JPA entityManager.merge将LocalDateTime转换为SQLServer 2012 DATETIME2

Question

我有一个DateTime列作为主键的表:

USE [idatest]
GO

CREATE TABLE [dbo].[DatesTbl](
    [creationDate] [datetime] NOT NULL
 CONSTRAINT [PK_DatesTbl] PRIMARY KEY CLUSTERED
(
    [creationDate] ASC
))
GO

当我在做entityManager.merge时,我得到重复的PK违规,因为datetime为milisec保存了3位数,但是hibernet将它转换为datetime2,它为milisec保存7位数.在java代码中,我使用LocaDatetime,它为milsec保存10位数.

我已经尝试过在Hibernate MSSQL datetime2映射中解释的解决方案, 但它不起作用:java代码如下所示:pom.xml

<parent>
    <groupId>org.springframework.boot</groupId>
    <artifactId>spring-boot-starter-parent</artifactId>
    <version>2.0.3.RELEASE</version>
</parent>

<groupId>com.example</groupId>
<artifactId>spring-jap-test</artifactId>
<version>1.0-SNAPSHOT</version>
<build>
    <plugins>
        <plugin>
            <groupId>org.apache.maven.plugins</groupId>
            <artifactId>maven-compiler-plugin</artifactId>
            <configuration>
                <source>1.8</source>
                <target>1.8</target>
            </configuration>
        </plugin>
    </plugins>
</build>

<dependencies>
    <dependency>
        <groupId>org.hibernate</groupId>
        <artifactId>hibernate-entitymanager</artifactId>
    </dependency>
    <dependency>
        <groupId>com.microsoft.sqlserver</groupId>
        <artifactId>mssql-jdbc</artifactId>
        <version>7.0.0.jre8</version>
    </dependency>
    <dependency>
        <groupId>org.springframework.data</groupId>
        <artifactId>spring-data-jpa</artifactId>
    </dependency>
    <dependency>
        <groupId>org.projectlombok</groupId>
        <artifactId>lombok</artifactId>
    </dependency>
</dependencies>

DatesTbl类

@Entity
@NoArgsConstructor
@AllArgsConstructor
public class DatesTbl {

    @Column(columnDefinition = "DATETIME", nullable = false)
    @Id
    private LocalDateTime creationDate;
}

主要课程

@EnableTransactionManagement
public class Main {

    public static void main(String[] args) {

        ApplicationContext context = new AnnotationConfigApplicationContext(Main.class);

        EntityManagerFactory entityManagerFactory = context.getBean(EntityManagerFactory.class);
        final EntityManager entityManager = entityManagerFactory.createEntityManager();
        final LocalDateTime creationDate = LocalDateTime.of(2018, 12, 26, 8, 10, 40, 340);
        entityManager.getTransaction().begin();
        final DatesTbl datesTbl = entityManager.merge(new DatesTbl(creationDate));
        entityManager.getTransaction().commit();

        System.out.println("test");
    }

    @Bean
    @Primary
    public DataSource getDataSource() {

        SQLServerDataSource ds = null;
        try {
            ds = new SQLServerDataSource();
            ds.setServerName("localhost");
            ds.setDatabaseName("idatest");
            ds.setIntegratedSecurity(true);
        } catch (Exception ex) {
            System.out.println(ex.getMessage());
        }
        return ds;
    }


    @Bean
    public JpaVendorAdapter jpaVendorAdapter() {
        HibernateJpaVendorAdapter hibernateJpaVendorAdapter = new HibernateJpaVendorAdapter();
        hibernateJpaVendorAdapter.setShowSql(true);
        hibernateJpaVendorAdapter.setGenerateDdl(true);
        hibernateJpaVendorAdapter.setDatabase(Database.SQL_SERVER);
        return hibernateJpaVendorAdapter;
    }


    @Bean
    public LocalContainerEntityManagerFactoryBean abstractEntityManagerFactoryBean(
            JpaVendorAdapter jpaVendorAdapter) {

        Properties properties = new Properties();
         properties.setProperty(FORMAT_SQL, String.valueOf(true));
        properties.setProperty(SHOW_SQL, String.valueOf(true));
        properties.setProperty(DIALECT, ModifiedSQLServerDialect.class.getTypeName());
        LocalContainerEntityManagerFactoryBean localContainerEntityManagerFactoryBean =
                new LocalContainerEntityManagerFactoryBean();

        localContainerEntityManagerFactoryBean.setDataSource(getDataSource());
        localContainerEntityManagerFactoryBean.setJpaVendorAdapter(jpaVendorAdapter);
        localContainerEntityManagerFactoryBean.setJpaProperties(properties);
        localContainerEntityManagerFactoryBean.setPackagesToScan("enteties");

        return localContainerEntityManagerFactoryBean;
    }


    @Bean
    public PlatformTransactionManager platformTransactionManager(EntityManagerFactory emf) {
        return new JpaTransactionManager(emf);
    }
}


public class ModifiedSQLServerDialect extends SQLServer2012Dialect {


    public ModifiedSQLServerDialect () {
        super();
        registerColumnType(Types.TIMESTAMP, "timestamp");
        registerColumnType(Types.DATE, "timestamp");
        registerColumnType(Types.TIME, "timestamp");
        registerHibernateType(Types.TIMESTAMP, "timestamp");
        registerHibernateType(Types.DATE, "timestamp");
        registerHibernateType(Types.TIME, "timestamp");
    }
}

但我仍然在SQLServer探查器中看到:

exec sp_executesql N'select datestbl0_.creationDate as creation1_0_0_ from DatesTbl datestbl0_ where datestbl0_.creationDate=@P0        ',N'@P0 `datetime2`','2018-12-26 08:10:40.0000003'

解决方案有什么问题？

Answer 1

解释

该问题与mssql-jdbc（版本4.x和6.x）中的一个问题有关，PreparedStatement.setTimestamp（index，timestamp，calendar）存在数据类型转换问题，它总是将LocalDateTime具有datetime2数据类型的参数发送到SQL服务器（忽略）表的列类型）。由于datetime（0.00333秒）和datetime2（100纳秒）的精度不同，并且datetime用作PK，Hibernate在这种情况下工作错误。

当我们运行主程序时，其creationDate值为2018-12-26 08:10:40.000000340，并且该值在数据库中保存为2018-12-26 08:10:40.000，因为 Hibernate 在数据库中找不到具有相同键的记录。当我们再次运行主程序时，Hibernate 首先检查是否有具有相同键的记录，使用

'从 DatesTbldatestbl0_ 选择datestbl0_.creationDate作为creation1_0_0_，其中datestbl0_.creationDate=@P0'，N'@P0'datetime2''，'2018-12-26 08:10:40.0000003'

看来SQL Server将datetime表中的值向上转换来datetime2进行比较，并且没有返回任何记录。因此Hibernate再次插入记录，并导致主键冲突。

解决方法

正如 Vlad Mihalcea 所建议的，使用 DATETIME 列作为 PK 并不是一个好主意。
但是，假设我们仍然需要该datetime列作为 PK，则以下解决方法应该有效。解决这个问题的关键是让datetime和之间的比较datetime2返回true。为了实现这一点，我们可以在传递给数据库之前将datetime2值截断/舍入为相应的datetime值。对主程序的以下更改使用 SQL Server 2012 Express 进行测试，没有错误。

public static void main(String[] args) {
    ApplicationContext context = new AnnotationConfigApplicationContext(Main.class);

    EntityManagerFactory entityManagerFactory = context.getBean(EntityManagerFactory.class);
    final EntityManager entityManager = entityManagerFactory.createEntityManager();

    LocalDateTime creationDate0 = LocalDateTime.of(2018, 12, 26, 8, 10, 40, 341340340);
    LocalDateTime creationDate3 = LocalDateTime.of(2018, 12, 26, 8, 10, 40, 343340340);
    LocalDateTime creationDate7 = LocalDateTime.of(2018, 12, 26, 8, 10, 40, 346670340);
    LocalDateTime creationDate10 = LocalDateTime.of(2018, 12, 26, 8, 10, 40, 349670340);
    entityManager.getTransaction().begin();
    final DatesTbl datesTbl0 = entityManager.merge(new DatesTbl(roundNanoSecForDateTime(creationDate0)));
    final DatesTbl datesTbl3 = entityManager.merge(new DatesTbl(roundNanoSecForDateTime(creationDate3)));
    final DatesTbl datesTbl7 = entityManager.merge(new DatesTbl(roundNanoSecForDateTime(creationDate7)));
    final DatesTbl datesTbl10 = entityManager.merge(new DatesTbl(roundNanoSecForDateTime(creationDate10)));
    entityManager.getTransaction().commit();
    System.out.println("test");
}

private static LocalDateTime roundNanoSecForDateTime(LocalDateTime localDateTime) {
    int nanoSec = localDateTime.getNano();
    // The rounding is based on following results on SQL server 2012 express
    // select cast(cast('2018-12-26 08:10:40.3414999' as datetime2) as datetime);
    // 2018-12-26 08:10:40.340
    // select cast(cast('2018-12-26 08:10:40.3415000' as datetime2) as datetime);
    // select cast(cast('2018-12-26 08:10:40.3444999' as datetime2) as datetime);
    // 2018-12-26 08:10:40.343
    // select cast(cast('2018-12-26 08:10:40.3445000' as datetime2) as datetime);
    // select cast(cast('2018-12-26 08:10:40.3484999' as datetime2) as datetime);
    // 2018-12-26 08:10:40.347
    // select cast(cast('2018-12-26 08:10:40.3485000' as datetime2) as datetime);
    // 2018-12-26 08:10:40.350
    int last7DigitOfNano = nanoSec - (nanoSec / 10000000) * 10000000;
    int roundedNanoSec = 0;
    if (last7DigitOfNano < 1500000) {
        roundedNanoSec = nanoSec - last7DigitOfNano;
    } else if (last7DigitOfNano < 4500000) {
        roundedNanoSec = nanoSec - last7DigitOfNano + 3000000;
    } else if (last7DigitOfNano < 8500000) {
        roundedNanoSec = nanoSec - last7DigitOfNano + 7000000;
    } else {
        roundedNanoSec = nanoSec - last7DigitOfNano + 10000000;
    }
    System.out.println("Before Rounding" + nanoSec);
    System.out.println("After Rounding" + roundedNanoSec);
    return localDateTime.withNano(roundedNanoSec);
}

参考：
1. SQL Server 中的 DateTime2 与 DateTime
2.日期和时间数据类型和函数 (Transact-SQL)