Spring JPA Ⅲ delete方法详解

一、deleteById 和 delete

为什么要把这两个方法放在一起呢？我们先看源码再说

deleteById(Id id)（通过id进行删除）

源码

@Transactional
@Override
public void deleteById(ID id) {

   Assert.notNull(id, ID_MUST_NOT_BE_NULL);

   delete(findById(id).orElseThrow(() -> new EmptyResultDataAccessException(
         String.format("No %s entity with id %s exists!", entityInformation.getJavaType(), id), 1)));
}

delete(T entity)（通过实体对象进行删除）

源码

@Override
@Transactional
@SuppressWarnings("unchecked")
public void delete(T entity) {

   Assert.notNull(entity, "Entity must not be null!");

   if (entityInformation.isNew(entity)) {
      return;
   }

   Class<?> type = ProxyUtils.getUserClass(entity);

   T existing = (T) em.find(type, entityInformation.getId(entity));

   // if the entity to be deleted doesn't exist, delete is a NOOP
   if (existing == null) {
      return;
   }

   em.remove(em.contains(entity) ? entity : em.merge(entity));
}

一目了然了吧！deleteById 先在方法体内通过 id 求出 entity 对象，然后调用了 delete 的方法。也就是说，这两个方法同根同源，使用起来差距不大，结果呢？也是一样的，就是单条删除。实际使用中呢，也是使用 deleteById 的情况比较多，废话少说，try it。

实例

service 层

添加deleteById方法（deleteById 是 JPA 自带接口不需要在dao层中添加）

@Transactional
public void deleteById(Integer id){
    userDao.deleteById(id);
}

control层

/**
 * 通过id进行删除数据
 * @param id
 */
@GetMapping("/deleteById")
public void deleteById(Integer id){
	userService.deleteById(id);
}

执行请求 /deleteById?id=2，控制台打印如下：

Hibernate: select user0_.id as id1_0_0_, user0_.age as age2_0_0_, user0_.name as name3_0_0_ from user user0_ where user0_.id=?
Hibernate: delete from user where id=?

结论

先通过 select 查询实体对象是否存在，然后再通过 id 进行删除。

二、deleteAllById 和 deleteAll

1、deleteAllById(Iterable<? extends ID> ids)（通过id进行批量删除）

源码

@Override
@Transactional
public void deleteAllById(Iterable<? extends ID> ids) {

   Assert.notNull(ids, "Ids must not be null!");

   for (ID id : ids) {
      deleteById(id);
   }
}

结论

通过源码可以看出，就是遍历 ids 然后循环调用上面的 deleteById(Id id) 方法。

2、deleteAll(Iterable<? extends T> entities)（通过实体对象进行批量删除）

源码

@Override
@Transactional
public void deleteAll(Iterable<? extends T> entities) {

   Assert.notNull(entities, "Entities must not be null!");

   for (T entity : entities) {
      delete(entity);
   }
}

结论

这个呢？也就是遍历 entities 然后循环调用上面的 delete(T entity) 方法

还有一个不传参数的deleteAll()方法来删除所有数据（慎用）

@Override
@Transactional
public void deleteAll() {

   for (T element : findAll()) {
      delete(element);
   }
}

就是通过findAll求出所有实体对象然后循环调用delete方法

综上所述，我们发现以上所有的删除事件都是调用了delete(T entity)方法，也就是差距不是很大，就是单条 和多条删除的区别。

实例

service 层

添加 deleteAllById 方法（deleteAllById 是三方件自带接口不需要在dao层中添加）

@Transactional
public void deleteAllById(Iterable ids){
	userDao.deleteAllById(ids);
}

control层

/**
 * 通过id进行批量删除
 * @param ids
 */
@GetMapping("/deleteAllById")
public void deleteAllById(Integer[] ids){
	userService.deleteAllById(Arrays.asList(ids));
}

浏览器测试成功 /deleteAllById?id=3,4
删除前：

删除后：

控制台打印如下：

Hibernate: select user0_.id as id1_0_0_, user0_.age as age2_0_0_, user0_.name as name3_0_0_ from user user0_ where user0_.id=?
Hibernate: select user0_.id as id1_0_0_, user0_.age as age2_0_0_, user0_.name as name3_0_0_ from user user0_ where user0_.id=?
Hibernate: delete from user where id=?
Hibernate: delete from user where id=?

由此可以看出，数据是一条一条的进行了删除。

三、deleteAllInBatch 和 deleteAllByIdInBatch

1、deleteAllInBatch(Iterable<T> entities)（通过实体对象进行批量删除）

源码

public static final String DELETE_ALL_QUERY_STRING = "delete from %s x";

@Override
@Transactional
public void deleteAllInBatch(Iterable<T> entities) {

   Assert.notNull(entities, "Entities must not be null!");

   if (!entities.iterator().hasNext()) {
      return;
   }

   applyAndBind(getQueryString(DELETE_ALL_QUERY_STRING, entityInformation.getEntityName()), entities, em)
         .executeUpdate();
}

/**
 * Creates a where-clause referencing the given entities and appends it to the given query string. Binds the given
 * entities to the query.
 *
 * @param <T> type of the entities.
 * @param queryString must not be {@literal null}.
 * @param entities must not be {@literal null}.
 * @param entityManager must not be {@literal null}.
 * @return Guaranteed to be not {@literal null}.
 */

public static <T> Query applyAndBind(String queryString, Iterable<T> entities, EntityManager entityManager) {

   Assert.notNull(queryString, "Querystring must not be null!");
   Assert.notNull(entities, "Iterable of entities must not be null!");
   Assert.notNull(entityManager, "EntityManager must not be null!");

   Iterator<T> iterator = entities.iterator();

   if (!iterator.hasNext()) {
      return entityManager.createQuery(queryString);
   }

   String alias = detectAlias(queryString);
   StringBuilder builder = new StringBuilder(queryString);
   builder.append(" where");

   int i = 0;

   while (iterator.hasNext()) {

      iterator.next();

      builder.append(String.format(" %s = ?%d", alias, ++i));

      if (iterator.hasNext()) {
         builder.append(" or");
      }
   }

   Query query = entityManager.createQuery(builder.toString());

   iterator = entities.iterator();
   i = 0;

   while (iterator.hasNext()) {
      query.setParameter(++i, iterator.next());
   }

   return query;
}

通过上面的源码，我们大体能猜测出deleteAllInBatch(Iterable<T> entities)的实现原理：
delete from %s where x=? or x=?
实际测试一下：http://localhost:7777/deleteAllInBatch?ids=14,15,16&names=a,b,c&ages=0,0,0
控制台打印如下：

Hibernate: delete from user where id=? or id=? or id=?

2、deleteAllByIdInBatch(Iterable<ID> ids)源码（通过ids批量删除）

public static final String DELETE_ALL_QUERY_BY_ID_STRING = "delete from %s x where %s in :ids";

@Override
@Transactional
public void deleteAllByIdInBatch(Iterable<ID> ids) {

   Assert.notNull(ids, "Ids must not be null!");

   if (!ids.iterator().hasNext()) {
      return;
   }

   if (entityInformation.hasCompositeId()) {

      List<T> entities = new ArrayList<>();
      // generate entity (proxies) without accessing the database.
      ids.forEach(id -> entities.add(getReferenceById(id)));
      deleteAllInBatch(entities);
   } else {

      String queryString = String.format(DELETE_ALL_QUERY_BY_ID_STRING, entityInformation.getEntityName(),
            entityInformation.getIdAttribute().getName());

      Query query = em.createQuery(queryString);
      /**
       * Some JPA providers require {@code ids} to be a {@link Collection} so we must convert if it's not already.
       */
      if (Collection.class.isInstance(ids)) {
         query.setParameter("ids", ids);
      } else {
         Collection<ID> idsCollection = StreamSupport.stream(ids.spliterator(), false)
               .collect(Collectors.toCollection(ArrayList::new));
         query.setParameter("ids", idsCollection);
      }

      query.executeUpdate();
   }
}

通过上面源码我们大体可以猜出deleteAllByIdInBatch(Iterable ids)的实现原理：
delete from %s where id in (?,?,?)
实际测试一下：http://localhost:7777/deleteAllByIdInBatch?ids=17,18,19 控制台打印如下：

Hibernate: delete from user where id in (? , ? , ?)

这里同样有个不带参数的deleteAllInBatch()的方法，源码如下：

@Override
@Transactional
public void deleteAllInBatch() {
   em.createQuery(getDeleteAllQueryString()).executeUpdate();
}

public static final String DELETE_ALL_QUERY_STRING = "delete from %s x";

private String getDeleteAllQueryString() {
   return getQueryString(DELETE_ALL_QUERY_STRING, entityInformation.getEntityName());
}

通过源码不难猜到实现原理吧，多的不说，直接给测试的控制台数据：
Hibernate: delete from user

最后结论：

从上面两种删除接口来看，第二种实现比起第一种更加的快捷；第一种就是一条一条的进行删除操作，如果有万级的数据，执行起来肯定非常耗时，所以如果数据量比较大的话，还是建议大家使用第二种。