Hibernate之list和iterator

在Hibernate3中将Session.find()，Session.iterator()换成：session.createQuery().list()，session.createQuery().iterator()，（实现机制并没有什么区别）；

两者都是根据指定条件查询并返回符合查询的条件的实体对象；

看下session.createQuery().list()：

List<TUser> users = session.createQuery(hql).list();
int len = users.size();
for (int i = 0; i < len; i++) {
    TUser user = users.get(i);
    System.out.println("User name:"+user.getName());
}

<strong><span style="font-size:18px;color:#000000;">		List<TUser> users = session.createQuery(hql).list();
		int len = users.size();
		for (int i = 0; i < len; i++) {
			TUser user = users.get(i);
			System.out.println("User name:"+user.getName());
		}</span></strong>

在控制台上只会输出：

Hibernate: select   tuser0_.userid as userid0_, tuser0_.name as name0_,tuser0_.age as age0_, tuser0_.info as info0_  from t_user tuser0_  where   tuser0_.age>316

<strong><span style="font-size:18px;color:#000000;">Hibernate: select   tuser0_.userid as userid0_, tuser0_.name as name0_,tuser0_.age as age0_, tuser0_.info as info0_  from t_user tuser0_  where   tuser0_.age>316  
</span></strong>

显示的结果为：

User name:keith 299

<strong><span style="font-size:18px;color:#000000;">User name:keith 299</span></strong>

 

在看下session.createQuery().iterate()：

Iterator it = session.createQuery(hql).iterate();
while (it.hasNext()) {
    TUser users = (TUser) it.next();
    System.out.println("user Name:"+users.getName());
}

<strong><span style="font-size:18px;color:#000000;">		Iterator it = session.createQuery(hql).iterate();
		while (it.hasNext()) {
			TUser users = (TUser) it.next();
			System.out.println("user Name:"+users.getName());
		}</span></strong>

 看下控制台上的输出：

Hibernate: select   tuser0_.userid as col_0_0_  from  t_user tuser0_ where   tuser0_.age>316
Hibernate: select   tuser0_.userid as userid0_0_,tuser0_.name as name0_0_, tuser0_.age as age0_0_,tuser0_.info as info0_0_   from    t_user tuser0_  where   tuser0_.userid=?

<strong><span style="font-size:18px;color:#000000;">Hibernate: select   tuser0_.userid as col_0_0_  from  t_user tuser0_ where   tuser0_.age>316
Hibernate: select   tuser0_.userid as userid0_0_,tuser0_.name as name0_0_, tuser0_.age as age0_0_,tuser0_.info as info0_0_   from    t_user tuser0_  where   tuser0_.userid=?</span></strong>

 所显示的结果为：

user Name:keith 299

<strong><span style="font-size:18px;color:#000000;">user Name:keith 299</span></strong>

 

两者相比较，list()只发一条语句将符合条件的数据全部查出，而iterator()却现将id查出来，然后根据id再将符合条件的数据查出，这就构成了N＋1的问题；既然list更高效，为什么hibernate还将iterator存在呢？

 

这个问题与Hibernate缓存机制密切相关，我们将两种方法写在一起：

List<TUser> users = session.createQuery(hql).list();
int len = users.size();
for (int i = 0; i < len; i++) {
    TUser user = users.get(i);
    System.out.println("user Name："+user.getName());
}

System.out.println("
 query by list is Over,and query by iterator is start...");

Iterator it = session.createQuery(hql).iterate();

while (it.hasNext()) {
    TUser user = (TUser) it.next();
    System.out.println("users name:"+user.getName());
}

<strong><span style="font-size:18px;color:#000000;">		List<TUser> users = session.createQuery(hql).list();
		int len = users.size();
		for (int i = 0; i < len; i++) {
			TUser user = users.get(i);
			System.out.println("user Name："+user.getName());
		}
		
		System.out.println("
 query by list is Over,and query by iterator is start...");
		
		Iterator it = session.createQuery(hql).iterate();
		
		while (it.hasNext()) {
			TUser user = (TUser) it.next();
			System.out.println("users name:"+user.getName());
		}</span></strong>

 看下控制台的输出：

Hibernate: select  tuser0_.userid as userid0_,tuser0_.name as name0_,tuser0_.age as age0_,tuser0_.info as info0_   from   t_user tuser0_   where   tuser0_.age>316

user Name：keith 299

 query by list is Over,and query by iterator is start...

Hibernate: select  tuser0_.userid as col_0_0_  from  t_user tuser0_ where  tuser0_.age>316

users name:keith 299

<strong><span style="font-size:18px;color:#000000;">Hibernate: select  tuser0_.userid as userid0_,tuser0_.name as name0_,tuser0_.age as age0_,tuser0_.info as info0_   from   t_user tuser0_   where   tuser0_.age>316

user Name：keith 299

 query by list is Over,and query by iterator is start...

Hibernate: select  tuser0_.userid as col_0_0_  from  t_user tuser0_ where  tuser0_.age>316

users name:keith 299</span></strong>

 只是将之前的list()和iterator()放在一起使用，这时的iterator只执行了一条SQL；其中的差异在于hibernate的缓存机制；

list()方法将执行Select SQL从数据库中获取所有符合满足条件的记录并构造相应的实体对象，实体对象构建完毕后，就将其纳入缓存；

这样等到iterator()执行时，首先会执行一条SQL来查询符合条件数据的id，随即，iterator方法首先在本地缓存内根据id查找对应的实体对象是否存在，如果缓存中已经存在对应的数据，则直接以此数据对象作为查询结果；如果没有找到，则再次执行Select语句获得对应数据库中的表记录（如果iterator在数据库中查到并构建了完整的数据对象，也会将其纳入缓存中）;

在上面这个实例中，list()将读取的数据放入缓存中，iterator()直接可以用于是出现了以上的结果；

 

当我们再次执行上面的程序，发现结果还是一样的；list()并没有去读取它自己放进缓存的数据，因为当我们list()查询完后，即使缓存中有一些符合条件的数据，但是我们也无法保证这些数据就是合法的，如果更改下条件，将大于316改成300；那么缓存中的数据只满足大于316不会满足大于300；所有list()方法还是需要执行一次Select SQL来保证结果的完整性；得出结论：list()实际上无法利用缓存，它对缓存只写不读，而iterator()可以充分利用缓存，如果目标数据只读或者读取相当频繁，可以使用iterator()来减少性能上的消耗；

 

但是会出现这样一个问题，如果是海量数据怎么办？如果超过10万条记录的话，会有很大的可能出发OutOfMemoryError；导致系统异常；那么就用iterator()和evict()来解决吧！将内存消耗保持在可以接受的范围内；比如：

Iterator it = session.createQuery(hql).iterate();
while (it.hasNext()) {
    TUser users = (TUser) it.next();
    //在一级缓存中清理
    session.evict(users);
    //在二级缓存中清理，二级缓存可以设定
    SessionFactory.evict(TUser.class, users.getUserid());
    System.out.println("user Name:"+users.getName());
}

<strong><span style="font-size:18px;color:#000000;">		Iterator it = session.createQuery(hql).iterate();
		while (it.hasNext()) {
			TUser users = (TUser) it.next();
			//在一级缓存中清理
			session.evict(users);
			//在二级缓存中清理，二级缓存可以设定
			SessionFactory.evict(TUser.class, users.getUserid());
			System.out.println("user Name:"+users.getName());
		}</span></strong>

 但是这样的话，由于JVM的异步内存回收机制，无效对象会不断的在内存中积累等待回收，如果数据量很大，必然会频繁的激发JVM的内存回收机制，导致系统性能急剧下降。因此对于这样的问题，最好使用SQL或者存储过程！