Specifications 的思路来自于“领域驱动设计”的概念,通过可编程的方式实现查询的where语句。我们今天就来写一个Specification的例子。

一、首先建立一个Spring Boot 工程#

我们使用的IDE是InteliJ IDEA,开发语言是Kotlin,
项目信息
Spring 的工程特性选择了JPA、Rest Repository和一种数据库支持(这里是MySql)
项目工程特性
新建项目没有什么可说的,接下来我们正式进入编码:

二、编写一个实体类来记录数据#

我们使用一个Record实体来保存数据,包含了名称和值,代码如下:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
package cn.techcave.demo.jpa3.domain

import javax.persistence.Entity
import javax.persistence.GeneratedValue
import javax.persistence.GenerationType
import javax.persistence.Id

@Entity
class Record(
/**
* 主键
*/
@Id
@GeneratedValue(strategy = GenerationType.IDENTITY)
var id:Long? = null,
/**
* 名称
*/
var name:String? = null,
/**
* 值
*/
var value:Double? = null
)

三、通过接口实现Specification的支持#

要使我们的Repository支持Specification查询,需要在Repository中继承JpaSpecificationExecutor接口,代码如下:

1
2
3
4
5
6
7
8
9
10
package cn.techcave.demo.jpa3.repository

import cn.techcave.demo.jpa3.domain.Record
import org.springframework.data.jpa.repository.JpaRepository
import org.springframework.data.jpa.repository.JpaSpecificationExecutor
import org.springframework.stereotype.Repository

@Repository
interface RecordRepo:JpaRepository<Record, Long>, JpaSpecificationExecutor<Record>{
}

四、通过JPA 2 的criteria API实现Specifications查询#

Repository支持了Specifications查询,我们在代码中实现查询,如下代码所示:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
package cn.techcave.demo.jpa3.service

import cn.techcave.demo.jpa3.domain.Record
import cn.techcave.demo.jpa3.repository.RecordRepo
import org.springframework.beans.factory.annotation.Autowired
import org.springframework.stereotype.Service
import java.time.LocalDate
import javax.persistence.criteria.CriteriaBuilder
import javax.persistence.criteria.Root
import org.springframework.data.jpa.domain.Specification
import org.springframework.data.jpa.domain.Specifications.where
import javax.persistence.criteria.CriteriaQuery
import javax.persistence.criteria.Predicate


@Service
class RecordService {
@Autowired
lateinit var recordRepo: RecordRepo

fun nameLike(name:String): Specification<Record> {
return object : Specification<Record> {
override fun toPredicate(root: Root<Record>, query: CriteriaQuery<*>,
builder: CriteriaBuilder): Predicate {
return builder.like(root.get("name"),"%$name%")
}
}
}

fun filterByName(name:String) = recordRepo.findAll(nameLike(name))

fun filterByNames(n1:String, n2:String) :List<Record> {
return recordRepo.findAll(where(nameLike(n1)).or(nameLike(n2)))
}
fun findAll() = recordRepo.findAll()
}

上面的代码中,我们首先编写nameLike方法,返回一个基于名称的Specification查询。然后我们在filterByName中执行此查询。Specification还可以进行组合条件的查询,例如filterByNames方法,我们用or组合了两个nameLike的Specification。

五、测试#

我们用REST发布我们的查询功能,代码如下:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
package cn.techcave.demo.jpa3.web

import cn.techcave.demo.jpa3.service.RecordService
import org.springframework.beans.factory.annotation.Autowired
import org.springframework.web.bind.annotation.GetMapping
import org.springframework.web.bind.annotation.RestController

@RestController
class RecordController {

@Autowired
lateinit var recordSvc:RecordService

@GetMapping("findAll")
fun findAll() = recordSvc.findAll()

@GetMapping("findByName")
fun findByName(name:String) = recordSvc.filterByName(name)

@GetMapping("findByNames")
fun findByNames(n1:String, n2:String) = recordSvc.filterByNames(n1, n2)
}

然后,我们在这里手工做一些数据用于测试:

测试数据

用Postman调用REST接口,首先是findAll,可以看到所有数据被查询出来了:

findAll结果

然后findByName,包含查询参数name的值被查询出来了:
findByName结果

findByNames,这里测试或验收的是组合查询,两个组合条件都是包含名称,当然也可以使用其他查询条件:
findByNames组合查询结果