Hadoop Serialize (一)

说hadoop的序列化与反序列化，怎么都离不开java的序列化。

        Block block = new Block();
        block.setId(0000001l);
        block.setName("block one");
        block.setPosition("offset 1");
        block.setSize(10000l);

        ObjectOutputStream outputStream = new ObjectOutputStream(System.out);
        outputStream.writeObject(block);

Block 如下：

public class Block implements Serializable, Comparable<Block> {

java对象的序列化是将对象转换未连续的 字节流，便于存储或者网络传输。如代码一，创建ObjectOutputStream对象，通过调用writeObject()方法就可达到目的。如果关注过RMI的同学可能注意过，当RMI调用发生异常时，异常信息通常是有ObjectOutputStream的信息的。

我觉得这里有必要贴一下 writeObject()方法的注释。

   /**
     * Write the specified object to the ObjectOutputStream.  The class of the
     * object, the signature of the class, and the values of the non-transient
     * and non-static fields of the class and all of its supertypes are
     * written.  Default serialization for a class can be overridden using the
     * writeObject and the readObject methods.  Objects referenced by this
     * object are written transitively so that a complete equivalent graph of
     * objects can be reconstructed by an ObjectInputStream.
     *
     * <p>Exceptions are thrown for problems with the OutputStream and for
     * classes that should not be serialized.  All exceptions are fatal to the
     * OutputStream, which is left in an indeterminate state, and it is up to
     * the caller to ignore or recover the stream state.
     *
     * @throws  InvalidClassException Something is wrong with a class used by
     *          serialization.
     * @throws  NotSerializableException Some object to be serialized does not
     *          implement the java.io.Serializable interface.
     * @throws  IOException Any exception thrown by the underlying
     *          OutputStream.
     */
    public final void writeObject(Object obj) throws IOException {

writeObject()会将 类的non-transient 和 non-static 字段的值以及它所有的父类都要被写入。序列化机制会自动访问对象的父类，以保证对象内容的一致性。同时，序列化机制不仅存储对象在内存中的原始数据，还会追踪通过该对象可以到达的其他对象的内部数据，并描述所有这些对象是如何被链接起来的。

java序列化机制中，反序列化的过程会不断的创建新的对象，这增加了 java对象的分配和回收，效率有些低。Hadoop中，反序列化可以复用对象，可以在同一个对象上得到多个反序列化结果。