Examples with GenericUDF org.apache.hadoop.hive.ql.udf.generic.GenericUDF used on opensource projects

Example 21 with GenericUDF

use of org.apache.hadoop.hive.ql.udf.generic.GenericUDF in project hive by apache.

the class TestVectorizationContext method testUnaryStringExpressions.

@Test
public void testUnaryStringExpressions() throws HiveException {
    ExprNodeGenericFuncDesc stringUnary = new ExprNodeGenericFuncDesc();
    stringUnary.setTypeInfo(TypeInfoFactory.stringTypeInfo);
    ExprNodeColumnDesc colDesc = new ExprNodeColumnDesc(String.class, "a", "table", false);
    List<ExprNodeDesc> children = new ArrayList<ExprNodeDesc>();
    children.add(colDesc);
    stringUnary.setChildren(children);
    List<String> columns = new ArrayList<String>();
    columns.add("b");
    columns.add("a");
    VectorizationContext vc = new VectorizationContext("name", columns);
    GenericUDF stringLower = new GenericUDFLower();
    stringUnary.setGenericUDF(stringLower);
    VectorExpression ve = vc.getVectorExpression(stringUnary);
    assertEquals(StringLower.class, ve.getClass());
    assertEquals(1, ((StringLower) ve).getColNum());
    assertEquals(2, ((StringLower) ve).getOutputColumn());
    vc = new VectorizationContext("name", columns);
    ExprNodeGenericFuncDesc anotherUnary = new ExprNodeGenericFuncDesc();
    anotherUnary.setTypeInfo(TypeInfoFactory.stringTypeInfo);
    List<ExprNodeDesc> children2 = new ArrayList<ExprNodeDesc>();
    children2.add(stringUnary);
    anotherUnary.setChildren(children2);
    GenericUDFBridge udfbridge = new GenericUDFBridge("ltrim", false, GenericUDFLTrim.class.getName());
    anotherUnary.setGenericUDF(udfbridge);
    ve = vc.getVectorExpression(anotherUnary);
    VectorExpression childVe = ve.getChildExpressions()[0];
    assertEquals(StringLower.class, childVe.getClass());
    assertEquals(1, ((StringLower) childVe).getColNum());
    assertEquals(2, ((StringLower) childVe).getOutputColumn());
    assertEquals(StringLTrim.class, ve.getClass());
    assertEquals(2, ((StringLTrim) ve).getInputColumn());
    assertEquals(3, ((StringLTrim) ve).getOutputColumn());
}

Example 22 with GenericUDF

use of org.apache.hadoop.hive.ql.udf.generic.GenericUDF in project hive by apache.

the class TestVectorSelectOperator method testSelectOperator.

@Test
public void testSelectOperator() throws HiveException {
    List<String> columns = new ArrayList<String>();
    columns.add("a");
    columns.add("b");
    columns.add("c");
    VectorizationContext vc = new VectorizationContext("name", columns);
    SelectDesc selDesc = new SelectDesc(false);
    List<ExprNodeDesc> colList = new ArrayList<ExprNodeDesc>();
    ExprNodeColumnDesc colDesc1 = new ExprNodeColumnDesc(Long.class, "a", "table", false);
    ExprNodeColumnDesc colDesc2 = new ExprNodeColumnDesc(Long.class, "b", "table", false);
    ExprNodeColumnDesc colDesc3 = new ExprNodeColumnDesc(Long.class, "c", "table", false);
    ExprNodeGenericFuncDesc plusDesc = new ExprNodeGenericFuncDesc();
    GenericUDF gudf = new GenericUDFOPPlus();
    plusDesc.setGenericUDF(gudf);
    List<ExprNodeDesc> children = new ArrayList<ExprNodeDesc>();
    children.add(colDesc1);
    children.add(colDesc2);
    plusDesc.setChildren(children);
    plusDesc.setTypeInfo(TypeInfoFactory.longTypeInfo);
    colList.add(plusDesc);
    colList.add(colDesc3);
    selDesc.setColList(colList);
    List<String> outputColNames = new ArrayList<String>();
    outputColNames.add("_col0");
    outputColNames.add("_col1");
    selDesc.setOutputColumnNames(outputColNames);
    // CONSIDER unwinding ValidatorVectorSelectOperator as a subclass of VectorSelectOperator.
    VectorSelectDesc vectorSelectDesc = new VectorSelectDesc();
    selDesc.setVectorDesc(vectorSelectDesc);
    List<ExprNodeDesc> selectColList = selDesc.getColList();
    VectorExpression[] vectorSelectExprs = new VectorExpression[selectColList.size()];
    for (int i = 0; i < selectColList.size(); i++) {
        ExprNodeDesc expr = selectColList.get(i);
        VectorExpression ve = vc.getVectorExpression(expr);
        vectorSelectExprs[i] = ve;
    }
    vectorSelectDesc.setSelectExpressions(vectorSelectExprs);
    vectorSelectDesc.setProjectedOutputColumns(new int[] { 3, 2 });
    ValidatorVectorSelectOperator vso = new ValidatorVectorSelectOperator(new CompilationOpContext(), vc, selDesc);
    VectorizedRowBatch vrg = VectorizedRowGroupGenUtil.getVectorizedRowBatch(VectorizedRowBatch.DEFAULT_SIZE, 4, 17);
    vso.process(vrg, 0);
}

Example 23 with GenericUDF

use of org.apache.hadoop.hive.ql.udf.generic.GenericUDF in project hive by apache.

the class TestFunctionRegistry method testIsPermanentFunction.

public void testIsPermanentFunction() throws Exception {
    // Setup exprNode
    GenericUDF udf = new GenericUDFCurrentTimestamp();
    List<ExprNodeDesc> children = new ArrayList<ExprNodeDesc>();
    ExprNodeGenericFuncDesc fnExpr = new ExprNodeGenericFuncDesc(TypeInfoFactory.timestampTypeInfo, udf, children);
    assertFalse("Function not added as permanent yet", FunctionRegistry.isPermanentFunction(fnExpr));
    // Now register as permanent function
    FunctionResource[] emptyResources = new FunctionResource[] {};
    FunctionRegistry.registerPermanentFunction("default.perm_current_timestamp", GenericUDFCurrentTimestamp.class.getName(), true, emptyResources);
    assertTrue("Function should now be recognized as permanent function", FunctionRegistry.isPermanentFunction(fnExpr));
}

Example 24 with GenericUDF

use of org.apache.hadoop.hive.ql.udf.generic.GenericUDF in project hive by apache.

the class TestColumnPrunerProcCtx method testGetSelectNestedColPathsFromChildren7.

// Test select pow(root.col1.b, root.col2) from table test(root
// struct<col1:struct<a:boolean,b:double>, col2:double>);
@Test
public void testGetSelectNestedColPathsFromChildren7() {
    ColumnPrunerProcCtx ctx = new ColumnPrunerProcCtx(null);
    ExprNodeDesc colDesc = new ExprNodeColumnDesc(col3Type, "root", "test", false);
    ExprNodeDesc col1 = new ExprNodeFieldDesc(col1Type, colDesc, "col1", false);
    ExprNodeDesc fieldDesc1 = new ExprNodeFieldDesc(TypeInfoFactory.doubleTypeInfo, col1, "b", false);
    colDesc = new ExprNodeColumnDesc(col3Type, "root", "test", false);
    ExprNodeDesc col2 = new ExprNodeFieldDesc(col2Type, colDesc, "col2", false);
    final List<FieldNode> paths = Arrays.asList(new FieldNode("_col0"));
    GenericUDF udf = mock(GenericUDFPower.class);
    List<ExprNodeDesc> list = new ArrayList<>();
    list.add(fieldDesc1);
    list.add(col2);
    ExprNodeDesc funcDesc = new ExprNodeGenericFuncDesc(TypeInfoFactory.doubleTypeInfo, udf, "pow", list);
    SelectOperator selectOperator = buildSelectOperator(Arrays.asList(funcDesc), paths);
    List<FieldNode> groups = ctx.getSelectColsFromChildren(selectOperator, paths);
    compareTestResults(groups, "root.col1.b", "root.col2");
}

Example 25 with GenericUDF

use of org.apache.hadoop.hive.ql.udf.generic.GenericUDF in project hive by apache.

the class VectorUDFAdaptor method init.

// Initialize transient fields. To be called after deserialization of other fields.
public void init() throws HiveException, UDFArgumentException {
    genericUDF = expr.getGenericUDF();
    deferredChildren = new GenericUDF.DeferredObject[expr.getChildren().size()];
    childrenOIs = new ObjectInspector[expr.getChildren().size()];
    writers = VectorExpressionWriterFactory.getExpressionWriters(expr.getChildren());
    for (int i = 0; i < childrenOIs.length; i++) {
        childrenOIs[i] = writers[i].getObjectInspector();
    }
    MapredContext context = MapredContext.get();
    if (context != null) {
        context.setup(genericUDF);
    }
    outputOI = VectorExpressionWriterFactory.genVectorExpressionWritable(expr).getObjectInspector();
    genericUDF.initialize(childrenOIs);
    // Initialize constant arguments
    for (int i = 0; i < argDescs.length; i++) {
        if (argDescs[i].isConstant()) {
            argDescs[i].prepareConstant();
        }
    }
}