use of org.apache.hadoop.hive.serde2.objectinspector.primitive.ShortObjectInspector in project hive by apache.
the class TypedBytesSerDe method serializeField.
private void serializeField(Object o, ObjectInspector oi, Object reuse) throws IOException {
switch(oi.getCategory()) {
case PRIMITIVE:
{
PrimitiveObjectInspector poi = (PrimitiveObjectInspector) oi;
switch(poi.getPrimitiveCategory()) {
case VOID:
{
return;
}
case BOOLEAN:
{
BooleanObjectInspector boi = (BooleanObjectInspector) poi;
BooleanWritable r = reuse == null ? new BooleanWritable() : (BooleanWritable) reuse;
r.set(boi.get(o));
tbOut.write(r);
return;
}
case BYTE:
{
ByteObjectInspector boi = (ByteObjectInspector) poi;
ByteWritable r = reuse == null ? new ByteWritable() : (ByteWritable) reuse;
r.set(boi.get(o));
tbOut.write(r);
return;
}
case SHORT:
{
ShortObjectInspector spoi = (ShortObjectInspector) poi;
ShortWritable r = reuse == null ? new ShortWritable() : (ShortWritable) reuse;
r.set(spoi.get(o));
tbOut.write(r);
return;
}
case INT:
{
IntObjectInspector ioi = (IntObjectInspector) poi;
IntWritable r = reuse == null ? new IntWritable() : (IntWritable) reuse;
r.set(ioi.get(o));
tbOut.write(r);
return;
}
case LONG:
{
LongObjectInspector loi = (LongObjectInspector) poi;
LongWritable r = reuse == null ? new LongWritable() : (LongWritable) reuse;
r.set(loi.get(o));
tbOut.write(r);
return;
}
case FLOAT:
{
FloatObjectInspector foi = (FloatObjectInspector) poi;
FloatWritable r = reuse == null ? new FloatWritable() : (FloatWritable) reuse;
r.set(foi.get(o));
tbOut.write(r);
return;
}
case DOUBLE:
{
DoubleObjectInspector doi = (DoubleObjectInspector) poi;
DoubleWritable r = reuse == null ? new DoubleWritable() : (DoubleWritable) reuse;
r.set(doi.get(o));
tbOut.write(r);
return;
}
case STRING:
{
StringObjectInspector soi = (StringObjectInspector) poi;
Text t = soi.getPrimitiveWritableObject(o);
tbOut.write(t);
return;
}
default:
{
throw new RuntimeException("Unrecognized type: " + poi.getPrimitiveCategory());
}
}
}
case LIST:
case MAP:
case STRUCT:
{
// For complex object, serialize to JSON format
String s = SerDeUtils.getJSONString(o, oi);
Text t = reuse == null ? new Text() : (Text) reuse;
// convert to Text and write it
t.set(s);
tbOut.write(t);
}
default:
{
throw new RuntimeException("Unrecognized type: " + oi.getCategory());
}
}
}
use of org.apache.hadoop.hive.serde2.objectinspector.primitive.ShortObjectInspector in project hive by apache.
the class DruidSerDe method serialize.
@Override
public Writable serialize(Object o, ObjectInspector objectInspector) throws SerDeException {
if (objectInspector.getCategory() != ObjectInspector.Category.STRUCT) {
throw new SerDeException(getClass().toString() + " can only serialize struct types, but we got: " + objectInspector.getTypeName());
}
// Prepare the field ObjectInspectors
StructObjectInspector soi = (StructObjectInspector) objectInspector;
List<? extends StructField> fields = soi.getAllStructFieldRefs();
List<Object> values = soi.getStructFieldsDataAsList(o);
// We deserialize the result
final Map<String, Object> value = new HashMap<>();
for (int i = 0; i < columns.length; i++) {
if (values.get(i) == null) {
// null, we just add it
value.put(columns[i], null);
continue;
}
final Object res;
switch(types[i].getPrimitiveCategory()) {
case TIMESTAMP:
res = ((TimestampObjectInspector) fields.get(i).getFieldObjectInspector()).getPrimitiveJavaObject(values.get(i)).toEpochMilli();
break;
case TIMESTAMPLOCALTZ:
res = ((TimestampLocalTZObjectInspector) fields.get(i).getFieldObjectInspector()).getPrimitiveJavaObject(values.get(i)).getZonedDateTime().toInstant().toEpochMilli();
break;
case BYTE:
res = ((ByteObjectInspector) fields.get(i).getFieldObjectInspector()).get(values.get(i));
break;
case SHORT:
res = ((ShortObjectInspector) fields.get(i).getFieldObjectInspector()).get(values.get(i));
break;
case INT:
res = ((IntObjectInspector) fields.get(i).getFieldObjectInspector()).get(values.get(i));
break;
case LONG:
res = ((LongObjectInspector) fields.get(i).getFieldObjectInspector()).get(values.get(i));
break;
case FLOAT:
res = ((FloatObjectInspector) fields.get(i).getFieldObjectInspector()).get(values.get(i));
break;
case DOUBLE:
res = ((DoubleObjectInspector) fields.get(i).getFieldObjectInspector()).get(values.get(i));
break;
case CHAR:
res = ((HiveCharObjectInspector) fields.get(i).getFieldObjectInspector()).getPrimitiveJavaObject(values.get(i)).getValue();
break;
case VARCHAR:
res = ((HiveVarcharObjectInspector) fields.get(i).getFieldObjectInspector()).getPrimitiveJavaObject(values.get(i)).getValue();
break;
case STRING:
res = ((StringObjectInspector) fields.get(i).getFieldObjectInspector()).getPrimitiveJavaObject(values.get(i));
break;
case BOOLEAN:
res = ((BooleanObjectInspector) fields.get(i).getFieldObjectInspector()).get(values.get(i)) ? 1L : 0L;
break;
default:
throw new SerDeException("Unsupported type: " + types[i].getPrimitiveCategory());
}
value.put(columns[i], res);
}
// Extract the partitions keys segments granularity and partition key if any
// First Segment Granularity has to be here.
final int granularityFieldIndex = columns.length;
assert values.size() > granularityFieldIndex;
Preconditions.checkArgument(fields.get(granularityFieldIndex).getFieldName().equals(Constants.DRUID_TIMESTAMP_GRANULARITY_COL_NAME));
Timestamp timestamp = ((TimestampObjectInspector) fields.get(granularityFieldIndex).getFieldObjectInspector()).getPrimitiveJavaObject(values.get(granularityFieldIndex));
Preconditions.checkNotNull(timestamp, "Timestamp column cannot have null value");
value.put(Constants.DRUID_TIMESTAMP_GRANULARITY_COL_NAME, timestamp.toEpochMilli());
if (values.size() == columns.length + 2) {
// Then partition number if any.
final int partitionNumPos = granularityFieldIndex + 1;
Preconditions.checkArgument(fields.get(partitionNumPos).getFieldName().equals(Constants.DRUID_SHARD_KEY_COL_NAME), String.format("expecting to encounter %s but was %s", Constants.DRUID_SHARD_KEY_COL_NAME, fields.get(partitionNumPos).getFieldName()));
value.put(Constants.DRUID_SHARD_KEY_COL_NAME, ((LongObjectInspector) fields.get(partitionNumPos).getFieldObjectInspector()).get(values.get(partitionNumPos)));
}
return new DruidWritable(value);
}
use of org.apache.hadoop.hive.serde2.objectinspector.primitive.ShortObjectInspector in project hive by apache.
the class ObjectInspectorUtils method hashCode.
public static int hashCode(Object o, ObjectInspector objIns) {
if (o == null) {
return 0;
}
switch(objIns.getCategory()) {
case PRIMITIVE:
{
PrimitiveObjectInspector poi = ((PrimitiveObjectInspector) objIns);
switch(poi.getPrimitiveCategory()) {
case VOID:
return 0;
case BOOLEAN:
return ((BooleanObjectInspector) poi).get(o) ? 1 : 0;
case BYTE:
return ((ByteObjectInspector) poi).get(o);
case SHORT:
return ((ShortObjectInspector) poi).get(o);
case INT:
return ((IntObjectInspector) poi).get(o);
case LONG:
{
long a = ((LongObjectInspector) poi).get(o);
return (int) ((a >>> 32) ^ a);
}
case FLOAT:
return Float.floatToIntBits(((FloatObjectInspector) poi).get(o));
case DOUBLE:
{
// This hash function returns the same result as Double.hashCode()
// while DoubleWritable.hashCode returns a different result.
long a = Double.doubleToLongBits(((DoubleObjectInspector) poi).get(o));
return (int) ((a >>> 32) ^ a);
}
case STRING:
{
// This hash function returns the same result as String.hashCode() when
// all characters are ASCII, while Text.hashCode() always returns a
// different result.
Text t = ((StringObjectInspector) poi).getPrimitiveWritableObject(o);
int r = 0;
for (int i = 0; i < t.getLength(); i++) {
r = r * 31 + t.getBytes()[i];
}
return r;
}
case CHAR:
return ((HiveCharObjectInspector) poi).getPrimitiveWritableObject(o).hashCode();
case VARCHAR:
return ((HiveVarcharObjectInspector) poi).getPrimitiveWritableObject(o).hashCode();
case BINARY:
return ((BinaryObjectInspector) poi).getPrimitiveWritableObject(o).hashCode();
case DATE:
return ((DateObjectInspector) poi).getPrimitiveWritableObject(o).hashCode();
case TIMESTAMP:
// Use old timestamp writable hash code for backwards compatibility
TimestampWritable ts = new TimestampWritable(java.sql.Timestamp.valueOf(((TimestampObjectInspector) poi).getPrimitiveWritableObject(o).toString()));
return ts.hashCode();
case TIMESTAMPLOCALTZ:
TimestampLocalTZWritable tstz = ((TimestampLocalTZObjectInspector) poi).getPrimitiveWritableObject(o);
return tstz.hashCode();
case INTERVAL_YEAR_MONTH:
HiveIntervalYearMonthWritable intervalYearMonth = ((HiveIntervalYearMonthObjectInspector) poi).getPrimitiveWritableObject(o);
return intervalYearMonth.hashCode();
case INTERVAL_DAY_TIME:
HiveIntervalDayTimeWritable intervalDayTime = ((HiveIntervalDayTimeObjectInspector) poi).getPrimitiveWritableObject(o);
return intervalDayTime.hashCode();
case DECIMAL:
// compatible hash code.
return ((HiveDecimalObjectInspector) poi).getPrimitiveWritableObject(o).hashCode();
default:
{
throw new RuntimeException("Unknown type: " + poi.getPrimitiveCategory());
}
}
}
case LIST:
{
int r = 0;
ListObjectInspector listOI = (ListObjectInspector) objIns;
ObjectInspector elemOI = listOI.getListElementObjectInspector();
for (int ii = 0; ii < listOI.getListLength(o); ++ii) {
r = 31 * r + hashCode(listOI.getListElement(o, ii), elemOI);
}
return r;
}
case MAP:
{
int r = 0;
MapObjectInspector mapOI = (MapObjectInspector) objIns;
ObjectInspector keyOI = mapOI.getMapKeyObjectInspector();
ObjectInspector valueOI = mapOI.getMapValueObjectInspector();
Map<?, ?> map = mapOI.getMap(o);
for (Map.Entry<?, ?> entry : map.entrySet()) {
r += hashCode(entry.getKey(), keyOI) ^ hashCode(entry.getValue(), valueOI);
}
return r;
}
case STRUCT:
int r = 0;
StructObjectInspector structOI = (StructObjectInspector) objIns;
List<? extends StructField> fields = structOI.getAllStructFieldRefs();
for (StructField field : fields) {
r = 31 * r + hashCode(structOI.getStructFieldData(o, field), field.getFieldObjectInspector());
}
return r;
case UNION:
UnionObjectInspector uOI = (UnionObjectInspector) objIns;
byte tag = uOI.getTag(o);
return hashCode(uOI.getField(o), uOI.getObjectInspectors().get(tag));
default:
throw new RuntimeException("Unknown type: " + objIns.getTypeName());
}
}
use of org.apache.hadoop.hive.serde2.objectinspector.primitive.ShortObjectInspector in project hive by apache.
the class LazyBinarySerDe method serialize.
/**
* A recursive function that serialize an object to a byte buffer based on its
* object inspector.
*
* @param byteStream
* the byte stream storing the serialization data
* @param obj
* the object to serialize
* @param objInspector
* the object inspector
* @param skipLengthPrefix a boolean indicating whether length prefix is
* needed for list/map/struct
* @param warnedOnceNullMapKey a boolean indicating whether a warning
* has been issued once already when encountering null map keys
*/
public static void serialize(RandomAccessOutput byteStream, Object obj, ObjectInspector objInspector, boolean skipLengthPrefix, BooleanRef warnedOnceNullMapKey) throws SerDeException {
// do nothing for null object
if (null == obj) {
return;
}
switch(objInspector.getCategory()) {
case PRIMITIVE:
{
PrimitiveObjectInspector poi = (PrimitiveObjectInspector) objInspector;
switch(poi.getPrimitiveCategory()) {
case VOID:
{
return;
}
case BOOLEAN:
{
boolean v = ((BooleanObjectInspector) poi).get(obj);
byteStream.write((byte) (v ? 1 : 0));
return;
}
case BYTE:
{
ByteObjectInspector boi = (ByteObjectInspector) poi;
byte v = boi.get(obj);
byteStream.write(v);
return;
}
case SHORT:
{
ShortObjectInspector spoi = (ShortObjectInspector) poi;
short v = spoi.get(obj);
byteStream.write((byte) (v >> 8));
byteStream.write((byte) (v));
return;
}
case INT:
{
IntObjectInspector ioi = (IntObjectInspector) poi;
int v = ioi.get(obj);
LazyBinaryUtils.writeVInt(byteStream, v);
return;
}
case LONG:
{
LongObjectInspector loi = (LongObjectInspector) poi;
long v = loi.get(obj);
LazyBinaryUtils.writeVLong(byteStream, v);
return;
}
case FLOAT:
{
FloatObjectInspector foi = (FloatObjectInspector) poi;
int v = Float.floatToIntBits(foi.get(obj));
byteStream.write((byte) (v >> 24));
byteStream.write((byte) (v >> 16));
byteStream.write((byte) (v >> 8));
byteStream.write((byte) (v));
return;
}
case DOUBLE:
{
DoubleObjectInspector doi = (DoubleObjectInspector) poi;
LazyBinaryUtils.writeDouble(byteStream, doi.get(obj));
return;
}
case STRING:
{
StringObjectInspector soi = (StringObjectInspector) poi;
Text t = soi.getPrimitiveWritableObject(obj);
serializeText(byteStream, t, skipLengthPrefix);
return;
}
case CHAR:
{
HiveCharObjectInspector hcoi = (HiveCharObjectInspector) poi;
Text t = hcoi.getPrimitiveWritableObject(obj).getTextValue();
serializeText(byteStream, t, skipLengthPrefix);
return;
}
case VARCHAR:
{
HiveVarcharObjectInspector hcoi = (HiveVarcharObjectInspector) poi;
Text t = hcoi.getPrimitiveWritableObject(obj).getTextValue();
serializeText(byteStream, t, skipLengthPrefix);
return;
}
case BINARY:
{
BinaryObjectInspector baoi = (BinaryObjectInspector) poi;
BytesWritable bw = baoi.getPrimitiveWritableObject(obj);
int length = bw.getLength();
if (!skipLengthPrefix) {
LazyBinaryUtils.writeVInt(byteStream, length);
} else {
if (length == 0) {
throw new RuntimeException("LazyBinaryColumnarSerde cannot serialize a non-null zero " + "length binary field. Consider using either LazyBinarySerde or ColumnarSerde.");
}
}
byteStream.write(bw.getBytes(), 0, length);
return;
}
case DATE:
{
DateWritableV2 d = ((DateObjectInspector) poi).getPrimitiveWritableObject(obj);
writeDateToByteStream(byteStream, d);
return;
}
case TIMESTAMP:
{
TimestampObjectInspector toi = (TimestampObjectInspector) poi;
TimestampWritableV2 t = toi.getPrimitiveWritableObject(obj);
t.writeToByteStream(byteStream);
return;
}
case TIMESTAMPLOCALTZ:
{
TimestampLocalTZWritable t = ((TimestampLocalTZObjectInspector) poi).getPrimitiveWritableObject(obj);
t.writeToByteStream(byteStream);
return;
}
case INTERVAL_YEAR_MONTH:
{
HiveIntervalYearMonthWritable intervalYearMonth = ((HiveIntervalYearMonthObjectInspector) poi).getPrimitiveWritableObject(obj);
intervalYearMonth.writeToByteStream(byteStream);
return;
}
case INTERVAL_DAY_TIME:
{
HiveIntervalDayTimeWritable intervalDayTime = ((HiveIntervalDayTimeObjectInspector) poi).getPrimitiveWritableObject(obj);
intervalDayTime.writeToByteStream(byteStream);
return;
}
case DECIMAL:
{
HiveDecimalObjectInspector bdoi = (HiveDecimalObjectInspector) poi;
HiveDecimalWritable t = bdoi.getPrimitiveWritableObject(obj);
if (t == null) {
return;
}
writeToByteStream(byteStream, t);
return;
}
default:
{
throw new RuntimeException("Unrecognized type: " + poi.getPrimitiveCategory());
}
}
}
case LIST:
{
ListObjectInspector loi = (ListObjectInspector) objInspector;
ObjectInspector eoi = loi.getListElementObjectInspector();
int byteSizeStart = 0;
int listStart = 0;
if (!skipLengthPrefix) {
// 1/ reserve spaces for the byte size of the list
// which is a integer and takes four bytes
byteSizeStart = byteStream.getLength();
byteStream.reserve(4);
listStart = byteStream.getLength();
}
// 2/ write the size of the list as a VInt
int size = loi.getListLength(obj);
LazyBinaryUtils.writeVInt(byteStream, size);
// 3/ write the null bytes
byte nullByte = 0;
for (int eid = 0; eid < size; eid++) {
// set the bit to 1 if an element is not null
if (null != loi.getListElement(obj, eid)) {
nullByte |= 1 << (eid % 8);
}
// if this is the last element
if (7 == eid % 8 || eid == size - 1) {
byteStream.write(nullByte);
nullByte = 0;
}
}
// 4/ write element by element from the list
for (int eid = 0; eid < size; eid++) {
serialize(byteStream, loi.getListElement(obj, eid), eoi, false, warnedOnceNullMapKey);
}
if (!skipLengthPrefix) {
// 5/ update the list byte size
int listEnd = byteStream.getLength();
int listSize = listEnd - listStart;
writeSizeAtOffset(byteStream, byteSizeStart, listSize);
}
return;
}
case MAP:
{
MapObjectInspector moi = (MapObjectInspector) objInspector;
ObjectInspector koi = moi.getMapKeyObjectInspector();
ObjectInspector voi = moi.getMapValueObjectInspector();
Map<?, ?> map = moi.getMap(obj);
int byteSizeStart = 0;
int mapStart = 0;
if (!skipLengthPrefix) {
// 1/ reserve spaces for the byte size of the map
// which is a integer and takes four bytes
byteSizeStart = byteStream.getLength();
byteStream.reserve(4);
mapStart = byteStream.getLength();
}
// 2/ write the size of the map which is a VInt
int size = map.size();
LazyBinaryUtils.writeVInt(byteStream, size);
// 3/ write the null bytes
int b = 0;
byte nullByte = 0;
for (Map.Entry<?, ?> entry : map.entrySet()) {
// set the bit to 1 if a key is not null
if (null != entry.getKey()) {
nullByte |= 1 << (b % 8);
} else if (warnedOnceNullMapKey != null) {
if (!warnedOnceNullMapKey.value) {
LOG.warn("Null map key encountered! Ignoring similar problems.");
}
warnedOnceNullMapKey.value = true;
}
b++;
// set the bit to 1 if a value is not null
if (null != entry.getValue()) {
nullByte |= 1 << (b % 8);
}
b++;
// or if this is the last key-value pair
if (0 == b % 8 || b == size * 2) {
byteStream.write(nullByte);
nullByte = 0;
}
}
// 4/ write key-value pairs one by one
for (Map.Entry<?, ?> entry : map.entrySet()) {
serialize(byteStream, entry.getKey(), koi, false, warnedOnceNullMapKey);
serialize(byteStream, entry.getValue(), voi, false, warnedOnceNullMapKey);
}
if (!skipLengthPrefix) {
// 5/ update the byte size of the map
int mapEnd = byteStream.getLength();
int mapSize = mapEnd - mapStart;
writeSizeAtOffset(byteStream, byteSizeStart, mapSize);
}
return;
}
case STRUCT:
case UNION:
{
int byteSizeStart = 0;
int typeStart = 0;
if (!skipLengthPrefix) {
// 1/ reserve spaces for the byte size of the struct
// which is a integer and takes four bytes
byteSizeStart = byteStream.getLength();
byteStream.reserve(4);
typeStart = byteStream.getLength();
}
if (ObjectInspector.Category.STRUCT.equals(objInspector.getCategory())) {
// 2/ serialize the struct
serializeStruct(byteStream, obj, (StructObjectInspector) objInspector, warnedOnceNullMapKey);
} else {
// 2/ serialize the union
serializeUnion(byteStream, obj, (UnionObjectInspector) objInspector, warnedOnceNullMapKey);
}
if (!skipLengthPrefix) {
// 3/ update the byte size of the struct
int typeEnd = byteStream.getLength();
int typeSize = typeEnd - typeStart;
writeSizeAtOffset(byteStream, byteSizeStart, typeSize);
}
return;
}
default:
{
throw new RuntimeException("Unrecognized type: " + objInspector.getCategory());
}
}
}
use of org.apache.hadoop.hive.serde2.objectinspector.primitive.ShortObjectInspector in project hive by apache.
the class DataWritableWriter method createWriter.
/**
* Creates a writer for the specific object inspector. The returned writer will be used
* to call Parquet API for the specific data type.
* @param inspector The object inspector used to get the correct value type.
* @param type Type that contains information about the type schema.
* @return A ParquetWriter object used to call the Parquet API fo the specific data type.
*/
private DataWriter createWriter(ObjectInspector inspector, Type type) {
if (type.isPrimitive()) {
checkInspectorCategory(inspector, ObjectInspector.Category.PRIMITIVE);
PrimitiveObjectInspector primitiveInspector = (PrimitiveObjectInspector) inspector;
switch(primitiveInspector.getPrimitiveCategory()) {
case BOOLEAN:
return new BooleanDataWriter((BooleanObjectInspector) inspector);
case BYTE:
return new ByteDataWriter((ByteObjectInspector) inspector);
case SHORT:
return new ShortDataWriter((ShortObjectInspector) inspector);
case INT:
return new IntDataWriter((IntObjectInspector) inspector);
case LONG:
return new LongDataWriter((LongObjectInspector) inspector);
case FLOAT:
return new FloatDataWriter((FloatObjectInspector) inspector);
case DOUBLE:
return new DoubleDataWriter((DoubleObjectInspector) inspector);
case STRING:
return new StringDataWriter((StringObjectInspector) inspector);
case CHAR:
return new CharDataWriter((HiveCharObjectInspector) inspector);
case VARCHAR:
return new VarcharDataWriter((HiveVarcharObjectInspector) inspector);
case BINARY:
return new BinaryDataWriter((BinaryObjectInspector) inspector);
case TIMESTAMP:
return new TimestampDataWriter((TimestampObjectInspector) inspector);
case DECIMAL:
return new DecimalDataWriter((HiveDecimalObjectInspector) inspector);
case DATE:
return new DateDataWriter((DateObjectInspector) inspector);
default:
throw new IllegalArgumentException("Unsupported primitive data type: " + primitiveInspector.getPrimitiveCategory());
}
} else {
GroupType groupType = type.asGroupType();
LogicalTypeAnnotation logicalType = type.getLogicalTypeAnnotation();
if (logicalType != null && logicalType instanceof ListLogicalTypeAnnotation) {
checkInspectorCategory(inspector, ObjectInspector.Category.LIST);
return new ListDataWriter((ListObjectInspector) inspector, groupType);
} else if (logicalType != null && logicalType instanceof MapLogicalTypeAnnotation) {
checkInspectorCategory(inspector, ObjectInspector.Category.MAP);
return new MapDataWriter((MapObjectInspector) inspector, groupType);
} else {
checkInspectorCategory(inspector, ObjectInspector.Category.STRUCT);
return new StructDataWriter((StructObjectInspector) inspector, groupType);
}
}
}
Aggregations