private void set(int[] pos, Type[] types, Tuple tuple, CoercibleType[] coercions) {
verifyModifiable();
if (pos.length != tuple.size())
throw new TupleException(
"given tuple not same size as position array: "
+ pos.length
+ ", tuple: "
+ tuple.print());
int count = 0;
for (int i : pos) {
Object element = tuple.elements.get(count);
if (types != null) {
Type type = types[i];
element = coercions[count].coerce(element, type);
}
elements.set(i, element);
count++;
}
}
/**
* Creates a new Tuple from the given positions, but sets the values in the current tuple to null.
*
* @param pos of type int[]
* @return Tuple
*/
Tuple extract(int[] pos) {
Tuple results = new Tuple();
for (int i : pos) results.add(elements.set(i, null));
return results;
}
public void operate(FlowProcess flowProcess, FunctionCall<NullContext> functionCall) {
String filename = functionCall.getArguments().getString("line");
Tuple result = new Tuple("");
BufferedReader reader;
try {
InputStream stream = new FileInputStream(filename);
reader = new BufferedReader(new InputStreamReader(stream, "us-ascii"));
} catch (UnsupportedEncodingException e) {
throw new RuntimeException("Impossible exception!", e);
} catch (Exception e) {
throw new RuntimeException(String.format("Exception splitting mbox file %s", filename), e);
}
StringBuilder email = new StringBuilder();
for (String curLine = safeReadLine(reader); curLine != null; curLine = safeReadLine(reader)) {
if (curLine.startsWith(MBOX_RECORD_DIVIDER)) {
if (email.length() > 0) {
result.setString(0, email.toString());
functionCall.getOutputCollector().add(result);
}
email.setLength(0);
}
email.append(curLine);
email.append('\n');
}
// Output the final record.
if (email.length() > 0) {
result.setString(0, email.toString());
functionCall.getOutputCollector().add(result);
}
}
/**
* Method append appends all the values of the given Tuple instances to a copy of this instance.
*
* @param tuples of type Tuple
* @return Tuple
*/
public Tuple append(Tuple... tuples) {
Tuple result = new Tuple(this);
for (Tuple tuple : tuples) result.addAll(tuple);
return result;
}
/**
* Method size returns a new Tuple instance of the given size with the given Comparable as its
* element values.
*
* @param size of type int
* @param value of type Comparable
* @return Tuple
*/
public static Tuple size(int size, Comparable value) {
Tuple result = new Tuple();
for (int i = 0; i < size; i++) result.add(value);
return result;
}
@Override
public boolean isRemove(
FlowProcess flowProcess, FilterCall<LinkedHashMap<Tuple, Object>> filterCall) {
// we assume its more painful to create lots of tuple copies vs comparisons
Tuple args = TupleHasher.wrapTuple(tupleHasher, filterCall.getArguments().getTuple());
switch (include) {
case ALL:
break;
case NO_NULLS:
if (Tuples.frequency(args, null) == args.size()) return true;
break;
}
if (filterCall.getContext().containsKey(args)) {
flowProcess.increment(Cache.Num_Keys_Hit, 1);
return true;
}
// only do the copy here
filterCall
.getContext()
.put(TupleHasher.wrapTuple(tupleHasher, filterCall.getArguments().getTupleCopy()), null);
flowProcess.increment(Cache.Num_Keys_Missed, 1);
return false;
}
/**
* Method size returns a new Tuple instance of the given size with the given Comparable as its
* element values.
*
* @param size of type int
* @param value of type Comparable
* @return Tuple
*/
public static Tuple size(int size, Comparable value) {
Tuple result = new Tuple(new ArrayList<Object>(size));
for (int i = 0; i < size; i++) result.add(value);
return result;
}
/**
* @param flowProcess
* @param sourceCall
* @return
* @throws IOException
*/
@Override
public boolean source(
FlowProcess<JobConf> flowProcess, SourceCall<BSONWritable[], RecordReader> sourceCall)
throws IOException {
Tuple result = new Tuple();
BSONWritable key = sourceCall.getContext()[0];
BSONWritable value = sourceCall.getContext()[1];
if (!sourceCall.getInput().next(key, value)) {
logger.info("Nothing left to read, exiting");
return false;
}
for (String columnFieldName : columnFieldNames) {
Object tupleEntry = value.get(columnFieldName);
if (tupleEntry != null) {
result.add(tupleEntry);
} else if (columnFieldName != this.keyColumnName) {
result.add("");
}
}
sourceCall.getIncomingEntry().setTuple(result);
return true;
}
@Override
public void complete(FlowProcess flowProcess, AggregatorCall aggregatorCall) {
Tuple result = new Tuple();
result.add(min);
result.add(max);
aggregatorCall.getOutputCollector().add(result);
}
Tuple nulledCopy(int[] pos) {
if (pos == null) return size(size());
Tuple results = new Tuple(this);
for (int i : pos) results.set(i, null);
return results;
}
/**
* Method get will return a new Tuple instace populated with element values from the given array
* of positions.
*
* @param pos of type int[]
* @return Tuple
*/
public Tuple get(int[] pos) {
if (pos == null || pos.length == 0) return new Tuple(this);
Tuple results = new Tuple();
for (int i : pos) results.add(elements.get(i));
return results;
}
/**
* Sets the values in the given positions to the values from the given Tuple.
*
* @param pos of type int[]
* @param tuple of type Tuple
*/
void set(int[] pos, Tuple tuple) {
verifyModifiable();
if (pos.length != tuple.size())
throw new TupleException(
"given tuple not same size as position array, tuple: " + tuple.print());
int count = 0;
for (int i : pos) elements.set(i, tuple.elements.get(count++));
}
/**
* Method is the inverse of {@link #remove(int[])}.
*
* @param pos of type int[]
* @return Tuple
*/
public Tuple leave(int[] pos) {
verifyModifiable();
Tuple results = remove(pos);
List<Comparable> temp = results.elements;
results.elements = this.elements;
this.elements = temp;
return results;
}
@Test
public void testArgumentSelector() {
Fields declarator = new Fields("1", "2", "3", "4");
Tuple incoming = new Tuple(1, 2, 3, 4);
Fields selector = new Fields("3", "2");
assertTuple(incoming.get(declarator, selector));
int[] pos = incoming.getPos(declarator, selector);
assertTuple(incoming.get(pos));
assertTuple(TupleViews.createNarrow(pos, incoming));
}
/**
* Method put places the values of the given tuple into the positions specified by the fields
* argument. The declarator Fields value declares the fields in this Tuple instance.
*
* @param declarator of type Fields
* @param fields of type Fields
* @param tuple of type Tuple
*/
public void put(Fields declarator, Fields fields, Tuple tuple) {
verifyModifiable();
int[] pos = getPos(declarator, fields);
for (int i = 0; i < pos.length; i++) internalSet(pos[i], tuple.getObject(i));
}
@Test
public void testReplace() throws IOException {
RegexReplace splitter = new RegexReplace(new Fields("words"), "\\s+", "-", true);
Tuple arguments = new Tuple("foo\t bar");
Fields resultFields = Fields.UNKNOWN;
TupleListCollector collector = invokeFunction(splitter, arguments, resultFields);
assertEquals("wrong size", 1, collector.size());
Iterator<Tuple> iterator = collector.iterator();
Tuple tuple = iterator.next();
assertEquals("not equal: tuple.get(0)", "foo-bar", tuple.getObject(0));
}
/**
* Method put places the values of the given tuple into the positions specified by the fields
* argument. The declarator Fields value declares the fields in this Tuple instance.
*
* @param declarator of type Fields
* @param fields of type Fields
* @param tuple of type Tuple
*/
public void put(Fields declarator, Fields fields, Tuple tuple) {
verifyModifiable();
int[] pos = declarator.getPos(fields, size());
for (int i = 0; i < pos.length; i++) elements.set(pos[i], tuple.get(i));
}
private void skipUrls(List<ScoredUrlDatum> urls, UrlStatus status, String traceMsg) {
for (ScoredUrlDatum datum : urls) {
FetchedDatum result = new FetchedDatum(datum);
Tuple tuple = result.getTuple();
tuple.add(status.toString());
_collector.add(BixoPlatform.clone(tuple, _flowProcess));
}
_flowProcess.increment(FetchCounters.URLS_SKIPPED, urls.size());
if (status == UrlStatus.SKIPPED_PER_SERVER_LIMIT) {
_flowProcess.increment(FetchCounters.URLS_SKIPPED_PER_SERVER_LIMIT, urls.size());
}
if ((traceMsg != null) && LOGGER.isTraceEnabled()) {
LOGGER.trace(String.format(traceMsg, urls.size()));
}
}
/**
* Create a new datum with field names defined by <fields>, and field values contained in <tuple>
*
* <p>WARNING - <tuple> will be kept as the data container, so don't call this with a tuple
* provided by a Cascading operation/iterator, as those get reused.
*
* @param fields Names of fields
* @param tuple Data for the datum
*/
public BaseDatum(Fields fields, Tuple tuple) {
if (fields.size() != tuple.size()) {
throw new IllegalArgumentException(
"Size of fields must be the same as the size of the tuple: " + fields + "/" + tuple);
}
_tupleEntry = new TupleEntry(fields, tuple);
}
public void setTuple(Tuple tuple) {
if (getFields().size() != tuple.size()) {
throw new IllegalArgumentException("Size of tuple doesn't match current fields");
}
_tupleEntry.setTuple(tuple);
reset();
}
@Test
public void testParserDeclared6() throws IOException {
RegexParser splitter = new RegexParser(new Fields("lhs"), "(\\S+)\\s+\\S+", new int[] {1});
Tuple arguments = new Tuple("foo\tbar");
Fields resultFields = Fields.size(1);
TupleListCollector collector = invokeFunction(splitter, arguments, resultFields);
assertEquals("wrong size", 1, collector.size());
Iterator<Tuple> iterator = collector.iterator();
Tuple tuple = iterator.next();
assertEquals("wrong tupel size", 1, tuple.size());
assertEquals("not equal: tuple.get(0)", "foo", tuple.getObject(0));
}
@Test
public void testSplitter() throws IOException {
RegexSplitter splitter = new RegexSplitter("\t");
Tuple arguments = new Tuple("foo\tbar");
Fields resultFields = Fields.UNKNOWN;
TupleListCollector collector = invokeFunction(splitter, arguments, resultFields);
assertEquals("wrong size", 1, collector.size());
Iterator<Tuple> iterator = collector.iterator();
Tuple tuple = iterator.next();
assertEquals("not equal: tuple.get(0)", "foo", tuple.getObject(0));
assertEquals("not equal: tuple.get(1)", "bar", tuple.getObject(1));
}
/** Contributed by gicode */
@Test
public void testParserDeclared5() throws IOException {
RegexParser splitter = new RegexParser(new Fields("bar"), "^GET /foo\\?bar=([^\\&]+)&");
Tuple arguments = new Tuple("GET /foo?bar=z123&baz=2");
Fields resultFields = Fields.size(1);
TupleListCollector collector = invokeFunction(splitter, arguments, resultFields);
assertEquals("wrong size", 1, collector.size());
Iterator<Tuple> iterator = collector.iterator();
Tuple tuple = iterator.next();
assertEquals("wrong tuple size", 1, tuple.size());
assertEquals("not equal: tuple.get(0)", "z123", tuple.getObject(0));
}
public static <T> void populateOutputTupleEntry(
CombinerDefinition<T> definition, TupleEntry output, Tuple resultTuple) {
// set the ID so we can differentiate later
output.setRaw(MultiCombiner.ID_FIELD, definition.getId());
// our tuples are of the form groupFields+outputFields, set the TupleEntry fields appropriately
Fields groupFields = definition.getGroupFields();
int index = 0;
for (int i = 0; i < groupFields.size(); i++) {
output.setRaw(groupFields.get(i), resultTuple.getObject(index));
index++;
}
Fields outputFields = definition.getOutputFields();
for (int i = 0; i < outputFields.size(); i++) {
output.setRaw(outputFields.get(i), resultTuple.getObject(index));
index++;
}
}
/**
* Method matchWholeTuple ...
*
* @param matcher
* @param input of type Tuple @return boolean
*/
protected boolean matchWholeTuple(Matcher matcher, Tuple input) {
matcher.reset(input.toString("\t", false));
boolean matchFound = matcher.find();
LOG.debug("pattern: {}, matches: {}", getPatternString(), matchFound);
return matchFound == negateMatch;
}
public Tuple source(Map<String, Object> settings, Object boxedKey, Object boxedColumns)
throws IOException {
SortedMap<ByteBuffer, IColumn> columns = (SortedMap<ByteBuffer, IColumn>) boxedColumns;
ByteBuffer key = (ByteBuffer) boxedKey;
Tuple result = new Tuple();
result.add(ByteBufferUtil.string(key));
Map<String, String> dataTypes = SettingsHelper.getTypes(settings);
List<String> sourceMappings = SettingsHelper.getSourceMappings(settings);
Map<String, IColumn> columnsByStringName = new HashMap<String, IColumn>();
for (ByteBuffer columnName : columns.keySet()) {
String stringName = ByteBufferUtil.string(columnName);
logger.debug("column name: {}", stringName);
IColumn col = columns.get(columnName);
logger.debug("column: {}", col);
columnsByStringName.put(stringName, col);
}
for (String columnName : sourceMappings) {
AbstractType columnValueType = SerializerHelper.inferType(dataTypes.get(columnName));
if (columnValueType != null) {
try {
IColumn column = columnsByStringName.get(columnName);
ByteBuffer serializedVal = column.value();
Object val = null;
if (serializedVal != null) {
val = SerializerHelper.deserialize(serializedVal, columnValueType);
}
logger.debug("Putting deserialized column: {}. {}", columnName, val);
result.add(val);
} catch (Exception e) {
throw new RuntimeException("Couldn't deserialize column: " + columnName, e);
}
} else {
throw new RuntimeException("no type given for column: " + columnName);
}
}
return result;
}
// where outgoing fields are ALL
@Test
public void testSelectorAll() {
// if( getOutputSelector().isAll() )
// return inputEntry.getTuple().append( output );
Fields incomingFields = new Fields("1", "2", "3", "4");
Tuple incoming = new Tuple(1, 2, 3, 4);
Fields resultFields = new Fields("5", "6", "7");
Tuple result = new Tuple(5, 6, 7);
Tuple view = TupleViews.createComposite(incoming, result);
assertEquals(new Tuple(1, 2, 3, 4, 5, 6, 7), view);
assertEquals(new Tuple(1, 2, 3, 4, 5, 6, 7), new Tuple(view));
Fields allFields = Fields.join(incomingFields, resultFields);
Fields selector = new Fields("3", "2");
assertTuple(view.get(allFields, selector));
}
/**
* Method remove removes the values specified by the given pos array and returns a new Tuple
* containing the removed values.
*
* @param pos of type int[]
* @return Tuple
*/
public Tuple remove(int[] pos) {
verifyModifiable();
// calculate offsets to apply when removing values from elements
int offset[] = new int[pos.length];
for (int i = 0; i < pos.length; i++) {
offset[i] = 0;
for (int j = 0; j < i; j++) {
if (pos[j] < pos[i]) offset[i]++;
}
}
Tuple results = new Tuple();
for (int i = 0; i < pos.length; i++) results.add(elements.remove(pos[i] - offset[i]));
return results;
}
/**
* Method parse will parse the {@link #print()} String representation of a Tuple instance and
* return a new Tuple instance.
*
* @param string of type String
* @return Tuple
*/
public static Tuple parse(String string) {
if (string == null || string.length() == 0) return null;
string = string.replaceAll("^ *\\[*", "");
string = string.replaceAll("\\]* *$", "");
Scanner scanner = new Scanner(new StringReader(string));
scanner.useDelimiter("(' *, *')|(^ *')|(' *$)");
Tuple result = new Tuple();
while (scanner.hasNext()) {
if (scanner.hasNextInt()) result.add(scanner.nextInt());
else if (scanner.hasNextDouble()) result.add(scanner.nextDouble());
else result.add(scanner.next());
}
scanner.close();
return result;
}
private StringBuffer print(StringBuffer buffer) {
buffer.append("[");
for (int i = 0; i < elements.size(); i++) {
Comparable element = elements.get(i);
if (element instanceof Tuple) ((Tuple) element).print(buffer);
else buffer.append("\'").append(element).append("\'");
if (i < elements.size() - 1) buffer.append(", ");
}
buffer.append("]");
return buffer;
}
