Examples with ShortWritable org.apache.hadoop.hive.serde2.io.ShortWritable used on opensource projects

Example 21 with ShortWritable

use of org.apache.hadoop.hive.serde2.io.ShortWritable in project hive by apache.

the class TestVectorGroupByOperator method testMultiKey.

private void testMultiKey(String aggregateName, FakeVectorRowBatchFromObjectIterables data, HashMap<Object, Object> expected) throws HiveException {
    Map<String, Integer> mapColumnNames = new HashMap<String, Integer>();
    ArrayList<String> outputColumnNames = new ArrayList<String>();
    ArrayList<ExprNodeDesc> keysDesc = new ArrayList<ExprNodeDesc>();
    Set<Object> keys = new HashSet<Object>();
    // The types array tells us the number of columns in the data
    final String[] columnTypes = data.getTypes();
    // Columns 0..N-1 are keys. Column N is the aggregate value input
    int i = 0;
    for (; i < columnTypes.length - 1; ++i) {
        String columnName = String.format("_col%d", i);
        mapColumnNames.put(columnName, i);
        outputColumnNames.add(columnName);
    }
    mapColumnNames.put("value", i);
    outputColumnNames.add("value");
    VectorizationContext ctx = new VectorizationContext("name", outputColumnNames);
    ArrayList<AggregationDesc> aggs = new ArrayList(1);
    aggs.add(buildAggregationDesc(ctx, aggregateName, GenericUDAFEvaluator.Mode.PARTIAL1, "value", TypeInfoFactory.getPrimitiveTypeInfo(columnTypes[i])));
    for (i = 0; i < columnTypes.length - 1; ++i) {
        String columnName = String.format("_col%d", i);
        keysDesc.add(buildColumnDesc(ctx, columnName, TypeInfoFactory.getPrimitiveTypeInfo(columnTypes[i])));
    }
    GroupByDesc desc = new GroupByDesc();
    VectorGroupByDesc vectorGroupByDesc = new VectorGroupByDesc();
    desc.setOutputColumnNames(outputColumnNames);
    desc.setAggregators(aggs);
    desc.setKeys(keysDesc);
    vectorGroupByDesc.setProcessingMode(ProcessingMode.HASH);
    CompilationOpContext cCtx = new CompilationOpContext();
    Operator<? extends OperatorDesc> groupByOp = OperatorFactory.get(cCtx, desc);
    VectorGroupByOperator vgo = (VectorGroupByOperator) Vectorizer.vectorizeGroupByOperator(groupByOp, ctx, vectorGroupByDesc);
    FakeCaptureVectorToRowOutputOperator out = FakeCaptureVectorToRowOutputOperator.addCaptureOutputChild(cCtx, vgo);
    vgo.initialize(hconf, null);
    out.setOutputInspector(new FakeCaptureVectorToRowOutputOperator.OutputInspector() {

        private int rowIndex;

        private String aggregateName;

        private Map<Object, Object> expected;

        private Set<Object> keys;

        @Override
        public void inspectRow(Object row, int tag) throws HiveException {
            assertTrue(row instanceof Object[]);
            Object[] fields = (Object[]) row;
            assertEquals(columnTypes.length, fields.length);
            ArrayList<Object> keyValue = new ArrayList<Object>(columnTypes.length - 1);
            for (int i = 0; i < columnTypes.length - 1; ++i) {
                Object key = fields[i];
                if (null == key) {
                    keyValue.add(null);
                } else if (key instanceof Text) {
                    Text txKey = (Text) key;
                    keyValue.add(txKey.toString());
                } else if (key instanceof ByteWritable) {
                    ByteWritable bwKey = (ByteWritable) key;
                    keyValue.add(bwKey.get());
                } else if (key instanceof ShortWritable) {
                    ShortWritable swKey = (ShortWritable) key;
                    keyValue.add(swKey.get());
                } else if (key instanceof IntWritable) {
                    IntWritable iwKey = (IntWritable) key;
                    keyValue.add(iwKey.get());
                } else if (key instanceof LongWritable) {
                    LongWritable lwKey = (LongWritable) key;
                    keyValue.add(lwKey.get());
                } else if (key instanceof TimestampWritableV2) {
                    TimestampWritableV2 twKey = (TimestampWritableV2) key;
                    keyValue.add(twKey.getTimestamp());
                } else if (key instanceof DoubleWritable) {
                    DoubleWritable dwKey = (DoubleWritable) key;
                    keyValue.add(dwKey.get());
                } else if (key instanceof FloatWritable) {
                    FloatWritable fwKey = (FloatWritable) key;
                    keyValue.add(fwKey.get());
                } else if (key instanceof BooleanWritable) {
                    BooleanWritable bwKey = (BooleanWritable) key;
                    keyValue.add(bwKey.get());
                } else {
                    Assert.fail(String.format("Not implemented key output type %s: %s", key.getClass().getName(), key));
                }
            }
            String keyAsString = Arrays.deepToString(keyValue.toArray());
            assertTrue(expected.containsKey(keyValue));
            Object expectedValue = expected.get(keyValue);
            Object value = fields[columnTypes.length - 1];
            Validator validator = getValidator(aggregateName);
            validator.validate(keyAsString, expectedValue, new Object[] { value });
            keys.add(keyValue);
        }

        private FakeCaptureVectorToRowOutputOperator.OutputInspector init(String aggregateName, Map<Object, Object> expected, Set<Object> keys) {
            this.aggregateName = aggregateName;
            this.expected = expected;
            this.keys = keys;
            return this;
        }
    }.init(aggregateName, expected, keys));
    for (VectorizedRowBatch unit : data) {
        vgo.process(unit, 0);
    }
    vgo.close(false);
    List<Object> outBatchList = out.getCapturedRows();
    assertNotNull(outBatchList);
    assertEquals(expected.size(), outBatchList.size());
    assertEquals(expected.size(), keys.size());
}

Example 22 with ShortWritable

use of org.apache.hadoop.hive.serde2.io.ShortWritable in project hive by apache.

the class TestVectorGroupByOperator method testKeyTypeAggregate.

private void testKeyTypeAggregate(String aggregateName, FakeVectorRowBatchFromObjectIterables data, Map<Object, Object> expected) throws HiveException {
    List<String> mapColumnNames = new ArrayList<String>();
    mapColumnNames.add("Key");
    mapColumnNames.add("Value");
    VectorizationContext ctx = new VectorizationContext("name", mapColumnNames);
    Set<Object> keys = new HashSet<Object>();
    AggregationDesc agg = buildAggregationDesc(ctx, aggregateName, GenericUDAFEvaluator.Mode.PARTIAL1, "Value", TypeInfoFactory.getPrimitiveTypeInfo(data.getTypes()[1]));
    ArrayList<AggregationDesc> aggs = new ArrayList<AggregationDesc>();
    aggs.add(agg);
    ArrayList<String> outputColumnNames = new ArrayList<String>();
    outputColumnNames.add("_col0");
    outputColumnNames.add("_col1");
    GroupByDesc desc = new GroupByDesc();
    VectorGroupByDesc vectorGroupByDesc = new VectorGroupByDesc();
    desc.setOutputColumnNames(outputColumnNames);
    desc.setAggregators(aggs);
    vectorGroupByDesc.setProcessingMode(ProcessingMode.HASH);
    ExprNodeDesc keyExp = buildColumnDesc(ctx, "Key", TypeInfoFactory.getPrimitiveTypeInfo(data.getTypes()[0]));
    ArrayList<ExprNodeDesc> keysDesc = new ArrayList<ExprNodeDesc>();
    keysDesc.add(keyExp);
    desc.setKeys(keysDesc);
    CompilationOpContext cCtx = new CompilationOpContext();
    Operator<? extends OperatorDesc> groupByOp = OperatorFactory.get(cCtx, desc);
    VectorGroupByOperator vgo = (VectorGroupByOperator) Vectorizer.vectorizeGroupByOperator(groupByOp, ctx, vectorGroupByDesc);
    if (vgo == null) {
        assertTrue(false);
    }
    FakeCaptureVectorToRowOutputOperator out = FakeCaptureVectorToRowOutputOperator.addCaptureOutputChild(cCtx, vgo);
    vgo.initialize(hconf, null);
    out.setOutputInspector(new FakeCaptureVectorToRowOutputOperator.OutputInspector() {

        private int rowIndex;

        private String aggregateName;

        private Map<Object, Object> expected;

        private Set<Object> keys;

        @Override
        public void inspectRow(Object row, int tag) throws HiveException {
            assertTrue(row instanceof Object[]);
            Object[] fields = (Object[]) row;
            assertEquals(2, fields.length);
            Object key = fields[0];
            Object keyValue = null;
            if (null == key) {
                keyValue = null;
            } else if (key instanceof ByteWritable) {
                ByteWritable bwKey = (ByteWritable) key;
                keyValue = bwKey.get();
            } else if (key instanceof ShortWritable) {
                ShortWritable swKey = (ShortWritable) key;
                keyValue = swKey.get();
            } else if (key instanceof IntWritable) {
                IntWritable iwKey = (IntWritable) key;
                keyValue = iwKey.get();
            } else if (key instanceof LongWritable) {
                LongWritable lwKey = (LongWritable) key;
                keyValue = lwKey.get();
            } else if (key instanceof TimestampWritableV2) {
                TimestampWritableV2 twKey = (TimestampWritableV2) key;
                keyValue = twKey.getTimestamp().toSqlTimestamp();
            } else if (key instanceof DoubleWritable) {
                DoubleWritable dwKey = (DoubleWritable) key;
                keyValue = dwKey.get();
            } else if (key instanceof FloatWritable) {
                FloatWritable fwKey = (FloatWritable) key;
                keyValue = fwKey.get();
            } else if (key instanceof BooleanWritable) {
                BooleanWritable bwKey = (BooleanWritable) key;
                keyValue = bwKey.get();
            } else if (key instanceof HiveDecimalWritable) {
                HiveDecimalWritable hdwKey = (HiveDecimalWritable) key;
                keyValue = hdwKey.getHiveDecimal();
            } else {
                Assert.fail(String.format("Not implemented key output type %s: %s", key.getClass().getName(), key));
            }
            String keyValueAsString = String.format("%s", keyValue);
            assertTrue(expected.containsKey(keyValue));
            Object expectedValue = expected.get(keyValue);
            Object value = fields[1];
            Validator validator = getValidator(aggregateName);
            validator.validate(keyValueAsString, expectedValue, new Object[] { value });
            keys.add(keyValue);
        }

        private FakeCaptureVectorToRowOutputOperator.OutputInspector init(String aggregateName, Map<Object, Object> expected, Set<Object> keys) {
            this.aggregateName = aggregateName;
            this.expected = expected;
            this.keys = keys;
            return this;
        }
    }.init(aggregateName, expected, keys));
    for (VectorizedRowBatch unit : data) {
        vgo.process(unit, 0);
    }
    vgo.close(false);
    List<Object> outBatchList = out.getCapturedRows();
    assertNotNull(outBatchList);
    assertEquals(expected.size(), outBatchList.size());
    assertEquals(expected.size(), keys.size());
}

Example 23 with ShortWritable

use of org.apache.hadoop.hive.serde2.io.ShortWritable in project hive by apache.

the class TestRCFile method testSimpleReadAndWrite.

@Test
public void testSimpleReadAndWrite() throws IOException, SerDeException {
    cleanup();
    byte[][] record_1 = { "123".getBytes(StandardCharsets.UTF_8), "456".getBytes(StandardCharsets.UTF_8), "789".getBytes(StandardCharsets.UTF_8), "1000".getBytes(StandardCharsets.UTF_8), "5.3".getBytes(StandardCharsets.UTF_8), "hive and hadoop".getBytes(StandardCharsets.UTF_8), new byte[0], "NULL".getBytes(StandardCharsets.UTF_8) };
    byte[][] record_2 = { "100".getBytes(StandardCharsets.UTF_8), "200".getBytes(StandardCharsets.UTF_8), "123".getBytes(StandardCharsets.UTF_8), "1000".getBytes(StandardCharsets.UTF_8), "5.3".getBytes(StandardCharsets.UTF_8), "hive and hadoop".getBytes(StandardCharsets.UTF_8), new byte[0], "NULL".getBytes(StandardCharsets.UTF_8) };
    RCFileOutputFormat.setColumnNumber(conf, expectedFieldsData.length);
    RCFile.Writer writer = new RCFile.Writer(fs, conf, file, null, RCFile.createMetadata(new Text("apple"), new Text("block"), new Text("cat"), new Text("dog")), new DefaultCodec());
    BytesRefArrayWritable bytes = new BytesRefArrayWritable(record_1.length);
    for (int i = 0; i < record_1.length; i++) {
        BytesRefWritable cu = new BytesRefWritable(record_1[i], 0, record_1[i].length);
        bytes.set(i, cu);
    }
    writer.append(bytes);
    bytes.clear();
    for (int i = 0; i < record_2.length; i++) {
        BytesRefWritable cu = new BytesRefWritable(record_2[i], 0, record_2[i].length);
        bytes.set(i, cu);
    }
    writer.append(bytes);
    writer.close();
    Object[] expectedRecord_1 = { new ByteWritable((byte) 123), new ShortWritable((short) 456), new IntWritable(789), new LongWritable(1000), new DoubleWritable(5.3), new Text("hive and hadoop"), null, null };
    Object[] expectedRecord_2 = { new ByteWritable((byte) 100), new ShortWritable((short) 200), new IntWritable(123), new LongWritable(1000), new DoubleWritable(5.3), new Text("hive and hadoop"), null, null };
    RCFile.Reader reader = new RCFile.Reader(fs, file, conf);
    assertEquals(new Text("block"), reader.getMetadata().get(new Text("apple")));
    assertEquals(new Text("block"), reader.getMetadataValueOf(new Text("apple")));
    assertEquals(new Text("dog"), reader.getMetadataValueOf(new Text("cat")));
    LongWritable rowID = new LongWritable();
    for (int i = 0; i < 2; i++) {
        reader.next(rowID);
        BytesRefArrayWritable cols = new BytesRefArrayWritable();
        reader.getCurrentRow(cols);
        cols.resetValid(8);
        Object row = serDe.deserialize(cols);
        StructObjectInspector oi = (StructObjectInspector) serDe.getObjectInspector();
        List<? extends StructField> fieldRefs = oi.getAllStructFieldRefs();
        assertEquals("Field size should be 8", 8, fieldRefs.size());
        for (int j = 0; j < fieldRefs.size(); j++) {
            Object fieldData = oi.getStructFieldData(row, fieldRefs.get(j));
            Object standardWritableData = ObjectInspectorUtils.copyToStandardObject(fieldData, fieldRefs.get(j).getFieldObjectInspector(), ObjectInspectorCopyOption.WRITABLE);
            if (i == 0) {
                assertEquals("Field " + i, standardWritableData, expectedRecord_1[j]);
            } else {
                assertEquals("Field " + i, standardWritableData, expectedRecord_2[j]);
            }
        }
    }
    reader.close();
}

Example 24 with ShortWritable

use of org.apache.hadoop.hive.serde2.io.ShortWritable in project hive by apache.

the class TestGenericUDFOPDivide method testByteDivideShort.

@Test
public void testByteDivideShort() throws HiveException {
    GenericUDFOPDivide udf = new GenericUDFOPDivide();
    ByteWritable left = new ByteWritable((byte) 4);
    ShortWritable right = new ShortWritable((short) 6);
    ObjectInspector[] inputOIs = { PrimitiveObjectInspectorFactory.writableByteObjectInspector, PrimitiveObjectInspectorFactory.writableShortObjectInspector };
    DeferredObject[] args = { new DeferredJavaObject(left), new DeferredJavaObject(right) };
    PrimitiveObjectInspector oi = (PrimitiveObjectInspector) udf.initialize(inputOIs);
    Assert.assertEquals(oi.getTypeInfo(), TypeInfoFactory.getDecimalTypeInfo(9, 6));
    HiveDecimalWritable res = (HiveDecimalWritable) udf.evaluate(args);
    Assert.assertEquals(HiveDecimal.create("0.666667"), res.getHiveDecimal());
}

Example 25 with ShortWritable

use of org.apache.hadoop.hive.serde2.io.ShortWritable in project hive by apache.

the class TestGenericUDFOPPlus method testBytePlusShort.

@Test
public void testBytePlusShort() throws HiveException {
    GenericUDFOPPlus udf = new GenericUDFOPPlus();
    // Byte
    ByteWritable left = new ByteWritable((byte) 4);
    ShortWritable right = new ShortWritable((short) 6);
    ObjectInspector[] inputOIs = { PrimitiveObjectInspectorFactory.writableByteObjectInspector, PrimitiveObjectInspectorFactory.writableShortObjectInspector };
    DeferredObject[] args = { new DeferredJavaObject(left), new DeferredJavaObject(right) };
    PrimitiveObjectInspector oi = (PrimitiveObjectInspector) udf.initialize(inputOIs);
    Assert.assertEquals(oi.getTypeInfo(), TypeInfoFactory.shortTypeInfo);
    ShortWritable res = (ShortWritable) udf.evaluate(args);
    Assert.assertEquals(10, res.get());
}