Hadoop 序列化

Hadoop提供了org.apache.hadoop.io.serializer.Serialization接口用于序列化的抽象

public interface Serialization<T> {
  
  boolean accept(Class<?> c);// 判断是否支持指定类的序列化
 
  Serializer<T> getSerializer(Class<T> c);// 返回序列化器

  Deserializer<T> getDeserializer(Class<T> c); }// 返回返序列化器

Hadoop默认提供了JavaSerialization、WritableSerialization以及Avro的实现，这三个实现类的accept的实现方式一致，都是根据对象是否有指定的注解来判断是否支持序列化与反序列化

  public boolean accept(Class<?> c) {
    return Serializable.class.isAssignableFrom(c);
  }

Serializer和Deserializer分别实现了open(用于准备资源)、close(用于关闭资源)和序列化/反序列化接口

public interface Serializer<T> {
  void open(OutputStream out) throws IOException;
  void serialize(T t) throws IOException;
  void close() throws IOException;
}

public interface Deserializer<T> {
  void open(InputStream in) throws IOException;
  T deserialize(T t) throws IOException;
  void close() throws IOException;
}

Hadoop实现了SerializationFactory工厂用于生产具体的实现类

  public SerializationFactory(Configuration conf) {
    super(conf);
    //根据配置io.serializations创建序列化实现类，如果没有配置，则为Hadoop默认配置的三个实现
    for (String serializerName : conf.getTrimmedStrings(
            CommonConfigurationKeys.IO_SERIALIZATIONS_KEY,
            new String[]{WritableSerialization.class.getName(),
                    AvroSpecificSerialization.class.getName(),
                    AvroReflectSerialization.class.getName()})) {
      add(conf, serializerName);
    }
  }

  // 通过反射创建具体的实现类，并配置相关参数private void add(Configuration conf, String serializationName) {
    try {
      Class<? extends Serialization> serializionClass =
        (Class<? extends Serialization>) conf.getClassByName(serializationName);
      serializations.add((Serialization)
      ReflectionUtils.newInstance(serializionClass, getConf()));
    } catch (ClassNotFoundException e) {
      LOG.warn("Serialization class not found: ", e);
    }
  }