Child1.java extends Root{String a;}
Child2.java extends Roor{String b;}
1.直接映射，2個類Child1，Child2分別映射2個表
<hibernate-mapping>
<class name="Root" table="CHILD1">
<id name="id" column="CHILD1_ID" type="long">
<generator class="native">
<param name="sequence">root_sequence</param>
<param name="parameters">START WITH 1000</param>
</generator>
</id>
<version name="version" column="VERSION" type="long"/>
<property name="code" type="string">
<column name="CODE" not-null="true" length="100" unique-key="UK_ROOT"/>
</property>
<property name="a" column="A" type="string"></property>
</class>
</hibernate-mapping>
<hibernate-mapping>
<class name="Root" table="CHILD2">
<id name="id" column="CHILD2_ID" type="long">
<generator class="native">
<param name="sequence">root_sequence</param>
<param name="parameters">START WITH 1000</param>
</generator>
</id>
<version name="version" column="VERSION" type="long"/>
<property name="code" type="string">
<column name="CODE" not-null="true" length="100" unique-key="UK_ROOT"/>
</property>
<property name="a" column="A" type="string"></property>
</class>
</hibernate-mapping>
這是最簡單的映射策略，隱式多態(tài)，查詢時面向?qū)ο?，查詢Root，會把其子類Child1，Child2的表都搜索
2。有joined-subclass，將公共字段放在一張表root上，child1表上存id，a；child2表存id，b；其中id字段與root上的字段關(guān)聯(lián)；共3張表
<hibernate-mapping>
<class name="Root" table="ROOT">
<id name="id" column="ROOT_ID" type="long">
<generator class="native">
<param name="sequence">root_sequence</param>
<param name="parameters">START WITH 1000</param>
</generator>
</id>
<version name="version" column="VERSION" type="long"/>
<property name="code" type="string">
<column name="CODE" not-null="true" length="100" unique-key="UK_ROOT"/>
</property>
<joined-subclass name="Child1" table="CHILD1">
<key column="id"></key>
<property name="a" type="string">
<column name="A" not-null="true" unique-key="UK_CHILD1"/>
</property>
</joined-subclass>
<joined-subclass name="Child2" table="CHILD2">
<key column="id"></key>
<property name="b" type="string">
<column name="B" not-null="true" unique-key="UK_CHILD2"/>
</property>
</joined-subclass>
</class>
</hibernate-mapping>
其中joined-subclass下的key表示外鍵，查詢Child1的時候，會將root表與child1聯(lián)合，效率也不高
3.采用discriminator，只有一張表
<hibernate-mapping>
<class name="Root" table="ROOT">
<id name="id" column="ROOT_ID" type="long">
<generator class="native">
<param name="sequence">root_sequence</param>
<param name="parameters">START WITH 1000</param>
</generator>
</id>
<discriminator column="TYPE" type="string"></discriminator>
<version name="version" column="VERSION" type="long"/>
<property name="code" type="string">
<column name="CODE" not-null="true" length="100" unique-key="UK_ROOT"/>
</property>
<subclass name="Child1" discriminator-value="1">
<property name="a" column="A" type="string"></property>
</subclass>
<subclass name="Child2" discriminator-value="2">
<property name="b" column="B" type="string"></property>
</subclass>
</class>
</hibernate-mapping>
注意的是discriminator 要在id和version之間，而且hbm文件都必須有dtd。這樣的查詢效率高多了，查詢child1的時候會自動根據(jù)條件type＝1
