/// <exception cref="System.IO.IOException"/>
public Decompressor GetDecompressor()
{
CompressionCodec codec = GetCodec();
if (codec != null)
{
Decompressor decompressor = CodecPool.GetDecompressor(codec);
if (decompressor != null)
{
if (decompressor.Finished())
{
// Somebody returns the decompressor to CodecPool but is still using
// it.
Log.Warn("Deompressor obtained from CodecPool already finished()");
}
else
{
if (Log.IsDebugEnabled())
{
Log.Debug("Got a decompressor: " + decompressor.GetHashCode());
}
}
decompressor.Reset();
}
return(decompressor);
}
return(null);
}
public virtual void TestMultipleClose()
{
Uri testFileUrl = GetType().GetClassLoader().GetResource("recordSpanningMultipleSplits.txt.bz2"
);
NUnit.Framework.Assert.IsNotNull("Cannot find recordSpanningMultipleSplits.txt.bz2"
, testFileUrl);
FilePath testFile = new FilePath(testFileUrl.GetFile());
Path testFilePath = new Path(testFile.GetAbsolutePath());
long testFileSize = testFile.Length();
Configuration conf = new Configuration();
conf.SetInt(LineRecordReader.MaxLineLength, int.MaxValue);
FileSplit split = new FileSplit(testFilePath, 0, testFileSize, (string[])null);
LineRecordReader reader = new LineRecordReader(conf, split);
LongWritable key = new LongWritable();
Text value = new Text();
//noinspection StatementWithEmptyBody
while (reader.Next(key, value))
{
}
reader.Close();
reader.Close();
BZip2Codec codec = new BZip2Codec();
codec.SetConf(conf);
ICollection <Decompressor> decompressors = new HashSet <Decompressor>();
for (int i = 0; i < 10; ++i)
{
decompressors.AddItem(CodecPool.GetDecompressor(codec));
}
NUnit.Framework.Assert.AreEqual(10, decompressors.Count);
}
/// <summary>Construct an IFile Reader.</summary>
/// <param name="conf">Configuration File</param>
/// <param name="in">The input stream</param>
/// <param name="length">
/// Length of the data in the stream, including the checksum
/// bytes.
/// </param>
/// <param name="codec">codec</param>
/// <param name="readsCounter">Counter for records read from disk</param>
/// <exception cref="System.IO.IOException"/>
public Reader(Configuration conf, FSDataInputStream @in, long length, CompressionCodec
codec, Counters.Counter readsCounter)
{
// Count records read from disk
// Possibly decompressed stream that we read
readRecordsCounter = readsCounter;
checksumIn = new IFileInputStream(@in, length, conf);
if (codec != null)
{
decompressor = CodecPool.GetDecompressor(codec);
if (decompressor != null)
{
this.@in = codec.CreateInputStream(checksumIn, decompressor);
}
else
{
Log.Warn("Could not obtain decompressor from CodecPool");
this.@in = checksumIn;
}
}
else
{
this.@in = checksumIn;
}
this.dataIn = new DataInputStream(this.@in);
this.fileLength = length;
if (conf != null)
{
bufferSize = conf.GetInt("io.file.buffer.size", DefaultBufferSize);
}
}
/// <exception cref="System.IO.IOException"/>
public override void Initialize(InputSplit genericSplit, TaskAttemptContext context
)
{
FileSplit split = (FileSplit)genericSplit;
Configuration job = context.GetConfiguration();
this.maxLineLength = job.GetInt(MaxLineLength, int.MaxValue);
start = split.GetStart();
end = start + split.GetLength();
Path file = split.GetPath();
// open the file and seek to the start of the split
FileSystem fs = file.GetFileSystem(job);
fileIn = fs.Open(file);
CompressionCodec codec = new CompressionCodecFactory(job).GetCodec(file);
if (null != codec)
{
isCompressedInput = true;
decompressor = CodecPool.GetDecompressor(codec);
if (codec is SplittableCompressionCodec)
{
SplitCompressionInputStream cIn = ((SplittableCompressionCodec)codec).CreateInputStream
(fileIn, decompressor, start, end, SplittableCompressionCodec.READ_MODE.Byblock);
@in = new CompressedSplitLineReader(cIn, job, this.recordDelimiterBytes);
start = cIn.GetAdjustedStart();
end = cIn.GetAdjustedEnd();
filePosition = cIn;
}
else
{
@in = new SplitLineReader(codec.CreateInputStream(fileIn, decompressor), job, this
.recordDelimiterBytes);
filePosition = fileIn;
}
}
else
{
fileIn.Seek(start);
@in = new UncompressedSplitLineReader(fileIn, job, this.recordDelimiterBytes, split
.GetLength());
filePosition = fileIn;
}
// If this is not the first split, we always throw away first record
// because we always (except the last split) read one extra line in
// next() method.
if (start != 0)
{
start += @in.ReadLine(new Text(), 0, MaxBytesToConsume(start));
}
this.pos = start;
}
/// <exception cref="System.IO.IOException"/>
public LineRecordReader(Configuration job, FileSplit split, byte[] recordDelimiter
)
{
this.maxLineLength = job.GetInt(LineRecordReader.MaxLineLength, int.MaxValue);
start = split.GetStart();
end = start + split.GetLength();
Path file = split.GetPath();
compressionCodecs = new CompressionCodecFactory(job);
codec = compressionCodecs.GetCodec(file);
// open the file and seek to the start of the split
FileSystem fs = file.GetFileSystem(job);
fileIn = fs.Open(file);
if (IsCompressedInput())
{
decompressor = CodecPool.GetDecompressor(codec);
if (codec is SplittableCompressionCodec)
{
SplitCompressionInputStream cIn = ((SplittableCompressionCodec)codec).CreateInputStream
(fileIn, decompressor, start, end, SplittableCompressionCodec.READ_MODE.Byblock);
@in = new CompressedSplitLineReader(cIn, job, recordDelimiter);
start = cIn.GetAdjustedStart();
end = cIn.GetAdjustedEnd();
filePosition = cIn;
}
else
{
// take pos from compressed stream
@in = new SplitLineReader(codec.CreateInputStream(fileIn, decompressor), job, recordDelimiter
);
filePosition = fileIn;
}
}
else
{
fileIn.Seek(start);
@in = new UncompressedSplitLineReader(fileIn, job, recordDelimiter, split.GetLength
());
filePosition = fileIn;
}
// If this is not the first split, we always throw away first record
// because we always (except the last split) read one extra line in
// next() method.
if (start != 0)
{
start += @in.ReadLine(new Text(), 0, MaxBytesToConsume(start));
}
this.pos = start;
}
// This is also called from the old FixedLengthRecordReader API implementation
/// <exception cref="System.IO.IOException"/>
public virtual void Initialize(Configuration job, long splitStart, long splitLength
, Path file)
{
start = splitStart;
end = start + splitLength;
long partialRecordLength = start % recordLength;
long numBytesToSkip = 0;
if (partialRecordLength != 0)
{
numBytesToSkip = recordLength - partialRecordLength;
}
// open the file and seek to the start of the split
FileSystem fs = file.GetFileSystem(job);
fileIn = fs.Open(file);
CompressionCodec codec = new CompressionCodecFactory(job).GetCodec(file);
if (null != codec)
{
isCompressedInput = true;
decompressor = CodecPool.GetDecompressor(codec);
CompressionInputStream cIn = codec.CreateInputStream(fileIn, decompressor);
filePosition = cIn;
inputStream = cIn;
numRecordsRemainingInSplit = long.MaxValue;
Log.Info("Compressed input; cannot compute number of records in the split");
}
else
{
fileIn.Seek(start);
filePosition = fileIn;
inputStream = fileIn;
long splitSize = end - start - numBytesToSkip;
numRecordsRemainingInSplit = (splitSize + recordLength - 1) / recordLength;
if (numRecordsRemainingInSplit < 0)
{
numRecordsRemainingInSplit = 0;
}
Log.Info("Expecting " + numRecordsRemainingInSplit + " records each with a length of "
+ recordLength + " bytes in the split with an effective size of " + splitSize +
" bytes");
}
if (numBytesToSkip != 0)
{
start += inputStream.Skip(numBytesToSkip);
}
this.pos = start;
}
public InMemoryMapOutput(Configuration conf, TaskAttemptID mapId, MergeManagerImpl
<K, V> merger, int size, CompressionCodec codec, bool primaryMapOutput)
: base(mapId, (long)size, primaryMapOutput)
{
// Decompression of map-outputs
this.conf = conf;
this.merger = merger;
this.codec = codec;
byteStream = new BoundedByteArrayOutputStream(size);
memory = byteStream.GetBuffer();
if (codec != null)
{
decompressor = CodecPool.GetDecompressor(codec);
}
else
{
decompressor = null;
}
}
public virtual void TestMultipleClose()
{
Uri testFileUrl = GetType().GetClassLoader().GetResource("recordSpanningMultipleSplits.txt.bz2"
);
NUnit.Framework.Assert.IsNotNull("Cannot find recordSpanningMultipleSplits.txt.bz2"
, testFileUrl);
FilePath testFile = new FilePath(testFileUrl.GetFile());
Path testFilePath = new Path(testFile.GetAbsolutePath());
long testFileSize = testFile.Length();
Configuration conf = new Configuration();
conf.SetInt(LineRecordReader.MaxLineLength, int.MaxValue);
TaskAttemptContext context = new TaskAttemptContextImpl(conf, new TaskAttemptID()
);
// read the data and check whether BOM is skipped
FileSplit split = new FileSplit(testFilePath, 0, testFileSize, null);
LineRecordReader reader = new LineRecordReader();
reader.Initialize(split, context);
//noinspection StatementWithEmptyBody
while (reader.NextKeyValue())
{
}
reader.Close();
reader.Close();
BZip2Codec codec = new BZip2Codec();
codec.SetConf(conf);
ICollection <Decompressor> decompressors = new HashSet <Decompressor>();
for (int i = 0; i < 10; ++i)
{
decompressors.AddItem(CodecPool.GetDecompressor(codec));
}
NUnit.Framework.Assert.AreEqual(10, decompressors.Count);
}