WritableFactories是个工厂类,通过它可以创建实例对象。原文中对此对象的注释是Factories for non-public writables,我的理解是处理那些不确定类型的对象进行实例化。例如在ObjectWritable的readObject方法中就调用了WritableFactories
public static Object readObject(DataInput in, ObjectWritable objectWritable, Configuration conf) throws IOException { ....... // Writable Class instanceClass = null; String str = UTF8.readString(in); instanceClass = loadClass(conf, str); Writable writable = WritableFactories.newInstance(instanceClass, conf); writable.readFields(in); instance = writable; ..... return instance; }
WritableFactories提供了注册机制,用户可以对某种类型自定义一个WritableFactory,然后调用WritableFactories的setFactory方法进行注册。WritableFactories.setFactory() 需要两个参数,分别是
注册类对应的类对象和能够构造注册类的WritableFactory 接口的实现
private static final Map<Class, WritableFactory> CLASS_TO_FACTORY = new ConcurrentHashMap<Class, WritableFactory>(); private WritableFactories() {} // singleton /** Define a factory for a class. */ public static void setFactory(Class c, WritableFactory factory) { CLASS_TO_FACTORY.put(c, factory); }
在HDFS的Block类中,就调用了此方法
public class Block implements Writable, Comparable<Block> { public static final String BLOCK_FILE_PREFIX = "blk_"; public static final String METADATA_EXTENSION = ".meta"; static { // register a ctor WritableFactories.setFactory (Block.class, new WritableFactory() { @Override public Writable newInstance() { return new Block(); } }); }
WritableFactories的核心方法当然是创建对接咯:newInstance。它的处理逻辑很简单,首先通过类型找到对应的工厂类,然后用工厂类创建实例对象,如果没有工厂类,则用Java的反射机制创建
/** Create a new instance of a class with a defined factory. */ public static Writable newInstance(Class<? extends Writable> c, Configuration conf) { //根据类型获取对应的工厂实现类 WritableFactory factory = WritableFactories.getFactory(c); if (factory != null) { //通过实现工厂创建实例 Writable result = factory.newInstance(); //如果实例是可配置的,注入Configuration对象 if (result instanceof Configurable) { ((Configurable) result).setConf(conf); } return result; } else { //如果没有对应实现类,则用反射创建实例 return ReflectionUtils.newInstance(c, conf); } }