IOException

_merge

protected void _merge(List<File> inputs,
                      DataWriter<T> writer)
               throws IOException

Throws:

IOException

_calculateRoundCount

protected static int _calculateRoundCount(int files,
                                          int mergeFactor)

_checkForCancel

protected boolean _checkForCancel()
                           throws IOException

Throws:

IOException

_checkForCancel

protected boolean _checkForCancel(Collection<File> tmpFilesToDelete)
                           throws IOException

Throws:

IOException

_rawComparator

protected Comparator<Object> _rawComparator()