public static ConstantObjectInspector getConstantObjectInspector( ObjectInspector oi, Object value) { if (oi instanceof ConstantObjectInspector) { return (ConstantObjectInspector) oi; } ObjectInspector writableOI = getStandardObjectInspector(oi, ObjectInspectorCopyOption.WRITABLE); Object writableValue = ObjectInspectorConverters.getConverter(oi, writableOI).convert(value); switch (writableOI.getCategory()) { case PRIMITIVE: PrimitiveObjectInspector poi = (PrimitiveObjectInspector) oi; return PrimitiveObjectInspectorFactory.getPrimitiveWritableConstantObjectInspector( poi.getTypeInfo(), writableValue); case LIST: ListObjectInspector loi = (ListObjectInspector) oi; return ObjectInspectorFactory.getStandardConstantListObjectInspector( getStandardObjectInspector( loi.getListElementObjectInspector(), ObjectInspectorCopyOption.WRITABLE), (List<?>) writableValue); case MAP: MapObjectInspector moi = (MapObjectInspector) oi; return ObjectInspectorFactory.getStandardConstantMapObjectInspector( getStandardObjectInspector( moi.getMapKeyObjectInspector(), ObjectInspectorCopyOption.WRITABLE), getStandardObjectInspector( moi.getMapValueObjectInspector(), ObjectInspectorCopyOption.WRITABLE), (Map<?, ?>) writableValue); default: throw new IllegalArgumentException( writableOI.getCategory() + " not yet supported for constant OI"); } }
/** * Whether comparison is supported for this type. Currently all types that references any map are * not comparable. */ public static boolean compareSupported(ObjectInspector oi) { switch (oi.getCategory()) { case PRIMITIVE: return true; case LIST: ListObjectInspector loi = (ListObjectInspector) oi; return compareSupported(loi.getListElementObjectInspector()); case STRUCT: StructObjectInspector soi = (StructObjectInspector) oi; List<? extends StructField> fields = soi.getAllStructFieldRefs(); for (int f = 0; f < fields.size(); f++) { if (!compareSupported(fields.get(f).getFieldObjectInspector())) { return false; } } return true; case MAP: return false; case UNION: UnionObjectInspector uoi = (UnionObjectInspector) oi; for (ObjectInspector eoi : uoi.getObjectInspectors()) { if (!compareSupported(eoi)) { return false; } } return true; default: return false; } }
private ObjectInspector solveOi(ObjectInspector arg) { switch (arg.getCategory()) { case PRIMITIVE: // VOID, BOOLEAN, BYTE, SHORT, INT, LONG, FLOAT, DOUBLE, STRING, TIMESTAMP, BINARY, DECIMAL, // UNKNOWN PrimitiveObjectInspector poi = (PrimitiveObjectInspector) arg; return PrimitiveObjectInspectorFactory.getPrimitiveJavaObjectInspector( poi.getPrimitiveCategory()); case LIST: return ObjectInspectorFactory.getStandardListObjectInspector( solveOi(((ListObjectInspector) arg).getListElementObjectInspector())); case MAP: return ObjectInspectorFactory.getStandardMapObjectInspector( solveOi(((MapObjectInspector) arg).getMapKeyObjectInspector()), solveOi(((MapObjectInspector) arg).getMapValueObjectInspector())); case STRUCT: StructObjectInspector soi = (StructObjectInspector) arg; int size = soi.getAllStructFieldRefs().size(); ArrayList<String> fnl = new ArrayList<String>(size); ArrayList<ObjectInspector> foil = new ArrayList<ObjectInspector>(size); for (StructField sf : ((StructObjectInspector) arg).getAllStructFieldRefs()) { fnl.add(sf.getFieldName()); foil.add(solveOi(sf.getFieldObjectInspector())); } return JsonStructObjectInspector.getJsonStructObjectInspector(fnl, foil); default: return arg; } }
public void printFileInfo() throws Exception { System.out.println("Reader: " + m_reader); System.out.println("# Rows: " + m_reader.getNumberOfRows()); System.out.println("# Types in the file: " + m_types.size()); for (int i = 0; i < m_types.size(); i++) { System.out.println("Type " + i + ": " + m_types.get(i).getKind()); } System.out.println("Compression: " + m_reader.getCompression()); if (m_reader.getCompression() != CompressionKind.NONE) { System.out.println("Compression size: " + m_reader.getCompressionSize()); } m_oi = (StructObjectInspector) m_reader.getObjectInspector(); System.out.println("object inspector type category: " + m_oi.getCategory()); System.out.println("object inspector type name : " + m_oi.getTypeName()); System.out.println("Number of columns in the table: " + m_fields.size()); // Print the type info: for (int i = 0; i < m_fields.size(); i++) { System.out.println("Column " + i + " name: " + m_fields.get(i).getFieldName()); ObjectInspector lv_foi = m_fields.get(i).getFieldObjectInspector(); System.out.println("Column " + i + " type category: " + lv_foi.getCategory()); System.out.println("Column " + i + " type name: " + lv_foi.getTypeName()); } }
public static ObjectInspector wrap(ObjectInspector inspector) { switch (inspector.getCategory()) { case PRIMITIVE: return inspector; case LIST: return new DelegatedListObjectInspector((ListObjectInspector) inspector); case MAP: return new DelegatedMapObjectInspector((MapObjectInspector) inspector); case STRUCT: return new DelegatedStructObjectInspector((StructObjectInspector) inspector); case UNION: return new DelegatedUnionObjectInspector((UnionObjectInspector) inspector); default: throw new RuntimeException("invalid category " + inspector.getCategory()); } }
/** * @param oi - Input object inspector * @param oiSettableProperties - Lookup map to cache the result.(If no caching, pass null) * @return - true if : (1) oi is an instance of settable<DataType>OI. (2) All the embedded object * inspectors are instances of settable<DataType>OI. If (1) or (2) is false, return false. */ public static boolean hasAllFieldsSettable( ObjectInspector oi, Map<ObjectInspector, Boolean> oiSettableProperties) { // If the result is already present in the cache, return it. if (!(oiSettableProperties == null) && oiSettableProperties.containsKey(oi)) { return oiSettableProperties.get(oi).booleanValue(); } // If the top-level object inspector is non-settable return false if (!(isInstanceOfSettableOI(oi))) { return setOISettablePropertiesMap(oi, oiSettableProperties, false); } Boolean returnValue = true; switch (oi.getCategory()) { case PRIMITIVE: break; case STRUCT: StructObjectInspector structOutputOI = (StructObjectInspector) oi; List<? extends StructField> listFields = structOutputOI.getAllStructFieldRefs(); for (StructField listField : listFields) { if (!hasAllFieldsSettable(listField.getFieldObjectInspector(), oiSettableProperties)) { returnValue = false; break; } } break; case LIST: ListObjectInspector listOutputOI = (ListObjectInspector) oi; returnValue = hasAllFieldsSettable( listOutputOI.getListElementObjectInspector(), oiSettableProperties); break; case MAP: MapObjectInspector mapOutputOI = (MapObjectInspector) oi; returnValue = hasAllFieldsSettable(mapOutputOI.getMapKeyObjectInspector(), oiSettableProperties) && hasAllFieldsSettable( mapOutputOI.getMapValueObjectInspector(), oiSettableProperties); break; case UNION: UnionObjectInspector unionOutputOI = (UnionObjectInspector) oi; List<ObjectInspector> unionListFields = unionOutputOI.getObjectInspectors(); for (ObjectInspector listField : unionListFields) { if (!hasAllFieldsSettable(listField, oiSettableProperties)) { returnValue = false; break; } } break; default: throw new RuntimeException( "Hive internal error inside hasAllFieldsSettable : " + oi.getTypeName() + " not supported yet."); } return setOISettablePropertiesMap(oi, oiSettableProperties, returnValue); }
public static boolean supportsConstantObjectInspector(ObjectInspector oi) { switch (oi.getCategory()) { case PRIMITIVE: case LIST: case MAP: return true; default: return false; } }
/** Ensures that an ObjectInspector is Writable. */ public static ObjectInspector getWritableObjectInspector(ObjectInspector oi) { // All non-primitive OIs are writable so we need only check this case. if (oi.getCategory() == Category.PRIMITIVE) { PrimitiveObjectInspector poi = (PrimitiveObjectInspector) oi; if (!(poi instanceof AbstractPrimitiveWritableObjectInspector)) { return PrimitiveObjectInspectorFactory.getPrimitiveWritableObjectInspector( poi.getTypeInfo()); } } return oi; }
public static ObjectInspector reset(ObjectInspector prev, ObjectInspector current) { switch (current.getCategory()) { case PRIMITIVE: break; case LIST: ((DelegatedListObjectInspector) prev).reset((ListObjectInspector) current); break; case MAP: ((DelegatedMapObjectInspector) prev).reset((MapObjectInspector) current); break; case STRUCT: ((DelegatedStructObjectInspector) prev).reset((StructObjectInspector) current); break; case UNION: ((DelegatedUnionObjectInspector) prev).reset((UnionObjectInspector) current); break; default: throw new RuntimeException("invalid category " + current.getCategory()); } return prev; }
private ObjectInspectorConverters.Converter getConverter(ObjectInspector arg) { switch (arg.getCategory()) { case PRIMITIVE: return ObjectInspectorConverters.getConverter(arg, arg); case LIST: case MAP: case STRUCT: return ObjectInspectorConverters.getConverter(arg, solveOi(arg)); default: return null; } }
public static void copyStructToArray( Object o, ObjectInspector oi, ObjectInspectorCopyOption objectInspectorOption, Object[] dest, int offset) throws SerDeException { if (o == null) { return; } if (oi.getCategory() != Category.STRUCT) { throw new SerDeException("Unexpected category " + oi.getCategory()); } StructObjectInspector soi = (StructObjectInspector) oi; List<? extends StructField> fields = soi.getAllStructFieldRefs(); for (int i = 0; i < fields.size(); ++i) { StructField f = fields.get(i); dest[offset + i] = copyToStandardObject( soi.getStructFieldData(o, f), f.getFieldObjectInspector(), objectInspectorOption); } }
private static boolean isInstanceOfSettableOI(ObjectInspector oi) { switch (oi.getCategory()) { case PRIMITIVE: return isInstanceOfSettablePrimitiveOI((PrimitiveObjectInspector) oi); case STRUCT: return oi instanceof SettableStructObjectInspector; case LIST: return oi instanceof SettableListObjectInspector; case MAP: return oi instanceof SettableMapObjectInspector; case UNION: return oi instanceof SettableUnionObjectInspector; default: throw new RuntimeException( "Hive internal error inside isAssignableFromSettableOI : " + oi.getTypeName() + " not supported yet."); } }
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: TimestampWritable t = ((TimestampObjectInspector) poi).getPrimitiveWritableObject(o); return t.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: 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()); } }
// This method is just for experimentation. public void testRead() throws Exception { m_reader = OrcFile.createReader(m_file_path, OrcFile.readerOptions(m_conf)); System.out.println("Reader: " + m_reader); System.out.println("# Rows: " + m_reader.getNumberOfRows()); m_types = m_reader.getTypes(); System.out.println("# Types in the file: " + m_types.size()); for (int i = 0; i < m_types.size(); i++) { System.out.println("Type " + i + ": " + m_types.get(i).getKind()); } System.out.println("Compression: " + m_reader.getCompression()); if (m_reader.getCompression() != CompressionKind.NONE) { System.out.println("Compression size: " + m_reader.getCompressionSize()); } StructObjectInspector m_oi = (StructObjectInspector) m_reader.getObjectInspector(); System.out.println("object inspector type category: " + m_oi.getCategory()); System.out.println("object inspector type name : " + m_oi.getTypeName()); m_fields = m_oi.getAllStructFieldRefs(); System.out.println("Number of columns in the table: " + m_fields.size()); RecordReader m_rr = m_reader.rows(); // Print the type info: for (int i = 0; i < m_fields.size(); i++) { System.out.println("Column " + i + " name: " + m_fields.get(i).getFieldName()); ObjectInspector lv_foi = m_fields.get(i).getFieldObjectInspector(); System.out.println("Column " + i + " type category: " + lv_foi.getCategory()); System.out.println("Column " + i + " type name: " + lv_foi.getTypeName()); // Object lv_column_val = m_oi.getStructFieldData(lv_row, m_fields.get(i)); // System.out.print("Column " + i + " value: " + lv_row.getFieldValue(i)); } OrcStruct lv_row = null; Object lv_field_val = null; StringBuilder lv_row_string = new StringBuilder(1024); while (m_rr.hasNext()) { lv_row = (OrcStruct) m_rr.next(lv_row); lv_row_string.setLength(0); for (int i = 0; i < m_fields.size(); i++) { lv_field_val = lv_row.getFieldValue(i); if (lv_field_val != null) { lv_row_string.append(lv_field_val); } lv_row_string.append('|'); } System.out.println(lv_row_string); } /** * Typecasting to appropriate type based on the 'kind' if (OrcProto.Type.Kind.INT == * m_types.get(1).getKind()) { IntWritable lvf_1_val = (IntWritable) lv_row.getFieldValue(0); * System.out.println("Column 1 value: " + lvf_1_val); } */ }
public static ObjectInspector getStandardObjectInspector( ObjectInspector oi, ObjectInspectorCopyOption objectInspectorOption) { ObjectInspector result = null; switch (oi.getCategory()) { case PRIMITIVE: { PrimitiveObjectInspector poi = (PrimitiveObjectInspector) oi; switch (objectInspectorOption) { case DEFAULT: { if (poi.preferWritable()) { result = PrimitiveObjectInspectorFactory.getPrimitiveWritableObjectInspector( poi.getTypeInfo()); } else { result = PrimitiveObjectInspectorFactory.getPrimitiveJavaObjectInspector( poi.getTypeInfo()); } break; } case JAVA: { result = PrimitiveObjectInspectorFactory.getPrimitiveJavaObjectInspector( poi.getTypeInfo()); break; } case WRITABLE: result = PrimitiveObjectInspectorFactory.getPrimitiveWritableObjectInspector( poi.getTypeInfo()); break; } break; } case LIST: { ListObjectInspector loi = (ListObjectInspector) oi; result = ObjectInspectorFactory.getStandardListObjectInspector( getStandardObjectInspector( loi.getListElementObjectInspector(), objectInspectorOption)); break; } case MAP: { MapObjectInspector moi = (MapObjectInspector) oi; result = ObjectInspectorFactory.getStandardMapObjectInspector( getStandardObjectInspector(moi.getMapKeyObjectInspector(), objectInspectorOption), getStandardObjectInspector( moi.getMapValueObjectInspector(), objectInspectorOption)); break; } case STRUCT: { StructObjectInspector soi = (StructObjectInspector) oi; List<? extends StructField> fields = soi.getAllStructFieldRefs(); List<String> fieldNames = new ArrayList<String>(fields.size()); List<ObjectInspector> fieldObjectInspectors = new ArrayList<ObjectInspector>(fields.size()); for (StructField f : fields) { fieldNames.add(f.getFieldName()); fieldObjectInspectors.add( getStandardObjectInspector(f.getFieldObjectInspector(), objectInspectorOption)); } result = ObjectInspectorFactory.getStandardStructObjectInspector( fieldNames, fieldObjectInspectors); break; } case UNION: { UnionObjectInspector uoi = (UnionObjectInspector) oi; List<ObjectInspector> ois = new ArrayList<ObjectInspector>(); for (ObjectInspector eoi : uoi.getObjectInspectors()) { ois.add(getStandardObjectInspector(eoi, objectInspectorOption)); } result = ObjectInspectorFactory.getStandardUnionObjectInspector(ois); break; } default: { throw new RuntimeException("Unknown ObjectInspector category!"); } } return result; }
public static int getStructSize(ObjectInspector oi) throws SerDeException { if (oi.getCategory() != Category.STRUCT) { throw new SerDeException("Unexpected category " + oi.getCategory()); } return ((StructObjectInspector) oi).getAllStructFieldRefs().size(); }
/** * Compares two types identified by the given object inspectors. This method compares the types as * follows: * * <ol> * <li>If the given inspectors do not belong to same category, the result is negative. * <li>If the given inspectors are for <code>PRIMITIVE</code> type, the result is the comparison * of their type names. * <li>If the given inspectors are for <code>LIST</code> type, then the result is recursive call * to compare the type of list elements. * <li>If the given inspectors are <code>MAP</code> type, then the result is a recursive call to * compare the map key and value types. * <li>If the given inspectors are <code>STRUCT</code> type, then the result is negative if they * do not have the same number of fields. If they do have the same number of fields, the * result is a recursive call to compare each of the field types. * <li>If none of the above, the result is negative. * </ol> * * @param o1 * @param o2 * @return true if the given object inspectors represent the same types. */ public static boolean compareTypes(ObjectInspector o1, ObjectInspector o2) { Category c1 = o1.getCategory(); Category c2 = o2.getCategory(); // Return false if categories are not equal if (!c1.equals(c2)) { return false; } // If both categories are primitive return the comparison of type names. if (c1.equals(Category.PRIMITIVE)) { return o1.getTypeName().equals(o2.getTypeName()); } // If lists, recursively compare the list element types if (c1.equals(Category.LIST)) { ObjectInspector child1 = ((ListObjectInspector) o1).getListElementObjectInspector(); ObjectInspector child2 = ((ListObjectInspector) o2).getListElementObjectInspector(); return compareTypes(child1, child2); } // If maps, recursively compare the key and value types if (c1.equals(Category.MAP)) { MapObjectInspector mapOI1 = (MapObjectInspector) o1; MapObjectInspector mapOI2 = (MapObjectInspector) o2; ObjectInspector childKey1 = mapOI1.getMapKeyObjectInspector(); ObjectInspector childKey2 = mapOI2.getMapKeyObjectInspector(); if (compareTypes(childKey1, childKey2)) { ObjectInspector childVal1 = mapOI1.getMapValueObjectInspector(); ObjectInspector childVal2 = mapOI2.getMapValueObjectInspector(); if (compareTypes(childVal1, childVal2)) { return true; } } return false; } // If structs, recursively compare the fields if (c1.equals(Category.STRUCT)) { StructObjectInspector structOI1 = (StructObjectInspector) o1; StructObjectInspector structOI2 = (StructObjectInspector) o2; List<? extends StructField> childFieldsList1 = structOI1.getAllStructFieldRefs(); List<? extends StructField> childFieldsList2 = structOI2.getAllStructFieldRefs(); if (childFieldsList1 == null && childFieldsList2 == null) { return true; } else if (childFieldsList1 == null || childFieldsList2 == null) { return false; } else if (childFieldsList1.size() != childFieldsList2.size()) { return false; } Iterator<? extends StructField> it1 = childFieldsList1.iterator(); Iterator<? extends StructField> it2 = childFieldsList2.iterator(); while (it1.hasNext()) { StructField field1 = it1.next(); StructField field2 = it2.next(); if (!compareTypes(field1.getFieldObjectInspector(), field2.getFieldObjectInspector())) { return false; } } return true; } if (c1.equals(Category.UNION)) { UnionObjectInspector uoi1 = (UnionObjectInspector) o1; UnionObjectInspector uoi2 = (UnionObjectInspector) o2; List<ObjectInspector> ois1 = uoi1.getObjectInspectors(); List<ObjectInspector> ois2 = uoi2.getObjectInspectors(); if (ois1 == null && ois2 == null) { return true; } else if (ois1 == null || ois2 == null) { return false; } else if (ois1.size() != ois2.size()) { return false; } Iterator<? extends ObjectInspector> it1 = ois1.iterator(); Iterator<? extends ObjectInspector> it2 = ois2.iterator(); while (it1.hasNext()) { if (!compareTypes(it1.next(), it2.next())) { return false; } } return true; } // Unknown category throw new RuntimeException("Unknown category encountered: " + c1); }
/** Compare two objects with their respective ObjectInspectors. */ public static int compare( Object o1, ObjectInspector oi1, Object o2, ObjectInspector oi2, MapEqualComparer mapEqualComparer) { if (oi1.getCategory() != oi2.getCategory()) { return oi1.getCategory().compareTo(oi2.getCategory()); } if (o1 == null) { return o2 == null ? 0 : -1; } else if (o2 == null) { return 1; } switch (oi1.getCategory()) { case PRIMITIVE: { PrimitiveObjectInspector poi1 = ((PrimitiveObjectInspector) oi1); PrimitiveObjectInspector poi2 = ((PrimitiveObjectInspector) oi2); if (poi1.getPrimitiveCategory() != poi2.getPrimitiveCategory()) { return poi1.getPrimitiveCategory().compareTo(poi2.getPrimitiveCategory()); } switch (poi1.getPrimitiveCategory()) { case VOID: return 0; case BOOLEAN: { int v1 = ((BooleanObjectInspector) poi1).get(o1) ? 1 : 0; int v2 = ((BooleanObjectInspector) poi2).get(o2) ? 1 : 0; return v1 - v2; } case BYTE: { int v1 = ((ByteObjectInspector) poi1).get(o1); int v2 = ((ByteObjectInspector) poi2).get(o2); return v1 - v2; } case SHORT: { int v1 = ((ShortObjectInspector) poi1).get(o1); int v2 = ((ShortObjectInspector) poi2).get(o2); return v1 - v2; } case INT: { int v1 = ((IntObjectInspector) poi1).get(o1); int v2 = ((IntObjectInspector) poi2).get(o2); return v1 > v2 ? 1 : (v1 < v2 ? -1 : 0); } case LONG: { long v1 = ((LongObjectInspector) poi1).get(o1); long v2 = ((LongObjectInspector) poi2).get(o2); return v1 > v2 ? 1 : (v1 < v2 ? -1 : 0); } case FLOAT: { float v1 = ((FloatObjectInspector) poi1).get(o1); float v2 = ((FloatObjectInspector) poi2).get(o2); return Float.compare(v1, v2); } case DOUBLE: { double v1 = ((DoubleObjectInspector) poi1).get(o1); double v2 = ((DoubleObjectInspector) poi2).get(o2); return Double.compare(v1, v2); } case STRING: { if (poi1.preferWritable() || poi2.preferWritable()) { Text t1 = (Text) poi1.getPrimitiveWritableObject(o1); Text t2 = (Text) poi2.getPrimitiveWritableObject(o2); return t1 == null ? (t2 == null ? 0 : -1) : (t2 == null ? 1 : t1.compareTo(t2)); } else { String s1 = (String) poi1.getPrimitiveJavaObject(o1); String s2 = (String) poi2.getPrimitiveJavaObject(o2); return s1 == null ? (s2 == null ? 0 : -1) : (s2 == null ? 1 : s1.compareTo(s2)); } } case CHAR: { HiveCharWritable t1 = ((HiveCharObjectInspector) poi1).getPrimitiveWritableObject(o1); HiveCharWritable t2 = ((HiveCharObjectInspector) poi2).getPrimitiveWritableObject(o2); return t1.compareTo(t2); } case VARCHAR: { HiveVarcharWritable t1 = ((HiveVarcharObjectInspector) poi1).getPrimitiveWritableObject(o1); HiveVarcharWritable t2 = ((HiveVarcharObjectInspector) poi2).getPrimitiveWritableObject(o2); return t1.compareTo(t2); } case BINARY: { BytesWritable bw1 = ((BinaryObjectInspector) poi1).getPrimitiveWritableObject(o1); BytesWritable bw2 = ((BinaryObjectInspector) poi2).getPrimitiveWritableObject(o2); return bw1.compareTo(bw2); } case DATE: { DateWritable d1 = ((DateObjectInspector) poi1).getPrimitiveWritableObject(o1); DateWritable d2 = ((DateObjectInspector) poi2).getPrimitiveWritableObject(o2); return d1.compareTo(d2); } case TIMESTAMP: { TimestampWritable t1 = ((TimestampObjectInspector) poi1).getPrimitiveWritableObject(o1); TimestampWritable t2 = ((TimestampObjectInspector) poi2).getPrimitiveWritableObject(o2); return t1.compareTo(t2); } case INTERVAL_YEAR_MONTH: { HiveIntervalYearMonthWritable i1 = ((HiveIntervalYearMonthObjectInspector) poi1).getPrimitiveWritableObject(o1); HiveIntervalYearMonthWritable i2 = ((HiveIntervalYearMonthObjectInspector) poi2).getPrimitiveWritableObject(o2); return i1.compareTo(i2); } case INTERVAL_DAY_TIME: { HiveIntervalDayTimeWritable i1 = ((HiveIntervalDayTimeObjectInspector) poi1).getPrimitiveWritableObject(o1); HiveIntervalDayTimeWritable i2 = ((HiveIntervalDayTimeObjectInspector) poi2).getPrimitiveWritableObject(o2); return i1.compareTo(i2); } case DECIMAL: { HiveDecimalWritable t1 = ((HiveDecimalObjectInspector) poi1).getPrimitiveWritableObject(o1); HiveDecimalWritable t2 = ((HiveDecimalObjectInspector) poi2).getPrimitiveWritableObject(o2); return t1.compareTo(t2); } default: { throw new RuntimeException("Unknown type: " + poi1.getPrimitiveCategory()); } } } case STRUCT: { StructObjectInspector soi1 = (StructObjectInspector) oi1; StructObjectInspector soi2 = (StructObjectInspector) oi2; List<? extends StructField> fields1 = soi1.getAllStructFieldRefs(); List<? extends StructField> fields2 = soi2.getAllStructFieldRefs(); int minimum = Math.min(fields1.size(), fields2.size()); for (int i = 0; i < minimum; i++) { int r = compare( soi1.getStructFieldData(o1, fields1.get(i)), fields1.get(i).getFieldObjectInspector(), soi2.getStructFieldData(o2, fields2.get(i)), fields2.get(i).getFieldObjectInspector(), mapEqualComparer); if (r != 0) { return r; } } return fields1.size() - fields2.size(); } case LIST: { ListObjectInspector loi1 = (ListObjectInspector) oi1; ListObjectInspector loi2 = (ListObjectInspector) oi2; int minimum = Math.min(loi1.getListLength(o1), loi2.getListLength(o2)); for (int i = 0; i < minimum; i++) { int r = compare( loi1.getListElement(o1, i), loi1.getListElementObjectInspector(), loi2.getListElement(o2, i), loi2.getListElementObjectInspector(), mapEqualComparer); if (r != 0) { return r; } } return loi1.getListLength(o1) - loi2.getListLength(o2); } case MAP: { if (mapEqualComparer == null) { throw new RuntimeException("Compare on map type not supported!"); } else { return mapEqualComparer.compare( o1, (MapObjectInspector) oi1, o2, (MapObjectInspector) oi2); } } case UNION: { UnionObjectInspector uoi1 = (UnionObjectInspector) oi1; UnionObjectInspector uoi2 = (UnionObjectInspector) oi2; byte tag1 = uoi1.getTag(o1); byte tag2 = uoi2.getTag(o2); if (tag1 != tag2) { return tag1 - tag2; } return compare( uoi1.getField(o1), uoi1.getObjectInspectors().get(tag1), uoi2.getField(o2), uoi2.getObjectInspectors().get(tag2), mapEqualComparer); } default: throw new RuntimeException("Compare on unknown type: " + oi1.getCategory()); } }
public static Object copyToStandardObject( Object o, ObjectInspector oi, ObjectInspectorCopyOption objectInspectorOption) { if (o == null) { return null; } Object result = null; switch (oi.getCategory()) { case PRIMITIVE: { PrimitiveObjectInspector loi = (PrimitiveObjectInspector) oi; if (objectInspectorOption == ObjectInspectorCopyOption.DEFAULT) { objectInspectorOption = loi.preferWritable() ? ObjectInspectorCopyOption.WRITABLE : ObjectInspectorCopyOption.JAVA; } switch (objectInspectorOption) { case JAVA: result = loi.getPrimitiveJavaObject(o); if (loi.getPrimitiveCategory() == PrimitiveObjectInspector.PrimitiveCategory.TIMESTAMP) { result = PrimitiveObjectInspectorFactory.javaTimestampObjectInspector.copyObject(result); } break; case WRITABLE: result = loi.getPrimitiveWritableObject(loi.copyObject(o)); break; } break; } case LIST: { ListObjectInspector loi = (ListObjectInspector) oi; int length = loi.getListLength(o); ArrayList<Object> list = new ArrayList<Object>(length); for (int i = 0; i < length; i++) { list.add( copyToStandardObject( loi.getListElement(o, i), loi.getListElementObjectInspector(), objectInspectorOption)); } result = list; break; } case MAP: { MapObjectInspector moi = (MapObjectInspector) oi; HashMap<Object, Object> map = new HashMap<Object, Object>(); Map<? extends Object, ? extends Object> omap = moi.getMap(o); for (Map.Entry<? extends Object, ? extends Object> entry : omap.entrySet()) { map.put( copyToStandardObject( entry.getKey(), moi.getMapKeyObjectInspector(), objectInspectorOption), copyToStandardObject( entry.getValue(), moi.getMapValueObjectInspector(), objectInspectorOption)); } result = map; break; } case STRUCT: { StructObjectInspector soi = (StructObjectInspector) oi; List<? extends StructField> fields = soi.getAllStructFieldRefs(); ArrayList<Object> struct = new ArrayList<Object>(fields.size()); for (StructField f : fields) { struct.add( copyToStandardObject( soi.getStructFieldData(o, f), f.getFieldObjectInspector(), objectInspectorOption)); } result = struct; break; } case UNION: { UnionObjectInspector uoi = (UnionObjectInspector) oi; List<ObjectInspector> objectInspectors = uoi.getObjectInspectors(); Object object = copyToStandardObject( uoi.getField(o), objectInspectors.get(uoi.getTag(o)), objectInspectorOption); result = object; break; } default: { throw new RuntimeException("Unknown ObjectInspector category!"); } } return result; }
/** Get the class names of the ObjectInspector hierarchy. Mainly used for debugging. */ public static String getObjectInspectorName(ObjectInspector oi) { switch (oi.getCategory()) { case PRIMITIVE: { return oi.getClass().getSimpleName(); } case LIST: { ListObjectInspector loi = (ListObjectInspector) oi; return oi.getClass().getSimpleName() + "<" + getObjectInspectorName(loi.getListElementObjectInspector()) + ">"; } case MAP: { MapObjectInspector moi = (MapObjectInspector) oi; return oi.getClass().getSimpleName() + "<" + getObjectInspectorName(moi.getMapKeyObjectInspector()) + "," + getObjectInspectorName(moi.getMapValueObjectInspector()) + ">"; } case STRUCT: { StringBuilder result = new StringBuilder(); result.append(oi.getClass().getSimpleName() + "<"); StructObjectInspector soi = (StructObjectInspector) oi; List<? extends StructField> fields = soi.getAllStructFieldRefs(); for (int i = 0; i < fields.size(); i++) { result.append(fields.get(i).getFieldName()); result.append(":"); result.append(getObjectInspectorName(fields.get(i).getFieldObjectInspector())); if (i == fields.size() - 1) { result.append(">"); } else { result.append(","); } } return result.toString(); } case UNION: { StringBuffer result = new StringBuffer(); result.append(oi.getClass().getSimpleName() + "<"); UnionObjectInspector uoi = (UnionObjectInspector) oi; List<ObjectInspector> ois = uoi.getObjectInspectors(); for (int i = 0; i < ois.size(); i++) { if (i > 0) { result.append(","); } result.append(getObjectInspectorName(ois.get(i))); } result.append(">"); return result.toString(); } default: { throw new RuntimeException("Unknown ObjectInspector category!"); } } }