本文介绍了Hibernate的四种抓取策略：连接抓取、查询抓取、子查询抓取和批量抓取，详细解析了每种策略的工作原理，并通过Java项目案例展示了如何在实践中应用这些策略。连接抓取通过外连接获取关联对象，查询抓取和子查询抓取在访问关联关系时执行额外的select语句，批量抓取则是一种优化方案，通过单条select语句获取一批对象。
抓取策略（fetching strategy）是指：当应用程序需要在（Hibernate实体对象图的）关联关系间进行导航的时候，Hibernate如何获取关联对象的策略。抓取策略可以在O/R映射的元数据中声明，也可以在特定的HQL或条件查询（Criteria Query）中重载声明。
Hibernate定义了如下几种抓取策略：
1、连接抓取（Join fetching）--Hibernate通过在select语句中使用outer join（外链接）来获得对象的关联实例或者关联集合。这时lazy无效。
2、查询抓取（select fetching）--另外发送一条select语句抓取当前对象的关联实体或集合。除非你显示的指定lazy=false禁止延迟抓取，否则只有当你真正访问关联关系的时候，才会执行第二条select语句。
3、子查询抓取（Subselect fetching）--另外发送一条select语句抓取在前面查询到（或者抓取到）的所有实体对象的关联集合。除非你显示的指定lazy=false禁止延迟抓取，否则只有当你真正访问关联关系的时候，才会执行第二条select语句。
4、批量抓取（Batch fetching）--对查询抓取的优化方案，通过指定一个主键或外键列表，Hibernate使用单条select语句获取一批对象实例或集合。
下面我们通过项目案例来简单了解下这几种策略
java项目结构如下：
该项目所有代码跟上一节《hibernate加载策略之lazy》一样，可以参考。
查询抓取（select fetching）
在多的一端设置fetch
Book.hbm.xml配置文件
<hibernate-mapping package="com.test.pojo">
<class name="Book" table="book">
<id name="id">
<generator class="identity" />
</id>
<many-to-one name="category" class="Category" column="category_id"
cascade="save-update" lazy="proxy" fetch="select"/>
<property name="author" />
<property name="name" column="book_name" />
<property name="price" />
<property name="pubDate" />
</class>
</hibernate-mapping>在HibernateTest类中添加代码
@Test
public void testLoad1(){
Session session=HibernateUtil.getSession();
Transaction tx=session.beginTransaction();
Book book=(Book) session.get(Book.class, 1);
System.out.println(book.getName());
System.out.println(book.getCategory().getName());
tx.commit();
HibernateUtil.closeSession();
}断点执行，当执行到
System.out.println(book.getName());
Hibernate:
select
book0_.id as id1_1_0_,
book0_.category_id as category2_1_0_,
book0_.author as author3_1_0_,
book0_.book_name as book_nam4_1_0_,
book0_.price as price5_1_0_,
book0_.pubDate as pubDate6_1_0_
from
book book0_
where
book0_.id=?当执行到
System.out.println(book.getCategory().getName());
Hibernate:
select
category0_.id as id1_0_0_,
category0_.name as name2_0_0_
from
Category category0_
where
category0_.id=?
<many-to-one name="category" class="Category" column="category_id"
cascade="save-update" lazy="false" fetch="select"/>
Book book=(Book) session.get(Book.class, 1);
Hibernate:
select
book0_.id as id1_1_0_,
book0_.category_id as category2_1_0_,
book0_.author as author3_1_0_,
book0_.book_name as book_nam4_1_0_,
book0_.price as price5_1_0_,
book0_.pubDate as pubDate6_1_0_
from
book book0_
where
book0_.id=?
Hibernate:
select
category0_.id as id1_0_0_,
category0_.name as name2_0_0_
from
Category category0_
where
category0_.id=?下面我们在一的一端设置fetch
Category.hbm.xml代码
<hibernate-mapping package="com.test.pojo">
<class name="Category" >
<id name="id" >
<generator class="identity" />
</id>
<property name="name" />
<set name="books" inverse="true" lazy="extra" fetch="select">
<key>
<column name="category_id" />
</key>
<one-to-many class="Book" />
</set>
</class>
</hibernate-mapping>断点执行HibernateTest类中的testGet方法
@Test
public void testGet(){
Session session=HibernateUtil.getSession();
Transaction tx=session.beginTransaction();
Category category=(Category) session.get(Category.class, 1);
System.out.println("分类名:"+category.getName());
for(Iterator<Book> iter=category.getBooks().iterator();iter.hasNext();){
System.out.println(iter.next().getName());
}
tx.commit();
HibernateUtil.closeSession();
}断点执行到
Category category=(Category) session.get(Category.class, 1);
Hibernate:
select
category0_.id as id1_0_0_,
category0_.name as name2_0_0_
from
Category category0_
where
category0_.id=?
System.out.println(iter.next().getName());
Hibernate:
select
books0_.category_id as category2_0_0_,
books0_.id as id1_1_0_,
books0_.id as id1_1_1_,
books0_.category_id as category2_1_1_,
books0_.author as author3_1_1_,
books0_.book_name as book_nam4_1_1_,
books0_.price as price5_1_1_,
books0_.pubDate as pubDate6_1_1_
from
book books0_
where
books0_.category_id=?连接抓取（Join fetching）
先看在多的一端设置fetch
将Book.hbm.xml中的fetch设为join，lazy设为proxy，如下
<many-to-one name="category" class="Category" column="category_id"
cascade="save-update" lazy="proxy" fetch="join"/>
Hibernate:
select
book0_.id as id1_1_0_,
book0_.category_id as category2_1_0_,
book0_.author as author3_1_0_,
book0_.book_name as book_nam4_1_0_,
book0_.price as price5_1_0_,
book0_.pubDate as pubDate6_1_0_,
category1_.id as id1_0_1_,
category1_.name as name2_0_1_
from
book book0_
left outer join
Category category1_
on book0_.category_id=category1_.id
where
book0_.id=?
读者
文学
再来从一的一端来查询
将Category.hbm.xml中的fetch设为join，lazy默认为true
<set name="books" inverse="true" fetch="join">
@Test
public void testLoad(){
Session session=HibernateUtil.getSession();
Transaction tx=session.beginTransaction();
Category category=(Category) session.get(Category.class, 1);
System.out.println("分类名:"+category.getName());
System.out.println("对应的书大小："+category.getBooks().size());
tx.commit();
HibernateUtil.closeSession();
}
Hibernate:
select
category0_.id as id1_0_0_,
category0_.name as name2_0_0_,
books1_.category_id as category2_0_1_,
books1_.id as id1_1_1_,
books1_.id as id1_1_2_,
books1_.category_id as category2_1_2_,
books1_.author as author3_1_2_,
books1_.book_name as book_nam4_1_2_,
books1_.price as price5_1_2_,
books1_.pubDate as pubDate6_1_2_
from
Category category0_
left outer join
book books1_
on category0_.id=books1_.category_id
where
category0_.id=?
分类名:文学
对应的书大小：1从一的一端也是使用left outer join方法把book表中的数据一起查询出来了。
接下来我们把Book.hbm.xml和Category.hbm.xml中的lazy和fetch都去掉，执行testLoad2，代码如下
@Test
public void testLoad2(){
Session session=HibernateUtil.getSession();
Transaction tx=session.beginTransaction();
List<Category> list=session.createCriteria(Category.class).list();
System.out.println("类型个数："+list.size());
for(Category cate:list){
System.out.println(cate.getName()+"---书籍本数："+cate.getBooks().size());
}
tx.commit();
HibernateUtil.closeSession();
}
Hibernate:
select
this_.id as id1_0_0_,
this_.name as name2_0_0_
from
Category this_
类型个数：6
Hibernate:
select
books0_.category_id as category2_0_0_,
books0_.id as id1_1_0_,
books0_.id as id1_1_1_,
books0_.category_id as category2_1_1_,
books0_.author as author3_1_1_,
books0_.book_name as book_nam4_1_1_,
books0_.price as price5_1_1_,
books0_.pubDate as pubDate6_1_1_
from
book books0_
where
books0_.category_id=?
文学---书籍本数：1
Hibernate:
select
books0_.category_id as category2_0_0_,
books0_.id as id1_1_0_,
books0_.id as id1_1_1_,
books0_.category_id as category2_1_1_,
books0_.author as author3_1_1_,
books0_.book_name as book_nam4_1_1_,
books0_.price as price5_1_1_,
books0_.pubDate as pubDate6_1_1_
from
book books0_
where
books0_.category_id=?
科幻---书籍本数：1
Hibernate:
select
books0_.category_id as category2_0_0_,
books0_.id as id1_1_0_,
books0_.id as id1_1_1_,
books0_.category_id as category2_1_1_,
books0_.author as author3_1_1_,
books0_.book_name as book_nam4_1_1_,
books0_.price as price5_1_1_,
books0_.pubDate as pubDate6_1_1_
from
book books0_
where
books0_.category_id=?
言情---书籍本数：1
Hibernate:
select
books0_.category_id as category2_0_0_,
books0_.id as id1_1_0_,
books0_.id as id1_1_1_,
books0_.category_id as category2_1_1_,
books0_.author as author3_1_1_,
books0_.book_name as book_nam4_1_1_,
books0_.price as price5_1_1_,
books0_.pubDate as pubDate6_1_1_
from
book books0_
where
books0_.category_id=?
历史---书籍本数：1
Hibernate:
select
books0_.category_id as category2_0_0_,
books0_.id as id1_1_0_,
books0_.id as id1_1_1_,
books0_.category_id as category2_1_1_,
books0_.author as author3_1_1_,
books0_.book_name as book_nam4_1_1_,
books0_.price as price5_1_1_,
books0_.pubDate as pubDate6_1_1_
from
book books0_
where
books0_.category_id=?
漫画---书籍本数：1
Hibernate:
select
books0_.category_id as category2_0_0_,
books0_.id as id1_1_0_,
books0_.id as id1_1_1_,
books0_.category_id as category2_1_1_,
books0_.author as author3_1_1_,
books0_.book_name as book_nam4_1_1_,
books0_.price as price5_1_1_,
books0_.pubDate as pubDate6_1_1_
from
book books0_
where
books0_.category_id=?
悬疑---书籍本数：1子查询抓取（Subselect fetching）
我们在Category.hbm.xml中加上fetch=subselect
<set name="books" inverse="true" fetch="subselect">
Hibernate:
select
this_.id as id1_0_0_,
this_.name as name2_0_0_
from
Category this_
类型个数：6
Hibernate:
select
books0_.category_id as category2_0_1_,
books0_.id as id1_1_1_,
books0_.id as id1_1_0_,
books0_.category_id as category2_1_0_,
books0_.author as author3_1_0_,
books0_.book_name as book_nam4_1_0_,
books0_.price as price5_1_0_,
books0_.pubDate as pubDate6_1_0_
from
book books0_
where
books0_.category_id in (
select
this_.id
from
Category this_
)
文学---书籍本数：1
科幻---书籍本数：1
言情---书籍本数：1
历史---书籍本数：1
漫画---书籍本数：1
悬疑---书籍本数：1一次性将所有的分类Category的数据全部查询出来。
假设现在我想查询Category的id是1,3,5的数据，那么方法testLoad2的代码如下
@Test
public void testLoad2(){
Session session=HibernateUtil.getSession();
Transaction tx=session.beginTransaction();
List<Category> list=session.createCriteria(Category.class)
.add(Restrictions.in("id", new Integer[]{1,3,5}))
.list();
System.out.println("类型个数："+list.size());
for(Category cate:list){
System.out.println(cate.getName()+"---书籍本数："+cate.getBooks().size());
}
tx.commit();
HibernateUtil.closeSession();
}
Hibernate:
select
this_.id as id1_0_0_,
this_.name as name2_0_0_
from
Category this_
where
this_.id in (
?, ?, ?
)
类型个数：3
Hibernate:
select
books0_.category_id as category2_0_1_,
books0_.id as id1_1_1_,
books0_.id as id1_1_0_,
books0_.category_id as category2_1_0_,
books0_.author as author3_1_0_,
books0_.book_name as book_nam4_1_0_,
books0_.price as price5_1_0_,
books0_.pubDate as pubDate6_1_0_
from
book books0_
where
books0_.category_id in (
select
this_.id
from
Category this_
where
this_.id in (
?, ?, ?
)
)
文学---书籍本数：1
言情---书籍本数：1
漫画---书籍本数：1
批量抓取（Batch fetching）
数据量大的时候采用此方法有意义，数据量小时，采用此方法效果不明显。
扫一扫
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
