熟悉了mybatis的灵活，可能你对他的动态查询很喜欢，表示各种sql都能胜任。初步接触jpa时，你会各种吐槽，不如mybatis来的方便。其实jpa也能帮你完成你的各种需求，至于编写的复杂度，那可能就仁者见仁智者见智了。习惯了，其实也一样了。

代码放github和码云了：spring-data/github spring-data/码云

save操作(含merge操作，即update也在save里)

• save方法会预检查该entity是否持久化，isNew会判断该对象的Id类型 是否实现Persistable或EntityInformation进行重写isNew方法，如果Id是Number类型，直接判断value==0 true 执行entityManager.persist 否则执行entityManager.merge()

  /**
   * save方法会预检查该entity是否持久化，isNew会判断该对象的Id类型 是否实现Persistable或EntityInformation进行
   * 重写isNew方法，如果Id是Number类型，直接判断value==0 true 执行entityManager.persist 否则执行entityManager.merge()
   */
  @Test
  public void insert() {
    Customer customer = new Customer();
    customer.setName("lk");
    customer.setEmail("spring.jpa@163.com");
    customer.setAddress("Shanghai PuDong Area XueYe Road");
    customer.setPhone("13699999999");
    //这里保存以后customer的id会被填充为保存后entity的id
    Customer savedEntity = customerRepository.save(customer);
    //保存并立即刷新数据库，由于customer以及提供id，会执行merge方法进行保存
//    Customer savedAndFlush = customerRepository.saveAndFlush(customer);
    List<Customer> batchCustomers = Arrays.asList(new Customer(), new Customer());
    //批量保存,saveAll是循环单挑插入，并不是batch操作，数据较大使用时请注意性能
//    List<Customer> batchSaves = customerRepository.saveAll(batchCustomers);
  }

delete操作

• delte操作会先执行查询(除了batch操作)，在执行删除，若查询不到结果，抛出异常(EmptyResultDataAccessException)不执行删除

  /**
   * 删除操作，除了batch操作，其他方法均先查询后删除
   */
  @Test
  public void delete() {
    //select * from customer where id=?;delete from customer where id=?;
    //同delete(entity)
    customerRepository.deleteById(38L);
    //select * from customer;循环遍历id单个删除...delete from customer where id=?...
    customerRepository.deleteAll();
    Customer customer = new Customer();
    customer.setId(Long.valueOf(42L));
    Customer customerOther = new Customer();
    customerOther.setId(41L);
    List<Customer> deleteAll = Arrays.asList(customer,customerOther);
    //循环执行delete(entity)
    customerRepository.deleteAll(deleteAll);
    //不查询直接：delete from customer;(风险较大清空表)
    customerRepository.deleteAllInBatch();
    //不查询直接：delete from customer where id=? or id=?
    customerRepository.deleteInBatch(deleteAll);
  }

最常用的query操作

jpa 官方查询关键字

单表字段查询

    //select * from customer;
    customerRepository.findAll();
    //select * from customer where id = 1;
    customerRepository.findById(1L);
    //select * from customer where address = "address";
    customerRepository.findCustomerByAddress("address");
    //select * from customer where name = "lk" and phone = "133";
    customerRepository.findCustomersNameAndPhone("133", "lk");
    //select * from customer where name like '%k';
    customerRepository.findCustomersNameLike("k");
    //select * from customer where name like 'k'; 如果需要模糊查询需要手动拼接 % 连接符
    customerRepository.findCustomersByNameLike("k");
    //select * from customer where name like "%l";
    customerRepository.findCustomersByNameStartingWith("l");
    //select * from customer where name like "%k%";
    customerRepository.findCustomersByNameContains("k");
    / /.....还有很多，不再一一列举......

分页，排序

    //select * from customer order by name desc;
    customerRepository.findAll(Sort.by(Direction.DESC, "name"));
    //select * from customer limit 0,10;
    customerRepository.findAll(PageRequest.of(0, 10));

example查询（场景较少）

    Customer customer = new Customer();
    customer.setAddress("address");
    //select * from customer where address ="address";
    customerRepository.findAll(Example.of(customer));
    customer.setName("lk");
    customer.setPhone("133");
    ExampleMatcher matcher = ExampleMatcher.matching()
        .withMatcher("name", match -> match.contains())
        .withMatcher("phone", match -> match.startsWith());
    //select * from customer where name like '%lk%" and phone like '133%' and address = "address";
    customerRepository.findOne(Example.of(customer, matcher));

namedQuery 也是自定义的@Query的一种

//entity
@Entity(name = "Customer")
@NamedQuery(name = "Customer.findByNameNQ", query = "select c from Customer c where name =?1")
public class Customer {}

//repository定义方法findByNameNQ
List<Customer> findByNameNQ(String name);

//test
//namedQuery:select * from customer where name = "lk";
customerRepository.findByNameNQ("lk");

@Query自定义JQL语句查询，语法跟sql类似，但注意基于entity的命名，如果属性nativeQuery为ture则，必须用原生sql语句

  @Query("select c from Customer c where name like %?1")
  List<Customer> findCustomersNameLike(String name);

  @Query("select c from Customer c where name = :name and phone = :phone")
  List<Customer> findCustomersNameAndPhone(@Param("phone") String phone,
      @Param("name") String name);

  @Query(value = "select * from customer where name =?1",nativeQuery = true)
  List<Customer> findByNameSql(String name);

@Modify 配合 @Query实现 修改部分字段

  @Modifying
  @Query("update Customer c set c.name = :name where c.id = :id")
  int modifyByPhone(@Param("name") String name,@Param("id") Long id);

  @Modifying
  @Query("delete from Customer c  where c.id = ?1")
  int deleteCustomer(Long id);

自定义返回值，基于JQL语法，在拼装返回结果集时，是根据构造函数进行组装的，可以基于接口或者类，要保证属性是entity内的属性。也可以借助@Query，使用 new map()返回map，或者new Class返回想要的结果。

  //repository

  List<NameOnlyI> findCustomersByName(String name);

  List<NameOnly> findByName(String name);

  @Query("select new com.spring.jpa.beans.NameOnly(name,address) from Customer where name = ?1")
  List<NameOnly> findByName4Obj(String name);

  @Query("select new map(name as myname,address as myaddress) from Customer where name = :name")
  List<Map<String, Object>> findByName4Map(@Param("name") String name);

//test

    //基于接口的返回值
    List<NameOnlyI> interfaces = customerRepository.findCustomersByName("lk1");
    //基于类的返回值，如果有两个构造函数会报错，无法解析转换
    List<NameOnly> nameOnlies = customerRepository.findByName("lk1");
    //基于类的返回值，@Query显式声明返回bean
    List<NameOnly> objs = customerRepository.findByName4Obj("lk1");
    //@Query返回map 用as做key，不用as默认key是0，1，2...
    List<Map<String, Object>> maps = customerRepository.findByName4Map("lk1");

复杂关联关系查询 @OneToOne @ManyToMany @ManyToOne @OneToMany

customer 顾客表，和customer_group 多对一
customer_group 顾客分组表
book 书籍表，customer是多对多。
book_detail 书籍详细表，和book是一对一

• 四张表没任何业务，假象出来的，单纯为了验证jpa查询方式

表关系注解参数（@OneToOne @ManyToMany @ManyToOne @OneToMany ）

• Fetch
  抓取是否延迟加载，默认情况一的方为立即加载，多的一方为延迟加载，可以手动指定Fetch.EAGER/Fetch.LAZY
• mappedBy
  关联关系由此方属性维护，可以理解成一对注解使用mappedBy的一方由另一方维护，且必须是注解作用下的属性名。

可以根据关联表的属性作为条件查询，结果同样是根据两次sql查询出来的。通过关联表的属性进行查询时，使用关联 entityName_columnName方式，或者直接使用 _columnName进行查询。如果不想级联查询时，在一方不适用注解即可。

@OneToOne

一对一关联关系，有三种形式存在：

• 两张表共享主键pk，使用@PrimaryKeyJoinColumn来建立关联关系

@Entity
public class A {
  @Id
  private Long id;
  @OneToOne(cascade = CascadeType.ALL)
  @PrimaryKeyJoinColumn
  private B b;
}

@Entity
public class B {
  @Id
  private Long id;
}

• 通过中间表建立关联关系，使用@JoinTable注解，joinColumns指定本表和关联表的外键，inverseJoinColumns指定关联关系另一方和关联表的外键

@Entity
public class A {
  @Id
  private Long id;
  @OneToOne(cascade = CascadeType.ALL)
  joinColumns = @JoinColumn(name="a_fk"),
  inverseJoinColumns = @JoinColumn(name="b_fk")
  private B b;
}

@Entity
public class B {
  @Id
  private Long id;
  @OneToOne(mappedBy="b")
  private A a;
}

• 通过外键，唯一约束指定关联关系，使用@JoinColumn注解，如果不写该注解，默认会在此表中自动创建连接列：主表属性_关联表主键名称

@Entity(name = "book")
public class Book {
  @Id
  @GeneratedValue(strategy = GenerationType.IDENTITY)
  private Long id;

  private String name;
  private Integer count;

  @OneToOne(cascade = CascadeType.ALL)
  @JoinColumn(name = "detail_id",referencedColumnName = "id")
//referencedColumnName 不写默认是主键，当不是主键是可以用此声明，但必须保证连接键在连接表是唯一约束的
  private BookDetail bookDetail;

//setter getter
}

@Entity(name = "book_detail")
public class BookDetail {

  @Id
  @GeneratedValue(strategy = GenerationType.IDENTITY)
  private Long id;
  private String size;
//不需要依赖book可以不写关联属性

//setter getter
}

public interface BookRepository extends JpaRepository<Book, Long> {

  //可以查出关联表的实体，结果是根据两次sql查询出来的，即select * from book where name = ?;->查出detail_id->select * from book_detail where id = ?;
  List<Book> findByName(String name);
  //可以根据关联表的属性作为条件查询，结果同样是根据两次sql查询出来的。通过关联表的属性进行查询时，使用关联 entityName_columnName方式，或者直接使用 _columnName进行查询。
  List<Book> findByNameAndBookDetail_Id(String name, Long id);

  //自定义返回值，一次查询返回结果
  @Query("select new com.spring.jpa.beans.BookResult(b.id as id,b.name as name ,d.size as size) from Book as b left join BookDetail as d on b.bookDetail = d.id")
  List<BookResult> findResults();

}

@ManyToMany

• 通过@ManyToMany 注解定义多对多关系，同时通过 @JoinTable 注解描述关联表和关联条件。

// 维护端注解
@Entity
public class A {
  @Id
  @GeneratedValue(strategy = GenerationType.IDENTITY)
  private Long id;
  @ManyToMany (cascade = CascadeType.REFRESH)
  @JoinTable (
           name =  "a_b" , //关联表名
           inverseJoinColumns =  @JoinColumn (name =  "a_id" ),//被维护端外键
           joinColumns =  @JoinColumn (name =  "b_id" ))//维护端外键被维护端注解
   private B b;

//setter getter
}

@Entity
public class B {
  @Id
  @GeneratedValue(strategy = GenerationType.IDENTITY)
  private Long id;
  @ManyToMany(cascade = CascadeType.REFRESH,
            mappedBy = "b",//通过维护端的属性关联
            fetch = FetchType.LAZY)
// 关系维护端删除时，如果中间表存在些纪录的关联信息，则会删除该关联信息;
// 关系被维护端删除时，如果中间表存在些纪录的关联信息，则会删除失败 .
  private A a;

//setter getter
}

• 默认joinColumn值：关联表名：主表表名 + 下划线 + 从表表名；关联表到主表的外键：主表表名 + 下划线 + 主表中主键列名；关联表到从表的外键名：主表中用于关联的属性名+ 下划线 + 从表的主键列名

@ManyToOne @OneToMany

• 注解和上面都差不多，无非就是谁关联谁

@Entity
@NamedQuery(name = "Customer.findByNameNQ", query = "select c from Customer c where name =?1")
public class Customer {

  @Id
  @GeneratedValue(strategy = GenerationType.IDENTITY)
  private Long id;
  private String name;
  private String email;
  private String address;
  private String phone;

//  //targetEntity 默认是关联实体，若使用接口作为关联实体时，应指明targetEntity的实现类，且接口应继承Serializable，否则无法被解析
//  @ManyToOne(fetch = FetchType.EAGER,targetEntity = CustomerGroup.class)
//  @JoinColumn(name = "group_type",referencedColumnName = "type")
//  private CustomerGroupInterface customerGroup;

  //manyToOne 单向关联或 oneToMany双向关联
//  //这里使用非主键作为外键关联，type在customer_group表中唯一约束，也可以使用@JoinTable 处理关联表做连接这里不再演示
  @ManyToOne(fetch = FetchType.EAGER,cascade = CascadeType.ALL)
  @JoinColumn(name = "group_type",referencedColumnName = "type")
  private CustomerGroup customerGroup;

  //customerGroup为一的一方，单向关联customer（多方）
//  @Column(name = "group_type")
//  private String type;

//setter getter
}

@Entity
@Table(name = "customer_group")
public class CustomerGroup implements CustomerGroupInterface {

  private static final long serialVersionUID = -6956725658881048590L;
  @Id
  @GeneratedValue(strategy = GenerationType.IDENTITY)
  private Long id;
  @Column(unique = true)
  private String type;
  private String name;
  private Integer level;

  //oneToMany 单向关联，customer中无需注解，注意joinColumn是table中的name
//  @OneToMany(fetch = FetchType.EAGER)
//  @JoinColumn(name = "group_type")

  //oneToMany 双向关联，customer中无需注解,mappedBy必须指向entity中的属性名，即标有@ManyToOne的属性名，且此处不可再使用@JoinColumn。
  @OneToMany(fetch = FetchType.EAGER, mappedBy = "customerGroup")
  private List<Customer> customers;

//setter getter
}

其他注解

最后写的比较急躁，代码放github和码云了：spring-data/github spring-data/码云