Search in sources :

Example 51 with DateWritableV2

use of org.apache.hadoop.hive.serde2.io.DateWritableV2 in project hive by apache.

the class RecordReaderImpl method nextDate.

static DateWritableV2 nextDate(ColumnVector vector, int row, Object previous) {
    if (vector.isRepeating) {
        row = 0;
    }
    if (vector.noNulls || !vector.isNull[row]) {
        DateWritableV2 result;
        if (previous == null || previous.getClass() != DateWritableV2.class) {
            result = new DateWritableV2();
        } else {
            result = (DateWritableV2) previous;
        }
        int date = (int) ((DateColumnVector) vector).vector[row];
        result.set(date);
        return result;
    } else {
        return null;
    }
}
Also used : DateWritableV2(org.apache.hadoop.hive.serde2.io.DateWritableV2) DateColumnVector(org.apache.hadoop.hive.ql.exec.vector.DateColumnVector)

Example 52 with DateWritableV2

use of org.apache.hadoop.hive.serde2.io.DateWritableV2 in project hive by apache.

the class ColumnStatisticsObjTranslator method unpackDateStats.

private static void unpackDateStats(ObjectInspector oi, Object o, ColumnStatsField csf, ColumnStatisticsObj statsObj) {
    switch(csf) {
        case COUNT_NULLS:
            long cn = ((LongObjectInspector) oi).get(o);
            statsObj.getStatsData().getDateStats().setNumNulls(cn);
            break;
        case MIN:
            DateWritableV2 min = ((DateObjectInspector) oi).getPrimitiveWritableObject(o);
            statsObj.getStatsData().getDateStats().setLowValue(new Date(min.getDays()));
            break;
        case MAX:
            DateWritableV2 max = ((DateObjectInspector) oi).getPrimitiveWritableObject(o);
            statsObj.getStatsData().getDateStats().setHighValue(new Date(max.getDays()));
            break;
        case NDV:
            long ndv = ((LongObjectInspector) oi).get(o);
            statsObj.getStatsData().getDateStats().setNumDVs(ndv);
            break;
        case BITVECTOR:
            PrimitiveObjectInspector poi = (PrimitiveObjectInspector) oi;
            byte[] buf = ((BinaryObjectInspector) poi).getPrimitiveJavaObject(o);
            statsObj.getStatsData().getDateStats().setBitVectors(buf);
            break;
        default:
            throw new RuntimeException("Unsupported column stat for DATE : " + csf);
    }
}
Also used : LongObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.LongObjectInspector) DateObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.DateObjectInspector) DateWritableV2(org.apache.hadoop.hive.serde2.io.DateWritableV2) BinaryObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.BinaryObjectInspector) PrimitiveObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.PrimitiveObjectInspector) Date(org.apache.hadoop.hive.metastore.api.Date)

Example 53 with DateWritableV2

use of org.apache.hadoop.hive.serde2.io.DateWritableV2 in project hive by apache.

the class VectorVerifyFast method doVerifyDeserializeRead.

public static void doVerifyDeserializeRead(DeserializeRead deserializeRead, TypeInfo typeInfo, Object object, boolean isNull) throws IOException {
    if (isNull) {
        if (object != null) {
            TestCase.fail("Field reports null but object is not null (class " + object.getClass().getName() + ", " + object.toString() + ")");
        }
        return;
    } else if (object == null) {
        TestCase.fail("Field report not null but object is null");
    }
    switch(typeInfo.getCategory()) {
        case PRIMITIVE:
            {
                PrimitiveTypeInfo primitiveTypeInfo = (PrimitiveTypeInfo) typeInfo;
                switch(primitiveTypeInfo.getPrimitiveCategory()) {
                    case BOOLEAN:
                        {
                            boolean value = deserializeRead.currentBoolean;
                            if (!(object instanceof BooleanWritable)) {
                                TestCase.fail("Boolean expected writable not Boolean");
                            }
                            boolean expected = ((BooleanWritable) object).get();
                            if (value != expected) {
                                TestCase.fail("Boolean field mismatch (expected " + expected + " found " + value + ")");
                            }
                        }
                        break;
                    case BYTE:
                        {
                            byte value = deserializeRead.currentByte;
                            if (!(object instanceof ByteWritable)) {
                                TestCase.fail("Byte expected writable not Byte");
                            }
                            byte expected = ((ByteWritable) object).get();
                            if (value != expected) {
                                TestCase.fail("Byte field mismatch (expected " + (int) expected + " found " + (int) value + ")");
                            }
                        }
                        break;
                    case SHORT:
                        {
                            short value = deserializeRead.currentShort;
                            if (!(object instanceof ShortWritable)) {
                                TestCase.fail("Short expected writable not Short");
                            }
                            short expected = ((ShortWritable) object).get();
                            if (value != expected) {
                                TestCase.fail("Short field mismatch (expected " + expected + " found " + value + ")");
                            }
                        }
                        break;
                    case INT:
                        {
                            int value = deserializeRead.currentInt;
                            if (!(object instanceof IntWritable)) {
                                TestCase.fail("Integer expected writable not Integer");
                            }
                            int expected = ((IntWritable) object).get();
                            if (value != expected) {
                                TestCase.fail("Int field mismatch (expected " + expected + " found " + value + ")");
                            }
                        }
                        break;
                    case LONG:
                        {
                            long value = deserializeRead.currentLong;
                            if (!(object instanceof LongWritable)) {
                                TestCase.fail("Long expected writable not Long");
                            }
                            Long expected = ((LongWritable) object).get();
                            if (value != expected) {
                                TestCase.fail("Long field mismatch (expected " + expected + " found " + value + ")");
                            }
                        }
                        break;
                    case FLOAT:
                        {
                            float value = deserializeRead.currentFloat;
                            if (!(object instanceof FloatWritable)) {
                                TestCase.fail("Float expected writable not Float");
                            }
                            float expected = ((FloatWritable) object).get();
                            if (value != expected) {
                                TestCase.fail("Float field mismatch (expected " + expected + " found " + value + ")");
                            }
                        }
                        break;
                    case DOUBLE:
                        {
                            double value = deserializeRead.currentDouble;
                            if (!(object instanceof DoubleWritable)) {
                                TestCase.fail("Double expected writable not Double");
                            }
                            double expected = ((DoubleWritable) object).get();
                            if (value != expected) {
                                TestCase.fail("Double field mismatch (expected " + expected + " found " + value + ")");
                            }
                        }
                        break;
                    case STRING:
                        {
                            byte[] stringBytes = Arrays.copyOfRange(deserializeRead.currentBytes, deserializeRead.currentBytesStart, deserializeRead.currentBytesStart + deserializeRead.currentBytesLength);
                            Text text = new Text(stringBytes);
                            String string = text.toString();
                            String expected = ((Text) object).toString();
                            if (!string.equals(expected)) {
                                TestCase.fail("String field mismatch (expected '" + expected + "' found '" + string + "')");
                            }
                        }
                        break;
                    case CHAR:
                        {
                            byte[] stringBytes = Arrays.copyOfRange(deserializeRead.currentBytes, deserializeRead.currentBytesStart, deserializeRead.currentBytesStart + deserializeRead.currentBytesLength);
                            Text text = new Text(stringBytes);
                            String string = text.toString();
                            HiveChar hiveChar = new HiveChar(string, ((CharTypeInfo) primitiveTypeInfo).getLength());
                            HiveChar expected = ((HiveCharWritable) object).getHiveChar();
                            if (!hiveChar.equals(expected)) {
                                TestCase.fail("Char field mismatch (expected '" + expected + "' found '" + hiveChar + "')");
                            }
                        }
                        break;
                    case VARCHAR:
                        {
                            byte[] stringBytes = Arrays.copyOfRange(deserializeRead.currentBytes, deserializeRead.currentBytesStart, deserializeRead.currentBytesStart + deserializeRead.currentBytesLength);
                            Text text = new Text(stringBytes);
                            String string = text.toString();
                            HiveVarchar hiveVarchar = new HiveVarchar(string, ((VarcharTypeInfo) primitiveTypeInfo).getLength());
                            HiveVarchar expected = ((HiveVarcharWritable) object).getHiveVarchar();
                            if (!hiveVarchar.equals(expected)) {
                                TestCase.fail("Varchar field mismatch (expected '" + expected + "' found '" + hiveVarchar + "')");
                            }
                        }
                        break;
                    case DECIMAL:
                        {
                            HiveDecimal value = deserializeRead.currentHiveDecimalWritable.getHiveDecimal();
                            if (value == null) {
                                TestCase.fail("Decimal field evaluated to NULL");
                            }
                            HiveDecimal expected = ((HiveDecimalWritable) object).getHiveDecimal();
                            if (!value.equals(expected)) {
                                DecimalTypeInfo decimalTypeInfo = (DecimalTypeInfo) primitiveTypeInfo;
                                int precision = decimalTypeInfo.getPrecision();
                                int scale = decimalTypeInfo.getScale();
                                TestCase.fail("Decimal field mismatch (expected " + expected.toString() + " found " + value.toString() + ") precision " + precision + ", scale " + scale);
                            }
                        }
                        break;
                    case DATE:
                        {
                            Date value = deserializeRead.currentDateWritable.get();
                            Date expected = ((DateWritableV2) object).get();
                            if (!value.equals(expected)) {
                                TestCase.fail("Date field mismatch (expected " + expected.toString() + " found " + value.toString() + ")");
                            }
                        }
                        break;
                    case TIMESTAMP:
                        {
                            Timestamp value = deserializeRead.currentTimestampWritable.getTimestamp();
                            Timestamp expected = ((TimestampWritableV2) object).getTimestamp();
                            if (!value.equals(expected)) {
                                TestCase.fail("Timestamp field mismatch (expected " + expected.toString() + " found " + value.toString() + ")");
                            }
                        }
                        break;
                    case INTERVAL_YEAR_MONTH:
                        {
                            HiveIntervalYearMonth value = deserializeRead.currentHiveIntervalYearMonthWritable.getHiveIntervalYearMonth();
                            HiveIntervalYearMonth expected = ((HiveIntervalYearMonthWritable) object).getHiveIntervalYearMonth();
                            if (!value.equals(expected)) {
                                TestCase.fail("HiveIntervalYearMonth field mismatch (expected " + expected.toString() + " found " + value.toString() + ")");
                            }
                        }
                        break;
                    case INTERVAL_DAY_TIME:
                        {
                            HiveIntervalDayTime value = deserializeRead.currentHiveIntervalDayTimeWritable.getHiveIntervalDayTime();
                            HiveIntervalDayTime expected = ((HiveIntervalDayTimeWritable) object).getHiveIntervalDayTime();
                            if (!value.equals(expected)) {
                                TestCase.fail("HiveIntervalDayTime field mismatch (expected " + expected.toString() + " found " + value.toString() + ")");
                            }
                        }
                        break;
                    case BINARY:
                        {
                            byte[] byteArray = Arrays.copyOfRange(deserializeRead.currentBytes, deserializeRead.currentBytesStart, deserializeRead.currentBytesStart + deserializeRead.currentBytesLength);
                            BytesWritable bytesWritable = (BytesWritable) object;
                            byte[] expected = Arrays.copyOfRange(bytesWritable.getBytes(), 0, bytesWritable.getLength());
                            if (byteArray.length != expected.length) {
                                TestCase.fail("Byte Array field mismatch (expected " + Arrays.toString(expected) + " found " + Arrays.toString(byteArray) + ")");
                            }
                            for (int b = 0; b < byteArray.length; b++) {
                                if (byteArray[b] != expected[b]) {
                                    TestCase.fail("Byte Array field mismatch (expected " + Arrays.toString(expected) + " found " + Arrays.toString(byteArray) + ")");
                                }
                            }
                        }
                        break;
                    default:
                        throw new Error("Unknown primitive category " + primitiveTypeInfo.getPrimitiveCategory());
                }
            }
            break;
        case LIST:
        case MAP:
        case STRUCT:
        case UNION:
            throw new Error("Complex types need to be handled separately");
        default:
            throw new Error("Unknown category " + typeInfo.getCategory());
    }
}
Also used : HiveChar(org.apache.hadoop.hive.common.type.HiveChar) DoubleWritable(org.apache.hadoop.hive.serde2.io.DoubleWritable) ShortWritable(org.apache.hadoop.hive.serde2.io.ShortWritable) Timestamp(org.apache.hadoop.hive.common.type.Timestamp) PrimitiveTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.PrimitiveTypeInfo) HiveDecimal(org.apache.hadoop.hive.common.type.HiveDecimal) LongWritable(org.apache.hadoop.io.LongWritable) ByteWritable(org.apache.hadoop.hive.serde2.io.ByteWritable) IntWritable(org.apache.hadoop.io.IntWritable) HiveIntervalDayTime(org.apache.hadoop.hive.common.type.HiveIntervalDayTime) HiveDecimalWritable(org.apache.hadoop.hive.serde2.io.HiveDecimalWritable) HiveCharWritable(org.apache.hadoop.hive.serde2.io.HiveCharWritable) HiveVarcharWritable(org.apache.hadoop.hive.serde2.io.HiveVarcharWritable) DateWritableV2(org.apache.hadoop.hive.serde2.io.DateWritableV2) Text(org.apache.hadoop.io.Text) HiveIntervalDayTimeWritable(org.apache.hadoop.hive.serde2.io.HiveIntervalDayTimeWritable) BytesWritable(org.apache.hadoop.io.BytesWritable) HiveVarchar(org.apache.hadoop.hive.common.type.HiveVarchar) HiveIntervalYearMonthWritable(org.apache.hadoop.hive.serde2.io.HiveIntervalYearMonthWritable) TimestampWritableV2(org.apache.hadoop.hive.serde2.io.TimestampWritableV2) Date(org.apache.hadoop.hive.common.type.Date) DecimalTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.DecimalTypeInfo) FloatWritable(org.apache.hadoop.io.FloatWritable) HiveIntervalYearMonth(org.apache.hadoop.hive.common.type.HiveIntervalYearMonth) BooleanWritable(org.apache.hadoop.io.BooleanWritable)

Example 54 with DateWritableV2

use of org.apache.hadoop.hive.serde2.io.DateWritableV2 in project hive by apache.

the class VectorVerifyFast 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[] bytes = value.getBytes();
                            int byteLength = value.getLength();
                            serializeWrite.writeString(bytes, 0, byteLength);
                        }
                        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();
                HashMap<Object, Object> hashMap = (HashMap<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());
    }
}
Also used : StandardUnionObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.StandardUnionObjectInspector) HashMap(java.util.HashMap) HiveChar(org.apache.hadoop.hive.common.type.HiveChar) ArrayList(java.util.ArrayList) DoubleWritable(org.apache.hadoop.hive.serde2.io.DoubleWritable) StructTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.StructTypeInfo) ShortWritable(org.apache.hadoop.hive.serde2.io.ShortWritable) Timestamp(org.apache.hadoop.hive.common.type.Timestamp) PrimitiveTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.PrimitiveTypeInfo) HiveDecimal(org.apache.hadoop.hive.common.type.HiveDecimal) ArrayList(java.util.ArrayList) List(java.util.List) LongWritable(org.apache.hadoop.io.LongWritable) ByteWritable(org.apache.hadoop.hive.serde2.io.ByteWritable) IntWritable(org.apache.hadoop.io.IntWritable) HiveIntervalDayTime(org.apache.hadoop.hive.common.type.HiveIntervalDayTime) HiveDecimalWritable(org.apache.hadoop.hive.serde2.io.HiveDecimalWritable) HiveCharWritable(org.apache.hadoop.hive.serde2.io.HiveCharWritable) HiveVarcharWritable(org.apache.hadoop.hive.serde2.io.HiveVarcharWritable) DateWritableV2(org.apache.hadoop.hive.serde2.io.DateWritableV2) Text(org.apache.hadoop.io.Text) HiveIntervalDayTimeWritable(org.apache.hadoop.hive.serde2.io.HiveIntervalDayTimeWritable) BytesWritable(org.apache.hadoop.io.BytesWritable) HiveVarchar(org.apache.hadoop.hive.common.type.HiveVarchar) HiveIntervalYearMonthWritable(org.apache.hadoop.hive.serde2.io.HiveIntervalYearMonthWritable) TimestampWritableV2(org.apache.hadoop.hive.serde2.io.TimestampWritableV2) MapTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.MapTypeInfo) ListTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.ListTypeInfo) StructTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.StructTypeInfo) PrimitiveTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.PrimitiveTypeInfo) DecimalTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.DecimalTypeInfo) TypeInfo(org.apache.hadoop.hive.serde2.typeinfo.TypeInfo) UnionTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.UnionTypeInfo) VarcharTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.VarcharTypeInfo) CharTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.CharTypeInfo) Date(org.apache.hadoop.hive.common.type.Date) DecimalTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.DecimalTypeInfo) FloatWritable(org.apache.hadoop.io.FloatWritable) HiveIntervalYearMonth(org.apache.hadoop.hive.common.type.HiveIntervalYearMonth) BooleanWritable(org.apache.hadoop.io.BooleanWritable) ListTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.ListTypeInfo) MapTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.MapTypeInfo) UnionTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.UnionTypeInfo)

Example 55 with DateWritableV2

use of org.apache.hadoop.hive.serde2.io.DateWritableV2 in project hive by apache.

the class BinarySortableSerDe method deserialize.

static Object deserialize(InputByteBuffer buffer, TypeInfo type, boolean invert, byte nullMarker, byte notNullMarker, Object reuse) throws IOException {
    // Is this field a null?
    byte isNull = buffer.read(invert);
    if (isNull == nullMarker) {
        return null;
    }
    assert (isNull == notNullMarker);
    switch(type.getCategory()) {
        case PRIMITIVE:
            {
                PrimitiveTypeInfo ptype = (PrimitiveTypeInfo) type;
                switch(ptype.getPrimitiveCategory()) {
                    case VOID:
                        {
                            return null;
                        }
                    case BOOLEAN:
                        {
                            BooleanWritable r = reuse == null ? new BooleanWritable() : (BooleanWritable) reuse;
                            byte b = buffer.read(invert);
                            assert (b == 1 || b == 2);
                            r.set(b == 2);
                            return r;
                        }
                    case BYTE:
                        {
                            ByteWritable r = reuse == null ? new ByteWritable() : (ByteWritable) reuse;
                            r.set((byte) (buffer.read(invert) ^ 0x80));
                            return r;
                        }
                    case SHORT:
                        {
                            ShortWritable r = reuse == null ? new ShortWritable() : (ShortWritable) reuse;
                            int v = buffer.read(invert) ^ 0x80;
                            v = (v << 8) + (buffer.read(invert) & 0xff);
                            r.set((short) v);
                            return r;
                        }
                    case INT:
                        {
                            IntWritable r = reuse == null ? new IntWritable() : (IntWritable) reuse;
                            r.set(deserializeInt(buffer, invert));
                            return r;
                        }
                    case LONG:
                        {
                            LongWritable r = reuse == null ? new LongWritable() : (LongWritable) reuse;
                            r.set(deserializeLong(buffer, invert));
                            return r;
                        }
                    case FLOAT:
                        {
                            FloatWritable r = reuse == null ? new FloatWritable() : (FloatWritable) reuse;
                            int v = 0;
                            for (int i = 0; i < 4; i++) {
                                v = (v << 8) + (buffer.read(invert) & 0xff);
                            }
                            if ((v & (1 << 31)) == 0) {
                                // negative number, flip all bits
                                v = ~v;
                            } else {
                                // positive number, flip the first bit
                                v = v ^ (1 << 31);
                            }
                            r.set(Float.intBitsToFloat(v));
                            return r;
                        }
                    case DOUBLE:
                        {
                            DoubleWritable r = reuse == null ? new DoubleWritable() : (DoubleWritable) reuse;
                            long v = 0;
                            for (int i = 0; i < 8; i++) {
                                v = (v << 8) + (buffer.read(invert) & 0xff);
                            }
                            if ((v & (1L << 63)) == 0) {
                                // negative number, flip all bits
                                v = ~v;
                            } else {
                                // positive number, flip the first bit
                                v = v ^ (1L << 63);
                            }
                            r.set(Double.longBitsToDouble(v));
                            return r;
                        }
                    case STRING:
                        {
                            Text r = reuse == null ? new Text() : (Text) reuse;
                            return deserializeText(buffer, invert, r);
                        }
                    case CHAR:
                        {
                            HiveCharWritable r = reuse == null ? new HiveCharWritable() : (HiveCharWritable) reuse;
                            // Use internal text member to read value
                            deserializeText(buffer, invert, r.getTextValue());
                            r.enforceMaxLength(getCharacterMaxLength(type));
                            return r;
                        }
                    case VARCHAR:
                        {
                            HiveVarcharWritable r = reuse == null ? new HiveVarcharWritable() : (HiveVarcharWritable) reuse;
                            // Use HiveVarchar's internal Text member to read the value.
                            deserializeText(buffer, invert, r.getTextValue());
                            // If we cache helper data for deserialization we could avoid having
                            // to call getVarcharMaxLength() on every deserialize call.
                            r.enforceMaxLength(getCharacterMaxLength(type));
                            return r;
                        }
                    case BINARY:
                        {
                            BytesWritable bw = new BytesWritable();
                            // Get the actual length first
                            int start = buffer.tell();
                            int length = 0;
                            do {
                                byte b = buffer.read(invert);
                                if (b == 0) {
                                    // end of string
                                    break;
                                }
                                if (b == 1) {
                                    // the last char is an escape char. read the actual char
                                    buffer.read(invert);
                                }
                                length++;
                            } while (true);
                            if (length == buffer.tell() - start) {
                                // No escaping happened, so we are already done.
                                bw.set(buffer.getData(), start, length);
                            } else {
                                // Escaping happened, we need to copy byte-by-byte.
                                // 1. Set the length first.
                                bw.set(buffer.getData(), start, length);
                                // 2. Reset the pointer.
                                buffer.seek(start);
                                // 3. Copy the data.
                                byte[] rdata = bw.getBytes();
                                for (int i = 0; i < length; i++) {
                                    byte b = buffer.read(invert);
                                    if (b == 1) {
                                        // The last char is an escape char, read the actual char.
                                        // The serialization format escape \0 to \1, and \1 to \2,
                                        // to make sure the string is null-terminated.
                                        b = (byte) (buffer.read(invert) - 1);
                                    }
                                    rdata[i] = b;
                                }
                                // 4. Read the null terminator.
                                byte b = buffer.read(invert);
                                assert (b == 0);
                            }
                            return bw;
                        }
                    case DATE:
                        {
                            DateWritableV2 d = reuse == null ? new DateWritableV2() : (DateWritableV2) reuse;
                            d.set(deserializeInt(buffer, invert));
                            return d;
                        }
                    case TIMESTAMP:
                        TimestampWritableV2 t = (reuse == null ? new TimestampWritableV2() : (TimestampWritableV2) reuse);
                        byte[] bytes = new byte[TimestampWritableV2.BINARY_SORTABLE_LENGTH];
                        for (int i = 0; i < bytes.length; i++) {
                            bytes[i] = buffer.read(invert);
                        }
                        t.setBinarySortable(bytes, 0);
                        return t;
                    case TIMESTAMPLOCALTZ:
                        TimestampLocalTZWritable tstz = (reuse == null ? new TimestampLocalTZWritable() : (TimestampLocalTZWritable) reuse);
                        byte[] data = new byte[TimestampLocalTZWritable.BINARY_SORTABLE_LENGTH];
                        for (int i = 0; i < data.length; i++) {
                            data[i] = buffer.read(invert);
                        }
                        // Across MR process boundary tz is normalized and stored in type
                        // and is not carried in data for each row.
                        tstz.fromBinarySortable(data, 0, ((TimestampLocalTZTypeInfo) type).timeZone());
                        return tstz;
                    case INTERVAL_YEAR_MONTH:
                        {
                            HiveIntervalYearMonthWritable i = reuse == null ? new HiveIntervalYearMonthWritable() : (HiveIntervalYearMonthWritable) reuse;
                            i.set(deserializeInt(buffer, invert));
                            return i;
                        }
                    case INTERVAL_DAY_TIME:
                        {
                            HiveIntervalDayTimeWritable i = reuse == null ? new HiveIntervalDayTimeWritable() : (HiveIntervalDayTimeWritable) reuse;
                            long totalSecs = deserializeLong(buffer, invert);
                            int nanos = deserializeInt(buffer, invert);
                            i.set(totalSecs, nanos);
                            return i;
                        }
                    case DECIMAL:
                        {
                            // See serialization of decimal for explanation (below)
                            HiveDecimalWritable bdw = (reuse == null ? new HiveDecimalWritable() : (HiveDecimalWritable) reuse);
                            int b = buffer.read(invert) - 1;
                            assert (b == 1 || b == -1 || b == 0);
                            boolean positive = b != -1;
                            int factor = buffer.read(invert) ^ 0x80;
                            for (int i = 0; i < 3; i++) {
                                factor = (factor << 8) + (buffer.read(invert) & 0xff);
                            }
                            if (!positive) {
                                factor = -factor;
                            }
                            int start = buffer.tell();
                            int length = 0;
                            do {
                                b = buffer.read(positive ? invert : !invert);
                                assert (b != 1);
                                if (b == 0) {
                                    // end of digits
                                    break;
                                }
                                length++;
                            } while (true);
                            final byte[] decimalBuffer = new byte[length];
                            buffer.seek(start);
                            for (int i = 0; i < length; ++i) {
                                decimalBuffer[i] = buffer.read(positive ? invert : !invert);
                            }
                            // read the null byte again
                            buffer.read(positive ? invert : !invert);
                            String digits = new String(decimalBuffer, 0, length, decimalCharSet);
                            BigInteger bi = new BigInteger(digits);
                            HiveDecimal bd = HiveDecimal.create(bi).scaleByPowerOfTen(factor - length);
                            if (!positive) {
                                bd = bd.negate();
                            }
                            bdw.set(bd);
                            return bdw;
                        }
                    default:
                        {
                            throw new RuntimeException("Unrecognized type: " + ptype.getPrimitiveCategory());
                        }
                }
            }
        case LIST:
            {
                ListTypeInfo ltype = (ListTypeInfo) type;
                TypeInfo etype = ltype.getListElementTypeInfo();
                // Create the list if needed
                ArrayList<Object> r = reuse == null ? new ArrayList<Object>() : (ArrayList<Object>) reuse;
                // Read the list
                int size = 0;
                while (true) {
                    int more = buffer.read(invert);
                    if (more == 0) {
                        // \0 to terminate
                        break;
                    }
                    // \1 followed by each element
                    assert (more == 1);
                    if (size == r.size()) {
                        r.add(null);
                    }
                    r.set(size, deserialize(buffer, etype, invert, nullMarker, notNullMarker, r.get(size)));
                    size++;
                }
                // Remove additional elements if the list is reused
                while (r.size() > size) {
                    r.remove(r.size() - 1);
                }
                return r;
            }
        case MAP:
            {
                MapTypeInfo mtype = (MapTypeInfo) type;
                TypeInfo ktype = mtype.getMapKeyTypeInfo();
                TypeInfo vtype = mtype.getMapValueTypeInfo();
                // Create the map if needed
                Map<Object, Object> r;
                if (reuse == null || reuse.getClass() != LinkedHashMap.class) {
                    r = new LinkedHashMap<Object, Object>();
                } else {
                    r = (Map<Object, Object>) reuse;
                    r.clear();
                }
                while (true) {
                    int more = buffer.read(invert);
                    if (more == 0) {
                        // \0 to terminate
                        break;
                    }
                    // \1 followed by each key and then each value
                    assert (more == 1);
                    Object k = deserialize(buffer, ktype, invert, nullMarker, notNullMarker, null);
                    Object v = deserialize(buffer, vtype, invert, nullMarker, notNullMarker, null);
                    r.put(k, v);
                }
                return r;
            }
        case STRUCT:
            {
                StructTypeInfo stype = (StructTypeInfo) type;
                List<TypeInfo> fieldTypes = stype.getAllStructFieldTypeInfos();
                int size = fieldTypes.size();
                // Create the struct if needed
                ArrayList<Object> r = reuse == null ? new ArrayList<Object>(size) : (ArrayList<Object>) reuse;
                assert (r.size() <= size);
                // Set the size of the struct
                while (r.size() < size) {
                    r.add(null);
                }
                // Read one field by one field
                for (int eid = 0; eid < size; eid++) {
                    r.set(eid, deserialize(buffer, fieldTypes.get(eid), invert, nullMarker, notNullMarker, r.get(eid)));
                }
                return r;
            }
        case UNION:
            {
                UnionTypeInfo utype = (UnionTypeInfo) type;
                StandardUnion r = reuse == null ? new StandardUnion() : (StandardUnion) reuse;
                // Read the tag
                byte tag = buffer.read(invert);
                r.setTag(tag);
                r.setObject(deserialize(buffer, utype.getAllUnionObjectTypeInfos().get(tag), invert, nullMarker, notNullMarker, null));
                return r;
            }
        default:
            {
                throw new RuntimeException("Unrecognized type: " + type.getCategory());
            }
    }
}
Also used : ArrayList(java.util.ArrayList) DoubleWritable(org.apache.hadoop.hive.serde2.io.DoubleWritable) StructTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.StructTypeInfo) ShortWritable(org.apache.hadoop.hive.serde2.io.ShortWritable) PrimitiveTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.PrimitiveTypeInfo) LinkedHashMap(java.util.LinkedHashMap) HiveDecimal(org.apache.hadoop.hive.common.type.HiveDecimal) List(java.util.List) ArrayList(java.util.ArrayList) LongWritable(org.apache.hadoop.io.LongWritable) ByteWritable(org.apache.hadoop.hive.serde2.io.ByteWritable) IntWritable(org.apache.hadoop.io.IntWritable) HiveDecimalWritable(org.apache.hadoop.hive.serde2.io.HiveDecimalWritable) HiveCharWritable(org.apache.hadoop.hive.serde2.io.HiveCharWritable) HiveVarcharWritable(org.apache.hadoop.hive.serde2.io.HiveVarcharWritable) DateWritableV2(org.apache.hadoop.hive.serde2.io.DateWritableV2) Text(org.apache.hadoop.io.Text) BytesWritable(org.apache.hadoop.io.BytesWritable) TimestampLocalTZTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.TimestampLocalTZTypeInfo) HiveIntervalDayTimeWritable(org.apache.hadoop.hive.serde2.io.HiveIntervalDayTimeWritable) HiveIntervalYearMonthWritable(org.apache.hadoop.hive.serde2.io.HiveIntervalYearMonthWritable) TimestampWritableV2(org.apache.hadoop.hive.serde2.io.TimestampWritableV2) TimestampLocalTZTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.TimestampLocalTZTypeInfo) MapTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.MapTypeInfo) StructTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.StructTypeInfo) PrimitiveTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.PrimitiveTypeInfo) ListTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.ListTypeInfo) TypeInfo(org.apache.hadoop.hive.serde2.typeinfo.TypeInfo) UnionTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.UnionTypeInfo) BaseCharTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.BaseCharTypeInfo) FloatWritable(org.apache.hadoop.io.FloatWritable) BooleanWritable(org.apache.hadoop.io.BooleanWritable) TimestampLocalTZWritable(org.apache.hadoop.hive.serde2.io.TimestampLocalTZWritable) ListTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.ListTypeInfo) MapTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.MapTypeInfo) StandardUnion(org.apache.hadoop.hive.serde2.objectinspector.StandardUnionObjectInspector.StandardUnion) BigInteger(java.math.BigInteger) Map(java.util.Map) LinkedHashMap(java.util.LinkedHashMap) UnionTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.UnionTypeInfo)

Aggregations

DateWritableV2 (org.apache.hadoop.hive.serde2.io.DateWritableV2)73 Test (org.junit.Test)36 ObjectInspector (org.apache.hadoop.hive.serde2.objectinspector.ObjectInspector)34 Text (org.apache.hadoop.io.Text)30 DeferredJavaObject (org.apache.hadoop.hive.ql.udf.generic.GenericUDF.DeferredJavaObject)29 DeferredObject (org.apache.hadoop.hive.ql.udf.generic.GenericUDF.DeferredObject)29 TimestampWritableV2 (org.apache.hadoop.hive.serde2.io.TimestampWritableV2)28 Date (org.apache.hadoop.hive.common.type.Date)24 HiveDecimalWritable (org.apache.hadoop.hive.serde2.io.HiveDecimalWritable)23 Timestamp (org.apache.hadoop.hive.common.type.Timestamp)21 PrimitiveTypeInfo (org.apache.hadoop.hive.serde2.typeinfo.PrimitiveTypeInfo)21 BytesWritable (org.apache.hadoop.io.BytesWritable)21 IntWritable (org.apache.hadoop.io.IntWritable)18 HiveDecimal (org.apache.hadoop.hive.common.type.HiveDecimal)16 HiveIntervalDayTimeWritable (org.apache.hadoop.hive.serde2.io.HiveIntervalDayTimeWritable)16 ByteWritable (org.apache.hadoop.hive.serde2.io.ByteWritable)15 DoubleWritable (org.apache.hadoop.hive.serde2.io.DoubleWritable)15 HiveIntervalYearMonthWritable (org.apache.hadoop.hive.serde2.io.HiveIntervalYearMonthWritable)15 ShortWritable (org.apache.hadoop.hive.serde2.io.ShortWritable)15 LongWritable (org.apache.hadoop.io.LongWritable)15