@Override
public Long exec(Tuple input) throws IOException {
try {
String tinput = "";
if (input == null || input.size() == 0) return null;
else {
if (input.getType(0) == DataType.CHARARRAY) tinput = (String) input.get(0);
else
throw new RuntimeException(
"Input type expected to be chararray but got: " + input.getType(0));
}
tinput = tinput.replaceAll("[-+.^:, ]", "");
if (tinput.length() > 14) return Long.parseLong(tinput.substring(0, 14));
else if (tinput.length() < 14)
return Long.parseLong(String.format("%-14s", tinput).replace(' ', '0'));
else return Long.parseLong(tinput);
} catch (ExecException exp) {
throw exp;
} catch (Exception e) {
int errCode = 2107;
String msg = "Error while computing date_format in " + this.getClass().getSimpleName();
throw new ExecException(msg, errCode, PigException.BUG, e);
}
}
@Override
protected void doHadoopWork() throws BuildException {
Tuple tuple = ContextManager.getCurrentTuple();
if (tuple == null) {
throw new BuildException(
this.getTaskName()
+ " should be put inside task container which provides tuple to execution context");
}
try {
if (tuple.getType(fieldNumber) != DataType.TUPLE
|| !(tuple.get(fieldNumber) instanceof Tuple)) {
throw new BuildException("Tuple field " + fieldNumber + " doesn't represent a Tuple");
}
ContextManager.setCurrentTupleContext((Tuple) tuple.get(fieldNumber));
try {
for (Task task : tasks) {
task.perform();
}
} finally {
ContextManager.resetCurrentTupleContext();
}
} catch (ExecException e) {
throw new BuildException("Failed to check type of tuple field " + fieldNumber, e);
}
}
@Override
public Tuple exec(Tuple input) throws IOException {
if (input == null || input.size() == 0) {
return null;
}
Object obj = null;
Integer limnum = null;
try {
obj = (DataByteArray) input.get(1);
} catch (ExecException e) {
logger.error("Error in reading field proto:", e);
throw e;
}
try {
limnum = (Integer) input.get(2);
} catch (ExecException e) {
logger.error("Error in reading baglimit:", e);
throw e;
}
DataByteArray dba = null;
try {
dba = (DataByteArray) obj;
} catch (ClassCastException e) {
logger.error("Error in casting Object (" + input.getType(1) + ") to DataByteArray:", e);
throw e;
}
DocumentMetadata dm = null;
try {
dm = DocumentMetadata.parseFrom(dba.get());
} catch (InvalidProtocolBufferException e) {
logger.error("Error in reading ByteArray to DocumentMetadata:", e);
throw e;
}
String key = dm.getKey();
DataBag db = new DefaultDataBag();
int bagsize = 0;
for (ClassifCode code : dm.getBasicMetadata().getClassifCodeList()) {
for (String co_str : code.getValueList()) {
bagsize++;
db.add(TupleFactory.getInstance().newTuple(co_str));
}
}
if (bagsize > limnum) {
Object[] to = new Object[] {key, db, bagsize};
return TupleFactory.getInstance().newTuple(Arrays.asList(to));
}
return null;
}
@Override
public byte getType(int fieldNum) throws ExecException {
return t.getType(fieldNum);
}