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

Example 66 with TypeInfo

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

the class GenericUDTFGetSplits method convertTypeString.

private TypeDesc convertTypeString(String typeString) throws HiveException {
    TypeDesc typeDesc;
    TypeInfo typeInfo = TypeInfoUtils.getTypeInfoFromTypeString(typeString);
    Preconditions.checkState(typeInfo.getCategory() == ObjectInspector.Category.PRIMITIVE, "Unsupported non-primitive type " + typeString);
    switch(((PrimitiveTypeInfo) typeInfo).getPrimitiveCategory()) {
        case BOOLEAN:
            typeDesc = new TypeDesc(TypeDesc.Type.BOOLEAN);
            break;
        case BYTE:
            typeDesc = new TypeDesc(TypeDesc.Type.TINYINT);
            break;
        case SHORT:
            typeDesc = new TypeDesc(TypeDesc.Type.SMALLINT);
            break;
        case INT:
            typeDesc = new TypeDesc(TypeDesc.Type.INT);
            break;
        case LONG:
            typeDesc = new TypeDesc(TypeDesc.Type.BIGINT);
            break;
        case FLOAT:
            typeDesc = new TypeDesc(TypeDesc.Type.FLOAT);
            break;
        case DOUBLE:
            typeDesc = new TypeDesc(TypeDesc.Type.DOUBLE);
            break;
        case STRING:
            typeDesc = new TypeDesc(TypeDesc.Type.STRING);
            break;
        case CHAR:
            CharTypeInfo charTypeInfo = (CharTypeInfo) typeInfo;
            typeDesc = new TypeDesc(TypeDesc.Type.CHAR, charTypeInfo.getLength());
            break;
        case VARCHAR:
            VarcharTypeInfo varcharTypeInfo = (VarcharTypeInfo) typeInfo;
            typeDesc = new TypeDesc(TypeDesc.Type.VARCHAR, varcharTypeInfo.getLength());
            break;
        case DATE:
            typeDesc = new TypeDesc(TypeDesc.Type.DATE);
            break;
        case TIMESTAMP:
            typeDesc = new TypeDesc(TypeDesc.Type.TIMESTAMP);
            break;
        case BINARY:
            typeDesc = new TypeDesc(TypeDesc.Type.BINARY);
            break;
        case DECIMAL:
            DecimalTypeInfo decimalTypeInfo = (DecimalTypeInfo) typeInfo;
            typeDesc = new TypeDesc(TypeDesc.Type.DECIMAL, decimalTypeInfo.getPrecision(), decimalTypeInfo.getScale());
            break;
        default:
            throw new HiveException("Unsupported type " + typeString);
    }
    return typeDesc;
}

Example 67 with TypeInfo

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

the class GenericUDFToChar method initialize.

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

Example 68 with TypeInfo

use of org.apache.hadoop.hive.serde2.typeinfo.TypeInfo 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 69 with TypeInfo

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

the class TestVectorUDFAdaptor method testMultiArgumentUDF.

@Test
public void testMultiArgumentUDF() {
    // create a syntax tree for a function call "testudf(col0, col1, col2)"
    ExprNodeGenericFuncDesc funcDesc;
    TypeInfo typeInfoStr = TypeInfoFactory.stringTypeInfo;
    TypeInfo typeInfoLong = TypeInfoFactory.longTypeInfo;
    TypeInfo typeInfoDbl = TypeInfoFactory.doubleTypeInfo;
    GenericUDFBridge genericUDFBridge = new GenericUDFBridge("testudf", false, ConcatTextLongDoubleUDF.class.getName());
    List<ExprNodeDesc> children = new ArrayList<ExprNodeDesc>();
    children.add(new ExprNodeColumnDesc(typeInfoStr, "col0", "tablename", false));
    children.add(new ExprNodeColumnDesc(typeInfoLong, "col1", "tablename", false));
    children.add(new ExprNodeColumnDesc(typeInfoDbl, "col2", "tablename", false));
    VectorUDFArgDesc[] argDescs = new VectorUDFArgDesc[3];
    for (int i = 0; i < 3; i++) {
        argDescs[i] = new VectorUDFArgDesc();
        argDescs[i].setVariable(i);
    }
    funcDesc = new ExprNodeGenericFuncDesc(typeInfoStr, genericUDFBridge, genericUDFBridge.getUdfName(), children);
    // create the adaptor for this function call to work in vector mode
    VectorUDFAdaptor vudf = null;
    try {
        vudf = new VectorUDFAdaptor(funcDesc, 3, "String", argDescs);
    } catch (HiveException e) {
        // We should never get here.
        assertTrue(false);
        throw new RuntimeException(e);
    }
    // with no nulls
    VectorizedRowBatch b = getBatchStrDblLongWithStrOut();
    vudf.evaluate(b);
    byte[] result = null;
    byte[] result2 = null;
    try {
        result = "red:1:1.0".getBytes("UTF-8");
        result2 = "blue:0:0.0".getBytes("UTF-8");
    } catch (Exception e) {
        ;
    }
    BytesColumnVector out = (BytesColumnVector) b.cols[3];
    int cmp = StringExpr.compare(result, 0, result.length, out.vector[1], out.start[1], out.length[1]);
    assertEquals(0, cmp);
    assertTrue(out.noNulls);
    // with nulls
    b = getBatchStrDblLongWithStrOut();
    b.cols[1].noNulls = false;
    vudf.evaluate(b);
    out = (BytesColumnVector) b.cols[3];
    assertFalse(out.noNulls);
    assertTrue(out.isNull[1]);
    // with all input columns repeating
    b = getBatchStrDblLongWithStrOut();
    b.cols[0].isRepeating = true;
    b.cols[1].isRepeating = true;
    b.cols[2].isRepeating = true;
    vudf.evaluate(b);
    out = (BytesColumnVector) b.cols[3];
    assertTrue(out.isRepeating);
    cmp = StringExpr.compare(result2, 0, result2.length, out.vector[0], out.start[0], out.length[0]);
    assertEquals(0, cmp);
    assertTrue(out.noNulls);
}

Example 70 with TypeInfo

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

the class TestVectorUDFAdaptor method testLongUDF.

@Test
public void testLongUDF() {
    // create a syntax tree for a simple function call "longudf(col0)"
    ExprNodeGenericFuncDesc funcDesc;
    TypeInfo typeInfo = TypeInfoFactory.longTypeInfo;
    GenericUDFBridge genericUDFBridge = new GenericUDFBridge("longudf", false, LongUDF.class.getName());
    List<ExprNodeDesc> children = new ArrayList<ExprNodeDesc>();
    ExprNodeColumnDesc colDesc = new ExprNodeColumnDesc(typeInfo, "col0", "tablename", false);
    children.add(colDesc);
    VectorUDFArgDesc[] argDescs = new VectorUDFArgDesc[1];
    argDescs[0] = new VectorUDFArgDesc();
    argDescs[0].setVariable(0);
    funcDesc = new ExprNodeGenericFuncDesc(typeInfo, genericUDFBridge, genericUDFBridge.getUdfName(), children);
    // create the adaptor for this function call to work in vector mode
    VectorUDFAdaptor vudf = null;
    try {
        vudf = new VectorUDFAdaptor(funcDesc, 1, "Long", argDescs);
    } catch (HiveException e) {
        // We should never get here.
        assertTrue(false);
    }
    VectorizedRowBatch b = getBatchLongInLongOut();
    vudf.evaluate(b);
    // verify output
    LongColumnVector out = (LongColumnVector) b.cols[1];
    assertEquals(1000, out.vector[0]);
    assertEquals(1001, out.vector[1]);
    assertEquals(1002, out.vector[2]);
    assertTrue(out.noNulls);
    assertFalse(out.isRepeating);
    // with nulls
    b = getBatchLongInLongOut();
    out = (LongColumnVector) b.cols[1];
    b.cols[0].noNulls = false;
    vudf.evaluate(b);
    assertFalse(out.noNulls);
    assertEquals(1000, out.vector[0]);
    assertEquals(1001, out.vector[1]);
    assertTrue(out.isNull[2]);
    assertFalse(out.isRepeating);
    // with repeating
    b = getBatchLongInLongOut();
    out = (LongColumnVector) b.cols[1];
    b.cols[0].isRepeating = true;
    vudf.evaluate(b);
    // The implementation may or may not set output it isRepeting.
    // That is implementation-defined.
    assertTrue(b.cols[1].isRepeating && out.vector[0] == 1000 || !b.cols[1].isRepeating && out.vector[2] == 1000);
    assertEquals(3, b.size);
}