Examples with LazySerDeParameters org.apache.hadoop.hive.serde2.lazy.LazySerDeParameters used on opensource projects

Example 21 with LazySerDeParameters

use of org.apache.hadoop.hive.serde2.lazy.LazySerDeParameters in project cdap by caskdata.

the class SimpleMapEqualComparerTest method testIncompatibleType.

@Test
public void testIncompatibleType() throws SerDeException, IOException {
    // empty maps
    StringTextMapHolder o1 = new StringTextMapHolder();
    StructObjectInspector oi1 = (StructObjectInspector) ObjectInspectorFactory.getReflectionObjectInspector(StringTextMapHolder.class);
    LazySimpleSerDe serde = new LazySimpleSerDe();
    Configuration conf = new Configuration();
    Properties tbl = new Properties();
    tbl.setProperty(serdeConstants.LIST_COLUMNS, ObjectInspectorUtils.getFieldNames(oi1));
    tbl.setProperty(serdeConstants.LIST_COLUMN_TYPES, ObjectInspectorUtils.getFieldTypes(oi1));
    LazySerDeParameters serdeParams = new LazySerDeParameters(conf, tbl, LazySimpleSerDe.class.getName());
    serde.initialize(conf, tbl);
    ObjectInspector oi2 = serde.getObjectInspector();
    Object o2 = serializeAndDeserialize(o1, oi1, serde, serdeParams);
    int rc = ObjectInspectorUtils.compare(o1, oi1, o2, oi2, new SimpleMapEqualComparer());
    Assert.assertEquals(0, rc);
    // equal maps
    o1.mMap.put("42", new Text("The answer to Life, Universe And Everything"));
    o1.mMap.put("1729", new Text("A taxi cab number"));
    o2 = serializeAndDeserialize(o1, oi1, serde, serdeParams);
    rc = ObjectInspectorUtils.compare(o1, oi1, o2, oi2, new SimpleMapEqualComparer());
    Assert.assertFalse(0 == rc);
}

Example 22 with LazySerDeParameters

use of org.apache.hadoop.hive.serde2.lazy.LazySerDeParameters in project cdap by caskdata.

the class SimpleMapEqualComparerTest method testCompatibleType.

@Test
public void testCompatibleType() throws SerDeException, IOException {
    // empty maps
    TextStringMapHolder o1 = new TextStringMapHolder();
    StructObjectInspector oi1 = (StructObjectInspector) ObjectInspectorFactory.getReflectionObjectInspector(TextStringMapHolder.class);
    LazySimpleSerDe serde = new LazySimpleSerDe();
    Configuration conf = new Configuration();
    Properties tbl = new Properties();
    tbl.setProperty(serdeConstants.LIST_COLUMNS, ObjectInspectorUtils.getFieldNames(oi1));
    tbl.setProperty(serdeConstants.LIST_COLUMN_TYPES, ObjectInspectorUtils.getFieldTypes(oi1));
    LazySerDeParameters serdeParams = new LazySerDeParameters(conf, tbl, LazySimpleSerDe.class.getName());
    serde.initialize(conf, tbl);
    ObjectInspector oi2 = serde.getObjectInspector();
    Object o2 = serializeAndDeserialize(o1, oi1, serde, serdeParams);
    int rc = ObjectInspectorUtils.compare(o1, oi1, o2, oi2, new SimpleMapEqualComparer());
    Assert.assertEquals(0, rc);
    // equal maps
    o1.mMap.put(new Text("42"), "The answer to Life, Universe And Everything");
    o1.mMap.put(new Text("1729"), "A taxi cab number");
    o2 = serializeAndDeserialize(o1, oi1, serde, serdeParams);
    rc = ObjectInspectorUtils.compare(o1, oi1, o2, oi2, new SimpleMapEqualComparer());
    Assert.assertEquals(0, rc);
    // unequal maps
    o1.mMap.put(new Text("1729"), "Hardy-Ramanujan Number");
    rc = ObjectInspectorUtils.compare(o1, oi1, o2, oi2, new SimpleMapEqualComparer());
    Assert.assertFalse(0 == rc);
}

Example 23 with LazySerDeParameters

use of org.apache.hadoop.hive.serde2.lazy.LazySerDeParameters in project hive by apache.

the class TestHiveAccumuloTableOutputFormat method testWriteMap.

@Test
public void testWriteMap() throws Exception {
    Instance inst = new MockInstance(test.getMethodName());
    Connector conn = inst.getConnector("root", new PasswordToken(""));
    HiveAccumuloTableOutputFormat outputFormat = new HiveAccumuloTableOutputFormat();
    String table = test.getMethodName();
    conn.tableOperations().create(table);
    JobConf conf = new JobConf();
    conf.set(AccumuloConnectionParameters.INSTANCE_NAME, inst.getInstanceName());
    conf.set(AccumuloConnectionParameters.USER_NAME, "root");
    conf.set(AccumuloConnectionParameters.USER_PASS, "");
    conf.setBoolean(AccumuloConnectionParameters.USE_MOCK_INSTANCE, true);
    conf.set(AccumuloConnectionParameters.TABLE_NAME, test.getMethodName());
    FileSystem local = FileSystem.getLocal(conf);
    outputFormat.checkOutputSpecs(local, conf);
    RecordWriter<Text, Mutation> recordWriter = outputFormat.getRecordWriter(local, conf, null, null);
    List<String> names = Arrays.asList("row", "col1");
    List<TypeInfo> types = Arrays.<TypeInfo>asList(TypeInfoFactory.stringTypeInfo, TypeInfoFactory.stringTypeInfo);
    Properties tableProperties = new Properties();
    tableProperties.setProperty(AccumuloSerDeParameters.COLUMN_MAPPINGS, ":rowID,cf:*");
    tableProperties.setProperty(serdeConstants.FIELD_DELIM, " ");
    tableProperties.setProperty(serdeConstants.LIST_COLUMNS, Joiner.on(',').join(names));
    tableProperties.setProperty(serdeConstants.LIST_COLUMN_TYPES, Joiner.on(',').join(types));
    AccumuloSerDeParameters accumuloSerDeParams = new AccumuloSerDeParameters(new Configuration(), tableProperties, AccumuloSerDe.class.getSimpleName());
    LazySerDeParameters serDeParams = accumuloSerDeParams.getSerDeParameters();
    AccumuloRowSerializer serializer = new AccumuloRowSerializer(0, serDeParams, accumuloSerDeParams.getColumnMappings(), AccumuloSerDeParameters.DEFAULT_VISIBILITY_LABEL, accumuloSerDeParams.getRowIdFactory());
    TypeInfo stringTypeInfo = TypeInfoFactory.getPrimitiveTypeInfo(serdeConstants.STRING_TYPE_NAME);
    LazyStringObjectInspector stringOI = (LazyStringObjectInspector) LazyFactory.createLazyObjectInspector(stringTypeInfo, new byte[] { 0 }, 0, serDeParams.getNullSequence(), serDeParams.isEscaped(), serDeParams.getEscapeChar());
    LazyMapObjectInspector mapOI = LazyObjectInspectorFactory.getLazySimpleMapObjectInspector(stringOI, stringOI, (byte) ',', (byte) ':', serDeParams.getNullSequence(), serDeParams.isEscaped(), serDeParams.getEscapeChar());
    LazySimpleStructObjectInspector structOI = (LazySimpleStructObjectInspector) LazyObjectInspectorFactory.getLazySimpleStructObjectInspector(Arrays.asList("row", "data"), Arrays.asList(stringOI, mapOI), (byte) ' ', serDeParams.getNullSequence(), serDeParams.isLastColumnTakesRest(), serDeParams.isEscaped(), serDeParams.getEscapeChar());
    LazyStruct struct = (LazyStruct) LazyFactory.createLazyObject(structOI);
    ByteArrayRef bytes = new ByteArrayRef();
    bytes.setData("row cq1:value1,cq2:value2".getBytes());
    struct.init(bytes, 0, bytes.getData().length);
    // Serialize the struct into a mutation
    Mutation m = serializer.serialize(struct, structOI);
    // Write the mutation
    recordWriter.write(new Text(table), m);
    // Close the writer
    recordWriter.close(null);
    Iterator<Entry<Key, Value>> iter = conn.createScanner(table, new Authorizations()).iterator();
    Assert.assertTrue("Iterator did not have an element as expected", iter.hasNext());
    Entry<Key, Value> entry = iter.next();
    Key k = entry.getKey();
    Value v = entry.getValue();
    Assert.assertEquals("row", k.getRow().toString());
    Assert.assertEquals("cf", k.getColumnFamily().toString());
    Assert.assertEquals("cq1", k.getColumnQualifier().toString());
    Assert.assertEquals(AccumuloSerDeParameters.DEFAULT_VISIBILITY_LABEL, k.getColumnVisibilityParsed());
    Assert.assertEquals("value1", new String(v.get()));
    Assert.assertTrue("Iterator did not have an element as expected", iter.hasNext());
    entry = iter.next();
    k = entry.getKey();
    v = entry.getValue();
    Assert.assertEquals("row", k.getRow().toString());
    Assert.assertEquals("cf", k.getColumnFamily().toString());
    Assert.assertEquals("cq2", k.getColumnQualifier().toString());
    Assert.assertEquals(AccumuloSerDeParameters.DEFAULT_VISIBILITY_LABEL, k.getColumnVisibilityParsed());
    Assert.assertEquals("value2", new String(v.get()));
    Assert.assertFalse("Iterator unexpectedly had more data", iter.hasNext());
}

Example 24 with LazySerDeParameters

use of org.apache.hadoop.hive.serde2.lazy.LazySerDeParameters in project hive by apache.

the class MultiDelimitSerDe method initialize.

@Override
public void initialize(Configuration conf, Properties tbl) throws SerDeException {
    // get the SerDe parameters
    super.initialize(conf, tbl);
    serdeParams = new LazySerDeParameters(conf, tbl, getClass().getName());
    fieldDelimited = tbl.getProperty(serdeConstants.FIELD_DELIM);
    if (fieldDelimited == null || fieldDelimited.isEmpty()) {
        throw new SerDeException("This table does not have serde property \"field.delim\"!");
    }
    // get the collection separator and map key separator
    // TODO: use serdeConstants.COLLECTION_DELIM when the typo is fixed
    collSep = LazyUtils.getByte(tbl.getProperty(COLLECTION_DELIM), DEFAULT_SEPARATORS[1]);
    keySep = LazyUtils.getByte(tbl.getProperty(serdeConstants.MAPKEY_DELIM), DEFAULT_SEPARATORS[2]);
    serdeParams.setSeparator(1, collSep);
    serdeParams.setSeparator(2, keySep);
    // Create the ObjectInspectors for the fields
    cachedObjectInspector = LazyFactory.createLazyStructInspector(serdeParams.getColumnNames(), serdeParams.getColumnTypes(), serdeParams.getSeparators(), serdeParams.getNullSequence(), serdeParams.isLastColumnTakesRest(), serdeParams.isEscaped(), serdeParams.getEscapeChar());
    cachedLazyStruct = (LazyStruct) LazyFactory.createLazyObject(cachedObjectInspector);
    assert serdeParams.getColumnNames().size() == serdeParams.getColumnTypes().size();
    numColumns = serdeParams.getColumnNames().size();
}

Example 25 with LazySerDeParameters

use of org.apache.hadoop.hive.serde2.lazy.LazySerDeParameters in project hive by apache.

the class TestVectorSerDeRow method testVectorDeserializeRow.

void testVectorDeserializeRow(Random r, SerializationType serializationType, boolean alternate1, boolean alternate2, boolean useExternalBuffer) throws HiveException, IOException, SerDeException {
    String[] emptyScratchTypeNames = new String[0];
    VectorRandomRowSource source = new VectorRandomRowSource();
    source.init(r);
    VectorizedRowBatchCtx batchContext = new VectorizedRowBatchCtx();
    batchContext.init(source.rowStructObjectInspector(), emptyScratchTypeNames);
    VectorizedRowBatch batch = batchContext.createVectorizedRowBatch();
    // junk the destination for the 1st pass
    for (ColumnVector cv : batch.cols) {
        Arrays.fill(cv.isNull, true);
    }
    PrimitiveTypeInfo[] primitiveTypeInfos = source.primitiveTypeInfos();
    int fieldCount = source.typeNames().size();
    DeserializeRead deserializeRead;
    SerializeWrite serializeWrite;
    switch(serializationType) {
        case BINARY_SORTABLE:
            boolean useColumnSortOrderIsDesc = alternate1;
            if (!useColumnSortOrderIsDesc) {
                deserializeRead = new BinarySortableDeserializeRead(source.primitiveTypeInfos(), useExternalBuffer);
                serializeWrite = new BinarySortableSerializeWrite(fieldCount);
            } else {
                boolean[] columnSortOrderIsDesc = new boolean[fieldCount];
                for (int i = 0; i < fieldCount; i++) {
                    columnSortOrderIsDesc[i] = r.nextBoolean();
                }
                deserializeRead = new BinarySortableDeserializeRead(source.primitiveTypeInfos(), useExternalBuffer, columnSortOrderIsDesc);
                byte[] columnNullMarker = new byte[fieldCount];
                byte[] columnNotNullMarker = new byte[fieldCount];
                for (int i = 0; i < fieldCount; i++) {
                    if (columnSortOrderIsDesc[i]) {
                        // Descending
                        // Null last (default for descending order)
                        columnNullMarker[i] = BinarySortableSerDe.ZERO;
                        columnNotNullMarker[i] = BinarySortableSerDe.ONE;
                    } else {
                        // Ascending
                        // Null first (default for ascending order)
                        columnNullMarker[i] = BinarySortableSerDe.ZERO;
                        columnNotNullMarker[i] = BinarySortableSerDe.ONE;
                    }
                }
                serializeWrite = new BinarySortableSerializeWrite(columnSortOrderIsDesc, columnNullMarker, columnNotNullMarker);
            }
            boolean useBinarySortableCharsNeedingEscape = alternate2;
            if (useBinarySortableCharsNeedingEscape) {
                source.addBinarySortableAlphabets();
            }
            break;
        case LAZY_BINARY:
            deserializeRead = new LazyBinaryDeserializeRead(source.primitiveTypeInfos(), useExternalBuffer);
            serializeWrite = new LazyBinarySerializeWrite(fieldCount);
            break;
        case LAZY_SIMPLE:
            {
                StructObjectInspector rowObjectInspector = source.rowStructObjectInspector();
                Configuration conf = new Configuration();
                Properties tbl = new Properties();
                tbl.setProperty(serdeConstants.FIELD_DELIM, "\t");
                tbl.setProperty(serdeConstants.LINE_DELIM, "\n");
                byte separator = (byte) '\t';
                boolean useLazySimpleEscapes = alternate1;
                if (useLazySimpleEscapes) {
                    tbl.setProperty(serdeConstants.QUOTE_CHAR, "'");
                    String escapeString = "\\";
                    tbl.setProperty(serdeConstants.ESCAPE_CHAR, escapeString);
                }
                LazySerDeParameters lazySerDeParams = getSerDeParams(conf, tbl, rowObjectInspector);
                if (useLazySimpleEscapes) {
                    // LazySimple seems to throw away everything but \n and \r.
                    boolean[] needsEscape = lazySerDeParams.getNeedsEscape();
                    StringBuilder sb = new StringBuilder();
                    if (needsEscape['\n']) {
                        sb.append('\n');
                    }
                    if (needsEscape['\r']) {
                        sb.append('\r');
                    }
                    // for (int i = 0; i < needsEscape.length; i++) {
                    //  if (needsEscape[i]) {
                    //    sb.append((char) i);
                    //  }
                    // }
                    String needsEscapeStr = sb.toString();
                    if (needsEscapeStr.length() > 0) {
                        source.addEscapables(needsEscapeStr);
                    }
                }
                deserializeRead = new LazySimpleDeserializeRead(source.primitiveTypeInfos(), useExternalBuffer, separator, lazySerDeParams);
                serializeWrite = new LazySimpleSerializeWrite(fieldCount, separator, lazySerDeParams);
            }
            break;
        default:
            throw new Error("Unknown serialization type " + serializationType);
    }
    VectorDeserializeRow vectorDeserializeRow = new VectorDeserializeRow(deserializeRead);
    vectorDeserializeRow.init();
    // junk the destination for the 1st pass
    for (ColumnVector cv : batch.cols) {
        Arrays.fill(cv.isNull, true);
        cv.noNulls = false;
    }
    VectorExtractRow vectorExtractRow = new VectorExtractRow();
    vectorExtractRow.init(source.typeNames());
    Object[][] randomRows = source.randomRows(100000);
    int firstRandomRowIndex = 0;
    for (int i = 0; i < randomRows.length; i++) {
        Object[] row = randomRows[i];
        Output output = serializeRow(row, source, serializeWrite);
        vectorDeserializeRow.setBytes(output.getData(), 0, output.getLength());
        try {
            vectorDeserializeRow.deserialize(batch, batch.size);
        } catch (Exception e) {
            throw new HiveException("\nDeserializeRead details: " + vectorDeserializeRow.getDetailedReadPositionString(), e);
        }
        batch.size++;
        if (batch.size == batch.DEFAULT_SIZE) {
            examineBatch(batch, vectorExtractRow, primitiveTypeInfos, randomRows, firstRandomRowIndex);
            firstRandomRowIndex = i + 1;
            batch.reset();
        }
    }
    if (batch.size > 0) {
        examineBatch(batch, vectorExtractRow, primitiveTypeInfos, randomRows, firstRandomRowIndex);
    }
}