JPA中自定义的插入、更新、删除方法为什么要添加@Modifying注解和@Transactional注解?
前几天,有个同事在使用JPA的自定义SQL方法时,程序一直报异常,捣鼓了半天也没能解决,咨询我的时候,我看了一眼他的程序,差不多是这个样子的:
1 @Repository 2 public interface UserRepository extends JpaRepository<User,Long> { 3 4 @Query(value = "delete from pro_user where id = ?1",nativeQuery = true) 5 void deleteUserById(Long id); 6 }
我告诉他,你的deleteUserById方法缺少了@Modifying注解和@Transactional注解,他半信半疑地试了一下,然后果然就解决了。其实,如果他查一下官方资料或许很快也就能找到答案。基于这个背景,本文详细讲解一下为何我们自定义的插入、更新、删除操作需要加@Modifying注解和@Transactional注解。
一、@Modifying注解
在官方资料中,给出了这样几句说明:
As the queries themselves are tied to the Java method that executes them, you can actually bind them directly by using the Spring Data JPA @Query annotation rather than annotating them to the domain class.
You can modify queries that only need parameter binding by annotating the query method with @Modifying
The @Modifying annotation is only relevant in combination with the @Query annotation. Derived query methods or custom methods do not require this Annotation.
Doing so triggers the query annotated to the method as an updating query instead of a selecting one.
如下:
@Modifying @Query("update User u set u.firstname = ?1 where u.lastname = ?2") int setFixedFirstnameFor(String firstname, String lastname);
第一句话的意思是可以用@Query注解来将自定义sql语句绑定到自定义方法上。
第二句话的意思时,可以用@Modifying注解来标注只需要绑定参数的自定义的更新类语句(更新、插入、删除)。
第三名话的意思是说@Modifying只与@Query联合使用,派生类的查询方法和自定义的方法不需要此注解,如:
1 @Repository 2 public interface UserRepository extends JpaRepository<User,Long> { 3 4 // 父类的保存方法 5 @Override 6 User save(User entity); 7 8 // 按照JPA语法规则自定义的查询方法 9 List<User> findFirst10ByLastname(String lastName, Pageable pageable); 10 }
第四句话的意思是,当加上@Modifying注解时,JPA会以更新类语句来执行,而不再是以查询语句执行。
也就是说,当我们要通过自已写的更新、插入、删除SQL语句来实现更新、插入、删除操作时,至少需要用两个步骤:
1)@Query来注入我们自定义的sql;
2)使用@Modifying来标注是一个更新类的自定义语句。
按照这个规则,修改同事的那个方法:
1 @Repository 2 public interface UserRepository extends JpaRepository<User,Long> { 3 4 @Modifying 5 @Query(value = "delete from pro_user where id = ?1",nativeQuery = true) 6 void deleteUserById(Long id); 7 }
但是,此时,该方法还不完整,执行时程序会报以下错误:
org.springframework.dao.InvalidDataAccessApiUsageException: Executing an update/delete query; nested exception is javax.persistence.TransactionRequiredException:
Executing an update/delete query at org.springframework.orm.jpa.EntityManagerFactoryUtils.convertJpaAccessExceptionIfPossible(EntityManagerFactoryUtils.java:402) at org.springframework.orm.jpa.vendor.HibernateJpaDialect.translateExceptionIfPossible(HibernateJpaDialect.java:255) ...... at com.intellij.rt.execution.junit.JUnitStarter.main(JUnitStarter.java:70) Caused by: javax.persistence.TransactionRequiredException: Executing an update/delete query at org.hibernate.internal.AbstractSharedSessionContract.checkTransactionNeededForUpdateOperation(AbstractSharedSessionContract.java:398) at org.hibernate.query.internal.AbstractProducedQuery.executeUpdate(AbstractProducedQuery.java:1585) .......
二、@Transactional注解
官方的说明:
readOnly
flag is set to true
. All others are configured with a plain @Transactional
so that default transaction configuration applies. For details, see JavaDoc of SimpleJpaRepository
. If you need to tweak transaction configuration for one of the methods declared in a repository, redeclare the method in your repository interface, as follows:
Example. Custom transaction configuration for CRUD
1 public interface UserRepository extends CrudRepository<User, Long> { 2 3 @Override 4 @Transactional(timeout = 10) 5 public List<User> findAll(); 6 7 // Further query method declarations 8 }
这句话的意思是,默认情况下,repository 接口中的CRUD方法都是被@Transactional注解修饰了的,对于读的操作方法,@Transactional注解的readOnly属性是被设置为true的,即只读;CRUD中的其他方法被@Transactional修饰,即非只读。如果你需要修改repository 接口中的某些方法的事务属性,可以在该方法上重新加上@Transactional注解,并设置需要的属性。
我们先来看一下,@Transactional注解的源码:
1 @Target({ElementType.METHOD, ElementType.TYPE}) 2 @Retention(RetentionPolicy.RUNTIME) 3 @Inherited 4 @Documented 5 public @interface Transactional { 6 7 Propagation propagation() default Propagation.REQUIRED; 8 9 Isolation isolation() default Isolation.DEFAULT; 10 11 int timeout() default -1; 12 13 boolean readOnly() default false; 14 15 // 其他省略 16 }
由上可见@Transactional注解的readOnly默认的属性的false,即非只读,当一个事务是非只读事务的时候,我们可以进行任何操作。
再看一下repository 接口的实现类SimpleJpaRepository的源码(只摘了部分源码):
1 @Repository 2 @Transactional( 3 readOnly = true 4 ) 5 public class SimpleJpaRepository<T, ID> implements JpaRepositoryImplementation<T, ID> { 6 7 @Transactional 8 public void deleteById(ID id) { 9 Assert.notNull(id, "The given id must not be null!"); 10 this.delete(this.findById(id).orElseThrow(() -> { 11 return new EmptyResultDataAccessException(String.format("No %s entity with id %s exists!", this.entityInformation.getJavaType(), id), 1); 12 })); 13 } 14 15 @Transactional 16 public void delete(T entity) { 17 Assert.notNull(entity, "The entity must not be null!"); 18 this.em.remove(this.em.contains(entity) ? entity : this.em.merge(entity)); 19 } 20 21 @Transactional 22 public void deleteAll(Iterable<? extends T> entities) { 23 Assert.notNull(entities, "The given Iterable of entities not be null!"); 24 Iterator var2 = entities.iterator(); 25 26 while(var2.hasNext()) { 27 T entity = var2.next(); 28 this.delete(entity); 29 } 30 } 31 32 public T getOne(ID id) { 33 Assert.notNull(id, "The given id must not be null!"); 34 return this.em.getReference(this.getDomainClass(), id); 35 } 36 37 public List<T> findAll() { 38 return this.getQuery((Specification)null, (Sort)Sort.unsorted()).getResultList(); 39 } 40 41 public List<T> findAll(@Nullable Specification<T> spec) { 42 return this.getQuery(spec, Sort.unsorted()).getResultList(); 43 } 44 45 public List<T> findAll(@Nullable Specification<T> spec, Sort sort) { 46 return this.getQuery(spec, sort).getResultList(); 47 } 48 49 public <S extends T> long count(Example<S> example) { 50 return executeCountQuery(this.getCountQuery(new SimpleJpaRepository.ExampleSpecification(example), example.getProbeType())); 51 } 52 53 public <S extends T> boolean exists(Example<S> example) { 54 return !this.getQuery(new SimpleJpaRepository.ExampleSpecification(example), example.getProbeType(), (Sort)Sort.unsorted()).getResultList().isEmpty(); 55 } 56 57 @Transactional 58 public <S extends T> S save(S entity) { 59 if (this.entityInformation.isNew(entity)) { 60 this.em.persist(entity); 61 return entity; 62 } else { 63 return this.em.merge(entity); 64 } 65 } 66 67 @Transactional 68 public <S extends T> S saveAndFlush(S entity) { 69 S result = this.save(entity); 70 this.flush(); 71 return result; 72 } 73 74 @Transactional 75 public void flush() { 76 this.em.flush(); 77 } 78 }
从SimpleJpaRepository源码中可以看出:
1)该类上注解了只读事务@Transactional(readOnly = true);
2)该类的所有查询类操作方法都与类相同,都拥有只读事务;
3)该类的所有保存、更新、删除操作方法都用@Transactional重新注解了(默认readOnly=false)。
说明JPA为我们提供的所有方法,包括JPA规则的自定义方法在其底层都为我们做好了事务处理,而我们自定义的方法需要自己来标注事务的类型是只读还是非只读。根据这个原理,再次修改开篇所列出的方法:
1 @Repository 2 public interface UserRepository extends JpaRepository<User,Long> { 3 4 @Transactional 5 @Modifying 6 @Query(value = "delete from pro_user where id = ?1",nativeQuery = true) 7 void deleteUserById(Long id); 8 }
至此,该方法按所期望的结果运行成功了。
三、@Modifying注解补充说明
1 @Retention(RetentionPolicy.RUNTIME) 2 @Target({ ElementType.METHOD, ElementType.ANNOTATION_TYPE }) 3 @Documented 4 public @interface Modifying { 5 6 boolean flushAutomatically() default false; 7 8 boolean clearAutomatically() default false; 9 }
该注解中有两个属性:flushAutomatically、clearAutomatically,从字面理解是自动刷新和自动清除。
自动刷新,即执行完语句后立即将变化内容刷新到磁盘,如果是insert语句操作,则与JPA的<S extends T> S saveAndFlush(S entity);方法效果相同;
自动清除,即执行完语句后自动清除掉已经过期的实体,比如,我们删除了一个实体,但是在还没有执行flush操作时,这个实体还存在于实体管理器EntityManager中,但这个实体已经过期没有任何用处,直到flush操作时才会被删除掉。如果希望在删除该实体时立即将该实体从实体管理器中删除,则可以将该属性设置为true,如:
1 @Modifying(clearAutomatically = true) 2 @Transactional 3 @Query(value = "delete from pro_user where id = ?1",nativeQuery = true) 4 void deleteUserById(Long id);