use of org.apache.hadoop.hive.serde2.io.TimestampWritable in project hive by apache.
the class GenericUDFTrunc method evaluateDate.
private Object evaluateDate(DeferredObject[] arguments) throws UDFArgumentLengthException, HiveException, UDFArgumentTypeException, UDFArgumentException {
if (arguments.length != 2) {
throw new UDFArgumentLengthException("trunc() requires 2 argument, got " + arguments.length);
}
if (arguments[0].get() == null || arguments[1].get() == null) {
return null;
}
if (textConverter2 != null) {
fmtInput = textConverter2.convert(arguments[1].get()).toString();
}
Date date;
switch(inputType1) {
case STRING:
String dateString = textConverter1.convert(arguments[0].get()).toString();
try {
date = formatter.parse(dateString.toString());
} catch (ParseException e) {
return null;
}
break;
case TIMESTAMP:
Timestamp ts = ((TimestampWritable) timestampConverter.convert(arguments[0].get())).getTimestamp();
date = ts;
break;
case DATE:
DateWritable dw = (DateWritable) dateWritableConverter.convert(arguments[0].get());
date = dw.get();
break;
default:
throw new UDFArgumentTypeException(0, "TRUNC() only takes STRING/TIMESTAMP/DATEWRITABLE types, got " + inputType1);
}
if (evalDate(date) == null) {
return null;
}
Date newDate = calendar.getTime();
output.set(formatter.format(newDate));
return output;
}
use of org.apache.hadoop.hive.serde2.io.TimestampWritable in project hive by apache.
the class GenericUDFFromUtcTimestamp method evaluate.
@Override
public Object evaluate(DeferredObject[] arguments) throws HiveException {
Object o0 = arguments[0].get();
if (o0 == null) {
return null;
}
Object o1 = arguments[1].get();
if (o1 == null) {
return null;
}
Object converted_o0 = timestampConverter.convert(o0);
if (converted_o0 == null) {
return null;
}
Timestamp inputTs = ((TimestampWritable) converted_o0).getTimestamp();
String tzStr = textConverter.convert(o1).toString();
TimeZone timezone = TimeZone.getTimeZone(tzStr);
TimeZone fromTz;
TimeZone toTz;
if (invert()) {
fromTz = timezone;
toTz = tzUTC;
} else {
fromTz = tzUTC;
toTz = timezone;
}
// inputTs is the year/month/day/hour/minute/second in the local timezone.
// For this UDF we want it in the timezone represented by fromTz
Timestamp fromTs = timestampFromString(inputTs.toString(), fromTz);
if (fromTs == null) {
return null;
}
// Now output this timestamp's millis value to the equivalent toTz.
dateFormat.setTimeZone(toTz);
Timestamp result = Timestamp.valueOf(dateFormat.format(fromTs));
if (inputTs.getNanos() != 0) {
result.setNanos(inputTs.getNanos());
}
return result;
}
use of org.apache.hadoop.hive.serde2.io.TimestampWritable in project hive by apache.
the class TestVectorSerDeRow method serializeRow.
private Output serializeRow(Object[] row, VectorRandomRowSource source, SerializeWrite serializeWrite) throws HiveException, IOException {
Output output = new Output();
serializeWrite.set(output);
PrimitiveTypeInfo[] primitiveTypeInfos = source.primitiveTypeInfos();
for (int i = 0; i < primitiveTypeInfos.length; i++) {
Object object = row[i];
PrimitiveCategory primitiveCategory = primitiveTypeInfos[i].getPrimitiveCategory();
switch(primitiveCategory) {
case BOOLEAN:
{
BooleanWritable expectedWritable = (BooleanWritable) object;
boolean value = expectedWritable.get();
serializeWrite.writeBoolean(value);
}
break;
case BYTE:
{
ByteWritable expectedWritable = (ByteWritable) object;
byte value = expectedWritable.get();
serializeWrite.writeByte(value);
}
break;
case SHORT:
{
ShortWritable expectedWritable = (ShortWritable) object;
short value = expectedWritable.get();
serializeWrite.writeShort(value);
}
break;
case INT:
{
IntWritable expectedWritable = (IntWritable) object;
int value = expectedWritable.get();
serializeWrite.writeInt(value);
}
break;
case LONG:
{
LongWritable expectedWritable = (LongWritable) object;
long value = expectedWritable.get();
serializeWrite.writeLong(value);
}
break;
case DATE:
{
DateWritable expectedWritable = (DateWritable) object;
Date value = expectedWritable.get();
serializeWrite.writeDate(value);
}
break;
case FLOAT:
{
FloatWritable expectedWritable = (FloatWritable) object;
float value = expectedWritable.get();
serializeWrite.writeFloat(value);
}
break;
case DOUBLE:
{
DoubleWritable expectedWritable = (DoubleWritable) object;
double value = expectedWritable.get();
serializeWrite.writeDouble(value);
}
break;
case STRING:
{
Text text = (Text) object;
serializeWrite.writeString(text.getBytes(), 0, text.getLength());
}
break;
case CHAR:
{
HiveCharWritable expectedWritable = (HiveCharWritable) object;
HiveChar value = expectedWritable.getHiveChar();
serializeWrite.writeHiveChar(value);
}
break;
case VARCHAR:
{
HiveVarcharWritable expectedWritable = (HiveVarcharWritable) object;
HiveVarchar value = expectedWritable.getHiveVarchar();
serializeWrite.writeHiveVarchar(value);
}
break;
case BINARY:
{
BytesWritable expectedWritable = (BytesWritable) object;
byte[] bytes = expectedWritable.getBytes();
int length = expectedWritable.getLength();
serializeWrite.writeBinary(bytes, 0, length);
}
break;
case TIMESTAMP:
{
TimestampWritable expectedWritable = (TimestampWritable) object;
Timestamp value = expectedWritable.getTimestamp();
serializeWrite.writeTimestamp(value);
}
break;
case INTERVAL_YEAR_MONTH:
{
HiveIntervalYearMonthWritable expectedWritable = (HiveIntervalYearMonthWritable) object;
HiveIntervalYearMonth value = expectedWritable.getHiveIntervalYearMonth();
serializeWrite.writeHiveIntervalYearMonth(value);
}
break;
case INTERVAL_DAY_TIME:
{
HiveIntervalDayTimeWritable expectedWritable = (HiveIntervalDayTimeWritable) object;
HiveIntervalDayTime value = expectedWritable.getHiveIntervalDayTime();
serializeWrite.writeHiveIntervalDayTime(value);
}
break;
case DECIMAL:
{
HiveDecimalWritable expectedWritable = (HiveDecimalWritable) object;
HiveDecimal value = expectedWritable.getHiveDecimal();
serializeWrite.writeHiveDecimal(value, ((DecimalTypeInfo) primitiveTypeInfos[i]).scale());
}
break;
default:
throw new HiveException("Unexpected primitive category " + primitiveCategory);
}
}
return output;
}
use of org.apache.hadoop.hive.serde2.io.TimestampWritable in project hive by apache.
the class TestVectorSerDeRow method deserializeAndVerify.
void deserializeAndVerify(Output output, DeserializeRead deserializeRead, VectorRandomRowSource source, Object[] expectedRow) throws HiveException, IOException {
deserializeRead.set(output.getData(), 0, output.getLength());
PrimitiveCategory[] primitiveCategories = source.primitiveCategories();
for (int i = 0; i < primitiveCategories.length; i++) {
Object expected = expectedRow[i];
PrimitiveCategory primitiveCategory = primitiveCategories[i];
PrimitiveTypeInfo primitiveTypeInfo = source.primitiveTypeInfos()[i];
if (!deserializeRead.readNextField()) {
throw new HiveException("Unexpected NULL when reading primitiveCategory " + primitiveCategory + " expected (" + expected.getClass().getName() + ", " + expected.toString() + ") " + " deserializeRead " + deserializeRead.getClass().getName());
}
switch(primitiveCategory) {
case BOOLEAN:
{
Boolean value = deserializeRead.currentBoolean;
BooleanWritable expectedWritable = (BooleanWritable) expected;
if (!value.equals(expectedWritable.get())) {
TestCase.fail("Boolean field mismatch (expected " + expected + " found " + value + ")");
}
}
break;
case BYTE:
{
Byte value = deserializeRead.currentByte;
ByteWritable expectedWritable = (ByteWritable) expected;
if (!value.equals(expectedWritable.get())) {
TestCase.fail("Byte field mismatch (expected " + (int) expected + " found " + (int) value + ")");
}
}
break;
case SHORT:
{
Short value = deserializeRead.currentShort;
ShortWritable expectedWritable = (ShortWritable) expected;
if (!value.equals(expectedWritable.get())) {
TestCase.fail("Short field mismatch (expected " + expected + " found " + value + ")");
}
}
break;
case INT:
{
Integer value = deserializeRead.currentInt;
IntWritable expectedWritable = (IntWritable) expected;
if (!value.equals(expectedWritable.get())) {
TestCase.fail("Int field mismatch (expected " + expected + " found " + value + ")");
}
}
break;
case LONG:
{
Long value = deserializeRead.currentLong;
LongWritable expectedWritable = (LongWritable) expected;
if (!value.equals(expectedWritable.get())) {
TestCase.fail("Long field mismatch (expected " + expected + " found " + value + ")");
}
}
break;
case DATE:
{
DateWritable value = deserializeRead.currentDateWritable;
DateWritable expectedWritable = (DateWritable) expected;
if (!value.equals(expectedWritable)) {
TestCase.fail("Date field mismatch (expected " + expected.toString() + " found " + value.toString() + ")");
}
}
break;
case FLOAT:
{
Float value = deserializeRead.currentFloat;
FloatWritable expectedWritable = (FloatWritable) expected;
if (!value.equals(expectedWritable.get())) {
TestCase.fail("Float field mismatch (expected " + expected + " found " + value + ")");
}
}
break;
case DOUBLE:
{
Double value = deserializeRead.currentDouble;
DoubleWritable expectedWritable = (DoubleWritable) expected;
if (!value.equals(expectedWritable.get())) {
TestCase.fail("Double field mismatch (expected " + expected + " found " + value + ")");
}
}
break;
case STRING:
case CHAR:
case VARCHAR:
case BINARY:
{
byte[] stringBytes = Arrays.copyOfRange(deserializeRead.currentBytes, deserializeRead.currentBytesStart, deserializeRead.currentBytesStart + deserializeRead.currentBytesLength);
Text text = new Text(stringBytes);
String string = text.toString();
switch(primitiveCategory) {
case STRING:
{
Text expectedWritable = (Text) expected;
if (!string.equals(expectedWritable.toString())) {
TestCase.fail("String field mismatch (expected '" + expectedWritable.toString() + "' found '" + string + "')");
}
}
break;
case CHAR:
{
HiveChar hiveChar = new HiveChar(string, ((CharTypeInfo) primitiveTypeInfo).getLength());
HiveCharWritable expectedWritable = (HiveCharWritable) expected;
if (!hiveChar.equals(expectedWritable.getHiveChar())) {
TestCase.fail("Char field mismatch (expected '" + expectedWritable.getHiveChar() + "' found '" + hiveChar + "')");
}
}
break;
case VARCHAR:
{
HiveVarchar hiveVarchar = new HiveVarchar(string, ((VarcharTypeInfo) primitiveTypeInfo).getLength());
HiveVarcharWritable expectedWritable = (HiveVarcharWritable) expected;
if (!hiveVarchar.equals(expectedWritable.getHiveVarchar())) {
TestCase.fail("Varchar field mismatch (expected '" + expectedWritable.getHiveVarchar() + "' found '" + hiveVarchar + "')");
}
}
break;
case BINARY:
{
BytesWritable expectedWritable = (BytesWritable) expected;
if (stringBytes.length != expectedWritable.getLength()) {
TestCase.fail("Byte Array field mismatch (expected " + expected + " found " + stringBytes + ")");
}
byte[] expectedBytes = expectedWritable.getBytes();
for (int b = 0; b < stringBytes.length; b++) {
if (stringBytes[b] != expectedBytes[b]) {
TestCase.fail("Byte Array field mismatch (expected " + expected + " found " + stringBytes + ")");
}
}
}
break;
default:
throw new HiveException("Unexpected primitive category " + primitiveCategory);
}
}
break;
case DECIMAL:
{
HiveDecimal value = deserializeRead.currentHiveDecimalWritable.getHiveDecimal();
if (value == null) {
TestCase.fail("Decimal field evaluated to NULL");
}
HiveDecimalWritable expectedWritable = (HiveDecimalWritable) expected;
if (!value.equals(expectedWritable.getHiveDecimal())) {
DecimalTypeInfo decimalTypeInfo = (DecimalTypeInfo) primitiveTypeInfo;
int precision = decimalTypeInfo.getPrecision();
int scale = decimalTypeInfo.getScale();
TestCase.fail("Decimal field mismatch (expected " + expectedWritable.getHiveDecimal() + " found " + value.toString() + ") precision " + precision + ", scale " + scale);
}
}
break;
case TIMESTAMP:
{
Timestamp value = deserializeRead.currentTimestampWritable.getTimestamp();
TimestampWritable expectedWritable = (TimestampWritable) expected;
if (!value.equals(expectedWritable.getTimestamp())) {
TestCase.fail("Timestamp field mismatch (expected " + expectedWritable.getTimestamp() + " found " + value.toString() + ")");
}
}
break;
case INTERVAL_YEAR_MONTH:
{
HiveIntervalYearMonth value = deserializeRead.currentHiveIntervalYearMonthWritable.getHiveIntervalYearMonth();
HiveIntervalYearMonthWritable expectedWritable = (HiveIntervalYearMonthWritable) expected;
HiveIntervalYearMonth expectedValue = expectedWritable.getHiveIntervalYearMonth();
if (!value.equals(expectedValue)) {
TestCase.fail("HiveIntervalYearMonth field mismatch (expected " + expectedValue + " found " + value.toString() + ")");
}
}
break;
case INTERVAL_DAY_TIME:
{
HiveIntervalDayTime value = deserializeRead.currentHiveIntervalDayTimeWritable.getHiveIntervalDayTime();
HiveIntervalDayTimeWritable expectedWritable = (HiveIntervalDayTimeWritable) expected;
HiveIntervalDayTime expectedValue = expectedWritable.getHiveIntervalDayTime();
if (!value.equals(expectedValue)) {
TestCase.fail("HiveIntervalDayTime field mismatch (expected " + expectedValue + " found " + value.toString() + ")");
}
}
break;
default:
throw new HiveException("Unexpected primitive category " + primitiveCategory);
}
}
TestCase.assertTrue(deserializeRead.isEndOfInputReached());
}
use of org.apache.hadoop.hive.serde2.io.TimestampWritable in project hive by apache.
the class TestVectorGroupByOperator method testMultiKey.
private void testMultiKey(String aggregateName, FakeVectorRowBatchFromObjectIterables data, HashMap<Object, Object> expected) throws HiveException {
Map<String, Integer> mapColumnNames = new HashMap<String, Integer>();
ArrayList<String> outputColumnNames = new ArrayList<String>();
ArrayList<ExprNodeDesc> keysDesc = new ArrayList<ExprNodeDesc>();
Set<Object> keys = new HashSet<Object>();
// The types array tells us the number of columns in the data
final String[] columnTypes = data.getTypes();
// Columns 0..N-1 are keys. Column N is the aggregate value input
int i = 0;
for (; i < columnTypes.length - 1; ++i) {
String columnName = String.format("_col%d", i);
mapColumnNames.put(columnName, i);
outputColumnNames.add(columnName);
}
mapColumnNames.put("value", i);
outputColumnNames.add("value");
VectorizationContext ctx = new VectorizationContext("name", outputColumnNames);
ArrayList<AggregationDesc> aggs = new ArrayList(1);
aggs.add(buildAggregationDesc(ctx, aggregateName, GenericUDAFEvaluator.Mode.PARTIAL1, "value", TypeInfoFactory.getPrimitiveTypeInfo(columnTypes[i])));
for (i = 0; i < columnTypes.length - 1; ++i) {
String columnName = String.format("_col%d", i);
keysDesc.add(buildColumnDesc(ctx, columnName, TypeInfoFactory.getPrimitiveTypeInfo(columnTypes[i])));
}
GroupByDesc desc = new GroupByDesc();
desc.setVectorDesc(new VectorGroupByDesc());
desc.setOutputColumnNames(outputColumnNames);
desc.setAggregators(aggs);
desc.setKeys(keysDesc);
((VectorGroupByDesc) desc.getVectorDesc()).setProcessingMode(ProcessingMode.HASH);
CompilationOpContext cCtx = new CompilationOpContext();
Operator<? extends OperatorDesc> groupByOp = OperatorFactory.get(cCtx, desc);
VectorGroupByOperator vgo = (VectorGroupByOperator) Vectorizer.vectorizeGroupByOperator(groupByOp, ctx);
FakeCaptureOutputOperator out = FakeCaptureOutputOperator.addCaptureOutputChild(cCtx, vgo);
vgo.initialize(hconf, null);
out.setOutputInspector(new FakeCaptureOutputOperator.OutputInspector() {
private int rowIndex;
private String aggregateName;
private Map<Object, Object> expected;
private Set<Object> keys;
@Override
public void inspectRow(Object row, int tag) throws HiveException {
assertTrue(row instanceof Object[]);
Object[] fields = (Object[]) row;
assertEquals(columnTypes.length, fields.length);
ArrayList<Object> keyValue = new ArrayList<Object>(columnTypes.length - 1);
for (int i = 0; i < columnTypes.length - 1; ++i) {
Object key = fields[i];
if (null == key) {
keyValue.add(null);
} else if (key instanceof Text) {
Text txKey = (Text) key;
keyValue.add(txKey.toString());
} else if (key instanceof ByteWritable) {
ByteWritable bwKey = (ByteWritable) key;
keyValue.add(bwKey.get());
} else if (key instanceof ShortWritable) {
ShortWritable swKey = (ShortWritable) key;
keyValue.add(swKey.get());
} else if (key instanceof IntWritable) {
IntWritable iwKey = (IntWritable) key;
keyValue.add(iwKey.get());
} else if (key instanceof LongWritable) {
LongWritable lwKey = (LongWritable) key;
keyValue.add(lwKey.get());
} else if (key instanceof TimestampWritable) {
TimestampWritable twKey = (TimestampWritable) key;
keyValue.add(twKey.getTimestamp());
} else if (key instanceof DoubleWritable) {
DoubleWritable dwKey = (DoubleWritable) key;
keyValue.add(dwKey.get());
} else if (key instanceof FloatWritable) {
FloatWritable fwKey = (FloatWritable) key;
keyValue.add(fwKey.get());
} else if (key instanceof BooleanWritable) {
BooleanWritable bwKey = (BooleanWritable) key;
keyValue.add(bwKey.get());
} else {
Assert.fail(String.format("Not implemented key output type %s: %s", key.getClass().getName(), key));
}
}
String keyAsString = Arrays.deepToString(keyValue.toArray());
assertTrue(expected.containsKey(keyValue));
Object expectedValue = expected.get(keyValue);
Object value = fields[columnTypes.length - 1];
Validator validator = getValidator(aggregateName);
validator.validate(keyAsString, expectedValue, new Object[] { value });
keys.add(keyValue);
}
private FakeCaptureOutputOperator.OutputInspector init(String aggregateName, Map<Object, Object> expected, Set<Object> keys) {
this.aggregateName = aggregateName;
this.expected = expected;
this.keys = keys;
return this;
}
}.init(aggregateName, expected, keys));
for (VectorizedRowBatch unit : data) {
vgo.process(unit, 0);
}
vgo.close(false);
List<Object> outBatchList = out.getCapturedRows();
assertNotNull(outBatchList);
assertEquals(expected.size(), outBatchList.size());
assertEquals(expected.size(), keys.size());
}
Aggregations