private List<Map.Entry> previewTextBatch(FileStatus fileStatus, int batchSize)
throws IOException, InterruptedException {
TextInputFormat textInputFormat = new TextInputFormat();
long fileLength = fileStatus.getLen();
// Hadoop does unsafe casting from long to int, so split length should not be greater than int
// max value
long splitLength = (fileLength < Integer.MAX_VALUE) ? fileLength : Integer.MAX_VALUE;
InputSplit fileSplit = new FileSplit(fileStatus.getPath(), 0, splitLength, null);
TaskAttemptContext taskAttemptContext =
new TaskAttemptContextImpl(
hadoopConf, TaskAttemptID.forName("attempt_1439420318532_0011_m_000000_0"));
RecordReader<LongWritable, Text> recordReader =
textInputFormat.createRecordReader(fileSplit, taskAttemptContext);
recordReader.initialize(fileSplit, taskAttemptContext);
boolean hasNext = recordReader.nextKeyValue();
List<Map.Entry> batch = new ArrayList<>();
while (hasNext && batch.size() < batchSize) {
batch.add(
new Pair(
fileStatus.getPath().toUri().getPath() + "::" + recordReader.getCurrentKey(),
String.valueOf(recordReader.getCurrentValue())));
hasNext = recordReader.nextKeyValue(); // not like iterator.hasNext, actually advances
}
return batch;
}
void checkFormat(Job job) throws Exception {
TaskAttemptContext attemptContext =
new TaskAttemptContext(job.getConfiguration(), new TaskAttemptID("123", 0, false, 1, 2));
MyClassMessagePackBase64LineInputFormat format = new MyClassMessagePackBase64LineInputFormat();
FileInputFormat.setInputPaths(job, workDir);
List<InputSplit> splits = format.getSplits(job);
for (int j = 0; j < splits.size(); j++) {
RecordReader<LongWritable, MyClassWritable> reader =
format.createRecordReader(splits.get(j), attemptContext);
reader.initialize(splits.get(j), attemptContext);
int count = 0;
try {
while (reader.nextKeyValue()) {
LongWritable key = reader.getCurrentKey();
MyClassWritable val = reader.getCurrentValue();
MyClass mc = val.get();
assertEquals(mc.v, count);
assertEquals(mc.s, Integer.toString(count));
count++;
}
} finally {
reader.close();
}
}
}
/**
* Generate random data, compress it, index and md5 hash the data. Then read it all back and md5
* that too, to verify that it all went ok.
*
* @param testWithIndex Should we index or not?
* @param charsToOutput How many characters of random data should we output.
* @throws IOException
* @throws NoSuchAlgorithmException
* @throws InterruptedException
*/
private void runTest(boolean testWithIndex, int charsToOutput)
throws IOException, NoSuchAlgorithmException, InterruptedException {
Configuration conf = new Configuration();
conf.setLong("fs.local.block.size", charsToOutput / 2);
// reducing block size to force a split of the tiny file
conf.set("io.compression.codecs", LzopCodec.class.getName());
Assume.assumeTrue(CoreTestUtil.okToRunLzoTests(conf));
FileSystem.getLocal(conf).close(); // remove cached filesystem (if any)
FileSystem localFs = FileSystem.getLocal(conf);
localFs.delete(outputDir_, true);
localFs.mkdirs(outputDir_);
Job job = new Job(conf);
TextOutputFormat.setCompressOutput(job, true);
TextOutputFormat.setOutputCompressorClass(job, LzopCodec.class);
TextOutputFormat.setOutputPath(job, outputDir_);
TaskAttemptContext attemptContext =
new TaskAttemptContext(job.getConfiguration(), new TaskAttemptID("123", 0, false, 1, 2));
// create some input data
byte[] expectedMd5 = createTestInput(outputDir_, localFs, attemptContext, charsToOutput);
if (testWithIndex) {
Path lzoFile = new Path(outputDir_, lzoFileName_);
LzoIndex.createIndex(localFs, lzoFile);
}
LzoTextInputFormat inputFormat = new LzoTextInputFormat();
TextInputFormat.setInputPaths(job, outputDir_);
List<InputSplit> is = inputFormat.getSplits(job);
// verify we have the right number of lzo chunks
if (testWithIndex && OUTPUT_BIG == charsToOutput) {
assertEquals(3, is.size());
} else {
assertEquals(1, is.size());
}
// let's read it all and calculate the md5 hash
for (InputSplit inputSplit : is) {
RecordReader<LongWritable, Text> rr =
inputFormat.createRecordReader(inputSplit, attemptContext);
rr.initialize(inputSplit, attemptContext);
while (rr.nextKeyValue()) {
Text value = rr.getCurrentValue();
md5_.update(value.getBytes(), 0, value.getLength());
}
rr.close();
}
localFs.close();
assertTrue(Arrays.equals(expectedMd5, md5_.digest()));
}
private E computeNextKey() throws IOException, InterruptedException {
while (unfiltered.nextKeyValue()) {
E element = unfiltered.getCurrentKey();
if (predicate.apply(element)) {
return element;
}
}
return null;
}
public boolean nextKeyValue() throws IOException, InterruptedException {
while ((curReader == null) || !curReader.nextKeyValue()) {
if (!initNextRecordReader()) {
return false;
}
}
return true;
}
/**
* Read all records from a RecordReader
*
* @param reader RecordReader
* @throws IOException I/O errors
* @throws InterruptedException thread errors
*/
private static long readFully(RecordReader<WritableComparable, HiveReadableRecord> reader)
throws IOException, InterruptedException {
long num = 0;
while (reader.nextKeyValue()) {
HiveReadableRecord record = reader.getCurrentValue();
parseLongLongDouble(record);
++num;
// if (num % 1000000 == 0) {
// System.out.println("Parsed " + num + " rows");
// }
}
return num;
}
@Test
public void testPreferredServerUnreachable() throws Exception {
InfinispanInputSplit invalidSplit = createInfinispanSplit();
Configuration configuration = miniHadoopCluster.getConfiguration();
TaskAttemptContextImpl fakeTaskContext =
new TaskAttemptContextImpl(configuration, new TaskAttemptID());
InfinispanInputFormat<Integer, WebPage> inputFormat = new InfinispanInputFormat<>();
RecordReader<Integer, WebPage> reader =
inputFormat.createRecordReader(invalidSplit, fakeTaskContext);
reader.initialize(invalidSplit, fakeTaskContext);
reader.nextKeyValue();
assertNotNull(reader.getCurrentKey());
}
@Test
public void testVbRecordReader() throws IOException, InterruptedException {
RecordReader reader =
getRecordReader(
testUtils.getCobolFileLocation(),
testUtils.getTestVbFileLocation(),
"0x01",
net.sf.JRecord.Common.Constants.IO_VB);
int counter = 0;
while (reader.nextKeyValue()) {
counter++;
System.out.println(reader.getCurrentKey() + "::\t" + reader.getCurrentValue());
}
assertEquals(testUtils.getTestDataLength(), counter);
}
@Override
public boolean hasNext() {
try {
boolean retVal = curRecReader.nextKeyValue();
if (retVal) {
return true;
}
// if its false, we need to close recordReader.
curRecReader.close();
return false;
} catch (IOException e) {
throw new RuntimeException(e);
} catch (InterruptedException e) {
throw new RuntimeException(e);
}
}
/**
* Randomize the split order, then take the specified number of keys from each split sampled,
* where each key is selected with the specified probability and possibly replaced by a
* subsequently selected key when the quota of keys from that split is satisfied.
*/
@SuppressWarnings("unchecked") // ArrayList::toArray doesn't preserve type
public K[] getSample(InputFormat<K, V> inf, Job job) throws IOException, InterruptedException {
List<InputSplit> splits = inf.getSplits(job);
ArrayList<K> samples = new ArrayList<K>(numSamples);
int splitsToSample = Math.min(maxSplitsSampled, splits.size());
Random r = new Random();
long seed = r.nextLong();
r.setSeed(seed);
// shuffle splits
for (int i = 0; i < splits.size(); ++i) {
InputSplit tmp = splits.get(i);
int j = r.nextInt(splits.size());
splits.set(i, splits.get(j));
splits.set(j, tmp);
}
// our target rate is in terms of the maximum number of sample splits,
// but we accept the possibility of sampling additional splits to hit
// the target sample keyset
for (int i = 0;
i < splitsToSample || (i < splits.size() && samples.size() < numSamples);
++i) {
RecordReader<K, V> reader =
inf.createRecordReader(
splits.get(i), new TaskAttemptContext(job.getConfiguration(), new TaskAttemptID()));
while (reader.nextKeyValue()) {
if (r.nextDouble() <= freq) {
if (samples.size() < numSamples) {
samples.add(reader.getCurrentKey());
} else {
// When exceeding the maximum number of samples, replace a
// random element with this one, then adjust the frequency
// to reflect the possibility of existing elements being
// pushed out
int ind = r.nextInt(numSamples);
if (ind != numSamples) {
samples.set(ind, reader.getCurrentKey());
}
freq *= (numSamples - 1) / (double) numSamples;
}
}
}
reader.close();
}
return (K[]) samples.toArray();
}
@Override
public Tuple getNext() throws IOException {
try {
if (!reader.nextKeyValue()) {
return null;
}
final PhoenixRecord phoenixRecord = reader.getCurrentValue();
if (phoenixRecord == null) {
return null;
}
final Tuple tuple = TypeUtil.transformToTuple(phoenixRecord, schema.getFields());
return tuple;
} catch (InterruptedException e) {
int errCode = 6018;
final String errMsg = "Error while reading input";
throw new ExecException(errMsg, errCode, PigException.REMOTE_ENVIRONMENT, e);
}
}
@Override
public Tuple getNext() throws IOException {
Tuple tuple = null;
try {
if (reader.nextKeyValue()) {
QuadWritable quad = reader.getCurrentValue();
tuple = tupleFactory.newTuple(4);
tuple.set(0, NodeEncoder.asString(quad.getQuad().getGraph()));
tuple.set(1, NodeEncoder.asString(quad.getQuad().getSubject()));
tuple.set(2, NodeEncoder.asString(quad.getQuad().getPredicate()));
tuple.set(3, NodeEncoder.asString(quad.getQuad().getObject()));
}
} catch (InterruptedException e) {
throw new IOException(String.format("Error while reading %s", location));
}
log.debug("getNext() --> {}", tuple);
return tuple;
}
@Override
public Tuple getNext() throws IOException {
try {
List values = new ArrayList();
if (!reader.nextKeyValue()) {
return null;
}
Text value = (Text) reader.getCurrentValue();
// TODO: parse record into component fields, add to values in order
// check that the appropriate index of requiredFields is true before adding
return tupleFactory.newTuple(values);
} catch (InterruptedException e) {
int errCode = 6018;
String errMsg = "Error while reading input";
throw new ExecException(errMsg, errCode, PigException.REMOTE_ENVIRONMENT, e);
}
}
@Override
public Tuple getNext() throws IOException {
try {
if (!reader.nextKeyValue()) {
return null;
}
Text value = (Text) reader.getCurrentValue();
String line = value.toString();
Tuple tuple = tupleFactory.newTuple(ranges.size());
for (int i = 0; i < ranges.size(); i++) {
Range range = ranges.get(i);
if (range.getEnd() > line.length()) {
LOG.warn(
String.format(
"Range end (%s) is longer than line length (%s)", range.getEnd(), line.length()));
continue;
}
tuple.set(i, new DataByteArray(range.getSubstring(line)));
}
return tuple;
} catch (InterruptedException e) {
throw new ExecException(e);
}
}
/**
* For each split sampled, emit when the ratio of the number of records retained to the total
* record count is less than the specified frequency.
*/
@SuppressWarnings("unchecked") // ArrayList::toArray doesn't preserve type
public K[] getSample(InputFormat<K, V> inf, Job job) throws IOException, InterruptedException {
List<InputSplit> splits = inf.getSplits(job);
ArrayList<K> samples = new ArrayList<K>();
int splitsToSample = Math.min(maxSplitsSampled, splits.size());
int splitStep = splits.size() / splitsToSample;
long records = 0;
long kept = 0;
for (int i = 0; i < splitsToSample; ++i) {
RecordReader<K, V> reader =
inf.createRecordReader(
splits.get(i * splitStep),
new TaskAttemptContext(job.getConfiguration(), new TaskAttemptID()));
while (reader.nextKeyValue()) {
++records;
if ((double) kept / records < freq) {
++kept;
samples.add(reader.getCurrentKey());
}
}
reader.close();
}
return (K[]) samples.toArray();
}
/**
* Read the next k,v pair into the head of this object; return true iff the RR and this are
* exhausted.
*/
private boolean next() throws IOException, InterruptedException {
empty = !rr.nextKeyValue();
key = rr.getCurrentKey();
value = rr.getCurrentValue();
return !empty;
}
public void testBinary() throws IOException, InterruptedException {
Configuration conf = new Configuration();
Job job = new Job(conf);
Path outdir = new Path(System.getProperty("test.build.data", "/tmp"), "outseq");
Random r = new Random();
long seed = r.nextLong();
r.setSeed(seed);
FileOutputFormat.setOutputPath(job, outdir);
SequenceFileAsBinaryOutputFormat.setSequenceFileOutputKeyClass(job, IntWritable.class);
SequenceFileAsBinaryOutputFormat.setSequenceFileOutputValueClass(job, DoubleWritable.class);
SequenceFileAsBinaryOutputFormat.setCompressOutput(job, true);
SequenceFileAsBinaryOutputFormat.setOutputCompressionType(job, CompressionType.BLOCK);
BytesWritable bkey = new BytesWritable();
BytesWritable bval = new BytesWritable();
TaskAttemptContext context =
MapReduceTestUtil.createDummyMapTaskAttemptContext(job.getConfiguration());
OutputFormat<BytesWritable, BytesWritable> outputFormat =
new SequenceFileAsBinaryOutputFormat();
OutputCommitter committer = outputFormat.getOutputCommitter(context);
committer.setupJob(job);
RecordWriter<BytesWritable, BytesWritable> writer = outputFormat.getRecordWriter(context);
IntWritable iwritable = new IntWritable();
DoubleWritable dwritable = new DoubleWritable();
DataOutputBuffer outbuf = new DataOutputBuffer();
LOG.info("Creating data by SequenceFileAsBinaryOutputFormat");
try {
for (int i = 0; i < RECORDS; ++i) {
iwritable = new IntWritable(r.nextInt());
iwritable.write(outbuf);
bkey.set(outbuf.getData(), 0, outbuf.getLength());
outbuf.reset();
dwritable = new DoubleWritable(r.nextDouble());
dwritable.write(outbuf);
bval.set(outbuf.getData(), 0, outbuf.getLength());
outbuf.reset();
writer.write(bkey, bval);
}
} finally {
writer.close(context);
}
committer.commitTask(context);
committer.commitJob(job);
InputFormat<IntWritable, DoubleWritable> iformat =
new SequenceFileInputFormat<IntWritable, DoubleWritable>();
int count = 0;
r.setSeed(seed);
SequenceFileInputFormat.setInputPaths(job, outdir);
LOG.info("Reading data by SequenceFileInputFormat");
for (InputSplit split : iformat.getSplits(job)) {
RecordReader<IntWritable, DoubleWritable> reader = iformat.createRecordReader(split, context);
MapContext<IntWritable, DoubleWritable, BytesWritable, BytesWritable> mcontext =
new MapContextImpl<IntWritable, DoubleWritable, BytesWritable, BytesWritable>(
job.getConfiguration(),
context.getTaskAttemptID(),
reader,
null,
null,
MapReduceTestUtil.createDummyReporter(),
split);
reader.initialize(split, mcontext);
try {
int sourceInt;
double sourceDouble;
while (reader.nextKeyValue()) {
sourceInt = r.nextInt();
sourceDouble = r.nextDouble();
iwritable = reader.getCurrentKey();
dwritable = reader.getCurrentValue();
assertEquals(
"Keys don't match: " + "*" + iwritable.get() + ":" + sourceInt + "*",
sourceInt,
iwritable.get());
assertTrue(
"Vals don't match: " + "*" + dwritable.get() + ":" + sourceDouble + "*",
Double.compare(dwritable.get(), sourceDouble) == 0);
++count;
}
} finally {
reader.close();
}
}
assertEquals("Some records not found", RECORDS, count);
}
@Override
public boolean nextKeyValue() throws IOException, InterruptedException {
return originalRR.nextKeyValue();
}
/* (non-Javadoc)
* @see org.apache.pig.builtin.PigStorage#getNext()
*/
@Override
public Tuple getNext() throws IOException {
// If SKIP_INPUT_HEADER and this is the first input split, skip header record
// We store its value as a string though, so we can compare
// further records to it. If they are the same (this would
// happen if multiple small files each with a header were combined
// into one split), we know to skip the duplicate header record as well.
if (loadingFirstRecord
&& headerTreatment == Headers.SKIP_INPUT_HEADER
&& (splitIndex == 0 || splitIndex == -1)) {
try {
if (!in.nextKeyValue()) return null;
header = ((Text) in.getCurrentValue()).toString();
} catch (InterruptedException e) {
int errCode = 6018;
String errMsg = "Error while reading input";
throw new ExecException(errMsg, errCode, PigException.REMOTE_ENVIRONMENT, e);
}
}
loadingFirstRecord = false;
mProtoTuple = new ArrayList<Object>();
getNextInQuotedField = false;
boolean evenQuotesSeen = true;
boolean sawEmbeddedRecordDelimiter = false;
byte[] buf = null;
if (!mRequiredColumnsInitialized) {
if (udfContextSignature != null) {
Properties p = UDFContext.getUDFContext().getUDFProperties(this.getClass());
mRequiredColumns =
(boolean[]) ObjectSerializer.deserialize(p.getProperty(udfContextSignature));
}
mRequiredColumnsInitialized = true;
}
// Note: we cannot factor out the check for nextKeyValue() being null,
// because that call overwrites buf with the new line, which is
// bad if we have a field with a newline.
try {
int recordLen = 0;
getNextFieldID = 0;
while (sawEmbeddedRecordDelimiter || getNextFieldID == 0) {
Text value = null;
if (sawEmbeddedRecordDelimiter) {
// Deal with pulling more records from the input, because
// a double quoted embedded newline was encountered in a field.
// Save the length of the record so far, plus one byte for the
// record delimiter (usually newline) that's embedded in the field
// we were working on before falling into this branch:
int prevLineLen = recordLen + 1;
// Save previous line (the one with the field that has the newline) in a new array.
// The last byte will be random; we'll fill in the embedded
// record delimiter (usually newline) below:
byte[] prevLineSaved = Arrays.copyOf(buf, prevLineLen);
prevLineSaved[prevLineLen - 1] = RECORD_DEL;
// Read the continuation of the record, unless EOF:
if (!in.nextKeyValue()) {
return null;
}
value = (Text) in.getCurrentValue();
recordLen = value.getLength();
// Grab the continuation's bytes:
buf = value.getBytes();
// Combine the previous line and the continuation into a new array.
// The following copyOf() does half the job: it allocates all the
// space, and also copies the previous line into that space:
byte[] prevLineAndContinuation = Arrays.copyOf(prevLineSaved, prevLineLen + recordLen);
// Now append the continuation. Parms: fromBuf, fromStartPos, toBuf, toStartPos,
// lengthToCopy:
System.arraycopy(buf, 0, prevLineAndContinuation, prevLineLen, recordLen);
// We'll work with the combination now:
buf = prevLineAndContinuation;
// Do the whole record over from the start:
mProtoTuple.clear();
getNextInQuotedField = false;
evenQuotesSeen = true;
getNextFieldID = 0;
recordLen = prevLineAndContinuation.length;
} else {
// Previous record finished cleanly: start with the next record,
// unless EOF:
if (!in.nextKeyValue()) {
return null;
}
value = (Text) in.getCurrentValue();
// if the line is a duplicate header and 'SKIP_INPUT_HEADER' is set, ignore it
// (this might happen if multiple files each with a header are combined into a single
// split)
if (headerTreatment == Headers.SKIP_INPUT_HEADER && value.toString().equals(header)) {
if (!in.nextKeyValue()) return null;
value = (Text) in.getCurrentValue();
}
buf = value.getBytes();
getNextFieldID = 0;
recordLen = value.getLength();
}
nextTupleSkipChar = false;
ByteBuffer fieldBuffer = ByteBuffer.allocate(recordLen);
sawEmbeddedRecordDelimiter =
processOneInRecord(evenQuotesSeen, buf, recordLen, fieldBuffer);
// The last field is never delimited by a FIELD_DEL, but by
// the end of the record. So we need to add that last field.
// The '!sawEmbeddedRecordDelimiter' handles the case of
// embedded newlines; we are amidst a field, not at
// the final record:
if (!sawEmbeddedRecordDelimiter) readField(fieldBuffer, getNextFieldID++);
} // end while
} catch (InterruptedException e) {
int errCode = 6018;
String errMsg = "Error while reading input";
throw new ExecException(errMsg, errCode, PigException.REMOTE_ENVIRONMENT, e);
}
Tuple t = mTupleFactory.newTupleNoCopy(mProtoTuple);
return t;
}
