use of org.apache.hadoop.hive.serde2.Serializer in project hive by apache.
the class CreateSequenceFile method main.
public static void main(String[] args) throws Exception {
// Read parameters
int lines = 10;
List<String> extraArgs = new ArrayList<String>();
for (int ai = 0; ai < args.length; ai++) {
if (args[ai].equals("-line") && ai + 1 < args.length) {
lines = Integer.parseInt(args[ai + 1]);
ai++;
} else {
extraArgs.add(args[ai]);
}
}
if (extraArgs.size() != 1) {
usage();
}
JobConf conf = new JobConf(CreateSequenceFile.class);
ThriftSerializer serializer = new ThriftSerializer();
// Open files
SequenceFile.Writer writer = new SequenceFile.Writer(FileSystem.get(conf), conf, new Path(extraArgs.get(0)), BytesWritable.class, BytesWritable.class);
// write to file
BytesWritable key = new BytesWritable();
Random rand = new Random(20081215);
for (int i = 0; i < lines; i++) {
ArrayList<Integer> alist = new ArrayList<Integer>();
alist.add(i);
alist.add(i * 2);
alist.add(i * 3);
ArrayList<String> slist = new ArrayList<String>();
slist.add("" + i * 10);
slist.add("" + i * 100);
slist.add("" + i * 1000);
ArrayList<IntString> islist = new ArrayList<IntString>();
islist.add(new IntString(i * i, "" + i * i * i, i));
HashMap<String, String> hash = new HashMap<String, String>();
hash.put("key_" + i, "value_" + i);
Map<String, Map<String, Map<String, PropValueUnion>>> unionMap = new HashMap<String, Map<String, Map<String, PropValueUnion>>>();
Map<String, Map<String, PropValueUnion>> erMap = new HashMap<String, Map<String, PropValueUnion>>();
Map<String, PropValueUnion> attrMap = new HashMap<String, PropValueUnion>();
erMap.put("erVal" + i, attrMap);
attrMap.put("value_" + i, PropValueUnion.doubleValue(1.0));
unionMap.put("key_" + i, erMap);
Complex complex = new Complex(rand.nextInt(), "record_" + (new Integer(i)).toString(), alist, slist, islist, hash, unionMap, PropValueUnion.stringValue("test" + i), PropValueUnion.unionMStringString(hash), PropValueUnion.lString(slist));
Writable value = serializer.serialize(complex);
writer.append(key, value);
}
// Add an all-null record
Complex complex = new Complex(0, null, null, null, null, null, null, null, null, null);
Writable value = serializer.serialize(complex);
writer.append(key, value);
// Close files
writer.close();
}
use of org.apache.hadoop.hive.serde2.Serializer in project hive by apache.
the class TestAccumuloRowSerializer method testBufferResetBeforeUse.
@Test
public void testBufferResetBeforeUse() throws IOException {
ByteStream.Output output = new ByteStream.Output();
PrimitiveObjectInspector fieldObjectInspector = Mockito.mock(StringObjectInspector.class);
ColumnMapping mapping = Mockito.mock(ColumnMapping.class);
// Write some garbage to the buffer that should be erased
output.write("foobar".getBytes());
// Stub out the serializer
AccumuloRowSerializer serializer = Mockito.mock(AccumuloRowSerializer.class);
String object = "hello";
Mockito.when(serializer.getSerializedValue(Mockito.any(ObjectInspector.class), Mockito.any(), Mockito.any(ByteStream.Output.class), Mockito.any(ColumnMapping.class))).thenCallRealMethod();
Mockito.when(fieldObjectInspector.getCategory()).thenReturn(ObjectInspector.Category.PRIMITIVE);
Mockito.when(fieldObjectInspector.getPrimitiveCategory()).thenReturn(PrimitiveCategory.STRING);
Mockito.when(fieldObjectInspector.getPrimitiveWritableObject(Mockito.any(Object.class))).thenReturn(new Text(object));
Mockito.when(mapping.getEncoding()).thenReturn(ColumnEncoding.STRING);
// Invoke the method
serializer.getSerializedValue(fieldObjectInspector, object, output, mapping);
// Verify the buffer was reset (real output doesn't happen because it was mocked)
Assert.assertEquals(0, output.size());
}
use of org.apache.hadoop.hive.serde2.Serializer in project hive by apache.
the class TestAccumuloRowSerializer method testBinarySerialization.
@Test
public void testBinarySerialization() throws IOException, SerDeException {
List<String> columns = Arrays.asList("row", "cq1", "cq2", "cq3");
List<TypeInfo> types = Arrays.<TypeInfo>asList(TypeInfoFactory.stringTypeInfo, TypeInfoFactory.intTypeInfo, TypeInfoFactory.intTypeInfo, TypeInfoFactory.stringTypeInfo);
List<String> typeNames = new ArrayList<String>(types.size());
for (TypeInfo type : types) {
typeNames.add(type.getTypeName());
}
Properties tableProperties = new Properties();
tableProperties.setProperty(AccumuloSerDeParameters.COLUMN_MAPPINGS, ":rowid,cf:cq1#b,cf:cq2#b,cf:cq3");
tableProperties.setProperty(serdeConstants.FIELD_DELIM, " ");
tableProperties.setProperty(serdeConstants.LIST_COLUMNS, Joiner.on(',').join(columns));
tableProperties.setProperty(serdeConstants.LIST_COLUMN_TYPES, Joiner.on(',').join(typeNames));
AccumuloSerDeParameters accumuloSerDeParams = new AccumuloSerDeParameters(new Configuration(), tableProperties, AccumuloSerDe.class.getSimpleName());
LazySerDeParameters serDeParams = accumuloSerDeParams.getSerDeParameters();
LazySimpleStructObjectInspector oi = (LazySimpleStructObjectInspector) LazyFactory.createLazyStructInspector(columns, types, serDeParams.getSeparators(), serDeParams.getNullSequence(), serDeParams.isLastColumnTakesRest(), serDeParams.isEscaped(), serDeParams.getEscapeChar());
AccumuloRowSerializer serializer = new AccumuloRowSerializer(0, serDeParams, accumuloSerDeParams.getColumnMappings(), new ColumnVisibility(), accumuloSerDeParams.getRowIdFactory());
// Create the LazyStruct from the LazyStruct...Inspector
LazyStruct obj = (LazyStruct) LazyFactory.createLazyObject(oi);
ByteArrayRef byteRef = new ByteArrayRef();
byteRef.setData(new byte[] { 'r', 'o', 'w', '1', ' ', '1', '0', ' ', '2', '0', ' ', 'v', 'a', 'l', 'u', 'e' });
obj.init(byteRef, 0, byteRef.getData().length);
Mutation m = (Mutation) serializer.serialize(obj, oi);
Assert.assertArrayEquals("row1".getBytes(), m.getRow());
List<ColumnUpdate> updates = m.getUpdates();
Assert.assertEquals(3, updates.size());
ByteArrayOutputStream baos = new ByteArrayOutputStream();
DataOutputStream out = new DataOutputStream(baos);
ColumnUpdate update = updates.get(0);
Assert.assertEquals("cf", new String(update.getColumnFamily()));
Assert.assertEquals("cq1", new String(update.getColumnQualifier()));
out.writeInt(10);
Assert.assertArrayEquals(baos.toByteArray(), update.getValue());
update = updates.get(1);
Assert.assertEquals("cf", new String(update.getColumnFamily()));
Assert.assertEquals("cq2", new String(update.getColumnQualifier()));
baos.reset();
out.writeInt(20);
Assert.assertArrayEquals(baos.toByteArray(), update.getValue());
update = updates.get(2);
Assert.assertEquals("cf", new String(update.getColumnFamily()));
Assert.assertEquals("cq3", new String(update.getColumnQualifier()));
Assert.assertEquals("value", new String(update.getValue()));
}
use of org.apache.hadoop.hive.serde2.Serializer in project hive by apache.
the class TestAccumuloSerDe method testStructOfMapSerialization.
@Test
public void testStructOfMapSerialization() throws IOException, SerDeException {
List<String> columns = Arrays.asList("row", "col");
List<String> structColNames = Arrays.asList("map1", "map2");
TypeInfo mapTypeInfo = TypeInfoFactory.getMapTypeInfo(TypeInfoFactory.stringTypeInfo, TypeInfoFactory.stringTypeInfo);
// struct<map1:map<string,string>,map2:map<string,string>>,string
List<TypeInfo> types = Arrays.<TypeInfo>asList(TypeInfoFactory.getStructTypeInfo(structColNames, Arrays.asList(mapTypeInfo, mapTypeInfo)), TypeInfoFactory.stringTypeInfo);
Properties tableProperties = new Properties();
tableProperties.setProperty(AccumuloSerDeParameters.COLUMN_MAPPINGS, ":rowid,cf:cq");
// Use the default separators [0, 1, 2, 3, ..., 7]
tableProperties.setProperty(serdeConstants.LIST_COLUMNS, Joiner.on(',').join(columns));
tableProperties.setProperty(serdeConstants.LIST_COLUMN_TYPES, Joiner.on(',').join(types));
AccumuloSerDeParameters accumuloSerDeParams = new AccumuloSerDeParameters(new Configuration(), tableProperties, AccumuloSerDe.class.getSimpleName());
LazySerDeParameters serDeParams = accumuloSerDeParams.getSerDeParameters();
byte[] seps = serDeParams.getSeparators();
// struct<map<k:v,k:v>_map<k:v,k:v>>>
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, seps[3], seps[4], serDeParams.getNullSequence(), serDeParams.isEscaped(), serDeParams.getEscapeChar());
LazySimpleStructObjectInspector rowStructOI = (LazySimpleStructObjectInspector) LazyObjectInspectorFactory.getLazySimpleStructObjectInspector(structColNames, Arrays.<ObjectInspector>asList(mapOI, mapOI), (byte) seps[2], serDeParams.getNullSequence(), serDeParams.isLastColumnTakesRest(), serDeParams.isEscaped(), serDeParams.getEscapeChar());
LazySimpleStructObjectInspector structOI = (LazySimpleStructObjectInspector) LazyObjectInspectorFactory.getLazySimpleStructObjectInspector(columns, Arrays.asList(rowStructOI, stringOI), seps[1], serDeParams.getNullSequence(), serDeParams.isLastColumnTakesRest(), serDeParams.isEscaped(), serDeParams.getEscapeChar());
AccumuloRowSerializer serializer = new AccumuloRowSerializer(0, serDeParams, accumuloSerDeParams.getColumnMappings(), new ColumnVisibility(), accumuloSerDeParams.getRowIdFactory());
Map<String, String> map1 = new HashMap<String, String>(), map2 = new HashMap<String, String>();
map1.put("key10", "value10");
map1.put("key11", "value11");
map2.put("key20", "value20");
map2.put("key21", "value21");
ByteArrayRef byteRef = new ByteArrayRef();
// Default separators are 1-indexed (instead of 0-indexed), thus the separator at offset 1 is
// (byte) 2
// The separator for the hive row is \x02, for the row Id struct, \x03, and the maps \x04 and
// \x05
String accumuloRow = "key10\5value10\4key11\5value11\3key20\5value20\4key21\5value21";
LazyStruct entireStruct = (LazyStruct) LazyFactory.createLazyObject(structOI);
byteRef.setData((accumuloRow + "\2foo").getBytes());
entireStruct.init(byteRef, 0, byteRef.getData().length);
Mutation m = serializer.serialize(entireStruct, structOI);
Assert.assertArrayEquals(accumuloRow.getBytes(), m.getRow());
Assert.assertEquals(1, m.getUpdates().size());
ColumnUpdate update = m.getUpdates().get(0);
Assert.assertEquals("cf", new String(update.getColumnFamily()));
Assert.assertEquals("cq", new String(update.getColumnQualifier()));
Assert.assertEquals("foo", new String(update.getValue()));
AccumuloHiveRow haRow = new AccumuloHiveRow(new String(m.getRow()));
haRow.add("cf", "cq", "foo".getBytes());
LazyAccumuloRow lazyAccumuloRow = new LazyAccumuloRow(structOI);
lazyAccumuloRow.init(haRow, accumuloSerDeParams.getColumnMappings(), accumuloSerDeParams.getRowIdFactory());
List<Object> objects = lazyAccumuloRow.getFieldsAsList();
Assert.assertEquals(2, objects.size());
Assert.assertEquals("foo", objects.get(1).toString());
LazyStruct rowStruct = (LazyStruct) objects.get(0);
List<Object> rowObjects = rowStruct.getFieldsAsList();
Assert.assertEquals(2, rowObjects.size());
LazyMap rowMap = (LazyMap) rowObjects.get(0);
Map<?, ?> actualMap = rowMap.getMap();
System.out.println("Actual map 1: " + actualMap);
Map<String, String> actualStringMap = new HashMap<String, String>();
for (Entry<?, ?> entry : actualMap.entrySet()) {
actualStringMap.put(entry.getKey().toString(), entry.getValue().toString());
}
Assert.assertEquals(map1, actualStringMap);
rowMap = (LazyMap) rowObjects.get(1);
actualMap = rowMap.getMap();
System.out.println("Actual map 2: " + actualMap);
actualStringMap = new HashMap<String, String>();
for (Entry<?, ?> entry : actualMap.entrySet()) {
actualStringMap.put(entry.getKey().toString(), entry.getValue().toString());
}
Assert.assertEquals(map2, actualStringMap);
}
use of org.apache.hadoop.hive.serde2.Serializer in project hive by apache.
the class TestAccumuloRowSerializer method testMapSerialization.
@Test
public void testMapSerialization() throws IOException, SerDeException {
List<String> columns = Arrays.asList("row", "col");
List<TypeInfo> types = Arrays.<TypeInfo>asList(TypeInfoFactory.stringTypeInfo, TypeInfoFactory.getMapTypeInfo(TypeInfoFactory.stringTypeInfo, TypeInfoFactory.stringTypeInfo));
List<String> typeNames = new ArrayList<String>(types.size());
for (TypeInfo type : types) {
typeNames.add(type.getTypeName());
}
Properties tableProperties = new Properties();
tableProperties.setProperty(AccumuloSerDeParameters.COLUMN_MAPPINGS, ":rowid,cf:*");
tableProperties.setProperty(serdeConstants.FIELD_DELIM, " ");
tableProperties.setProperty(serdeConstants.COLLECTION_DELIM, ",");
tableProperties.setProperty(serdeConstants.MAPKEY_DELIM, ":");
tableProperties.setProperty(serdeConstants.LIST_COLUMNS, Joiner.on(',').join(columns));
tableProperties.setProperty(serdeConstants.LIST_COLUMN_TYPES, Joiner.on(',').join(typeNames));
AccumuloSerDeParameters accumuloSerDeParams = new AccumuloSerDeParameters(new Configuration(), tableProperties, AccumuloSerDe.class.getSimpleName());
LazySerDeParameters serDeParams = accumuloSerDeParams.getSerDeParameters();
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(columns, Arrays.asList(stringOI, mapOI), (byte) ' ', serDeParams.getNullSequence(), serDeParams.isLastColumnTakesRest(), serDeParams.isEscaped(), serDeParams.getEscapeChar());
AccumuloRowSerializer serializer = new AccumuloRowSerializer(0, serDeParams, accumuloSerDeParams.getColumnMappings(), new ColumnVisibility(), accumuloSerDeParams.getRowIdFactory());
// Create the LazyStruct from the LazyStruct...Inspector
LazyStruct obj = (LazyStruct) LazyFactory.createLazyObject(structOI);
ByteArrayRef byteRef = new ByteArrayRef();
byteRef.setData("row1 cq1:10,cq2:20,cq3:value".getBytes());
obj.init(byteRef, 0, byteRef.getData().length);
Mutation m = (Mutation) serializer.serialize(obj, structOI);
Assert.assertArrayEquals("row1".getBytes(), m.getRow());
List<ColumnUpdate> updates = m.getUpdates();
Assert.assertEquals(3, updates.size());
ColumnUpdate update = updates.get(0);
Assert.assertEquals("cf", new String(update.getColumnFamily()));
Assert.assertEquals("cq1", new String(update.getColumnQualifier()));
Assert.assertEquals("10", new String(update.getValue()));
update = updates.get(1);
Assert.assertEquals("cf", new String(update.getColumnFamily()));
Assert.assertEquals("cq2", new String(update.getColumnQualifier()));
Assert.assertEquals("20", new String(update.getValue()));
update = updates.get(2);
Assert.assertEquals("cf", new String(update.getColumnFamily()));
Assert.assertEquals("cq3", new String(update.getColumnQualifier()));
Assert.assertEquals("value", new String(update.getValue()));
}
Aggregations