com.bigdata.bfs
Class FileVersionOutputStream

java.lang.Object
  java.io.OutputStream
      com.bigdata.bfs.FileVersionOutputStream

All Implemented Interfaces:

Closeable, Flushable

public class FileVersionOutputStream
extends OutputStream

Class buffers up to a block of data at a time and flushes blocks using an atomic append operation on the identifier file version.

Version:

$Id: FileVersionOutputStream.java 2265 2009-10-26 12:51:06Z thompsonbry $

Author:

Bryan Thompson

TODO:

this would benefit from asynchronous write-behind of the last block so that caller's do not wait for the RPC that writes the block onto the data index. use a blocking queue of buffers to be written so that the caller can not get far ahead of the database. a queue capacity of 1 or 2 should be sufficient.

Field Summary

protected String	id
protected BigdataFileSystem	repo
protected int	version

Constructor Summary

FileVersionOutputStream(BigdataFileSystem repo, String id, int version)
Create an output stream that will atomically append blocks of data to the specified file version.

Method Summary

void	close() Flushes the buffer.
long	copyStream(InputStream is) Consumes the input stream, writing blocks onto the file version.
void	flush() If there is data data accumulated in the buffer then it is written on the file version using an atomic append (empty buffers are NOT flushed).
long	getBlockCount() #of blocks written onto the file version.
long	getByteCount() #of bytes written onto this output stream.
String	getId() The file identifier.
int	getVersion() The file version identifer.
void	write(int b) Buffers the byte.

Methods inherited from class java.io.OutputStream

write, write

Methods inherited from class java.lang.Object

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

Field Detail

repo

protected final BigdataFileSystem repo

id

protected final String id

version

protected final int version

Constructor Detail

FileVersionOutputStream

public FileVersionOutputStream(BigdataFileSystem repo,
                               String id,
                               int version)

Create an output stream that will atomically append blocks of data to the specified file version.

Parameters:

id - The file identifier.

version - The version identifier.

Method Detail

getId

public String getId()

The file identifier.

getVersion

public int getVersion()

The file version identifer.

getByteCount

public long getByteCount()

#of bytes written onto this output stream.

TODO:

handle overflow of long - leave counter at Long.MAX_VALUE.

getBlockCount

public long getBlockCount()

#of blocks written onto the file version.

write

public void write(int b)
           throws IOException

Buffers the byte. If the buffer would overflow then it is flushed.

Specified by:

write in class OutputStream

Throws:

IOException

flush

public void flush()
           throws IOException

If there is data data accumulated in the buffer then it is written on the file version using an atomic append (empty buffers are NOT flushed).

Specified by:

flush in interface Flushable

Overrides:

flush in class OutputStream

Throws:

IOException

close

public void close()
           throws IOException

Flushes the buffer.

Specified by:

close in interface Closeable

Overrides:

close in class OutputStream

Throws:

IOException

copyStream

public long copyStream(InputStream is)
                throws IOException

Consumes the input stream, writing blocks onto the file version. The output stream is NOT flushed.

Parameters:

is - The input stream (closed iff it is fully consumed).

Returns:

The #of bytes copied from the input stream.

Throws:

IOException