private static List<Text> readSplit(TextInputFormat format, InputSplit split, JobConf job) throws IOException { List<Text> result = new ArrayList<Text>(); RecordReader<LongWritable, Text> reader = format.getRecordReader(split, job, voidReporter); LongWritable key = reader.createKey(); Text value = reader.createValue(); while (reader.next(key, value)) { result.add(value); value = reader.createValue(); } reader.close(); return result; }
public void testInputFormat() { try { JobConf conf = new JobConf(); String TMP_DIR = System.getProperty("test.build.data", "/tmp"); Path filename = new Path("file:///" + TMP_DIR + "/tmpSeqFile"); SequenceFile.Writer sfw = SequenceFile.createWriter( FileSystem.getLocal(conf), conf, filename, ChukwaArchiveKey.class, ChunkImpl.class, SequenceFile.CompressionType.NONE, Reporter.NULL); StringBuilder buf = new StringBuilder(); int offsets[] = new int[lines.length]; for (int i = 0; i < lines.length; ++i) { buf.append(lines[i]); buf.append("\n"); offsets[i] = buf.length() - 1; } ChukwaArchiveKey key = new ChukwaArchiveKey(0, "datatype", "sname", 0); ChunkImpl val = new ChunkImpl("datatype", "sname", 0, buf.toString().getBytes(), null); val.setRecordOffsets(offsets); sfw.append(key, val); sfw.append(key, val); // write it twice sfw.close(); long len = FileSystem.getLocal(conf).getFileStatus(filename).getLen(); InputSplit split = new FileSplit(filename, 0, len, (String[]) null); ChukwaInputFormat in = new ChukwaInputFormat(); RecordReader<LongWritable, Text> r = in.getRecordReader(split, conf, Reporter.NULL); LongWritable l = r.createKey(); Text line = r.createValue(); for (int i = 0; i < lines.length * 2; ++i) { boolean succeeded = r.next(l, line); assertTrue(succeeded); assertEquals(i, l.get()); assertEquals(lines[i % lines.length], line.toString()); System.out.println("read line: " + l.get() + " " + line); } boolean succeeded = r.next(l, line); assertFalse(succeeded); } catch (IOException e) { e.printStackTrace(); fail("IO exception " + e); } }
public V createValue() { return rawIn.createValue(); }