Examples with ObjectInspector org.apache.hadoop.hive.serde2.objectinspector.ObjectInspector used on opensource projects

Example 11 with ObjectInspector

use of org.apache.hadoop.hive.serde2.objectinspector.ObjectInspector in project hive by apache.

the class GenericUDFToUnixTimeStamp method initializeInput.

protected void initializeInput(ObjectInspector[] arguments) throws UDFArgumentException {
    if (arguments.length < 1) {
        throw new UDFArgumentLengthException("The function " + getName().toUpperCase() + "requires at least one argument");
    }
    for (ObjectInspector argument : arguments) {
        if (arguments[0].getCategory() != Category.PRIMITIVE) {
            throw new UDFArgumentException(getName().toUpperCase() + " only takes string/date/timestamp types, got " + argument.getTypeName());
        }
    }
    PrimitiveObjectInspector arg1OI = (PrimitiveObjectInspector) arguments[0];
    switch(arg1OI.getPrimitiveCategory()) {
        case CHAR:
        case VARCHAR:
        case STRING:
            inputTextConverter = ObjectInspectorConverters.getConverter(arg1OI, PrimitiveObjectInspectorFactory.javaStringObjectInspector);
            if (arguments.length > 1) {
                PrimitiveObjectInspector arg2OI = (PrimitiveObjectInspector) arguments[1];
                if (PrimitiveObjectInspectorUtils.getPrimitiveGrouping(arg2OI.getPrimitiveCategory()) != PrimitiveGrouping.STRING_GROUP) {
                    throw new UDFArgumentException("The time pattern for " + getName().toUpperCase() + " should be string type");
                }
                patternConverter = ObjectInspectorConverters.getConverter(arg2OI, PrimitiveObjectInspectorFactory.javaStringObjectInspector);
            }
            break;
        case DATE:
            inputDateOI = (DateObjectInspector) arguments[0];
            break;
        case TIMESTAMP:
            inputTimestampOI = (TimestampObjectInspector) arguments[0];
            break;
        default:
            throw new UDFArgumentException("The function " + getName().toUpperCase() + " takes only string/date/timestamp types");
    }
}

Example 12 with ObjectInspector

use of org.apache.hadoop.hive.serde2.objectinspector.ObjectInspector in project hive by apache.

the class GenericUDFToVarchar method initialize.

@Override
public ObjectInspector initialize(ObjectInspector[] arguments) throws UDFArgumentException {
    if (arguments.length != 1) {
        throw new UDFArgumentException("VARCHAR cast requires a value argument");
    }
    try {
        argumentOI = (PrimitiveObjectInspector) arguments[0];
    } catch (ClassCastException e) {
        throw new UDFArgumentException("The function VARCHAR takes only primitive types");
    }
    // Check if this UDF has been provided with type params for the output varchar type
    SettableHiveVarcharObjectInspector outputOI;
    outputOI = (SettableHiveVarcharObjectInspector) PrimitiveObjectInspectorFactory.getPrimitiveWritableObjectInspector(typeInfo);
    converter = new HiveVarcharConverter(argumentOI, outputOI);
    return outputOI;
}

Example 13 with ObjectInspector

use of org.apache.hadoop.hive.serde2.objectinspector.ObjectInspector in project hive by apache.

the class MatchPath method createSelectListOI.

protected static StructObjectInspector createSelectListOI(MatchPath evaluator, PTFInputDef inpDef) {
    StructObjectInspector inOI = inpDef.getOutputShape().getOI();
    ArrayList<String> inputColumnNames = new ArrayList<String>();
    ArrayList<String> selectListNames = new ArrayList<String>();
    ArrayList<ObjectInspector> fieldOIs = new ArrayList<ObjectInspector>();
    for (StructField f : inOI.getAllStructFieldRefs()) {
        String inputColName = evaluator.inputColumnNamesMap.get(f.getFieldName());
        if (inputColName != null) {
            inputColumnNames.add(inputColName);
            selectListNames.add(f.getFieldName());
            fieldOIs.add(f.getFieldObjectInspector());
        }
    }
    StandardListObjectInspector pathAttrOI = ObjectInspectorFactory.getStandardListObjectInspector(ObjectInspectorFactory.getStandardStructObjectInspector(inputColumnNames, fieldOIs));
    ArrayList<ObjectInspector> selectFieldOIs = new ArrayList<ObjectInspector>();
    selectFieldOIs.addAll(fieldOIs);
    selectFieldOIs.add(pathAttrOI);
    selectListNames.add(MatchPath.PATHATTR_NAME);
    return ObjectInspectorFactory.getStandardStructObjectInspector(selectListNames, selectFieldOIs);
}

Example 14 with ObjectInspector

use of org.apache.hadoop.hive.serde2.objectinspector.ObjectInspector in project hive by apache.

the class VectorRandomRowSource method getWritableObject.

public static Object getWritableObject(int column, Object object, List<ObjectInspector> primitiveObjectInspectorList, PrimitiveCategory[] primitiveCategories, PrimitiveTypeInfo[] primitiveTypeInfos) {
    ObjectInspector objectInspector = primitiveObjectInspectorList.get(column);
    PrimitiveCategory primitiveCategory = primitiveCategories[column];
    PrimitiveTypeInfo primitiveTypeInfo = primitiveTypeInfos[column];
    switch(primitiveCategory) {
        case BOOLEAN:
            return ((WritableBooleanObjectInspector) objectInspector).create((boolean) object);
        case BYTE:
            return ((WritableByteObjectInspector) objectInspector).create((byte) object);
        case SHORT:
            return ((WritableShortObjectInspector) objectInspector).create((short) object);
        case INT:
            return ((WritableIntObjectInspector) objectInspector).create((int) object);
        case LONG:
            return ((WritableLongObjectInspector) objectInspector).create((long) object);
        case DATE:
            return ((WritableDateObjectInspector) objectInspector).create((Date) object);
        case FLOAT:
            return ((WritableFloatObjectInspector) objectInspector).create((float) object);
        case DOUBLE:
            return ((WritableDoubleObjectInspector) objectInspector).create((double) object);
        case STRING:
            return ((WritableStringObjectInspector) objectInspector).create((String) object);
        case CHAR:
            {
                WritableHiveCharObjectInspector writableCharObjectInspector = new WritableHiveCharObjectInspector((CharTypeInfo) primitiveTypeInfo);
                return writableCharObjectInspector.create((HiveChar) object);
            }
        case VARCHAR:
            {
                WritableHiveVarcharObjectInspector writableVarcharObjectInspector = new WritableHiveVarcharObjectInspector((VarcharTypeInfo) primitiveTypeInfo);
                return writableVarcharObjectInspector.create((HiveVarchar) object);
            }
        case BINARY:
            return PrimitiveObjectInspectorFactory.writableBinaryObjectInspector.create((byte[]) object);
        case TIMESTAMP:
            return ((WritableTimestampObjectInspector) objectInspector).create((Timestamp) object);
        case INTERVAL_YEAR_MONTH:
            return ((WritableHiveIntervalYearMonthObjectInspector) objectInspector).create((HiveIntervalYearMonth) object);
        case INTERVAL_DAY_TIME:
            return ((WritableHiveIntervalDayTimeObjectInspector) objectInspector).create((HiveIntervalDayTime) object);
        case DECIMAL:
            {
                WritableHiveDecimalObjectInspector writableDecimalObjectInspector = new WritableHiveDecimalObjectInspector((DecimalTypeInfo) primitiveTypeInfo);
                HiveDecimalWritable result = (HiveDecimalWritable) writableDecimalObjectInspector.create((HiveDecimal) object);
                return result;
            }
        default:
            throw new Error("Unknown primitive category " + primitiveCategory);
    }
}

Example 15 with ObjectInspector

use of org.apache.hadoop.hive.serde2.objectinspector.ObjectInspector in project hive by apache.

the class OrcFileGenerator method generateOrcFile.

/**
   * Generates an orc file based on the provided record class in the specified file system
   * at the output path.
   *
   * @param conf the configuration used to initialize the orc writer
   * @param fs the file system to which will contain the generated orc file
   * @param outputPath the path where the generated orc will be placed
   * @param recordClass a class the defines the record format for the generated orc file, this
   * class must have exactly one constructor.
   */
public static void generateOrcFile(Configuration conf, FileSystem fs, Path outputPath, Class recordClass) throws IOException, InstantiationException, IllegalAccessException, InvocationTargetException {
    ObjectInspector inspector;
    synchronized (TestVectorizedORCReader.class) {
        inspector = ObjectInspectorFactory.getReflectionObjectInspector(recordClass, ObjectInspectorFactory.ObjectInspectorOptions.JAVA);
    }
    Writer writer = OrcFile.createWriter(fs, outputPath, conf, inspector, 100000, CompressionKind.ZLIB, 10000, 10000);
    try {
        Constructor[] constructors = recordClass.getConstructors();
        if (constructors.length != 1) {
            throw new UnsupportedOperationException("The provided recordClass must have exactly one constructor.");
        }
        BatchDataDistribution[] dataDist = BatchDataDistribution.values();
        Class[] columns = constructors[0].getParameterTypes();
        for (int i = 0; i < dataDist.length * 3; i++) {
            Object[][] rows = new Object[columns.length][VectorizedRowBatch.DEFAULT_SIZE];
            for (int c = 0; c < columns.length; c++) {
                if (!TYPE_TO_BATCH_GEN_MAP.containsKey(columns[c])) {
                    throw new UnsupportedOperationException("No batch generator defined for type " + columns[c].getName());
                }
                rows[c] = TYPE_TO_BATCH_GEN_MAP.get(columns[c]).generateBatch(dataDist[(i + c) % dataDist.length]);
            }
            for (int r = 0; r < VectorizedRowBatch.DEFAULT_SIZE; r++) {
                Object[] row = new Object[columns.length];
                for (int c = 0; c < columns.length; c++) {
                    row[c] = rows[c][r];
                }
                writer.addRow(constructors[0].newInstance(row));
            }
        }
    } finally {
        writer.close();
    }
}