Question

我有更新查询：

@Modifying
@Transactional
@Query("UPDATE Admin SET firstname = :firstname, lastname = :lastname, login = :login, superAdmin = :superAdmin, preferenceAdmin = :preferenceAdmin, address =  :address, zipCode = :zipCode, city = :city, country = :country, email = :email, profile = :profile, postLoginUrl = :postLoginUrl WHERE id = :id")
public void update(@Param("firstname") String firstname, @Param("lastname") String lastname, @Param("login") String login, @Param("superAdmin") boolean superAdmin, @Param("preferenceAdmin") boolean preferenceAdmin, @Param("address") String address, @Param("zipCode") String zipCode, @Param("city") String city, @Param("country") String country, @Param("email") String email, @Param("profile") String profile, @Param("postLoginUrl") String postLoginUrl, @Param("id") Long id);

我正在尝试在集成测试中使用它：

adminRepository.update("Toto", "LeHeros", admin0.getLogin(), admin0.getSuperAdmin(), admin0.getPreferenceAdmin(), admin0.getAddress(), admin0.getZipCode(), admin0.getCity(), admin0.getCountry(), admin0.getEmail(), admin0.getProfile(), admin0.getPostLoginUrl(), admin0.getId());
Admin loadedAdmin = adminRepository.findOne(admin0.getId());
assertEquals("Toto", loadedAdmin.getFirstname());
assertEquals("LeHeros", loadedAdmin.getLastname());

但字段未更新并保留其初始值，因此测试失败。

我尝试在findOne查询之前添加一个flush：

adminRepository.flush();

但失败的断言仍然相同。

我可以在日志中看到更新sql语句：

update admin set firstname='Toto', lastname='LeHeros', login='stephane', super_admin=0, preference_admin=0,
address=NULL, zip_code=NULL, city=NULL, country=NULL, email='stephane@thalasoft.com', profile=NULL,
post_login_url=NULL where id=2839

但是日志显示没有可能与finder相关的sql：

Admin loadedAdmin = adminRepository.findOne(admin0.getId());
The finder sql statement is not making its way to the database.

是否因某些缓存原因而被忽略？

如果我然后添加对findByEmail和findByLogin finders的调用，如下所示：

adminRepository.update("Toto", "LeHeros", "qwerty", admin0.getSuperAdmin(), admin0.getPreferenceAdmin(), admin0.getAddress(), admin0.getZipCode(), admin0.getCity(), admin0.getCountry(), admin0.getEmail(), admin0.getProfile(), admin0.getPostLoginUrl(), admin0.getId());
Admin loadedAdmin = adminRepository.findOne(admin0.getId());
Admin myadmin = adminRepository.findByEmail(admin0.getEmail());
Admin anadmin = adminRepository.findByLogin("qwerty");
assertEquals("Toto", anadmin.getFirstname());
assertEquals("Toto", myadmin.getFirstname());
assertEquals("Toto", loadedAdmin.getFirstname());
assertEquals("LeHeros", loadedAdmin.getLastname());

然后我可以在日志中看到正在生成的sql语句：

但断言：

assertEquals("Toto", myadmin.getFirstname());

即使跟踪显示检索到相同的域对象，

仍然失败：

TRACE [BasicExtractor] found [1037] as column [id14_]

另一个令我困惑的另一个问题是它显示了一个限制2子句，即使它只能返回一个Admin对象。

我认为返回一个域对象时总会有一个限制1。这是对Spring Data的错误假设吗？

当在MySQL客户端粘贴时，控制台日志中显示的sql语句，逻辑工作正常：

mysql> insert into admin (version, address, city, country, email, firstname, lastname, login, password, 
-> password_salt, post_login_url, preference_admin, profile, super_admin, zip_code) values (0,
-> NULL, NULL, NULL, 'zemail@thalasoft.com039', 'zfirstname039', 'zlastname039', 'zlogin039',
-> 'zpassword039', '', NULL, 0, NULL, 1, NULL);
Query OK, 1 row affected (0.07 sec)

mysql> select * from admin;
+------+---------+---------------+--------------+-----------+--------------+---------------+-------------+------------------+---------+----------+------+---------+-------------------------+---------+----------------+
| id | version | firstname | lastname | login | password | password_salt | super_admin | preference_admin | address | zip_code | city | country | email | profile | post_login_url |
+------+---------+---------------+--------------+-----------+--------------+---------------+-------------+------------------+---------+----------+------+---------+-------------------------+---------+----------------+
| 1807 | 0 | zfirstname039 | zlastname039 | zlogin039 | zpassword039 | | 1 | 0 | NULL | NULL | NULL | NULL | zemail@thalasoft.com039 | NULL | NULL | 
+------+---------+---------------+--------------+-----------+--------------+---------------+-------------+------------------+---------+----------+------+---------+-------------------------+---------+----------------+
1 row in set (0.00 sec)

mysql> update admin set firstname='Toto', lastname='LeHeros', login='qwerty', super_admin=0, preference_admin=0, address=NULL, zip_code=NULL, city=NULL, country=NULL, email='stephane@thalasoft.com', profile=NULL, post_login_url=NULL where id=1807;
Query OK, 1 row affected (0.07 sec)
Rows matched: 1 Changed: 1 Warnings: 0

mysql> select * from admin; +------+---------+-----------+----------+--------+--------------+---------------+-------------+------------------+---------+----------+------+---------+------------------------+---------+----------------+
| id | version | firstname | lastname | login | password | password_salt | super_admin | preference_admin | address | zip_code | city | country | email | profile | post_login_url |
+------+---------+-----------+----------+--------+--------------+---------------+-------------+------------------+---------+----------+------+---------+------------------------+---------+----------------+
| 1807 | 0 | Toto | LeHeros | qwerty | zpassword039 | | 0 | 0 | NULL | NULL | NULL | NULL | stephane@thalasoft.com | NULL | NULL | 
+------+---------+-----------+----------+--------+--------------+---------------+-------------+------------------+---------+----------+------+---------+------------------------+---------+----------------+
1 row in set (0.00 sec)

mysql> select admin0_.id as id14_, admin0_.version as version14_, admin0_.address as address14_, admin0_.city as city14_, admin0_.country as country14_, admin0_.email as email14_, admin0_.firstname as firstname14_, admin0_.lastname as lastname14_, admin0_.login as login14_, admin0_.password as password14_, admin0_.password_salt as password11_14_, admin0_.post_login_url as post12_14_, admin0_.preference_admin as preference13_14_, admin0_.profile as profile14_, admin0_.super_admin as super15_14_, admin0_.zip_code as zip16_14_ from admin admin0_ where admin0_.email='stephane@thalasoft.com' limit 2;
+-------+------------+------------+---------+------------+------------------------+--------------+-------------+----------+--------------+----------------+------------+------------------+------------+-------------+-----------+
| id14_ | version14_ | address14_ | city14_ | country14_ | email14_ | firstname14_ | lastname14_ | login14_ | password14_ | password11_14_ | post12_14_ | preference13_14_ | profile14_ | super15_14_ | zip16_14_ |
+-------+------------+------------+---------+------------+------------------------+--------------+-------------+----------+--------------+----------------+------------+------------------+------------+-------------+-----------+
| 1807 | 0 | NULL | NULL | NULL | stephane@thalasoft.com | Toto | LeHeros | qwerty | zpassword039 | | NULL | 0 | NULL | 0 | NULL | 
+-------+------------+------------+---------+------------+------------------------+--------------+-------------+----------+--------------+----------------+------------+------------------+------------+-------------+-----------+
1 row in set (0.00 sec)

mysql> select admin0_.id as id14_, admin0_.version as version14_, admin0_.address as address14_, admin0_.city as city14_, admin0_.country as country14_, admin0_.email as email14_, admin0_.firstname as firstname14_, admin0_.lastname as lastname14_, admin0_.login as login14_, admin0_.password as password14_, admin0_.password_salt as password11_14_, admin0_.post_login_url as post12_14_, admin0_.preference_admin as preference13_14_, admin0_.profile as profile14_, admin0_.super_admin as super15_14_, admin0_.zip_code as zip16_14_ from admin admin0_ where admin0_.login='qwerty' limit 2;
+-------+------------+------------+---------+------------+------------------------+--------------+-------------+----------+--------------+----------------+------------+------------------+------------+-------------+-----------+
| id14_ | version14_ | address14_ | city14_ | country14_ | email14_ | firstname14_ | lastname14_ | login14_ | password14_ | password11_14_ | post12_14_ | preference13_14_ | profile14_ | super15_14_ | zip16_14_ |
+-------+------------+------------+---------+------------+------------------------+--------------+-------------+----------+--------------+----------------+------------+------------------+------------+-------------+-----------+
| 1807 | 0 | NULL | NULL | NULL | stephane@thalasoft.com | Toto | LeHeros | qwerty | zpassword039 | | NULL | 0 | NULL | 0 | NULL | 
+-------+------------+------------+---------+------------+------------------------+--------------+-------------+----------+--------------+----------------+------------+------------------+------------+-------------+-----------+
1 row in set (0.00 sec)

那么为什么这不会反映在Java级别呢？

Answer 1

默认情况下，EntityManager不会自动刷新更改。您应该在查询语句中使用以下选项：

@Modifying(clearAutomatically = true)
@Query("update RssFeedEntry feedEntry set feedEntry.read =:isRead where feedEntry.id =:entryId")
void markEntryAsRead(@Param("entryId") Long rssFeedEntryId, @Param("isRead") boolean isRead);

Answer 2

我终于明白了发生了什么。

在保存对象的语句上创建集成测试时，建议刷新实体管理器以避免任何误报，即避免测试运行正常，但在生产中运行时操作会失败。实际上，测试可能运行正常，因为没有刷新第一级缓存并且没有写入数据库。为了避免这种假阴性积分测试，在测试体中使用显式刷新。请注意，生产代码永远不需要使用任何显式刷新，因为ORM的角色决定何时刷新。

在update语句上创建集成测试时，可能需要清除实体管理器以重新加载第一级缓存。实际上，更新语句完全绕过第一级缓存并直接写入数据库。然后，第一级缓存不同步并反映更新对象的旧值。要避免对象的这种陈旧状态，请在测试主体中使用显式清除。请注意，生产代码永远不需要使用任何明确的清除，因为ORM的角色决定何时清除。

我的测试现在运行得很好。

Answer 3

我能够让它发挥作用。我将在这里描述我的应用程序和集成测试。

示例应用

示例应用程序有两个类和一个与此问题相关的接口：

应用程序上下文配置类
实体类
存储库界面

以下描述了这些类和存储库接口。

PersistenceContext类的源代码如下所示：

import com.jolbox.bonecp.BoneCPDataSource;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.context.annotation.Bean;
import org.springframework.context.annotation.Configuration;
import org.springframework.context.annotation.PropertySource;
import org.springframework.core.env.Environment;
import org.springframework.data.jpa.repository.config.EnableJpaRepositories;
import org.springframework.orm.jpa.JpaTransactionManager;
import org.springframework.orm.jpa.LocalContainerEntityManagerFactoryBean;
import org.springframework.orm.jpa.vendor.HibernateJpaVendorAdapter;
import org.springframework.transaction.annotation.EnableTransactionManagement;

import javax.sql.DataSource;
import java.util.Properties;

@Configuration
@EnableTransactionManagement
@EnableJpaRepositories(basePackages = "net.petrikainulainen.spring.datajpa.todo.repository")
@PropertySource("classpath:application.properties")
public class PersistenceContext {

    protected static final String PROPERTY_NAME_DATABASE_DRIVER = "db.driver";
    protected static final String PROPERTY_NAME_DATABASE_PASSWORD = "db.password";
    protected static final String PROPERTY_NAME_DATABASE_URL = "db.url";
    protected static final String PROPERTY_NAME_DATABASE_USERNAME = "db.username";

    private static final String PROPERTY_NAME_HIBERNATE_DIALECT = "hibernate.dialect";
    private static final String PROPERTY_NAME_HIBERNATE_FORMAT_SQL = "hibernate.format_sql";
    private static final String PROPERTY_NAME_HIBERNATE_HBM2DDL_AUTO = "hibernate.hbm2ddl.auto";
    private static final String PROPERTY_NAME_HIBERNATE_NAMING_STRATEGY = "hibernate.ejb.naming_strategy";
    private static final String PROPERTY_NAME_HIBERNATE_SHOW_SQL = "hibernate.show_sql";

    private static final String PROPERTY_PACKAGES_TO_SCAN = "net.petrikainulainen.spring.datajpa.todo.model";

    @Autowired
    private Environment environment;

    @Bean
    public DataSource dataSource() {
        BoneCPDataSource dataSource = new BoneCPDataSource();

        dataSource.setDriverClass(environment.getRequiredProperty(PROPERTY_NAME_DATABASE_DRIVER));
        dataSource.setJdbcUrl(environment.getRequiredProperty(PROPERTY_NAME_DATABASE_URL));
        dataSource.setUsername(environment.getRequiredProperty(PROPERTY_NAME_DATABASE_USERNAME));
        dataSource.setPassword(environment.getRequiredProperty(PROPERTY_NAME_DATABASE_PASSWORD));

        return dataSource;
    }

    @Bean
    public JpaTransactionManager transactionManager() {
        JpaTransactionManager transactionManager = new JpaTransactionManager();

        transactionManager.setEntityManagerFactory(entityManagerFactory().getObject());

        return transactionManager;
    }

    @Bean
    public LocalContainerEntityManagerFactoryBean entityManagerFactory() {
        LocalContainerEntityManagerFactoryBean entityManagerFactoryBean = new LocalContainerEntityManagerFactoryBean();

        entityManagerFactoryBean.setDataSource(dataSource());
        entityManagerFactoryBean.setJpaVendorAdapter(new HibernateJpaVendorAdapter());
        entityManagerFactoryBean.setPackagesToScan(PROPERTY_PACKAGES_TO_SCAN);

        Properties jpaProperties = new Properties();
        jpaProperties.put(PROPERTY_NAME_HIBERNATE_DIALECT, environment.getRequiredProperty(PROPERTY_NAME_HIBERNATE_DIALECT));
        jpaProperties.put(PROPERTY_NAME_HIBERNATE_FORMAT_SQL, environment.getRequiredProperty(PROPERTY_NAME_HIBERNATE_FORMAT_SQL));
        jpaProperties.put(PROPERTY_NAME_HIBERNATE_HBM2DDL_AUTO, environment.getRequiredProperty(PROPERTY_NAME_HIBERNATE_HBM2DDL_AUTO));
        jpaProperties.put(PROPERTY_NAME_HIBERNATE_NAMING_STRATEGY, environment.getRequiredProperty(PROPERTY_NAME_HIBERNATE_NAMING_STRATEGY));
        jpaProperties.put(PROPERTY_NAME_HIBERNATE_SHOW_SQL, environment.getRequiredProperty(PROPERTY_NAME_HIBERNATE_SHOW_SQL));

        entityManagerFactoryBean.setJpaProperties(jpaProperties);

        return entityManagerFactoryBean;
    }
}

假设我们有一个名为Todo的简单实体，其源代码如下所示：

@Entity
@Table(name="todos")
public class Todo {

    public static final int MAX_LENGTH_DESCRIPTION = 500;
    public static final int MAX_LENGTH_TITLE = 100;

    @Id
    @GeneratedValue(strategy = GenerationType.AUTO)
    private Long id;

    @Column(name = "description", nullable = true, length = MAX_LENGTH_DESCRIPTION)
    private String description;

    @Column(name = "title", nullable = false, length = MAX_LENGTH_TITLE)
    private String title;

    @Version
    private long version;
}

我们的存储库接口有一个名为updateTitle()的方法，用于更新待办事项的标题。 TodoRepository接口的源代码如下所示：

import net.petrikainulainen.spring.datajpa.todo.model.Todo;
import org.springframework.data.jpa.repository.JpaRepository;
import org.springframework.data.jpa.repository.Modifying;
import org.springframework.data.jpa.repository.Query;
import org.springframework.data.repository.query.Param;

import java.util.List;

public interface TodoRepository extends JpaRepository<Todo, Long> {

    @Modifying
    @Query("Update Todo t SET t.title=:title WHERE t.id=:id")
    public void updateTitle(@Param("id") Long id, @Param("title") String title);
}

updateTitle()方法未使用@Transactional注释进行注释，因为我认为最好将服务层用作事务边界。

整合测试

集成测试使用DbUnit，Spring Test和Spring-Test-DBUnit。它有三个与此问题相关的组件：

DbUnit数据集，用于在执行测试之前将数据库初始化为已知状态。
DbUnit数据集，用于验证实体的标题是否已更新。
集成测试。

以下详细介绍了这些组件。

用于将数据库初始化为已知状态的DbUnit数据集文件的名称是 toDoData.xml ，其内容如下所示：

<dataset>
    <todos id="1" description="Lorem ipsum" title="Foo" version="0"/>
    <todos id="2" description="Lorem ipsum" title="Bar" version="0"/>
</dataset>

用于验证todo条目标题更新的DbUnit数据集的名称称为 toDoData-update.xml ，其内容如下所示（由于某种原因， todo条目没有更新，但标题是。任何想法为什么？）：

<dataset>
    <todos id="1" description="Lorem ipsum" title="FooBar" version="0"/>
    <todos id="2" description="Lorem ipsum" title="Bar" version="0"/>
</dataset>

实际集成测试的源代码如下所示（请记住使用@Transactional注释注释测试方法）：

import com.github.springtestdbunit.DbUnitTestExecutionListener;
import com.github.springtestdbunit.TransactionDbUnitTestExecutionListener;
import com.github.springtestdbunit.annotation.DatabaseSetup;
import com.github.springtestdbunit.annotation.ExpectedDatabase;
import org.junit.Test;
import org.junit.runner.RunWith;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.test.annotation.Rollback;
import org.springframework.test.context.ContextConfiguration;
import org.springframework.test.context.TestExecutionListeners;
import org.springframework.test.context.junit4.SpringJUnit4ClassRunner;
import org.springframework.test.context.support.DependencyInjectionTestExecutionListener;
import org.springframework.test.context.support.DirtiesContextTestExecutionListener;
import org.springframework.test.context.transaction.TransactionalTestExecutionListener;
import org.springframework.transaction.annotation.Transactional;

@RunWith(SpringJUnit4ClassRunner.class)
@ContextConfiguration(classes = {PersistenceContext.class})
@TestExecutionListeners({ DependencyInjectionTestExecutionListener.class,
        DirtiesContextTestExecutionListener.class,
        TransactionalTestExecutionListener.class,
        DbUnitTestExecutionListener.class })
@DatabaseSetup("todoData.xml")
public class ITTodoRepositoryTest {

    @Autowired
    private TodoRepository repository;

    @Test
    @Transactional
    @ExpectedDatabase("toDoData-update.xml")
    public void updateTitle_ShouldUpdateTitle() {
        repository.updateTitle(1L, "FooBar");
    }
}

运行集成测试后，测试通过并更新todo条目的标题。我遇到的唯一问题是版本字段未更新。有什么想法吗？

我不认为这种描述有点模糊。如果您想获得有关为Spring Data JPA存储库编写集成测试的更多信息，可以阅读my blog post about it。

Answer 4

这里的根本问题是JPA的一级缓存。从JPA规范2.2版第3.1节开始。强调是我的：

EntityManager实例与持久性上下文相关联。持久性上下文是一组实体实例，其中对于任何持久性实体标识，都有一个唯一的实体实例。

这很重要，因为JPA会跟踪对该实体的更改，以便将其刷新到数据库。作为副作用，这还意味着在单个持久性上下文中，实体仅加载一次。这就是为什么重新加载更改后的实体没有任何作用的原因。

您可以通过多种方式处理此问题：

从EntityManager驱逐实体。这可以通过调用EntityManager.detach并用@Modifying(clearAutomatically = true)注释退出所有实体的更新方法来完成。确保首先刷新对这些实体的更改，否则最终可能会丢失更改。
使用EntityManager.refresh()。
使用其他持久性上下文加载实体。最简单的方法是在单独的事务中进行。在Spring中，可以通过在未使用@Transactional进行注释的Bean上调用单独的方法来使用@Transactional进行注释的方法来实现。另一种方法是使用TransactionTemplate，它在使事务边界非常明显的测试中效果特别好。

Answer 5

我在尝试执行与您一样的更新查询时遇到了同样的问题

@Modifying
@Transactional
@Query(value = "UPDATE SAMPLE_TABLE st SET st.status=:flag WHERE se.referenceNo in :ids")
public int updateStatus(@Param("flag")String flag, @Param("ids")List<String> references);

如果您在主类上添加了@EnableTransactionManagement注释，这将起作用。 Spring 3.1引入了@EnableTransactionManagement类中要使用的@Configuration批注，并启用了事务支持。

Spring Data JPA Update @Query没有更新？

5 个答案: