前面的几篇文章分别介绍了CURD中的增删改,接下来进入最最常见的查询篇,看一下使用jpa进行db的记录查询时,可以怎么玩
本篇将介绍一些基础的查询使用姿势,主要包括根据字段查询,and/or/in/like/between
语句,数字比较,排序以及分页
I. 环境准备
在开始之前,当然得先准备好基础环境,如安装测试使用mysql,创建SpringBoot项目工程,设置好配置信息等,关于搭建项目的详情可以参考前一篇文章
下面简单的看一下演示添加记录的过程中,需要的配置
1. 表准备
沿用前一篇的表,结构如下
CREATE TABLE `money` (
`id` int(11) unsigned NOT NULL AUTO_INCREMENT,
`name` varchar(20) NOT NULL DEFAULT '' COMMENT '用户名',
`money` int(26) NOT NULL DEFAULT '0' COMMENT '钱',
`is_deleted` tinyint(1) NOT NULL DEFAULT '0',
`create_at` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP COMMENT '创建时间',
`update_at` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP COMMENT '更新时间',
PRIMARY KEY (`id`),
KEY `name` (`name`)
) ENGINE=InnoDB AUTO_INCREMENT=1 DEFAULT CHARSET=utf8mb4;
2. 项目配置
配置信息,与之前有一点点区别,我们新增了更详细的日志打印;本篇主要目标集中在添加记录的使用姿势,对于配置说明,后面单独进行说明
## DataSource
spring.datasource.url=jdbc:mysql://127.0.0.1:3306/story?useUnicode=true&characterEncoding=UTF-8&useSSL=false
spring.datasource.driver-class-name=com.mysql.jdbc.Driver
spring.datasource.username=root
spring.datasource.password=
## jpa相关配置
spring.jpa.database=MYSQL
spring.jpa.hibernate.ddl-auto=none
spring.jpa.show-sql=true
spring.jackson.serialization.indent_output=true
spring.jpa.hibernate.naming.physical-strategy=org.hibernate.boot.model.naming.PhysicalNamingStrategyStandardImpl
3. 数据准备
数据修改嘛,所以我们先向表里面插入两条数据,用于后面的操作
INSERT INTO `money` (`id`, `name`, `money`, `is_deleted`, `create_at`, `update_at`)
VALUES
(1, '一灰灰blog', 100, 0, '2019-04-18 17:01:40', '2019-04-18 17:01:40'),
(2, '一灰灰2', 200, 0, '2019-04-18 17:01:40', '2019-04-18 17:01:40'),
(3, '一灰灰3', 300, 0, '2019-04-18 17:01:40', '2019-04-18 17:01:40'),
(4, '一灰灰4', 400, 0, '2019-04-18 17:01:40', '2019-04-18 17:01:40'),
(5, '一灰灰5', 500, 0, '2019-04-18 17:01:40', '2019-04-18 17:01:40'),
(6, 'Batch 一灰灰blog', 100, 0, '2019-04-18 17:01:40', '2019-04-18 17:01:40'),
(7, 'Batch 一灰灰blog 2', 100, 0, '2019-04-18 17:01:40', '2019-04-18 17:01:40'),
(8, 'Batch 一灰灰 3', 200, 0, '2019-04-18 17:01:40', '2019-04-18 17:01:40'),
(9, 'Batch 一灰灰 4', 200, 0, '2019-04-18 17:01:40', '2019-04-18 17:01:40'),
(10, 'batch 一灰灰5', 1498, 0, '2019-04-18 17:01:40', '2019-04-18 17:01:58'),
(11, 'batch 一灰灰6', 1498, 0, '2019-04-18 17:01:40', '2019-04-18 17:01:58'),
(12, 'batch 一灰灰7', 400, 0, '2019-04-18 17:01:40', '2019-04-18 17:01:40'),
(13, 'batch 一灰灰8', 400, 0, '2019-04-18 17:01:40', '2019-04-18 17:01:40');
II. Query基本使用姿势
下面进入简单的查询操作姿势介绍,单表的简单and/or/in/compare查询方式
1. 表关联POJO
查询返回的记录与一个实体类POJO进行绑定,借助前面的分析结果,如下
@Data
@DynamicUpdate
@DynamicInsert
@Entity
@Table(name = "money")
public class MoneyPO {
@Id
// 如果是auto,则会报异常 Table 'mysql.hibernate_sequence' doesn't exist
// @GeneratedValue(strategy = GenerationType.AUTO)
@GeneratedValue(strategy = GenerationType.IDENTITY)
@Column(name = "id")
private Integer id;
@Column(name = "name")
private String name;
@Column(name = "money")
private Long money;
@Column(name = "is_deleted")
private Byte isDeleted;
@Column(name = "create_at")
@CreatedDate
private Timestamp createAt;
@Column(name = "update_at")
@CreatedDate
private Timestamp updateAt;
}
上面类中的几个注解,说明如下
@Data
属于lombok注解,与jpa无关,自动生成getter/setter/equals/hashcode/tostring
等方法@Entity
,@Table
jpa注解,表示这个类与db的表关联,具体匹配的是表money
@Id
@GeneratedValue
作用与自增主键@Column
表明这个属性与表中的某列对应@CreateDate
根据当前时间来生成默认的时间戳
2. Repository API声明
接下来我们新建一个api继承自CurdRepository
,然后通过这个api来与数据库打交道,后面会在这个类中添加较多的查询方法
public interface MoneyBaseQueryRepository extends CrudRepository<MoneyPO, Integer> {
}
3. 使用姿势
a. 根据id查询
CrudRepository
已经提供的功能,根据主键id进行查询,对于使用者而言,没有什么需要额外操作的,直接访问即可
private void queryById() {
// 根据主键查询,直接使用接口即可
Optional<MoneyPO> res = moneyCurdRepository.findById(1);
System.out.println("queryById return: " + res.get());
}
b. 根据字段查询
除了根据主键查询,实际的业务场景中,根据某个字段进行查询的case,简直不要更多,在jpa中可以怎么做呢?
- 在
Repository
接口中声明一个方法,命名规则为 - findByXXX 或者 queryByXXX (注意这里的xxx用POJO中的成员名替换,表示根据这个成员进行查询)
一个简单的case,如果我希望实现根据name进行查询,那么在MoneyBaseQueryRepository
中添加下面两个方法中的任意一个都可以
/**
* 根据用户名查询
*
* @param name
* @return
*/
List<MoneyPO> findByName(String name);
List<MoneyPO> queryByName(String name);
如果需要多个成员的查询呢?也简单,形如findByXxxAndYyyy
相当于sql中的where xxxx=? and yyy=?
如我们也可以增加下面两个方法(一个and、一个or查询)
/**
* 根据用户名 + money查询
*
* @param name
* @param money
* @return
*/
List<MoneyPO> findByNameAndMoney(String name, Long money);
/**
* 根据用户名 or id查询
*
* @param name
* @param id
* @return
*/
List<MoneyPO> findByNameOrId(String name, Integer id);
一个简单的测试case可以如下
private void queryByField() {
// 根据内部成员进行查询,需要自己定义新的接口
String name = "一灰灰blog";
Iterable<MoneyPO> res = moneyCurdRepository.findByName(name);
System.out.println("findByName return: " + res);
res = moneyCurdRepository.queryByName(name);
System.out.println("queryByName return: " + res);
Long money = 100L;
res = moneyCurdRepository.findByNameAndMoney(name, money);
System.out.println("findByNameAndMoney return: " + res);
Integer id = 5;
res = moneyCurdRepository.findByNameOrId(name, id);
System.out.println("findByNameOrId return: " + res);
}
执行之后输出结果如下,下面也包括了对应的sql,便于理解
Hibernate: select moneypo0_.id as id1_0_, moneypo0_.create_at as create_a2_0_, moneypo0_.is_deleted as is_delet3_0_, moneypo0_.money as money4_0_, moneypo0_.name as name5_0_, moneypo0_.update_at as update_a6_0_ from money moneypo0_ where moneypo0_.name=?
findByName return: [MoneyPO(id=1, name=一灰灰blog, money=100, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0)]
-------- 人工拆分 -----------
Hibernate: select moneypo0_.id as id1_0_, moneypo0_.create_at as create_a2_0_, moneypo0_.is_deleted as is_delet3_0_, moneypo0_.money as money4_0_, moneypo0_.name as name5_0_, moneypo0_.update_at as update_a6_0_ from money moneypo0_ where moneypo0_.name=?
queryByName return: [MoneyPO(id=1, name=一灰灰blog, money=100, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0)]
-------- 人工拆分 -----------
Hibernate: select moneypo0_.id as id1_0_, moneypo0_.create_at as create_a2_0_, moneypo0_.is_deleted as is_delet3_0_, moneypo0_.money as money4_0_, moneypo0_.name as name5_0_, moneypo0_.update_at as update_a6_0_ from money moneypo0_ where moneypo0_.name=? and moneypo0_.money=?
findByNameAndMoney return: [MoneyPO(id=1, name=一灰灰blog, money=100, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0)]
-------- 人工拆分 -----------
Hibernate: select moneypo0_.id as id1_0_, moneypo0_.create_at as create_a2_0_, moneypo0_.is_deleted as is_delet3_0_, moneypo0_.money as money4_0_, moneypo0_.name as name5_0_, moneypo0_.update_at as update_a6_0_ from money moneypo0_ where moneypo0_.name=? or moneypo0_.id=?
findByNameOrId return: [MoneyPO(id=1, name=一灰灰blog, money=100, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0), MoneyPO(id=5, name=一灰灰5, money=500, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0)]
c. like查询
上面的查询方式为等值查询,当在sql中除了等值查询(即=查询)之外,还有各种比较查询,不等查询以及like语句,在jpa中也比较简单,在repository
定义的方法名,加一个like即可
/**
* like查询
*
* @param name
* @return
*/
List<MoneyPO> findByNameLike(String name);
使用的时候,需要稍微注意一下,根据实际情况决定要不要加上 '%'
private void queryByLike() {
// like 语句查询
String name = "一灰灰%";
Iterable<MoneyPO> res = moneyCurdRepository.findByNameLike(name);
System.out.println("findByName like: " + res);
}
输出结果为
Hibernate: select moneypo0_.id as id1_0_, moneypo0_.create_at as create_a2_0_, moneypo0_.is_deleted as is_delet3_0_, moneypo0_.money as money4_0_, moneypo0_.name as name5_0_, moneypo0_.update_at as update_a6_0_ from money moneypo0_ where moneypo0_.name like ?
findByName like: [MoneyPO(id=1, name=一灰灰blog, money=100, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0), MoneyPO(id=2, name=一灰灰2, money=200, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0), MoneyPO(id=3, name=一灰灰3, money=300, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0), MoneyPO(id=4, name=一灰灰4, money=400, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0), MoneyPO(id=5, name=一灰灰5, money=500, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0)]
d. in查询
对于in查询,CurdRepository
提供了根据主键id的查询方式,直接调用findAllById
即可,如果是其他的,可以通过声明一个接口的方式来支持
/**
* in查询
*
* @param moneys
* @return
*/
List<MoneyPO> findByMoneyIn(List<Long> moneys);
测试case如下
// in 查询
List<Integer> ids = Arrays.asList(1, 2, 3);
Iterable<MoneyPO> res = moneyCurdRepository.findAllById(ids);
System.out.println("findByIds return: " + res);
res = moneyCurdRepository.findByMoneyIn(Arrays.asList(400L, 300L));
System.out.println("findByMoneyIn return: " + res);
输出结果
Hibernate: select moneypo0_.id as id1_0_, moneypo0_.create_at as create_a2_0_, moneypo0_.is_deleted as is_delet3_0_, moneypo0_.money as money4_0_, moneypo0_.name as name5_0_, moneypo0_.update_at as update_a6_0_ from money moneypo0_ where moneypo0_.id in (? , ? , ?)
findByIds return: [MoneyPO(id=1, name=一灰灰blog, money=100, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0), MoneyPO(id=2, name=一灰灰2, money=200, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0), MoneyPO(id=3, name=一灰灰3, money=300, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0)]
------ 手动拆分 ----------
Hibernate: select moneypo0_.id as id1_0_, moneypo0_.create_at as create_a2_0_, moneypo0_.is_deleted as is_delet3_0_, moneypo0_.money as money4_0_, moneypo0_.name as name5_0_, moneypo0_.update_at as update_a6_0_ from money moneypo0_ where moneypo0_.money in (? , ?)
findByMoneyIn return: [MoneyPO(id=3, name=一灰灰3, money=300, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0), MoneyPO(id=4, name=一灰灰4, money=400, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0), MoneyPO(id=12, name=batch 一灰灰7, money=400, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0), MoneyPO(id=13, name=batch 一灰灰8, money=400, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0)]
e. 比较查询
数字的比较查询,比如大于等于,大于,小于,小于等于,between,下面的三个方法声明,应该能直观表示这种方式可以如何写
/**
* 查询大于or等于指定id的所有记录
*
* @param id
* @return
*/
List<MoneyPO> findByIdGreaterThanEqual(Integer id);
/**
* 查询小于or等于指定id的所有记录
*
* @param id
* @return
*/
List<MoneyPO> findByIdLessThanEqual(Integer id);
/**
* between查询
*
* @param low
* @param high
* @return
*/
List<MoneyPO> findByIdIsBetween(Integer low, Integer high);
下面是简单的映射关系
>
:xxGreaterThan
>=
:xxGreaterThanEqual
<
:xxLessThan
<=
:xxLessThanEqual
!=
:xxNot
between a and b
:xxIsBetween
测试case如下
private void queryByCompare() {
Integer id1 = 3;
Iterable<MoneyPO> res = moneyCurdRepository.findByIdLessThanEqual(id1);
System.out.println("findByIdLessThan 3 return: " + res);
Integer id2 = 10;
res = moneyCurdRepository.findByIdGreaterThanEqual(id2);
System.out.println("findByIdGreaterThan 10 return: " + res);
id1 = 4;
id2 = 6;
res = moneyCurdRepository.findByIdIsBetween(id1, id2);
System.out.println("findByIdsWBetween 3, 10 return: " + res);
}
输出结果为
Hibernate: select moneypo0_.id as id1_0_, moneypo0_.create_at as create_a2_0_, moneypo0_.is_deleted as is_delet3_0_, moneypo0_.money as money4_0_, moneypo0_.name as name5_0_, moneypo0_.update_at as update_a6_0_ from money moneypo0_ where moneypo0_.id<=?
findByIdLessThan 3 return: [MoneyPO(id=1, name=一灰灰blog, money=100, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0), MoneyPO(id=2, name=一灰灰2, money=200, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0), MoneyPO(id=3, name=一灰灰3, money=300, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0)]
------ 手动拆分 ----------
Hibernate: select moneypo0_.id as id1_0_, moneypo0_.create_at as create_a2_0_, moneypo0_.is_deleted as is_delet3_0_, moneypo0_.money as money4_0_, moneypo0_.name as name5_0_, moneypo0_.update_at as update_a6_0_ from money moneypo0_ where moneypo0_.id>=?
findByIdGreaterThan 10 return: [MoneyPO(id=10, name=batch 一灰灰5, money=1498, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:58.0), MoneyPO(id=11, name=batch 一灰灰6, money=1498, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:58.0), MoneyPO(id=12, name=batch 一灰灰7, money=400, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0), MoneyPO(id=13, name=batch 一灰灰8, money=400, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0)]
------ 手动拆分 ----------
Hibernate: select moneypo0_.id as id1_0_, moneypo0_.create_at as create_a2_0_, moneypo0_.is_deleted as is_delet3_0_, moneypo0_.money as money4_0_, moneypo0_.name as name5_0_, moneypo0_.update_at as update_a6_0_ from money moneypo0_ where moneypo0_.id between ? and ?
findByIdsWBetween 3, 10 return: [MoneyPO(id=4, name=一灰灰4, money=400, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0), MoneyPO(id=5, name=一灰灰5, money=500, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0), MoneyPO(id=6, name=Batch 一灰灰blog, money=100, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0)]
f. 排序
排序也属于基本查询的case了,jpa的实现中,通过加上OrderByXxxAsc/Desc
的方式来决定根据什么进行升序or降序
/**
* 根据money查询,并将最终的结果根据id进行倒排
*
* @param money
* @return
*/
List<MoneyPO> findByMoneyOrderByIdDesc(Long money);
/**
* 根据多个条件进行排序
*
* @param id
* @return
*/
List<MoneyPO> queryByIdGreaterThanEqualOrderByMoneyDescIdAsc(Integer id);
在根据多个列进行排序时,需要注意的是不能写多个 OrderBy
而是直接在OrderBy
后面加上对应的xxxAscyyyDesc
测试代码如
private void queryWithSort() {
// 排序
Long money = 400L;
Iterable<MoneyPO> res = moneyCurdRepository.findByMoneyOrderByIdDesc(money);
System.out.println("findByMoneyAndOrderByIdDesc return: " + res);
Integer startId = 7;
res = moneyCurdRepository.queryByIdGreaterThanEqualOrderByMoneyDescIdAsc(startId);
System.out.println("queryByIdGreaterThanEqualOrderByMoneyDescIdAsc return: " + res);
}
输出结果如下
Hibernate: select moneypo0_.id as id1_0_, moneypo0_.create_at as create_a2_0_, moneypo0_.is_deleted as is_delet3_0_, moneypo0_.money as money4_0_, moneypo0_.name as name5_0_, moneypo0_.update_at as update_a6_0_ from money moneypo0_ where moneypo0_.money=? order by moneypo0_.id desc
findByMoneyAndOrderByIdDesc return: [MoneyPO(id=13, name=batch 一灰灰8, money=400, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0), MoneyPO(id=12, name=batch 一灰灰7, money=400, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0), MoneyPO(id=4, name=一灰灰4, money=400, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0)]
------------- 人工拆分 --------
Hibernate: select moneypo0_.id as id1_0_, moneypo0_.create_at as create_a2_0_, moneypo0_.is_deleted as is_delet3_0_, moneypo0_.money as money4_0_, moneypo0_.name as name5_0_, moneypo0_.update_at as update_a6_0_ from money moneypo0_ where moneypo0_.id>=? order by moneypo0_.money desc, moneypo0_.id asc
queryByIdGreaterThanEqualOrderByMoneyDescIdAsc return: [MoneyPO(id=10, name=batch 一灰灰5, money=1498, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:58.0), MoneyPO(id=11, name=batch 一灰灰6, money=1498, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:58.0), MoneyPO(id=12, name=batch 一灰灰7, money=400, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0), MoneyPO(id=13, name=batch 一灰灰8, money=400, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0), MoneyPO(id=8, name=Batch 一灰灰 3, money=200, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0), MoneyPO(id=9, name=Batch 一灰灰 4, money=200, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0), MoneyPO(id=7, name=Batch 一灰灰blog 2, money=100, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0)]
g. 分页查询
分页有两种方式,一个是查询最大的多少条数据,一个是正常的limit/offset方式,下面是一个简单的实例demo
/**
* 分页查询,获取前面三个数据
*
* @param id
* @return
*/
List<MoneyPO> findTop3ByIdGreaterThan(Integer id);
/**
* 分页查询
*
* @param id
* @param pageable page 从0开始表示查询第0页,即返回size个正好>id数量的数据
* @return
*/
List<MoneyPO> findByIdGreaterThan(Integer id, Pageable pageable);
对于分页而言,通过传入参数Pageable来表明即可
测试case如
private void queryWithPageSize() {
// 分页查询
Iterable<MoneyPO> res = moneyCurdRepository.findTop3ByIdGreaterThan(3);
System.out.println("findTop3ByIdGreaterThan 3 return: " + res);
// id>3,第2页,每页3条,如果id递增时,则返回的第一条id=4 + 2 * 3 = 10
res = moneyCurdRepository.findByIdGreaterThan(3, PageRequest.of(2, 3));
System.out.println("findByIdGreaterThan 3 pageIndex 2 size 3 return: " + res);
}
输出结果为
Hibernate: select moneypo0_.id as id1_0_, moneypo0_.create_at as create_a2_0_, moneypo0_.is_deleted as is_delet3_0_, moneypo0_.money as money4_0_, moneypo0_.name as name5_0_, moneypo0_.update_at as update_a6_0_ from money moneypo0_ where moneypo0_.id>? limit ?
findTop3ByIdGreaterThan 3 return: [MoneyPO(id=4, name=一灰灰4, money=400, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0), MoneyPO(id=5, name=一灰灰5, money=500, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0), MoneyPO(id=6, name=Batch 一灰灰blog, money=100, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0)]
---------- 人工拆分 ------------
Hibernate: select moneypo0_.id as id1_0_, moneypo0_.create_at as create_a2_0_, moneypo0_.is_deleted as is_delet3_0_, moneypo0_.money as money4_0_, moneypo0_.name as name5_0_, moneypo0_.update_at as update_a6_0_ from money moneypo0_ where moneypo0_.id>? limit ?, ?
findByIdGreaterThan 3 pageIndex 2 size 3 return: [MoneyPO(id=10, name=batch 一灰灰5, money=1498, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:58.0), MoneyPO(id=11, name=batch 一灰灰6, money=1498, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:58.0), MoneyPO(id=12, name=batch 一灰灰7, money=400, isDeleted=0, createAt=2019-04-18 17:01:40.0, updateAt=2019-04-18 17:01:40.0)]
4. 小结
主要介绍了jpa的查询的最基本使用方式,主要是根据规则定义方法名的方式来实现sql的效果, 下表示一个简单的对比小结
方法名 | 说明 | 等效sql |
---|---|---|
findByXxx | 表示根据列Xxx 等于传参构建sql | where xxx= ? |
findByXxxAndYyy | 根据多个列进行查询 | where xxx=? and yyy=? |
findByXxxOrYyy | 根据多个列实现or查询 | where xxx=? or yyy=? |
findByXxxLike | like查询,需要注意查询条件中加% | where xxx like |
findByXxxIn | in查询 | where Xxx in () |
findByXxxGreaterThan | 大于 | where xxx > ? |
findByXxxGreaterThanEqual | 大于等于 | where xxx >= ? |
findByXxxLessThan | 小于 | where xxx < ? |
findByXxxLessThanEqual | 小于等于 | where xxx <= ? |
findByXxxNot | 不等于 | where xxx != ? |
findByXxxIsBetween | between查询 | where xxx between ? and ? |
OrderByXxxDesc | 排序 | order by xxx desc |
topN | 分页,表示获取最前面的n条 | limit n |
此外还有一个分页的方式是传参Pageable
,来指定具体的分页
我们常见的查询操作中,除了上面的一些case之外,还有一些是我们没有提到的,如下面的一些使用姿势,则会在后面的文章中引入
group by
distinct
join
- 各种函数的支持(sum,max,min,avg...)
- 查询部分表中部分字段时
- 统计查询
II. 其他
0. 源码&相关博文
源码
- 工程:https://github.com/liuyueyi/spring-boot-demo
- module: https://github.com/liuyueyi/spring-boot-demo/blob/master/spring-boot/102-jpa
相关博文
- mysql之锁与事务详解
- Spring学习之事务的使用姿势
- Spring学习之事务管理与传播属性
- 190612-SpringBoot系列教程JPA之基础环境搭建
- 190614-SpringBoot系列教程JPA之新增记录使用姿势
- 190623-SpringBoot系列教程JPA之update使用姿势
- 190702-SpringBoot系列教程JPA之delete使用姿势详解
1. 一灰灰Blog
尽信书则不如,以上内容,纯属一家之言,因个人能力有限,难免有疏漏和错误之处,如发现bug或者有更好的建议,欢迎批评指正,不吝感激
下面一灰灰的个人博客,记录所有学习和工作中的博文,欢迎大家前去逛逛
- 一灰灰Blog个人博客 https://blog.hhui.top
- 一灰灰Blog-Spring专题博客 http://spring.hhui.top