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

Example 46 with ObjectInspectorFactory.getStandardStructObjectInspector

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

the class GenericUDFStruct method initialize.

@Override
public ObjectInspector initialize(ObjectInspector[] arguments) throws UDFArgumentException {
    int numFields = arguments.length;
    ret = new Object[numFields];
    ArrayList<String> fname = new ArrayList<String>(numFields);
    for (int f = 1; f <= numFields; f++) {
        fname.add("col" + f);
    }
    boolean constantStruct = true;
    for (int i = 0; i < arguments.length; i++) {
        ObjectInspector oi = arguments[i];
        constantStruct &= (oi.getCategory() == Category.PRIMITIVE) && (oi instanceof ConstantObjectInspector);
        if (constantStruct) {
            // nested complex types trigger Kryo issue #216 in plan deserialization
            ret[i] = ((ConstantObjectInspector) oi).getWritableConstantValue();
        }
    }
    if (constantStruct) {
        return ObjectInspectorFactory.getStandardConstantStructObjectInspector(fname, Arrays.asList(arguments), Arrays.asList(ret));
    } else {
        return ObjectInspectorFactory.getStandardStructObjectInspector(fname, Arrays.asList(arguments));
    }
}

Example 47 with ObjectInspectorFactory.getStandardStructObjectInspector

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

the class RegexSerDe method initialize.

@Override
public void initialize(Configuration conf, Properties tbl) throws SerDeException {
    // We can get the table definition from tbl.
    // Read the configuration parameters
    inputRegex = tbl.getProperty(INPUT_REGEX);
    outputFormatString = tbl.getProperty(OUTPUT_FORMAT_STRING);
    String columnNameProperty = tbl.getProperty(serdeConstants.LIST_COLUMNS);
    String columnTypeProperty = tbl.getProperty(serdeConstants.LIST_COLUMN_TYPES);
    boolean inputRegexIgnoreCase = "true".equalsIgnoreCase(tbl.getProperty(INPUT_REGEX_CASE_SENSITIVE));
    // Parse the configuration parameters
    if (inputRegex != null) {
        inputPattern = Pattern.compile(inputRegex, Pattern.DOTALL + (inputRegexIgnoreCase ? Pattern.CASE_INSENSITIVE : 0));
    } else {
        inputPattern = null;
    }
    final String columnNameDelimiter = tbl.containsKey(serdeConstants.COLUMN_NAME_DELIMITER) ? tbl.getProperty(serdeConstants.COLUMN_NAME_DELIMITER) : String.valueOf(SerDeUtils.COMMA);
    List<String> columnNames = Arrays.asList(columnNameProperty.split(columnNameDelimiter));
    List<TypeInfo> columnTypes = TypeInfoUtils.getTypeInfosFromTypeString(columnTypeProperty);
    assert columnNames.size() == columnTypes.size();
    numColumns = columnNames.size();
    // All columns have to be of type STRING.
    for (int c = 0; c < numColumns; c++) {
        if (!columnTypes.get(c).equals(TypeInfoFactory.stringTypeInfo)) {
            throw new SerDeException(getClass().getName() + " only accepts string columns, but column[" + c + "] named " + columnNames.get(c) + " has type " + columnTypes.get(c));
        }
    }
    // Constructing the row ObjectInspector:
    // The row consists of some string columns, each column will be a java
    // String object.
    List<ObjectInspector> columnOIs = new ArrayList<ObjectInspector>(columnNames.size());
    for (int c = 0; c < numColumns; c++) {
        columnOIs.add(PrimitiveObjectInspectorFactory.javaStringObjectInspector);
    }
    // StandardStruct uses ArrayList to store the row.
    rowOI = ObjectInspectorFactory.getStandardStructObjectInspector(columnNames, columnOIs);
    // Constructing the row object, etc, which will be reused for all rows.
    row = new ArrayList<String>(numColumns);
    for (int c = 0; c < numColumns; c++) {
        row.add(null);
    }
    outputFields = new Object[numColumns];
    outputRowText = new Text();
}

Example 48 with ObjectInspectorFactory.getStandardStructObjectInspector

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

the class TypedBytesSerDe method initialize.

@Override
public void initialize(Configuration conf, Properties tbl) throws SerDeException {
    // We can get the table definition from tbl.
    serializeBytesWritable = new BytesWritable();
    barrStr = new NonSyncDataOutputBuffer();
    tbOut = new TypedBytesWritableOutput(barrStr);
    inBarrStr = new NonSyncDataInputBuffer();
    tbIn = new TypedBytesWritableInput(inBarrStr);
    // Read the configuration parameters
    String columnNameProperty = tbl.getProperty(serdeConstants.LIST_COLUMNS);
    String columnTypeProperty = tbl.getProperty(serdeConstants.LIST_COLUMN_TYPES);
    final String columnNameDelimiter = tbl.containsKey(serdeConstants.COLUMN_NAME_DELIMITER) ? tbl.getProperty(serdeConstants.COLUMN_NAME_DELIMITER) : String.valueOf(SerDeUtils.COMMA);
    columnNames = Arrays.asList(columnNameProperty.split(columnNameDelimiter));
    columnTypes = null;
    if (columnTypeProperty.length() == 0) {
        columnTypes = new ArrayList<TypeInfo>();
    } else {
        columnTypes = TypeInfoUtils.getTypeInfosFromTypeString(columnTypeProperty);
    }
    assert columnNames.size() == columnTypes.size();
    numColumns = columnNames.size();
    // All columns have to be primitive.
    for (int c = 0; c < numColumns; c++) {
        if (columnTypes.get(c).getCategory() != Category.PRIMITIVE) {
            throw new SerDeException(getClass().getName() + " only accepts primitive columns, but column[" + c + "] named " + columnNames.get(c) + " has category " + columnTypes.get(c).getCategory());
        }
    }
    // Constructing the row ObjectInspector:
    // The row consists of some string columns, each column will be a java
    // String object.
    List<ObjectInspector> columnOIs = new ArrayList<ObjectInspector>(columnNames.size());
    for (int c = 0; c < numColumns; c++) {
        columnOIs.add(TypeInfoUtils.getStandardWritableObjectInspectorFromTypeInfo(columnTypes.get(c)));
    }
    // StandardStruct uses ArrayList to store the row.
    rowOI = ObjectInspectorFactory.getStandardStructObjectInspector(columnNames, columnOIs);
    // Constructing the row object, etc, which will be reused for all rows.
    row = new ArrayList<Object>(numColumns);
    for (int c = 0; c < numColumns; c++) {
        row.add(null);
    }
}

Example 49 with ObjectInspectorFactory.getStandardStructObjectInspector

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

the class TestDruidSerDe method serializeObject.

private static void serializeObject(Properties properties, DruidSerDe serDe, Object[] rowObject, DruidWritable druidWritable) throws SerDeException {
    // Build OI with timestamp granularity column
    final List<String> columnNames = new ArrayList<>();
    final List<PrimitiveTypeInfo> columnTypes = new ArrayList<>();
    List<ObjectInspector> inspectors = new ArrayList<>();
    columnNames.addAll(Utilities.getColumnNames(properties));
    columnNames.add(Constants.DRUID_TIMESTAMP_GRANULARITY_COL_NAME);
    columnTypes.addAll(Lists.transform(Utilities.getColumnTypes(properties), new Function<String, PrimitiveTypeInfo>() {

        @Override
        public PrimitiveTypeInfo apply(String type) {
            return TypeInfoFactory.getPrimitiveTypeInfo(type);
        }
    }));
    columnTypes.add(TypeInfoFactory.getPrimitiveTypeInfo("timestamp"));
    inspectors.addAll(Lists.transform(columnTypes, new Function<PrimitiveTypeInfo, ObjectInspector>() {

        @Override
        public ObjectInspector apply(PrimitiveTypeInfo type) {
            return PrimitiveObjectInspectorFactory.getPrimitiveWritableObjectInspector(type);
        }
    }));
    ObjectInspector inspector = ObjectInspectorFactory.getStandardStructObjectInspector(columnNames, inspectors);
    // Serialize
    DruidWritable writable = (DruidWritable) serDe.serialize(rowObject, inspector);
    // Check result
    assertEquals(DRUID_WRITABLE.getValue().size(), writable.getValue().size());
    for (Entry<String, Object> e : DRUID_WRITABLE.getValue().entrySet()) {
        assertEquals(e.getValue(), writable.getValue().get(e.getKey()));
    }
}

Example 50 with ObjectInspectorFactory.getStandardStructObjectInspector

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

the class GenericUDFNamedStruct method initialize.

@Override
public ObjectInspector initialize(ObjectInspector[] arguments) throws UDFArgumentException {
    int numFields = arguments.length;
    if (numFields % 2 == 1) {
        throw new UDFArgumentLengthException("NAMED_STRUCT expects an even number of arguments.");
    }
    ret = new Object[numFields / 2];
    ArrayList<String> fname = new ArrayList<String>(numFields / 2);
    ArrayList<ObjectInspector> retOIs = new ArrayList<ObjectInspector>(numFields / 2);
    for (int f = 0; f < numFields; f += 2) {
        if (!(arguments[f] instanceof ConstantObjectInspector)) {
            throw new UDFArgumentTypeException(f, "Even arguments" + " to NAMED_STRUCT must be a constant STRING." + arguments[f].toString());
        }
        ConstantObjectInspector constantOI = (ConstantObjectInspector) arguments[f];
        fname.add(constantOI.getWritableConstantValue().toString());
        retOIs.add(arguments[f + 1]);
    }
    StructObjectInspector soi = ObjectInspectorFactory.getStandardStructObjectInspector(fname, retOIs);
    return soi;
}