Hibernate的检索策略

一， Hibernate的检索策略包括类级别检索策略和关联级别检索策略。Hibernate的检索策略包括类级别检索策略和关联级别检索策略。

二， 类级别检索策略有立即检索和延迟检索，默认的检索策略是立即检索。

三， 立即检索

在映射文件中设置<hibernate-mapping
package="cn.csdn.producets.domain" default-lazy="false">

<classname="Customers" table="Customers" lazy=”false”>

<setname="orders" lazy="false">

四， 延迟检索

设置lazy=”false”不采用延迟检索

<set name="orders" lazy=”false”>

Customers c=(Customers) session.get(Customers.class, 2);
System.out.println(c.getEmail());

运行结果：Hibernate: selectcustomers0_.id as id0_0_, customers0_.realName as realName0_0_, customers0_.passas pass0_0_, customers0_.sex as sex0_0_, customers0_.petName as petName0_0_,customers0_.email as email0_0_, customers0_.rdate as rdate0_0_
from Customerscustomers0_ where customers0_.id=?
Hibernate: select orders0_.cid as cid0_1_,orders0_.id as id1_, orders0_.id as id1_0_, orders0_.number as number1_0_,orders0_.address as address1_0_, orders0_.phone as phone1_0_, orders0_.odate asodate1_0_, orders0_.sum as sum1_0_, orders0_.status
as status1_0_, orders0_.cidas cid1_0_ from orders orders0_ where orders0_.cid=?

设置lazy=”true”采用延迟加载

<set name="orders" lazy=”false”>

Customers c=(Customers) session.get(Customers.class, 2);
System.out.println(c.getEmail());

运行结果：Hibernate: select customers0_.id as id0_0_,customers0_.realName as realName0_0_, customers0_.pass as pass0_0_,customers0_.sex as sex0_0_, customers0_.petName as petName0_0_,customers0_.email as email0_0_, customers0_.rdate as rdate0_0_ from Customerscustomers0_
where customers0_.id=?

总结：采用懒加载是会少输出一条sql查询语句；

五，关联级别检索策略有立即检索、延迟检索和迫切左外连接检索。

Fech=”join”;

第一种测试方式：<set name="orders" fetch="join" lazy=”true”>

Customers c=(Customers) session.get(Customers.class, 2);
System.out.println(c.getEmail());

代码执行结果：

Hibernate: select customers0_.id as id0_1_,customers0_.realName as realName0_1_, customers0_.pass as pass0_1_,customers0_.sex as sex0_1_, customers0_.petName as petName0_1_,customers0_.email as email0_1_, customers0_.rdate as rdate0_1_, orders1_.cid ascid0_3_,
orders1_.id as id3_, orders1_.id as id1_0_, orders1_.number asnumber1_0_, orders1_.address as address1_0_, orders1_.phone as phone1_0_,orders1_.odate as odate1_0_, orders1_.sum as sum1_0_, orders1_.status asstatus1_0_, orders1_.cid as cid1_0_ from Customers
customers0_ left outer joinorders orders1_ on customers0_.id=orders1_.cid where customers0_.id=?

可以从运行结果中得知：采用迫切左外链接

第二种测试方式：<set name="orders" fetch="join" lazy=”false”>

Customers c=(Customers) session.get(Customers.class, 2);
System.out.println(c.getEmail());

代码执行结果：

Hibernate: select customers0_.id as id0_1_, customers0_.realNameas realName0_1_, customers0_.pass as pass0_1_, customers0_.sex as sex0_1_,customers0_.petName as petName0_1_, customers0_.email as email0_1_,customers0_.rdate as rdate0_1_, orders1_.cid as cid0_3_,
orders1_.id as id3_,orders1_.id as id1_0_, orders1_.number as number1_0_, orders1_.address asaddress1_0_, orders1_.phone as phone1_0_, orders1_.odate as odate1_0_,orders1_.sum as sum1_0_, orders1_.status as status1_0_, orders1_.cid as cid1_0_from Customers
customers0_ left outer join orders orders1_ oncustomers0_.id=orders1_.cid where customers0_.id=?

可以从运行结果中得知：采用迫切左外链接

第三种测试方式：<set name="orders" fetch="join" lazy=”extra”>

Customers c=(Customers) session.get(Customers.class, 2);
System.out.println(c.getEmail());

代码执行结果：

Hibernate: select customers0_.id as id0_1_,customers0_.realName as realName0_1_, customers0_.pass as pass0_1_,customers0_.sex as sex0_1_, customers0_.petName as petName0_1_,customers0_.email as email0_1_, customers0_.rdate as rdate0_1_, orders1_.cid ascid0_3_,
orders1_.id as id3_, orders1_.id as id1_0_, orders1_.number asnumber1_0_, orders1_.address as address1_0_, orders1_.phone as phone1_0_,orders1_.odate as odate1_0_, orders1_.sum as sum1_0_, orders1_.status asstatus1_0_, orders1_.cid as cid1_0_ from Customers
customers0_ left outer joinorders orders1_ on customers0_.id=orders1_.cid where customers0_.id=?

可以从运行结果中得知：采用迫切左外链接

总结：若把fech属性设置为join则lazy属性就不起作用

二．Fech=”select”

第一种测试方式：<set name="orders" fetch="select" lazy=”true”>

Customersc=(Customers) session.get(Customers.class, 2);
System.out.println(c.getEmail());
System.out.println(c.getOrders().size());

运行结果：Hibernate: select customers0_.id as id0_0_, customers0_.realName asrealName0_0_, customers0_.pass as pass0_0_, customers0_.sex as sex0_0_, customers0_.petNameas petName0_0_, customers0_.email as email0_0_, customers0_.rdate as rdate0_0_from
Customers customers0_ where customers0_.id=?
null
Hibernate:select orders0_.cid as cid0_1_, orders0_.id as id1_, orders0_.id as id1_0_,orders0_.number as number1_0_, orders0_.address as address1_0_, orders0_.phoneas phone1_0_, orders0_.odate as odate1_0_, orders0_.sum as sum1_0_,orders0_.status
as status1_0_, orders0_.cid as cid1_0_ from orders orders0_where orders0_.cid=?
1结论：采用延迟加载；

第二种测试方式：<set name="orders" fetch="select" lazy=”false”>

Customers c=(Customers) session.get(Customers.class, 2);

运行结果：Hibernate: selectcustomers0_.id as id0_0_, customers0_.realName as realName0_0_,customers0_.pass as pass0_0_, customers0_.sex as sex0_0_, customers0_.petNameas petName0_0_, customers0_.email as email0_0_, customers0_.rdate as rdate0_0_from
Customers customers0_ where customers0_.id=?
Hibernate:select orders0_.cid as cid0_1_, orders0_.id as id1_, orders0_.id as id1_0_,orders0_.number as number1_0_, orders0_.address as address1_0_, orders0_.phoneas phone1_0_, orders0_.odate as odate1_0_, orders0_.sum as sum1_0_,orders0_.status
as status1_0_, orders0_.cid as cid1_0_ from orders orders0_where orders0_.cid=?
结论：采用立即检索

第三种测试方式：<setname="orders" fetch="select"lazy=”extra”>

Customersc=(Customers) session.get(Customers.class, 2);
System.out.println(c.getEmail());
System.out.println(c.getOrders().size());
运行结果：Hibernate: selectcustomers0_.id as id0_0_, customers0_.realName as realName0_0_,customers0_.pass as pass0_0_, customers0_.sex as sex0_0_, customers0_.petNameas petName0_0_, customers0_.email as email0_0_, customers0_.rdate as rdate0_0_from
Customers customers0_ where customers0_.id=?
null
Hibernate:select count(id) from orders where cid =?
1
结论：采用延迟加载
总结:听从lazy属性的

三．Fech=”subselect”;

第一种测试方式：<setname="orders" fetch="subselect"lazy=”true”>

Customers c=(Customers)session.get(Customers.class, 2);
System.out.println(c.getEmail());
System.out.println(c.getOrders().size());
Set<Orders>orders=c.getOrders();
for(Orders o:orders){
System.out.println(o.getNumber());
}
session.close();
}

运行结果：Hibernate: select customers0_.id as id0_0_, customers0_.realName asrealName0_0_, customers0_.pass as pass0_0_, customers0_.sex as sex0_0_,customers0_.petName as petName0_0_, customers0_.email as email0_0_,customers0_.rdate as rdate0_0_
from Customers customers0_ wherecustomers0_.id=?
null
Hibernate: select orders0_.cid as cid0_1_,orders0_.id as id1_, orders0_.id as id1_0_, orders0_.number as number1_0_,orders0_.address as address1_0_, orders0_.phone as phone1_0_, orders0_.odate asodate1_0_, orders0_.sum as sum1_0_, orders0_.status
as status1_0_, orders0_.cidas cid1_0_ from orders orders0_ where orders0_.cid=?
1
2

结论：采用延迟加载；

第二种测试方式：<setname="orders" fetch="subselect"lazy=”false”>

Customers c=(Customers)session.get(Customers.class, 2);
System.out.println(c.getEmail());
System.out.println(c.getOrders().size());
Set<Orders>orders=c.getOrders();
for(Orders o:orders){
System.out.println(o.getNumber());
}
session.close();
}

运行结果：Hibernate: select customers0_.id as id0_0_, customers0_.realName asrealName0_0_, customers0_.pass as pass0_0_, customers0_.sex as sex0_0_,customers0_.petName as petName0_0_, customers0_.email as email0_0_,customers0_.rdate as rdate0_0_
from Customers customers0_ wherecustomers0_.id=?
Hibernate: select orders0_.cid as cid0_1_,orders0_.id as id1_, orders0_.id as id1_0_, orders0_.number as number1_0_,orders0_.address as address1_0_, orders0_.phone as phone1_0_, orders0_.odate asodate1_0_, orders0_.sum as sum1_0_, orders0_.status
as status1_0_, orders0_.cidas cid1_0_ from orders orders0_ where orders0_.cid=?
null
1
2
结论：采用立即检索

第三种测试方式：<setname="orders" fetch="subselect"lazy=”extra”>

Customers c=(Customers)session.get(Customers.class, 2);
System.out.println(c.getEmail());
System.out.println(c.getOrders().size());
Set<Orders>orders=c.getOrders();
for(Orders o:orders){
System.out.println(o.getNumber());
}
session.close();
}

运行结果：Hibernate: select customers0_.id as id0_0_, customers0_.realName asrealName0_0_, customers0_.pass as pass0_0_, customers0_.sex as sex0_0_,customers0_.petName as petName0_0_, customers0_.email as email0_0_,customers0_.rdate as rdate0_0_
from Customers customers0_ wherecustomers0_.id=?
null
Hibernate: select count(id) from orders wherecid =?
1
Hibernate: select orders0_.cid as cid0_1_,orders0_.id as id1_, orders0_.id as id1_0_, orders0_.number as number1_0_,orders0_.address as address1_0_, orders0_.phone as phone1_0_, orders0_.odate asodate1_0_, orders0_.sum as sum1_0_, orders0_.status
as status1_0_, orders0_.cidas cid1_0_ from orders orders0_ where orders0_.cid=?
2

结论：采用延迟检索

总结; Lazy属性决定检索策略

对于检索策略，需要根据实际情况进行选择。对于立即检索和延迟检索，它们的优点在于select语句简单（每张表一条语句）、查询速度快，缺点在于关联表时需要多条select语句，增加了访问数据库的频率。因此在选择即检索和延迟检索时，可以考虑使用批量检索策略来减少select语句的数量（配置batch-size属性）。对于切左外连接检索，优点在于select较少，但缺点是select语句的复杂度提高，多表之间的关联会是很耗时的操作。另外，配置文件是死的，但程序是活的，可以根据需要在程序里显示的指定检索策略（可能经常需要在程序中显示指定迫切左外连接检索）。为了清楚检索策略的配置效果如何，可以配置show_sql属性查看程序运行时Hibernate执行的sql语句。