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

Example 1 with TypeInfoUtils.getTypeInfoFromTypeString

use of org.apache.hadoop.hive.serde2.typeinfo.TypeInfoUtils.getTypeInfoFromTypeString 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 2 with TypeInfoUtils.getTypeInfoFromTypeString

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

the class SerdeRandomRowSource method chooseSchema.

private void chooseSchema() {
    HashSet hashSet = null;
    boolean allTypes;
    boolean onlyOne = (r.nextInt(100) == 7);
    if (onlyOne) {
        columnCount = 1;
        allTypes = false;
    } else {
        allTypes = r.nextBoolean();
        if (allTypes) {
            // One of each type.
            columnCount = possibleHiveTypeNames.length;
            hashSet = new HashSet<Integer>();
        } else {
            columnCount = 1 + r.nextInt(20);
        }
    }
    typeNames = new ArrayList<String>(columnCount);
    primitiveCategories = new PrimitiveCategory[columnCount];
    primitiveTypeInfos = new PrimitiveTypeInfo[columnCount];
    primitiveObjectInspectorList = new ArrayList<ObjectInspector>(columnCount);
    List<String> columnNames = new ArrayList<String>(columnCount);
    for (int c = 0; c < columnCount; c++) {
        columnNames.add(String.format("col%d", c));
        String typeName;
        if (onlyOne) {
            typeName = possibleHiveTypeNames[r.nextInt(possibleHiveTypeNames.length)];
        } else {
            int typeNum;
            if (allTypes) {
                while (true) {
                    typeNum = r.nextInt(possibleHiveTypeNames.length);
                    Integer typeNumInteger = new Integer(typeNum);
                    if (!hashSet.contains(typeNumInteger)) {
                        hashSet.add(typeNumInteger);
                        break;
                    }
                }
            } else {
                typeNum = r.nextInt(possibleHiveTypeNames.length);
            }
            typeName = possibleHiveTypeNames[typeNum];
        }
        if (typeName.equals("char")) {
            int maxLength = 1 + r.nextInt(100);
            typeName = String.format("char(%d)", maxLength);
        } else if (typeName.equals("varchar")) {
            int maxLength = 1 + r.nextInt(100);
            typeName = String.format("varchar(%d)", maxLength);
        } else if (typeName.equals("decimal")) {
            typeName = String.format("decimal(%d,%d)", HiveDecimal.SYSTEM_DEFAULT_PRECISION, HiveDecimal.SYSTEM_DEFAULT_SCALE);
        }
        PrimitiveTypeInfo primitiveTypeInfo = (PrimitiveTypeInfo) TypeInfoUtils.getTypeInfoFromTypeString(typeName);
        primitiveTypeInfos[c] = primitiveTypeInfo;
        PrimitiveCategory primitiveCategory = primitiveTypeInfo.getPrimitiveCategory();
        primitiveCategories[c] = primitiveCategory;
        primitiveObjectInspectorList.add(PrimitiveObjectInspectorFactory.getPrimitiveWritableObjectInspector(primitiveTypeInfo));
        typeNames.add(typeName);
    }
    rowStructObjectInspector = ObjectInspectorFactory.getStandardStructObjectInspector(columnNames, primitiveObjectInspectorList);
}

Example 3 with TypeInfoUtils.getTypeInfoFromTypeString

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

the class TestVectorMapJoinFastRowHashMap method addAndVerifyRows.

private void addAndVerifyRows(VectorRandomRowSource valueSource, Object[][] rows, VectorMapJoinFastHashTable map, HashTableKeyType hashTableKeyType, VerifyFastRowHashMap verifyTable, String[] keyTypeNames, boolean doClipping, boolean useExactBytes) throws HiveException, IOException, SerDeException {
    final int keyCount = keyTypeNames.length;
    PrimitiveTypeInfo[] keyPrimitiveTypeInfos = new PrimitiveTypeInfo[keyCount];
    PrimitiveCategory[] keyPrimitiveCategories = new PrimitiveCategory[keyCount];
    ArrayList<ObjectInspector> keyPrimitiveObjectInspectorList = new ArrayList<ObjectInspector>(keyCount);
    for (int i = 0; i < keyCount; i++) {
        PrimitiveTypeInfo primitiveTypeInfo = (PrimitiveTypeInfo) TypeInfoUtils.getTypeInfoFromTypeString(keyTypeNames[i]);
        keyPrimitiveTypeInfos[i] = primitiveTypeInfo;
        PrimitiveCategory primitiveCategory = primitiveTypeInfo.getPrimitiveCategory();
        keyPrimitiveCategories[i] = primitiveCategory;
        keyPrimitiveObjectInspectorList.add(PrimitiveObjectInspectorFactory.getPrimitiveWritableObjectInspector(primitiveTypeInfo));
    }
    boolean[] keyColumnSortOrderIsDesc = new boolean[keyCount];
    Arrays.fill(keyColumnSortOrderIsDesc, false);
    byte[] keyColumnNullMarker = new byte[keyCount];
    Arrays.fill(keyColumnNullMarker, BinarySortableSerDe.ZERO);
    byte[] keyColumnNotNullMarker = new byte[keyCount];
    Arrays.fill(keyColumnNotNullMarker, BinarySortableSerDe.ONE);
    BinarySortableSerializeWrite keySerializeWrite = new BinarySortableSerializeWrite(keyColumnSortOrderIsDesc, keyColumnNullMarker, keyColumnNotNullMarker);
    PrimitiveTypeInfo[] valuePrimitiveTypeInfos = valueSource.primitiveTypeInfos();
    final int columnCount = valuePrimitiveTypeInfos.length;
    SerializeWrite valueSerializeWrite = new LazyBinarySerializeWrite(columnCount);
    final int count = rows.length;
    for (int i = 0; i < count; i++) {
        Object[] valueRow = rows[i];
        Output valueOutput = new Output();
        ((LazyBinarySerializeWrite) valueSerializeWrite).set(valueOutput);
        for (int index = 0; index < columnCount; index++) {
            Writable writable = (Writable) valueRow[index];
            VerifyFastRow.serializeWrite(valueSerializeWrite, valuePrimitiveTypeInfos[index], writable);
        }
        byte[] value = Arrays.copyOf(valueOutput.getData(), valueOutput.getLength());
        // Add a new key or add a value to an existing key?
        byte[] key;
        if (random.nextBoolean() || verifyTable.getCount() == 0) {
            Object[] keyRow = VectorRandomRowSource.randomRow(keyCount, random, keyPrimitiveObjectInspectorList, keyPrimitiveCategories, keyPrimitiveTypeInfos);
            Output keyOutput = new Output();
            keySerializeWrite.set(keyOutput);
            for (int index = 0; index < keyCount; index++) {
                Writable writable = (Writable) keyRow[index];
                VerifyFastRow.serializeWrite(keySerializeWrite, keyPrimitiveTypeInfos[index], writable);
            }
            key = Arrays.copyOf(keyOutput.getData(), keyOutput.getLength());
            verifyTable.add(key, keyRow, value, valueRow);
        } else {
            key = verifyTable.addRandomExisting(value, valueRow, random);
        }
        // Serialize keyRow into key bytes.
        BytesWritable keyWritable = new BytesWritable(key);
        BytesWritable valueWritable = new BytesWritable(value);
        map.putRow(keyWritable, valueWritable);
    // verifyTable.verify(map);
    }
    verifyTable.verify(map, hashTableKeyType, valuePrimitiveTypeInfos, doClipping, useExactBytes, random);
}

Example 4 with TypeInfoUtils.getTypeInfoFromTypeString

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

the class TestOrcStruct method testInspectorFromTypeInfo.

@Test
public void testInspectorFromTypeInfo() throws Exception {
    TypeInfo typeInfo = TypeInfoUtils.getTypeInfoFromTypeString("struct<c1:boolean,c2:tinyint" + ",c3:smallint,c4:int,c5:bigint,c6:float,c7:double,c8:binary," + "c9:string,c10:struct<c1:int>,c11:map<int,int>,c12:uniontype<int>" + ",c13:array<timestamp>>");
    StructObjectInspector inspector = (StructObjectInspector) OrcStruct.createObjectInspector(typeInfo);
    assertEquals("struct<c1:boolean,c2:tinyint,c3:smallint,c4:int,c5:" + "bigint,c6:float,c7:double,c8:binary,c9:string,c10:struct<" + "c1:int>,c11:map<int,int>,c12:uniontype<int>,c13:array<timestamp>>", inspector.getTypeName());
    assertEquals(null, inspector.getAllStructFieldRefs().get(0).getFieldComment());
    assertEquals(null, inspector.getStructFieldRef("UNKNOWN"));
    OrcStruct s1 = new OrcStruct(13);
    for (int i = 0; i < 13; ++i) {
        s1.setFieldValue(i, i);
    }
    List<Object> list = new ArrayList<Object>();
    list.addAll(Arrays.asList(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12));
    assertEquals(list, inspector.getStructFieldsDataAsList(s1));
    ListObjectInspector listOI = (ListObjectInspector) inspector.getAllStructFieldRefs().get(12).getFieldObjectInspector();
    assertEquals(ObjectInspector.Category.LIST, listOI.getCategory());
    assertEquals(10, listOI.getListElement(list, 10));
    assertEquals(null, listOI.getListElement(list, -1));
    assertEquals(null, listOI.getListElement(list, 13));
    assertEquals(13, listOI.getListLength(list));
    Map<Integer, Integer> map = new HashMap<Integer, Integer>();
    map.put(1, 2);
    map.put(2, 4);
    map.put(3, 6);
    MapObjectInspector mapOI = (MapObjectInspector) inspector.getAllStructFieldRefs().get(10).getFieldObjectInspector();
    assertEquals(3, mapOI.getMapSize(map));
    assertEquals(4, mapOI.getMapValueElement(map, 2));
}

Example 5 with TypeInfoUtils.getTypeInfoFromTypeString

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

the class MyTestPrimitiveClass method getPrimitiveTypeInfo.

public static PrimitiveTypeInfo getPrimitiveTypeInfo(int index, ExtraTypeInfo extraTypeInfo) {
    PrimitiveCategory primitiveCategory = getPrimitiveCategory(index);
    String typeName;
    switch(primitiveCategory) {
        case BYTE:
            typeName = "tinyint";
            break;
        case SHORT:
            typeName = "smallint";
            break;
        case LONG:
            typeName = "bigint";
            break;
        case CHAR:
            typeName = String.format("char(%d)", extraTypeInfo.hiveCharMaxLength);
            break;
        case VARCHAR:
            typeName = String.format("varchar(%d)", extraTypeInfo.hiveVarcharMaxLength);
            break;
        case DECIMAL:
            typeName = String.format("decimal(%d,%d)", extraTypeInfo.precision, extraTypeInfo.scale);
            break;
        default:
            // No type name difference or adornment.
            typeName = primitiveCategory.name().toLowerCase();
            break;
    }
    PrimitiveTypeInfo primitiveTypeInfo = (PrimitiveTypeInfo) TypeInfoUtils.getTypeInfoFromTypeString(typeName);
    return primitiveTypeInfo;
}