use of org.apache.hadoop.hive.serde2.typeinfo.UnionTypeInfo in project hive by apache.
the class VerifyFastRow method serializeWrite.
public static void serializeWrite(SerializeWrite serializeWrite, TypeInfo typeInfo, Object object) throws IOException {
if (object == null) {
serializeWrite.writeNull();
return;
}
switch(typeInfo.getCategory()) {
case PRIMITIVE:
{
PrimitiveTypeInfo primitiveTypeInfo = (PrimitiveTypeInfo) typeInfo;
switch(primitiveTypeInfo.getPrimitiveCategory()) {
case BOOLEAN:
{
boolean value = ((BooleanWritable) object).get();
serializeWrite.writeBoolean(value);
}
break;
case BYTE:
{
byte value = ((ByteWritable) object).get();
serializeWrite.writeByte(value);
}
break;
case SHORT:
{
short value = ((ShortWritable) object).get();
serializeWrite.writeShort(value);
}
break;
case INT:
{
int value = ((IntWritable) object).get();
serializeWrite.writeInt(value);
}
break;
case LONG:
{
long value = ((LongWritable) object).get();
serializeWrite.writeLong(value);
}
break;
case FLOAT:
{
float value = ((FloatWritable) object).get();
serializeWrite.writeFloat(value);
}
break;
case DOUBLE:
{
double value = ((DoubleWritable) object).get();
serializeWrite.writeDouble(value);
}
break;
case STRING:
{
Text value = (Text) object;
byte[] stringBytes = value.getBytes();
int stringLength = stringBytes.length;
serializeWrite.writeString(stringBytes, 0, stringLength);
}
break;
case CHAR:
{
HiveChar value = ((HiveCharWritable) object).getHiveChar();
serializeWrite.writeHiveChar(value);
}
break;
case VARCHAR:
{
HiveVarchar value = ((HiveVarcharWritable) object).getHiveVarchar();
serializeWrite.writeHiveVarchar(value);
}
break;
case DECIMAL:
{
HiveDecimal value = ((HiveDecimalWritable) object).getHiveDecimal();
DecimalTypeInfo decTypeInfo = (DecimalTypeInfo) primitiveTypeInfo;
serializeWrite.writeHiveDecimal(value, decTypeInfo.scale());
}
break;
case DATE:
{
Date value = ((DateWritableV2) object).get();
serializeWrite.writeDate(value);
}
break;
case TIMESTAMP:
{
Timestamp value = ((TimestampWritableV2) object).getTimestamp();
serializeWrite.writeTimestamp(value);
}
break;
case INTERVAL_YEAR_MONTH:
{
HiveIntervalYearMonth value = ((HiveIntervalYearMonthWritable) object).getHiveIntervalYearMonth();
serializeWrite.writeHiveIntervalYearMonth(value);
}
break;
case INTERVAL_DAY_TIME:
{
HiveIntervalDayTime value = ((HiveIntervalDayTimeWritable) object).getHiveIntervalDayTime();
serializeWrite.writeHiveIntervalDayTime(value);
}
break;
case BINARY:
{
BytesWritable byteWritable = (BytesWritable) object;
byte[] binaryBytes = byteWritable.getBytes();
int length = byteWritable.getLength();
serializeWrite.writeBinary(binaryBytes, 0, length);
}
break;
default:
throw new Error("Unknown primitive category " + primitiveTypeInfo.getPrimitiveCategory().name());
}
}
break;
case LIST:
{
ListTypeInfo listTypeInfo = (ListTypeInfo) typeInfo;
TypeInfo elementTypeInfo = listTypeInfo.getListElementTypeInfo();
ArrayList<Object> elements = (ArrayList<Object>) object;
serializeWrite.beginList(elements);
boolean isFirst = true;
for (Object elementObject : elements) {
if (isFirst) {
isFirst = false;
} else {
serializeWrite.separateList();
}
if (elementObject == null) {
serializeWrite.writeNull();
} else {
serializeWrite(serializeWrite, elementTypeInfo, elementObject);
}
}
serializeWrite.finishList();
}
break;
case MAP:
{
MapTypeInfo mapTypeInfo = (MapTypeInfo) typeInfo;
TypeInfo keyTypeInfo = mapTypeInfo.getMapKeyTypeInfo();
TypeInfo valueTypeInfo = mapTypeInfo.getMapValueTypeInfo();
Map<Object, Object> hashMap = (Map<Object, Object>) object;
serializeWrite.beginMap(hashMap);
boolean isFirst = true;
for (Map.Entry<Object, Object> entry : hashMap.entrySet()) {
if (isFirst) {
isFirst = false;
} else {
serializeWrite.separateKeyValuePair();
}
if (entry.getKey() == null) {
serializeWrite.writeNull();
} else {
serializeWrite(serializeWrite, keyTypeInfo, entry.getKey());
}
serializeWrite.separateKey();
if (entry.getValue() == null) {
serializeWrite.writeNull();
} else {
serializeWrite(serializeWrite, valueTypeInfo, entry.getValue());
}
}
serializeWrite.finishMap();
}
break;
case STRUCT:
{
StructTypeInfo structTypeInfo = (StructTypeInfo) typeInfo;
List<TypeInfo> fieldTypeInfos = structTypeInfo.getAllStructFieldTypeInfos();
List<Object> fieldValues = (List<Object>) object;
final int size = fieldValues.size();
serializeWrite.beginStruct(fieldValues);
boolean isFirst = true;
for (int i = 0; i < size; i++) {
if (isFirst) {
isFirst = false;
} else {
serializeWrite.separateStruct();
}
serializeWrite(serializeWrite, fieldTypeInfos.get(i), fieldValues.get(i));
}
serializeWrite.finishStruct();
}
break;
case UNION:
{
UnionTypeInfo unionTypeInfo = (UnionTypeInfo) typeInfo;
List<TypeInfo> fieldTypeInfos = unionTypeInfo.getAllUnionObjectTypeInfos();
final int size = fieldTypeInfos.size();
StandardUnionObjectInspector.StandardUnion standardUnion = (StandardUnionObjectInspector.StandardUnion) object;
byte tag = standardUnion.getTag();
serializeWrite.beginUnion(tag);
serializeWrite(serializeWrite, fieldTypeInfos.get(tag), standardUnion.getObject());
serializeWrite.finishUnion();
}
break;
default:
throw new Error("Unknown category " + typeInfo.getCategory().name());
}
}
use of org.apache.hadoop.hive.serde2.typeinfo.UnionTypeInfo in project hive by apache.
the class DeserializeRead method allocateCurrentWritable.
/*
* This class is used to read one field at a time. Simple fields like long, double, int are read
* into to primitive current* members; the non-simple field types like Date, Timestamp, etc, are
* read into a current object that this method will allocate.
*
* This method handles complex type fields by recursively calling this method.
*/
private void allocateCurrentWritable(TypeInfo typeInfo) {
switch(typeInfo.getCategory()) {
case PRIMITIVE:
switch(((PrimitiveTypeInfo) typeInfo).getPrimitiveCategory()) {
case DATE:
if (currentDateWritable == null) {
currentDateWritable = new DateWritableV2();
}
break;
case TIMESTAMP:
if (currentTimestampWritable == null) {
currentTimestampWritable = new TimestampWritableV2();
}
break;
case INTERVAL_YEAR_MONTH:
if (currentHiveIntervalYearMonthWritable == null) {
currentHiveIntervalYearMonthWritable = new HiveIntervalYearMonthWritable();
}
break;
case INTERVAL_DAY_TIME:
if (currentHiveIntervalDayTimeWritable == null) {
currentHiveIntervalDayTimeWritable = new HiveIntervalDayTimeWritable();
}
break;
case DECIMAL:
if (currentHiveDecimalWritable == null) {
currentHiveDecimalWritable = new HiveDecimalWritable();
}
break;
default:
}
break;
case LIST:
allocateCurrentWritable(((ListTypeInfo) typeInfo).getListElementTypeInfo());
break;
case MAP:
allocateCurrentWritable(((MapTypeInfo) typeInfo).getMapKeyTypeInfo());
allocateCurrentWritable(((MapTypeInfo) typeInfo).getMapValueTypeInfo());
break;
case STRUCT:
for (TypeInfo fieldTypeInfo : ((StructTypeInfo) typeInfo).getAllStructFieldTypeInfos()) {
allocateCurrentWritable(fieldTypeInfo);
}
break;
case UNION:
for (TypeInfo fieldTypeInfo : ((UnionTypeInfo) typeInfo).getAllUnionObjectTypeInfos()) {
allocateCurrentWritable(fieldTypeInfo);
}
break;
default:
throw new RuntimeException("Unexpected category " + typeInfo.getCategory());
}
}
use of org.apache.hadoop.hive.serde2.typeinfo.UnionTypeInfo in project hive by apache.
the class BinarySortableDeserializeRead method createField.
private Field createField(TypeInfo typeInfo) {
final Field field = new Field();
final Category category = typeInfo.getCategory();
field.category = category;
field.typeInfo = typeInfo;
switch(category) {
case PRIMITIVE:
field.primitiveCategory = ((PrimitiveTypeInfo) typeInfo).getPrimitiveCategory();
break;
case LIST:
field.children = new Field[1];
field.children[0] = createField(((ListTypeInfo) typeInfo).getListElementTypeInfo());
break;
case MAP:
field.children = new Field[2];
field.children[0] = createField(((MapTypeInfo) typeInfo).getMapKeyTypeInfo());
field.children[1] = createField(((MapTypeInfo) typeInfo).getMapValueTypeInfo());
break;
case STRUCT:
StructTypeInfo structTypeInfo = (StructTypeInfo) typeInfo;
List<TypeInfo> fieldTypeInfos = structTypeInfo.getAllStructFieldTypeInfos();
field.count = fieldTypeInfos.size();
field.children = createFields(fieldTypeInfos.toArray(new TypeInfo[fieldTypeInfos.size()]));
break;
case UNION:
UnionTypeInfo unionTypeInfo = (UnionTypeInfo) typeInfo;
List<TypeInfo> objectTypeInfos = unionTypeInfo.getAllUnionObjectTypeInfos();
field.count = 2;
field.children = createFields(objectTypeInfos.toArray(new TypeInfo[objectTypeInfos.size()]));
break;
default:
throw new RuntimeException();
}
return field;
}
use of org.apache.hadoop.hive.serde2.typeinfo.UnionTypeInfo in project hive by apache.
the class SerdeRandomRowSource method getObjectInspector.
private ObjectInspector getObjectInspector(TypeInfo typeInfo) {
ObjectInspector objectInspector;
switch(typeInfo.getCategory()) {
case PRIMITIVE:
{
final PrimitiveTypeInfo primitiveType = (PrimitiveTypeInfo) typeInfo;
objectInspector = PrimitiveObjectInspectorFactory.getPrimitiveWritableObjectInspector(primitiveType);
}
break;
case MAP:
{
final MapTypeInfo mapType = (MapTypeInfo) typeInfo;
final MapObjectInspector mapInspector = ObjectInspectorFactory.getStandardMapObjectInspector(getObjectInspector(mapType.getMapKeyTypeInfo()), getObjectInspector(mapType.getMapValueTypeInfo()));
objectInspector = mapInspector;
}
break;
case LIST:
{
final ListTypeInfo listType = (ListTypeInfo) typeInfo;
final ListObjectInspector listInspector = ObjectInspectorFactory.getStandardListObjectInspector(getObjectInspector(listType.getListElementTypeInfo()));
objectInspector = listInspector;
}
break;
case STRUCT:
{
final StructTypeInfo structType = (StructTypeInfo) typeInfo;
final List<TypeInfo> fieldTypes = structType.getAllStructFieldTypeInfos();
final List<ObjectInspector> fieldInspectors = new ArrayList<ObjectInspector>();
for (TypeInfo fieldType : fieldTypes) {
fieldInspectors.add(getObjectInspector(fieldType));
}
final StructObjectInspector structInspector = ObjectInspectorFactory.getStandardStructObjectInspector(structType.getAllStructFieldNames(), fieldInspectors);
objectInspector = structInspector;
}
break;
case UNION:
{
final UnionTypeInfo unionType = (UnionTypeInfo) typeInfo;
final List<TypeInfo> fieldTypes = unionType.getAllUnionObjectTypeInfos();
final List<ObjectInspector> fieldInspectors = new ArrayList<ObjectInspector>();
for (TypeInfo fieldType : fieldTypes) {
fieldInspectors.add(getObjectInspector(fieldType));
}
final UnionObjectInspector unionInspector = ObjectInspectorFactory.getStandardUnionObjectInspector(fieldInspectors);
objectInspector = unionInspector;
}
break;
default:
throw new RuntimeException("Unexpected category " + typeInfo.getCategory());
}
Preconditions.checkState(objectInspector != null);
return objectInspector;
}
use of org.apache.hadoop.hive.serde2.typeinfo.UnionTypeInfo in project hive by apache.
the class TestTypeInfoToSchema method createAvroUnionSchemaWithNull.
@Test
public void createAvroUnionSchemaWithNull() {
UnionTypeInfo unionTypeInfo = new UnionTypeInfo();
unionTypeInfo.setAllUnionObjectTypeInfos(Arrays.asList(INT, FLOAT, STRING, VOID));
final String specificSchema = Schema.createUnion(Arrays.asList(Schema.create(Schema.Type.NULL), Schema.create(Schema.Type.INT), Schema.create(Schema.Type.FLOAT), Schema.create(Schema.Type.STRING))).toString();
String expectedSchema = genSchemaWithoutNull(specificSchema);
Assert.assertEquals("Test for union's avro schema failed", expectedSchema, getAvroSchemaString(unionTypeInfo));
}
Aggregations