从源码分析 Spring 基于注解的事务

在spring引入基于注解的事务(@Transactional)之前，我们一般都是如下这样进行拦截事务的配置：

    <!-- 拦截器方式配置事务 -->
    <tx:advice id="transactionAdvice" transaction-manager="transactionManager">
        <tx:attributes>
            <tx:method name="add*" propagation="REQUIRED" />
            <tx:method name="append*" propagation="REQUIRED" />
            <tx:method name="insert*" propagation="REQUIRED" />
            <tx:method name="save*" propagation="REQUIRED" />
            <tx:method name="update*" propagation="REQUIRED" />
            <tx:method name="modify*" propagation="REQUIRED" />
            <tx:method name="edit*" propagation="REQUIRED" />
            <tx:method name="delete*" propagation="REQUIRED" />
            <tx:method name="remove*" propagation="REQUIRED" />
            <tx:method name="repair" propagation="REQUIRED" />
            <tx:method name="delAndRepair" propagation="REQUIRED" />

            <tx:method name="get*" propagation="SUPPORTS" />
            <tx:method name="find*" propagation="SUPPORTS" />
            <tx:method name="load*" propagation="SUPPORTS" />
            <tx:method name="search*" propagation="SUPPORTS" />
            <tx:method name="datagrid*" propagation="SUPPORTS" />

            <tx:method name="*" propagation="SUPPORTS" />
        </tx:attributes>
    </tx:advice>
    <aop:config>
        <aop:pointcut id="transactionPointcut" expression="execution(* net.aazj.service..*Impl.*(..))" />
        <aop:advisor pointcut-ref="transactionPointcut" advice-ref="transactionAdvice" />
    </aop:config>

这种方式明显的缺点是，不太容易理解，并且限定了service层的方法名称的前缀，没有模板的话写起来也很难，很容易写错。

因此在spring中引入了基于注解的事务配置方法之后，我们应该抛弃这种事务配置方法了。基于注解 @Transactional 的事务配置具有简单，灵活的优点。下面看一个例子：

@Service("userService")
@Transactional
public class UserServiceImpl implements UserService{
    @Autowired
    private UserMapper userMapper;

    @Transactional (propagation=Propagation.REQUIRED,isolation=Isolation.DEFAULT,readOnly=true)
    public User getUser(int userId) {
        return userMapper.getUser(userId);
    }

    @Transactional
    public void addUser(String username){
        userMapper.addUser(username);
//      int i = 1/0;    // 测试事务的回滚
    }

    @Transactional (rollbackFor = BaseBusinessException.class)
    public void addAndDeleteUser(String username, int id) throws BaseBusinessException{
        userMapper.addUser(username);
        this.m1();
        userMapper.deleteUserById(id);
    }

    private void m1() throws BaseBusinessException {
        throw new BaseBusinessException("xxx");
    }
}

首先在service类上声明了@Transactional，表明类中的所有方法都需要运行在事务中，然后在方法中可以指定具体的事务特性，方法中的@Transactional会覆盖类上的@Transactional。

下面我们从源码的角度(从源码的学习可以给我们实打实的比较深入理解，而且不会出错，二手资料总是会有时延的)来探究一下它们：

public @interface Transactional {

    /**
     * A qualifier value for the specified transaction.
     * <p>May be used to determine the target transaction manager,
     * matching the qualifier value (or the bean name) of a specific
     * {@link org.springframework.transaction.PlatformTransactionManager}
     * bean definition.
     */
    String value() default "";

    /**
     * The transaction propagation type.
     * Defaults to {@link Propagation#REQUIRED}.
     * @see org.springframework.transaction.interceptor.TransactionAttribute#getPropagationBehavior()
     */
    Propagation propagation() default Propagation.REQUIRED;

    /**
     * The transaction isolation level.
     * Defaults to {@link Isolation#DEFAULT}.
     * @see org.springframework.transaction.interceptor.TransactionAttribute#getIsolationLevel()
     */
    Isolation isolation() default Isolation.DEFAULT;

    /**
     * The timeout for this transaction.
     * Defaults to the default timeout of the underlying transaction system.
     * @see org.springframework.transaction.interceptor.TransactionAttribute#getTimeout()
     */
    int timeout() default TransactionDefinition.TIMEOUT_DEFAULT;

    /**
     * {@code true} if the transaction is read-only.
     * Defaults to {@code false}.
     * <p>This just serves as a hint for the actual transaction subsystem;
     * it will <i>not necessarily</i> cause failure of write access attempts.
     * A transaction manager which cannot interpret the read-only hint will
     * <i>not</i> throw an exception when asked for a read-only transaction.
     * @see org.springframework.transaction.interceptor.TransactionAttribute#isReadOnly()
     */
    boolean readOnly() default false;

    /**
     * Defines zero (0) or more exception {@link Class classes}, which must be a
     * subclass of {@link Throwable}, indicating which exception types must cause
     * a transaction rollback.
     * <p>This is the preferred way to construct a rollback rule, matching the
     * exception class and subclasses.
     * <p>Similar to {@link org.springframework.transaction.interceptor.RollbackRuleAttribute#RollbackRuleAttribute(Class clazz)}
     */
    Class<? extends Throwable>[] rollbackFor() default {};

    /**
     * Defines zero (0) or more exception names (for exceptions which must be a
     * subclass of {@link Throwable}), indicating which exception types must cause
     * a transaction rollback.
     * <p>This can be a substring, with no wildcard support at present.
     * A value of "ServletException" would match
     * {@link javax.servlet.ServletException} and subclasses, for example.
     * <p><b>NB: </b>Consider carefully how specific the pattern is, and whether
     * to include package information (which isn't mandatory). For example,
     * "Exception" will match nearly anything, and will probably hide other rules.
     * "java.lang.Exception" would be correct if "Exception" was meant to define
     * a rule for all checked exceptions. With more unusual {@link Exception}
     * names such as "BaseBusinessException" there is no need to use a FQN.
     * <p>Similar to {@link org.springframework.transaction.interceptor.RollbackRuleAttribute#RollbackRuleAttribute(String exceptionName)}
     */
    String[] rollbackForClassName() default {};

    /**
     * Defines zero (0) or more exception {@link Class Classes}, which must be a
     * subclass of {@link Throwable}, indicating which exception types must <b>not</b>
     * cause a transaction rollback.
     * <p>This is the preferred way to construct a rollback rule, matching the
     * exception class and subclasses.
     * <p>Similar to {@link org.springframework.transaction.interceptor.NoRollbackRuleAttribute#NoRollbackRuleAttribute(Class clazz)}
     */
    Class<? extends Throwable>[] noRollbackFor() default {};

    /**
     * Defines zero (0) or more exception names (for exceptions which must be a
     * subclass of {@link Throwable}) indicating which exception types must <b>not</b>
     * cause a transaction rollback.
     * <p>See the description of {@link #rollbackForClassName()} for more info on how
     * the specified names are treated.
     * <p>Similar to {@link org.springframework.transaction.interceptor.NoRollbackRuleAttribute#NoRollbackRuleAttribute(String exceptionName)}
     */
    String[] noRollbackForClassName() default {};

}

public @interface Transactional

注解@Transactional的属性有：propagation, isolation, timeout, readOnly, rollbackFor, rollbackForClassName, noRollbackFor, noRollbackForClassName

propagation, isolation, timeout, readOnly都有默认值，而rollbackFor, rollbackForClassName, noRollbackFor, noRollbackForClassName默认值都是空的。

我们具体看下我们可能会用到的属性：propagation, isolation, readOnly, rollbackFor

1）propagation指定事务的传播属性：

public enum Propagation {

    /**
     * Support a current transaction, create a new one if none exists.
     * Analogous to EJB transaction attribute of the same name.
     * <p>This is the default setting of a transaction annotation.
     */
    REQUIRED(TransactionDefinition.PROPAGATION_REQUIRED),

    /**
     * Support a current transaction, execute non-transactionally if none exists.
     * Analogous to EJB transaction attribute of the same name.
     * <p>Note: For transaction managers with transaction synchronization,
     * PROPAGATION_SUPPORTS is slightly different from no transaction at all,
     * as it defines a transaction scope that synchronization will apply for.
     * As a consequence, the same resources (JDBC Connection, Hibernate Session, etc)
     * will be shared for the entire specified scope. Note that this depends on
     * the actual synchronization configuration of the transaction manager.
     * @see org.springframework.transaction.support.AbstractPlatformTransactionManager#setTransactionSynchronization
     */
    SUPPORTS(TransactionDefinition.PROPAGATION_SUPPORTS),

    /**
     * Support a current transaction, throw an exception if none exists.
     * Analogous to EJB transaction attribute of the same name.
     */
    MANDATORY(TransactionDefinition.PROPAGATION_MANDATORY),

    /**
     * Create a new transaction, and suspend the current transaction if one exists.
     * Analogous to the EJB transaction attribute of the same name.
     * <p>Note: Actual transaction suspension will not work out-of-the-box on
     * all transaction managers. This in particular applies to JtaTransactionManager,
     * which requires the {@code javax.transaction.TransactionManager} to be
     * made available it to it (which is server-specific in standard J2EE).
     * @see org.springframework.transaction.jta.JtaTransactionManager#setTransactionManager
     */
    REQUIRES_NEW(TransactionDefinition.PROPAGATION_REQUIRES_NEW),

    /**
     * Execute non-transactionally, suspend the current transaction if one exists.
     * Analogous to EJB transaction attribute of the same name.
     * <p>Note: Actual transaction suspension will not work on out-of-the-box
     * on all transaction managers. This in particular applies to JtaTransactionManager,
     * which requires the {@code javax.transaction.TransactionManager} to be
     * made available it to it (which is server-specific in standard J2EE).
     * @see org.springframework.transaction.jta.JtaTransactionManager#setTransactionManager
     */
    NOT_SUPPORTED(TransactionDefinition.PROPAGATION_NOT_SUPPORTED),

    /**
     * Execute non-transactionally, throw an exception if a transaction exists.
     * Analogous to EJB transaction attribute of the same name.
     */
    NEVER(TransactionDefinition.PROPAGATION_NEVER),

    /**
     * Execute within a nested transaction if a current transaction exists,
     * behave like PROPAGATION_REQUIRED else. There is no analogous feature in EJB.
     * <p>Note: Actual creation of a nested transaction will only work on specific
     * transaction managers. Out of the box, this only applies to the JDBC
     * DataSourceTransactionManager when working on a JDBC 3.0 driver.
     * Some JTA providers might support nested transactions as well.
     * @see org.springframework.jdbc.datasource.DataSourceTransactionManager
     */
    NESTED(TransactionDefinition.PROPAGATION_NESTED);

    private final int value;

    Propagation(int value) { this.value = value; }

    public int value() { return this.value; }

}

public enum Propagation

propagation可以取如下的值：REQUIRED, SUPPORTS, MANDATORY, REQUIRES_NEW, NOT_SUPPORTED, NEVER, NESTED

我们一般会只用：REQUIRED（默认值）, SUPPORTS，其它的取值基本上不使用(具体可以参考上面源码中的注释，已经很详细了)。

propagation=Propagation.REQUIRED：表示该方法或类必须要事务的支持，如果已经是在一个事务中被调用，那么就使用该事务，如果没有在一个事务中，那么就新建一个事务。

propagation=Propagation.SUPPORTS：表示该方法或类支持事务，如果已经是在一个事务中被调用，那么就使用该事务，如果没有在一个事务中，也可以。

2）isolation指定事务的隔离级别：

public enum Isolation {

    /**
     * Use the default isolation level of the underlying datastore.
     * All other levels correspond to the JDBC isolation levels.
     * @see java.sql.Connection
     */
    DEFAULT(TransactionDefinition.ISOLATION_DEFAULT),

    /**
     * A constant indicating that dirty reads, non-repeatable reads and phantom reads
     * can occur. This level allows a row changed by one transaction to be read by
     * another transaction before any changes in that row have been committed
     * (a "dirty read"). If any of the changes are rolled back, the second
     * transaction will have retrieved an invalid row.
     * @see java.sql.Connection#TRANSACTION_READ_UNCOMMITTED
     */
    READ_UNCOMMITTED(TransactionDefinition.ISOLATION_READ_UNCOMMITTED),

    /**
     * A constant indicating that dirty reads are prevented; non-repeatable reads
     * and phantom reads can occur. This level only prohibits a transaction
     * from reading a row with uncommitted changes in it.
     * @see java.sql.Connection#TRANSACTION_READ_COMMITTED
     */
    READ_COMMITTED(TransactionDefinition.ISOLATION_READ_COMMITTED),

    /**
     * A constant indicating that dirty reads and non-repeatable reads are
     * prevented; phantom reads can occur. This level prohibits a transaction
     * from reading a row with uncommitted changes in it, and it also prohibits
     * the situation where one transaction reads a row, a second transaction
     * alters the row, and the first transaction rereads the row, getting
     * different values the second time (a "non-repeatable read").
     * @see java.sql.Connection#TRANSACTION_REPEATABLE_READ
     */
    REPEATABLE_READ(TransactionDefinition.ISOLATION_REPEATABLE_READ),

    /**
     * A constant indicating that dirty reads, non-repeatable reads and phantom
     * reads are prevented. This level includes the prohibitions in
     * {@code ISOLATION_REPEATABLE_READ} and further prohibits the situation
     * where one transaction reads all rows that satisfy a {@code WHERE}
     * condition, a second transaction inserts a row that satisfies that
     * {@code WHERE} condition, and the first transaction rereads for the
     * same condition, retrieving the additional "phantom" row in the second read.
     * @see java.sql.Connection#TRANSACTION_SERIALIZABLE
     */
    SERIALIZABLE(TransactionDefinition.ISOLATION_SERIALIZABLE);

    private final int value;

    Isolation(int value) { this.value = value; }

    public int value() { return this.value; }

}

public enum Isolation

isolation可以取如下的值：DEFAULT, READ_UNCOMMITTED, READ_COMMITTED, REPEATABLE_READ, SERIALIZABLE

我们一般只使用：DEFAULT（默认值），因为我们一般是直接在数据库的层面上来设置事务的隔离级别，很少会在应用层来设置隔离基本。

isolation=Isolation.DEFAULT：表示使用下层数据库指定的隔离级别

3）readOnly为只读，利于数据库优化器进行优化。默认值为 false。所以对于只对数据库进行读取的方法，我们可以如下指定：

    @Transactional (propagation=Propagation.SUPPORTS)
    public User getUser(int userId) {
        return userMapper.getUser(userId);
    }

表示：有事务则使用当前事务，没有事务则不使用事务。最大限度的利于数据库的优化器进行优化。

如果一定要使用事务的话，也可以这样使用readOnly=true来优化：

    @Transactional (propagation=Propagation.REQUIRED,readOnly=true)
    public User getUser(int userId) {
        return userMapper.getUser(userId);
    }

这就是 readOnly 的作用。表示有事务则使用当前事务，如果没有事务，则新建一个只读事务。其实 上面的 propagation=Propagation.REQUIRED也是可以去掉的，因为他是默认值！

    @Transactional (readOnly=true)
    public User getUser(int userId) {
        return userMapper.getUser(userId);
    }

表示强制事务，并且是 只读事务。readOnly 要注意有点的是，readOnly只有在 有事务时，才会生效，如果没有事务，那么 readOnly 是会被忽略的。比如:

@Transactional(propagation=Propagation.SUPPORTS, readonly=true)

表示的就是 没有事务 也是可以的，有事务的话，就一定是 只读事务。

4）rollbackFor，其实该属性也很少使用，而且经常被误用。表示抛出什么异常时，会回滚事务。异常分为受检异常(必须进行处理或者重新抛出)和非受检异常(可以不进行处理)。在遇到非受检异常时，事务是一定会进行回滚的。rollbackFor用于指定对于何种受检异常发生时，进行回滚。因为受检异常，我们必须进行处理或者重新抛出，所以只有一种情况下我们要使用rollbackFor来指定，就是我们不处理异常，直接抛出受检异常，并且我们需要方法在抛出该异常时，进行回滚，如上面的例子：

    @Transactional (rollbackFor = BaseBusinessException.class)
    public void addAndDeleteUser(String username, int id) throws BaseBusinessException{
        userMapper.addUser(username);
        this.m1();
        userMapper.deleteUserById(id);
    }
    private void m1() throws BaseBusinessException {
        throw new BaseBusinessException("xxx");
    }

因为 m1 方法抛出受检异常，我们在 addAndDeleteUser 方法中不对该异常进行处理，而是直接抛出，如果我们希望 userMapper.addUser(username) 和 userMapper.deleteUserById(id) 要么都成功，要么都失败，此时我们则应该指定：rollbackFor = BaseBusinessException.class ，进行回滚。

所以只有我们的方法声明要抛出一个受检异常时，我们才应该使用 rollbackFor 属性来进行处理。如果我们在 addAndDeleteUser  方法中对 m1 方法的受检异常进行了处理，那么就没有必要使用 rollbackFor 了：

    public void addAndDeleteUser(String username, int id){
        userMapper.addUser(username);
        try{
            this.m1();
        }catch(BaseBusinessException e){
            // 处理异常，比如记录进日志文件等
        }
        userMapper.deleteUserById(id);
    }

因为我们处理了 m1 方法的异常，那么就不会有受检异常导致 userMapper.addUser(username) 和 userMapper.deleteUserById(id) 这两个方法一个执行成功，一个没有执行。而非受检异常默认就会回滚。

受检异常是必须进行处理或者重新声明抛出的。只有声明重新抛出受检异常时，才会需要使用 rollbackFor 属性。所以下面的方式就属于滥用 rollbackFor 了：

    @Transactional (rollbackFor=Exception.class)
    public void addAndDeleteUser2(String username, int id){
        userMapper.addUser(username);
        userMapper.deleteUserById(id);
    }

因为：受检异常是必须进行处理或者重新声明抛出，而我们既没有进行处理，也没有重新抛出，就说明他绝对不可能会抛出受检异常了。而只会抛出未受检异常，而未受检异常，默认就会回滚，所以上面的 @Transactional (rollbackFor=Exception.class) 完全是多余的。

总结：

1）@Transactional 的默认值为：@Transactional (propagation=Propagation.REQUIRED,isolation=Isolation.DEFAULT,readOnly=false,

timeout=TransactionDefinition.TIMEOUT_DEFAULT)，默认值已经适合绝大多数情况，所以我们一般使用 @Transactional 进行注解就够了。

2）只有当默认值不符合我们的需要时才给@Transactional的属性指定值，一般也就指定：propagation=Propagation.SUPPORTS 和 readOnly=true，其它的属性和值一般很少使用非默认值。所以我们前面的UserServiceImpl类可以重构如下：

@Service("userService")
@Transactional
public class UserServiceImpl implements UserService{
    @Autowired
    private UserMapper userMapper;

    @Transactional (readOnly=true)
    public User getUser(int userId) {
        return userMapper.getUser(userId);
    }

    public void addUser(String username){
        userMapper.addUser(username);
        int i = 1/0;    // 测试事务的回滚
    }

    public void deleteUser(int id){
        userMapper.deleteUserById(id);
//      int i = 1/0;    // 测试事务的回滚
    }

    @Transactional (rollbackFor = BaseBusinessException.class)
    public void addAndDeleteUser(String username, int id) throws BaseBusinessException{
        userMapper.addUser(username);
        this.m1();
        userMapper.deleteUserById(id);
    }

    private void m1() throws BaseBusinessException {
        throw new BaseBusinessException("xxx");
    }
}

addUser 和 deleteUser 因为会继承上的 @Transactional ，所以无需另外指定了，只有当类上指定的 @Transactional 不适合时，才需要另外在方法上进行指定。

3）所以我们只实际情况中，我们只需要使用下面三者来进行注解事务的配置：

@Transactional，

@Transactional (readOnly=true)，

@Transactional (propagation=Propagation.SUPPORTS, readOnly=true)，

@Transactional (rollbackFor = xxException.class)，其它都可以保持默认值，其它的非默认值极少使用。

另外：

要使用@Transactional来进行注解事务配置，必须要在spring的配置文件中加入下面的配置说明，启用基于注解的配置：

<bean id="transactionManager" class="org.springframework.jdbc.datasource.DataSourceTransactionManager">
        <property name="dataSource" ref="dataSource" />
    </bean>
    
    <!-- 使用annotation定义事务 -->
    <tx:annotation-driven transaction-manager="transactionManager" />

补充说明：

也许你会觉得奇怪，userMapper.getUser(userId); 这些涉及到数据库访问的方法，为什么不会抛出 SQLException 受检异常呢？如果手动注册驱动，然后获取链接，进行数据库操作时，是会有许多的受检异常要处理的：

        try {
            Class.forName("com.mysql.jdbc.Driver");
        } catch (ClassNotFoundException e) {
            // ...
        }
        String url="jdbc:mysql://localhost:3306/databasename";
         try {
            Connection conn = (Connection) DriverManager.getConnection(url,"username","password");
        } catch (SQLException e) {
            // ...
        } 

哪为什么在这里userMapper.getUser(userId); 就不需要处理受检异常呢？其实这是spring的功能，spring为了将我们从那些十分麻烦的受检异常比如SQLException中解救处理，将所有的数据库访问层的受检异常转嫁到 Spring 的 非受检异常RuntimeException 体系中来——DataAccessException。