List of usage examples for org.apache.hadoop.io.compress CompressionCodec createInputStream
CompressionInputStream createInputStream(InputStream in, Decompressor decompressor) throws IOException;
From source file:ca.sparkera.adapters.mapreduce.MainframeVBRecordReader.java
License:Apache License
public void initialize(Configuration job, long splitStart, long splitLength, Path file) throws IOException { start = splitStart;/* w w w .ja v a 2s .co m*/ end = start + splitLength; LOG.info("Start of the split:" + start + "-End of split:" + end); LOG.debug("VLR initialize started: start pos:" + start + "endpos:" + end); // open the file and seek to the start of the split final 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; LOG.info("Compressed input; cannot compute number of records in the split"); } else { fileIn.seek(start); filePosition = fileIn; inputStream = fileIn; numBytesRemainingInSplit = splitLength; LOG.info("Variable length input; cannot compute number of records in the split"); } this.pos = start; }
From source file:com.bonc.mr_roamRecognition_hjpt.comm.PathRecordReader.java
License:Apache License
public void initialize(InputSplit genericSplit, TaskAttemptContext context) throws IOException { FileSplit split = (FileSplit) genericSplit; Configuration job = context.getConfiguration(); this.maxLineLength = job.getInt(MAX_LINE_LENGTH, Integer.MAX_VALUE); start = split.getStart();//w w w. j av a 2 s . c om end = start + split.getLength(); final Path file = split.getPath(); path = split.getPath().toString(); // open the file and seek to the start of the split final 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 instanceof SplittableCompressionCodec) { final 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 SplitLineReader(fileIn, job, this.recordDelimiterBytes); 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; }
From source file:com.cloudera.ByteBufferRecordReader.java
License:Apache License
private void initialize(Configuration job, long splitStart, long splitLength, Path file) throws IOException { start = splitStart;//ww w . ja v a 2 s. com end = start + splitLength; pos = start; // open the file and seek to the start of the split final FileSystem fs = file.getFileSystem(job); fileIn = fs.open(file); this.readStats = new ReadStatistics(); this.bufferPool = new ElasticByteBufferPool(); boolean skipChecksums = job.getBoolean("bytecount.skipChecksums", false); this.readOption = skipChecksums ? EnumSet.of(ReadOption.SKIP_CHECKSUMS) : EnumSet.noneOf(ReadOption.class); 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; LOG.info("Compressed input; cannot compute number of records in the split"); } else { fileIn.seek(start); filePosition = fileIn; inputStream = fileIn; LOG.info("Split pos = " + start + " length " + splitLength); } }
From source file:com.cloudera.sqoop.TestExport.java
License:Apache License
private void verifyCompressedFile(Path f, int expectedNumLines) throws IOException { Configuration conf = new Configuration(); if (!BaseSqoopTestCase.isOnPhysicalCluster()) { conf.set(CommonArgs.FS_DEFAULT_NAME, CommonArgs.LOCAL_FS); }/*from ww w . ja v a2s.co m*/ FileSystem fs = FileSystem.get(conf); InputStream is = fs.open(f); CompressionCodecFactory ccf = new CompressionCodecFactory(conf); CompressionCodec codec = ccf.getCodec(f); LOG.info("gzip check codec is " + codec); Decompressor decompressor = CodecPool.getDecompressor(codec); if (null == decompressor) { LOG.info("Verifying gzip sanity with null decompressor"); } else { LOG.info("Verifying gzip sanity with decompressor: " + decompressor.toString()); } is = codec.createInputStream(is, decompressor); BufferedReader r = new BufferedReader(new InputStreamReader(is)); int numLines = 0; while (true) { String ln = r.readLine(); if (ln == null) { break; } numLines++; } r.close(); assertEquals("Did not read back correct number of lines", expectedNumLines, numLines); LOG.info("gzip sanity check returned " + numLines + " lines; ok."); }
From source file:com.datascience.hadoop.CsvInputFormat.java
License:Apache License
@Override public RecordReader<LongWritable, ListWritable<Text>> getRecordReader(InputSplit inputSplit, JobConf conf, Reporter reporter) throws IOException { String charsetName = conf.get(CHARSET); Charset charset = charsetName != null ? Charset.forName(charsetName) : StandardCharsets.UTF_8; FileSplit split = (FileSplit) inputSplit; Path path = split.getPath();//w w w .j ava2 s . c o m FileSystem fs = path.getFileSystem(conf); InputStream is = fs.open(path); // If the input is compressed, load the compression codec. CompressionCodecFactory codecFactory = new CompressionCodecFactory(conf); CompressionCodec codec = codecFactory.getCodec(path); if (codec != null) { Decompressor decompressor = CodecPool.getDecompressor(codec); is = codec.createInputStream(is, decompressor); } return new CsvRecordReader(new InputStreamReader(is, charset), createFormat(conf), split.getLength(), conf.getBoolean(STRICT_MODE, true)); }
From source file:com.flipkart.fdp.migration.distcp.core.MirrorUtils.java
License:Apache License
public static InputStream getCodecInputStream(Configuration conf, String path, InputStream in) throws IOException { CompressionCodecFactory compressionCodecs = new CompressionCodecFactory(conf); CompressionCodec codec = compressionCodecs.getCodec(new Path(path)); if (codec == null) return in; System.out.println("Getting InputStream : " + codec.getDefaultExtension()); System.out.println("Getting InputStream : " + codec); Decompressor compressor = codec.createDecompressor(); in = codec.createInputStream(in, compressor); return in;/*from w ww .j a va 2 s. c o m*/ }
From source file:com.hadoop.compression.lzo.LzoIndex.java
License:Open Source License
/** * Index an lzo file to allow the input format to split them into separate map * jobs./*from ww w . jav a 2 s. co m*/ * * @param fs File system that contains the file. * @param lzoFile the lzo file to index. For filename.lzo, the created index file will be * filename.lzo.index. * @throws IOException */ public static void createIndex(FileSystem fs, Path lzoFile) throws IOException { Configuration conf = fs.getConf(); CompressionCodecFactory factory = new CompressionCodecFactory(conf); CompressionCodec codec = factory.getCodec(lzoFile); if (null == codec) { throw new IOException("Could not find codec for file " + lzoFile + " - you may need to add the LZO codec to your io.compression.codecs " + "configuration in core-site.xml"); } ((Configurable) codec).setConf(conf); FSDataInputStream is = null; FSDataOutputStream os = null; Path outputFile = lzoFile.suffix(LZO_INDEX_SUFFIX); Path tmpOutputFile = lzoFile.suffix(LZO_TMP_INDEX_SUFFIX); // Track whether an exception was thrown or not, so we know to either // delete the tmp index file on failure, or rename it to the new index file on success. boolean indexingSucceeded = false; try { is = fs.open(lzoFile); os = fs.create(tmpOutputFile); LzopDecompressor decompressor = (LzopDecompressor) codec.createDecompressor(); // Solely for reading the header codec.createInputStream(is, decompressor); int numCompressedChecksums = decompressor.getCompressedChecksumsCount(); int numDecompressedChecksums = decompressor.getDecompressedChecksumsCount(); while (true) { // read and ignore, we just want to get to the next int int uncompressedBlockSize = is.readInt(); if (uncompressedBlockSize == 0) { break; } else if (uncompressedBlockSize < 0) { throw new EOFException(); } int compressedBlockSize = is.readInt(); if (compressedBlockSize <= 0) { throw new IOException("Could not read compressed block size"); } // See LzopInputStream.getCompressedData boolean isUncompressedBlock = (uncompressedBlockSize == compressedBlockSize); int numChecksumsToSkip = isUncompressedBlock ? numDecompressedChecksums : numDecompressedChecksums + numCompressedChecksums; long pos = is.getPos(); // write the pos of the block start os.writeLong(pos - 8); // seek to the start of the next block, skip any checksums is.seek(pos + compressedBlockSize + (4 * numChecksumsToSkip)); } // If we're here, indexing was successful. indexingSucceeded = true; } finally { // Close any open streams. if (is != null) { is.close(); } if (os != null) { os.close(); } if (!indexingSucceeded) { // If indexing didn't succeed (i.e. an exception was thrown), clean up after ourselves. fs.delete(tmpOutputFile, false); } else { // Otherwise, rename filename.lzo.index.tmp to filename.lzo.index. fs.rename(tmpOutputFile, outputFile); } } }
From source file:com.hadoop.mapreduce.LzoTextInputFormat.java
License:Open Source License
/** * Index an lzo file to allow the input format to split them into separate map * jobs.// w w w .j av a 2s .c o m * * @param fs * File system that contains the file. * @param lzoFile * the lzo file to index. * @throws IOException */ public static void createIndex(FileSystem fs, Path lzoFile) throws IOException { Configuration conf = fs.getConf(); CompressionCodecFactory factory = new CompressionCodecFactory(fs.getConf()); CompressionCodec codec = factory.getCodec(lzoFile); ((Configurable) codec).setConf(conf); InputStream lzoIs = null; FSDataOutputStream os = null; Path outputFile = new Path(lzoFile.toString() + LzoTextInputFormat.LZO_INDEX_SUFFIX); Path tmpOutputFile = outputFile.suffix(".tmp"); try { FSDataInputStream is = fs.open(lzoFile); os = fs.create(tmpOutputFile); LzopDecompressor decompressor = (LzopDecompressor) codec.createDecompressor(); // for reading the header lzoIs = codec.createInputStream(is, decompressor); int numChecksums = decompressor.getChecksumsCount(); while (true) { // read and ignore, we just want to get to the next int int uncompressedBlockSize = is.readInt(); if (uncompressedBlockSize == 0) { break; } else if (uncompressedBlockSize < 0) { throw new EOFException(); } int compressedBlockSize = is.readInt(); if (compressedBlockSize <= 0) { throw new IOException("Could not read compressed block size"); } long pos = is.getPos(); // write the pos of the block start os.writeLong(pos - 8); // seek to the start of the next block, skip any checksums is.seek(pos + compressedBlockSize + (4 * numChecksums)); } } finally { if (lzoIs != null) { lzoIs.close(); } if (os != null) { os.close(); } } fs.rename(tmpOutputFile, outputFile); }
From source file:com.hadoop.mapreduce.TestLzoLazyLoading.java
License:Open Source License
public static String readFile(String name) throws IOException { Path file = new Path(TEST_ROOT_DIR + "/" + name); CompressionCodec codec = new CompressionCodecFactory(conf).getCodec(file); InputStream f;//from w w w. java 2 s . co m Decompressor decompressor = null; if (codec == null) { f = localFs.open(file); } else { decompressor = CodecPool.getDecompressor(codec); f = codec.createInputStream(localFs.open(file), decompressor); } BufferedReader b = new BufferedReader(new InputStreamReader(f)); StringBuilder result = new StringBuilder(); String line = b.readLine(); while (line != null) { result.append(line); result.append('\n'); line = b.readLine(); } b.close(); if (decompressor != null) { CodecPool.returnDecompressor(decompressor); } return result.toString(); }
From source file:com.mycompany.keywordsearch.LineRecordReaderV2.java
public void initialize(InputSplit genericSplit, TaskAttemptContext context) throws IOException { FileSplit split = (FileSplit) genericSplit; Configuration job = context.getConfiguration(); this.maxLineLength = job.getInt(MAX_LINE_LENGTH, Integer.MAX_VALUE); start = split.getStart();//from www . j a v a2 s .co m end = start + split.getLength(); final Path file = split.getPath(); // open the file and seek to the start of the split final FileSystem fs = file.getFileSystem(job); fileIn = fs.open(file); locationKey.set(file.toString()); CompressionCodec codec = new CompressionCodecFactory(job).getCodec(file); if (null != codec) { isCompressedInput = true; decompressor = CodecPool.getDecompressor(codec); if (codec instanceof SplittableCompressionCodec) { final 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 SplitLineReader(fileIn, job, this.recordDelimiterBytes); 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; }