Examples with TypeInfo org.apache.hadoop.hive.serde2.typeinfo.TypeInfo used on opensource projects

Example 56 with TypeInfo

use of org.apache.hadoop.hive.serde2.typeinfo.TypeInfo in project mongo-hadoop by mongodb.

the class BSONSerDe method deserialize.

/**
     * Given a Writable object of BSON, turn it into a Hive table row
     */
@Override
public //CHECKSTYLE:OFF
Object deserialize(final Writable writable) throws SerDeException {
    //CHECKSTYLE:ON
    BSONObject doc;
    row.clear();
    // Make sure it's a BSONWritable object
    if (writable instanceof BSONWritable) {
        doc = ((BSONWritable) writable).getDoc();
    } else {
        throw new SerDeException(format("%srequires a BSONWritable object, not%s", getClass(), writable.getClass()));
    }
    // For each field, cast it to a HIVE type and add to the current row
    Object value;
    List<String> structFieldNames = docTypeInfo.getAllStructFieldNames();
    for (String fieldName : structFieldNames) {
        try {
            TypeInfo fieldTypeInfo = docTypeInfo.getStructFieldTypeInfo(fieldName);
            // get the corresponding field name in MongoDB
            String mongoMapping;
            if (hiveToMongo == null) {
                mongoMapping = fieldName;
            } else {
                mongoMapping = hiveToMongo.containsKey(fieldName) ? hiveToMongo.get(fieldName) : fieldName;
            }
            value = deserializeField(getValue(doc, mongoMapping), fieldTypeInfo, fieldName);
        } catch (Exception e) {
            LOG.warn("Could not find the appropriate field for name " + fieldName);
            value = null;
        }
        row.add(value);
    }
    return row;
}

Example 57 with TypeInfo

use of org.apache.hadoop.hive.serde2.typeinfo.TypeInfo in project mongo-hadoop by mongodb.

the class BSONSerDe method deserializeMap.

/**
     * Also deserialize a Map with the same mapElemTypeInfo
     * @param value the value for which to get the Hive representation
     * @param valueTypeInfo a description of the value's type
     * @param ext the field name
     * @return the Hive representation of the value
     */
private Object deserializeMap(final Object value, final MapTypeInfo valueTypeInfo, final String ext) {
    BasicBSONObject b = (BasicBSONObject) value;
    TypeInfo mapValueTypeInfo = valueTypeInfo.getMapValueTypeInfo();
    for (Entry<String, Object> entry : b.entrySet()) {
        b.put(entry.getKey(), deserializeField(entry.getValue(), mapValueTypeInfo, ext));
    }
    return b.toMap();
}

Example 58 with TypeInfo

use of org.apache.hadoop.hive.serde2.typeinfo.TypeInfo in project hive by apache.

the class GenericUDFLower method initialize.

@Override
public ObjectInspector initialize(ObjectInspector[] arguments) throws UDFArgumentException {
    if (arguments.length != 1) {
        throw new UDFArgumentLengthException("LOWER requires 1 argument, got " + arguments.length);
    }
    if (arguments[0].getCategory() != Category.PRIMITIVE) {
        throw new UDFArgumentException("LOWER only takes primitive types, got " + argumentOI.getTypeName());
    }
    argumentOI = (PrimitiveObjectInspector) arguments[0];
    stringConverter = new PrimitiveObjectInspectorConverter.StringConverter(argumentOI);
    PrimitiveCategory inputType = argumentOI.getPrimitiveCategory();
    ObjectInspector outputOI = null;
    BaseCharTypeInfo typeInfo;
    switch(inputType) {
        case CHAR:
            // return type should have same length as the input.
            returnType = inputType;
            typeInfo = TypeInfoFactory.getCharTypeInfo(GenericUDFUtils.StringHelper.getFixedStringSizeForType(argumentOI));
            outputOI = PrimitiveObjectInspectorFactory.getPrimitiveWritableObjectInspector(typeInfo);
            break;
        case VARCHAR:
            // return type should have same length as the input.
            returnType = inputType;
            typeInfo = TypeInfoFactory.getVarcharTypeInfo(GenericUDFUtils.StringHelper.getFixedStringSizeForType(argumentOI));
            outputOI = PrimitiveObjectInspectorFactory.getPrimitiveWritableObjectInspector(typeInfo);
            break;
        default:
            returnType = PrimitiveCategory.STRING;
            outputOI = PrimitiveObjectInspectorFactory.writableStringObjectInspector;
            break;
    }
    returnHelper = new GenericUDFUtils.StringHelper(returnType);
    return outputOI;
}

Example 59 with TypeInfo

use of org.apache.hadoop.hive.serde2.typeinfo.TypeInfo in project hive by apache.

the class GenericUDFOPDTIPlus method initialize.

@Override
public ObjectInspector initialize(ObjectInspector[] arguments) throws UDFArgumentException {
    if (arguments.length != 2) {
        throw new UDFArgumentException(opName + " requires two arguments.");
    }
    PrimitiveObjectInspector resultOI = null;
    for (int i = 0; i < 2; i++) {
        Category category = arguments[i].getCategory();
        if (category != Category.PRIMITIVE) {
            throw new UDFArgumentTypeException(i, "The " + GenericUDFUtils.getOrdinal(i + 1) + " argument of " + opName + "  is expected to a " + Category.PRIMITIVE.toString().toLowerCase() + " type, but " + category.toString().toLowerCase() + " is found");
        }
    }
    inputOIs = new PrimitiveObjectInspector[] { (PrimitiveObjectInspector) arguments[0], (PrimitiveObjectInspector) arguments[1] };
    PrimitiveObjectInspector leftOI = inputOIs[0];
    PrimitiveObjectInspector rightOI = inputOIs[1];
    // IntervalDayTime + Timestamp = Timestamp (operands reversible)
    if (checkArgs(PrimitiveCategory.INTERVAL_YEAR_MONTH, PrimitiveCategory.INTERVAL_YEAR_MONTH)) {
        plusOpType = OperationType.INTERVALYM_PLUS_INTERVALYM;
        intervalArg1Idx = 0;
        intervalArg2Idx = 1;
        resultOI = PrimitiveObjectInspectorFactory.getPrimitiveWritableObjectInspector(TypeInfoFactory.intervalYearMonthTypeInfo);
    } else if (checkArgs(PrimitiveCategory.DATE, PrimitiveCategory.INTERVAL_YEAR_MONTH)) {
        plusOpType = OperationType.INTERVALYM_PLUS_DATE;
        dtArgIdx = 0;
        intervalArg1Idx = 1;
        resultOI = PrimitiveObjectInspectorFactory.getPrimitiveWritableObjectInspector(TypeInfoFactory.dateTypeInfo);
    } else if (checkArgs(PrimitiveCategory.INTERVAL_YEAR_MONTH, PrimitiveCategory.DATE)) {
        plusOpType = OperationType.INTERVALYM_PLUS_DATE;
        intervalArg1Idx = 0;
        dtArgIdx = 1;
        resultOI = PrimitiveObjectInspectorFactory.getPrimitiveWritableObjectInspector(TypeInfoFactory.dateTypeInfo);
    } else if (checkArgs(PrimitiveCategory.TIMESTAMP, PrimitiveCategory.INTERVAL_YEAR_MONTH)) {
        plusOpType = OperationType.INTERVALYM_PLUS_TIMESTAMP;
        dtArgIdx = 0;
        intervalArg1Idx = 1;
        resultOI = PrimitiveObjectInspectorFactory.getPrimitiveWritableObjectInspector(TypeInfoFactory.timestampTypeInfo);
    } else if (checkArgs(PrimitiveCategory.INTERVAL_YEAR_MONTH, PrimitiveCategory.TIMESTAMP)) {
        plusOpType = OperationType.INTERVALYM_PLUS_TIMESTAMP;
        intervalArg1Idx = 0;
        dtArgIdx = 1;
        resultOI = PrimitiveObjectInspectorFactory.getPrimitiveWritableObjectInspector(TypeInfoFactory.timestampTypeInfo);
    } else if (checkArgs(PrimitiveCategory.INTERVAL_DAY_TIME, PrimitiveCategory.INTERVAL_DAY_TIME)) {
        plusOpType = OperationType.INTERVALDT_PLUS_INTERVALDT;
        intervalArg1Idx = 0;
        intervalArg2Idx = 1;
        resultOI = PrimitiveObjectInspectorFactory.getPrimitiveWritableObjectInspector(TypeInfoFactory.intervalDayTimeTypeInfo);
    } else if (checkArgs(PrimitiveCategory.INTERVAL_DAY_TIME, PrimitiveCategory.DATE) || checkArgs(PrimitiveCategory.INTERVAL_DAY_TIME, PrimitiveCategory.TIMESTAMP)) {
        plusOpType = OperationType.INTERVALDT_PLUS_TIMESTAMP;
        intervalArg1Idx = 0;
        dtArgIdx = 1;
        resultOI = PrimitiveObjectInspectorFactory.getPrimitiveWritableObjectInspector(TypeInfoFactory.timestampTypeInfo);
        dtConverter = ObjectInspectorConverters.getConverter(leftOI, resultOI);
    } else if (checkArgs(PrimitiveCategory.DATE, PrimitiveCategory.INTERVAL_DAY_TIME) || checkArgs(PrimitiveCategory.TIMESTAMP, PrimitiveCategory.INTERVAL_DAY_TIME)) {
        plusOpType = OperationType.INTERVALDT_PLUS_TIMESTAMP;
        intervalArg1Idx = 1;
        dtArgIdx = 0;
        resultOI = PrimitiveObjectInspectorFactory.getPrimitiveWritableObjectInspector(TypeInfoFactory.timestampTypeInfo);
        dtConverter = ObjectInspectorConverters.getConverter(leftOI, resultOI);
    } else {
        // Unsupported types - error
        List<TypeInfo> argTypeInfos = new ArrayList<TypeInfo>(2);
        argTypeInfos.add(leftOI.getTypeInfo());
        argTypeInfos.add(rightOI.getTypeInfo());
        throw new NoMatchingMethodException(this.getClass(), argTypeInfos, null);
    }
    return resultOI;
}

Example 60 with TypeInfo

use of org.apache.hadoop.hive.serde2.typeinfo.TypeInfo in project hive by apache.

the class TestVectorExpressionWriters method testStructLong.

private void testStructLong(TypeInfo type) throws HiveException {
    LongColumnVector icv = VectorizedRowGroupGenUtil.generateLongColumnVector(true, false, vectorSize, new Random(10));
    icv.isNull[3] = true;
    LongColumnVector bcv = VectorizedRowGroupGenUtil.generateLongColumnVector(true, false, vectorSize, new Random(10));
    bcv.isNull[2] = true;
    ArrayList<Object>[] values = (ArrayList<Object>[]) new ArrayList[this.vectorSize];
    StructObjectInspector soi = genStructOI();
    VectorExpressionWriter[] vew = VectorExpressionWriterFactory.getExpressionWriters(soi);
    for (int i = 0; i < vectorSize; i++) {
        values[i] = new ArrayList<Object>(2);
        values[i].add(null);
        values[i].add(null);
        vew[0].setValue(values[i], icv, i);
        vew[1].setValue(values[i], bcv, i);
        Object theInt = values[i].get(0);
        if (theInt == null) {
            Assert.assertTrue(icv.isNull[i]);
        } else {
            IntWritable w = (IntWritable) theInt;
            Assert.assertEquals((int) icv.vector[i], w.get());
        }
        Object theBool = values[i].get(1);
        if (theBool == null) {
            Assert.assertTrue(bcv.isNull[i]);
        } else {
            BooleanWritable w = (BooleanWritable) theBool;
            Assert.assertEquals(bcv.vector[i] == 0 ? false : true, w.get());
        }
    }
}