Search in sources :

Example 96 with ObjectInspector

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

the class TestGenericUDFStringToMap method initGenericUDF.

private void initGenericUDF(GenericUDFStringToMap udf) throws UDFArgumentException {
    ObjectInspector valueOI0 = PrimitiveObjectInspectorFactory.javaStringObjectInspector;
    ObjectInspector valueOI1 = PrimitiveObjectInspectorFactory.javaStringObjectInspector;
    ObjectInspector valueOI2 = PrimitiveObjectInspectorFactory.javaStringObjectInspector;
    ObjectInspector[] arguments = { valueOI0, valueOI1, valueOI2 };
    udf.initialize(arguments);
}
Also used : ObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.ObjectInspector)

Example 97 with ObjectInspector

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

the class TestGenericUDFSubstringIndex method testSubstringIndex.

public void testSubstringIndex() throws HiveException {
    GenericUDFSubstringIndex udf = new GenericUDFSubstringIndex();
    ObjectInspector valueOI0 = PrimitiveObjectInspectorFactory.writableStringObjectInspector;
    ObjectInspector valueOI1 = PrimitiveObjectInspectorFactory.writableStringObjectInspector;
    ObjectInspector valueOI2 = PrimitiveObjectInspectorFactory.writableIntObjectInspector;
    ObjectInspector[] arguments = { valueOI0, valueOI1, valueOI2 };
    udf.initialize(arguments);
    runAndVerify("www.apache.org", ".", 3, "www.apache.org", udf);
    runAndVerify("www.apache.org", ".", 2, "www.apache", udf);
    runAndVerify("www.apache.org", ".", 1, "www", udf);
    runAndVerify("www.apache.org", ".", 0, "", udf);
    runAndVerify("www.apache.org", ".", -1, "org", udf);
    runAndVerify("www.apache.org", ".", -2, "apache.org", udf);
    runAndVerify("www.apache.org", ".", -3, "www.apache.org", udf);
    // str is empty string
    runAndVerify("", ".", 1, "", udf);
    // empty string delim
    runAndVerify("www.apache.org", "", 1, "", udf);
    // delim does not exist in str
    runAndVerify("www.apache.org", "-", 2, "www.apache.org", udf);
    // delim is 2 chars
    runAndVerify("www||apache||org", "||", 2, "www||apache", udf);
    // null
    runAndVerify(null, ".", 2, null, udf);
    runAndVerify("www.apache.org", null, 2, null, udf);
    runAndVerify("www.apache.org", ".", null, null, udf);
}
Also used : ObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.ObjectInspector)

Example 98 with ObjectInspector

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

the class TestGenericUDFToUnixTimestamp method testDate.

public void testDate() throws HiveException {
    GenericUDFToUnixTimeStamp udf = new GenericUDFToUnixTimeStamp();
    ObjectInspector valueOI = PrimitiveObjectInspectorFactory.writableDateObjectInspector;
    ObjectInspector[] arguments = { valueOI };
    udf.initialize(arguments);
    Date date = Date.valueOf("1970-01-01");
    runAndVerify(udf, new DateWritable(date), new LongWritable(date.getTime() / 1000));
    // test null values
    runAndVerify(udf, null, null);
}
Also used : ObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.ObjectInspector) DateWritable(org.apache.hadoop.hive.serde2.io.DateWritable) LongWritable(org.apache.hadoop.io.LongWritable) Date(java.sql.Date)

Example 99 with ObjectInspector

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

the class ObjectInspectorConverters method getConvertedOI.

/**
   * Utility function to convert from one object inspector type to another.
   * The output object inspector type should have all fields as settableOI type.
   * The above condition can be violated only if equalsCheck is true and inputOI is
   * equal to outputOI.
   * @param inputOI : input object inspector
   * @param outputOI : output object inspector
   * @param oiSettableProperties : The object inspector to isSettable mapping used to cache
   *                               intermediate results.
   * @param equalsCheck : Do we need to check if the inputOI and outputOI are the same?
   *                      true : If they are the same, we return the object inspector directly.
   *                      false : Do not perform an equality check on inputOI and outputOI
   * @return : The output object inspector containing all settable fields. The return value
   *           can contain non-settable fields only if inputOI equals outputOI and equalsCheck is
   *           true.
   */
public static ObjectInspector getConvertedOI(ObjectInspector inputOI, ObjectInspector outputOI, Map<ObjectInspector, Boolean> oiSettableProperties, boolean equalsCheck) {
    // 2. If the outputOI has all fields settable, return it
    if ((equalsCheck && inputOI.equals(outputOI)) || ObjectInspectorUtils.hasAllFieldsSettable(outputOI, oiSettableProperties) == true) {
        return outputOI;
    }
    // T is settable recursively i.e all the nested fields are also settable.
    switch(outputOI.getCategory()) {
        case PRIMITIVE:
            // Create a writable object inspector for primitive type and return it.
            PrimitiveObjectInspector primOutputOI = (PrimitiveObjectInspector) outputOI;
            return PrimitiveObjectInspectorFactory.getPrimitiveWritableObjectInspector(primOutputOI.getTypeInfo());
        case STRUCT:
            StructObjectInspector structOutputOI = (StructObjectInspector) outputOI;
            // create a standard settable struct object inspector.
            List<? extends StructField> listFields = structOutputOI.getAllStructFieldRefs();
            List<String> structFieldNames = new ArrayList<String>(listFields.size());
            List<ObjectInspector> structFieldObjectInspectors = new ArrayList<ObjectInspector>(listFields.size());
            for (StructField listField : listFields) {
                structFieldNames.add(listField.getFieldName());
                // We need to make sure that the underlying fields are settable as well.
                // Hence, the recursive call for each field.
                // Note that equalsCheck is false while invoking getConvertedOI() because
                // we need to bypass the initial inputOI.equals(outputOI) check.
                structFieldObjectInspectors.add(getConvertedOI(listField.getFieldObjectInspector(), listField.getFieldObjectInspector(), oiSettableProperties, false));
            }
            return ObjectInspectorFactory.getStandardStructObjectInspector(structFieldNames, structFieldObjectInspectors);
        case LIST:
            ListObjectInspector listOutputOI = (ListObjectInspector) outputOI;
            // We need to make sure that the list element type is settable.
            return ObjectInspectorFactory.getStandardListObjectInspector(getConvertedOI(listOutputOI.getListElementObjectInspector(), listOutputOI.getListElementObjectInspector(), oiSettableProperties, false));
        case MAP:
            MapObjectInspector mapOutputOI = (MapObjectInspector) outputOI;
            // We need to make sure that the key type and the value types are settable.
            return ObjectInspectorFactory.getStandardMapObjectInspector(getConvertedOI(mapOutputOI.getMapKeyObjectInspector(), mapOutputOI.getMapKeyObjectInspector(), oiSettableProperties, false), getConvertedOI(mapOutputOI.getMapValueObjectInspector(), mapOutputOI.getMapValueObjectInspector(), oiSettableProperties, false));
        case UNION:
            UnionObjectInspector unionOutputOI = (UnionObjectInspector) outputOI;
            // create a standard settable union object inspector
            List<ObjectInspector> unionListFields = unionOutputOI.getObjectInspectors();
            List<ObjectInspector> unionFieldObjectInspectors = new ArrayList<ObjectInspector>(unionListFields.size());
            for (ObjectInspector listField : unionListFields) {
                // We need to make sure that all the field associated with the union are settable.
                unionFieldObjectInspectors.add(getConvertedOI(listField, listField, oiSettableProperties, false));
            }
            return ObjectInspectorFactory.getStandardUnionObjectInspector(unionFieldObjectInspectors);
        default:
            // Unsupported in-memory structure.
            throw new RuntimeException("Hive internal error: conversion of " + inputOI.getTypeName() + " to " + outputOI.getTypeName() + " not supported yet.");
    }
}
Also used : SettableDoubleObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableDoubleObjectInspector) VoidObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.VoidObjectInspector) SettableBooleanObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableBooleanObjectInspector) SettableHiveCharObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableHiveCharObjectInspector) SettableShortObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableShortObjectInspector) SettableTimestampObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableTimestampObjectInspector) SettableHiveIntervalDayTimeObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableHiveIntervalDayTimeObjectInspector) SettableLongObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableLongObjectInspector) SettableBinaryObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableBinaryObjectInspector) SettableFloatObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableFloatObjectInspector) SettableHiveDecimalObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableHiveDecimalObjectInspector) JavaStringObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.JavaStringObjectInspector) SettableHiveIntervalYearMonthObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableHiveIntervalYearMonthObjectInspector) SettableHiveVarcharObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableHiveVarcharObjectInspector) SettableIntObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableIntObjectInspector) SettableDateObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableDateObjectInspector) WritableStringObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.WritableStringObjectInspector) SettableByteObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableByteObjectInspector) ArrayList(java.util.ArrayList)

Example 100 with ObjectInspector

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

the class ObjectInspectorFactory method getReflectionObjectInspectorNoCache.

private static ObjectInspector getReflectionObjectInspectorNoCache(Type t, ObjectInspectorOptions options, boolean ensureInited) {
    if (t instanceof GenericArrayType) {
        GenericArrayType at = (GenericArrayType) t;
        return getStandardListObjectInspector(getReflectionObjectInspector(at.getGenericComponentType(), options, ensureInited));
    }
    if (t instanceof ParameterizedType) {
        ParameterizedType pt = (ParameterizedType) t;
        // List?
        if (List.class.isAssignableFrom((Class<?>) pt.getRawType()) || Set.class.isAssignableFrom((Class<?>) pt.getRawType())) {
            return getStandardListObjectInspector(getReflectionObjectInspector(pt.getActualTypeArguments()[0], options, ensureInited));
        }
        // Map?
        if (Map.class.isAssignableFrom((Class<?>) pt.getRawType())) {
            return getStandardMapObjectInspector(getReflectionObjectInspector(pt.getActualTypeArguments()[0], options, ensureInited), getReflectionObjectInspector(pt.getActualTypeArguments()[1], options, ensureInited));
        }
        // Otherwise convert t to RawType so we will fall into the following if
        // block.
        t = pt.getRawType();
    }
    // Must be a class.
    if (!(t instanceof Class)) {
        throw new RuntimeException(ObjectInspectorFactory.class.getName() + " internal error:" + t);
    }
    Class<?> c = (Class<?>) t;
    // Java Primitive Type?
    if (PrimitiveObjectInspectorUtils.isPrimitiveJavaType(c)) {
        return PrimitiveObjectInspectorFactory.getPrimitiveJavaObjectInspector(PrimitiveObjectInspectorUtils.getTypeEntryFromPrimitiveJavaType(c).primitiveCategory);
    }
    // Java Primitive Class?
    if (PrimitiveObjectInspectorUtils.isPrimitiveJavaClass(c)) {
        return PrimitiveObjectInspectorFactory.getPrimitiveJavaObjectInspector(PrimitiveObjectInspectorUtils.getTypeEntryFromPrimitiveJavaClass(c).primitiveCategory);
    }
    // Primitive Writable class?
    if (PrimitiveObjectInspectorUtils.isPrimitiveWritableClass(c)) {
        return PrimitiveObjectInspectorFactory.getPrimitiveWritableObjectInspector(PrimitiveObjectInspectorUtils.getTypeEntryFromPrimitiveWritableClass(c).primitiveCategory);
    }
    // Enum class?
    if (Enum.class.isAssignableFrom(c)) {
        return PrimitiveObjectInspectorFactory.getPrimitiveJavaObjectInspector(PrimitiveObjectInspector.PrimitiveCategory.STRING);
    }
    // Must be struct because List and Map need to be ParameterizedType
    assert (!List.class.isAssignableFrom(c));
    assert (!Map.class.isAssignableFrom(c));
    // Create StructObjectInspector
    ReflectionStructObjectInspector oi;
    switch(options) {
        case JAVA:
            oi = new ReflectionStructObjectInspector();
            break;
        case THRIFT:
            oi = TUnion.class.isAssignableFrom(c) ? new ThriftUnionObjectInspector() : new ThriftStructObjectInspector();
            break;
        case PROTOCOL_BUFFERS:
            oi = new ProtocolBuffersStructObjectInspector();
            break;
        default:
            throw new RuntimeException(ObjectInspectorFactory.class.getName() + ": internal error.");
    }
    // put it into the cache BEFORE it is initialized to make sure we can catch
    // recursive types.
    ReflectionStructObjectInspector prev = (ReflectionStructObjectInspector) objectInspectorCache.putIfAbsent(t, oi);
    if (prev != null) {
        oi = prev;
    } else {
        try {
            oi.init(t, c, options);
        } finally {
            if (!oi.inited) {
                // Failed to init, remove it from cache
                objectInspectorCache.remove(t, oi);
            }
        }
    }
    return oi;
}
Also used : ParameterizedType(java.lang.reflect.ParameterizedType) Set(java.util.Set) HashSet(java.util.HashSet) ArrayList(java.util.ArrayList) List(java.util.List) GenericArrayType(java.lang.reflect.GenericArrayType) PrimitiveObjectInspectorFactory(org.apache.hadoop.hive.serde2.objectinspector.primitive.PrimitiveObjectInspectorFactory)

Aggregations

ObjectInspector (org.apache.hadoop.hive.serde2.objectinspector.ObjectInspector)700 PrimitiveObjectInspector (org.apache.hadoop.hive.serde2.objectinspector.PrimitiveObjectInspector)299 StructObjectInspector (org.apache.hadoop.hive.serde2.objectinspector.StructObjectInspector)261 Test (org.junit.Test)241 DeferredJavaObject (org.apache.hadoop.hive.ql.udf.generic.GenericUDF.DeferredJavaObject)163 ArrayList (java.util.ArrayList)159 DeferredObject (org.apache.hadoop.hive.ql.udf.generic.GenericUDF.DeferredObject)159 ListObjectInspector (org.apache.hadoop.hive.serde2.objectinspector.ListObjectInspector)107 Text (org.apache.hadoop.io.Text)103 MapObjectInspector (org.apache.hadoop.hive.serde2.objectinspector.MapObjectInspector)89 StringObjectInspector (org.apache.hadoop.hive.serde2.objectinspector.primitive.StringObjectInspector)75 TypeInfo (org.apache.hadoop.hive.serde2.typeinfo.TypeInfo)73 StructField (org.apache.hadoop.hive.serde2.objectinspector.StructField)66 IntWritable (org.apache.hadoop.io.IntWritable)65 LongObjectInspector (org.apache.hadoop.hive.serde2.objectinspector.primitive.LongObjectInspector)58 LongWritable (org.apache.hadoop.io.LongWritable)54 DoubleWritable (org.apache.hadoop.hive.serde2.io.DoubleWritable)53 IntObjectInspector (org.apache.hadoop.hive.serde2.objectinspector.primitive.IntObjectInspector)53 DoubleObjectInspector (org.apache.hadoop.hive.serde2.objectinspector.primitive.DoubleObjectInspector)51 BytesWritable (org.apache.hadoop.io.BytesWritable)51