如何使用Reator和R2dbc压缩嵌套列表 [英] How to zip nested lists with reactor and R2dbc
本文介绍了如何使用Reator和R2dbc压缩嵌套列表的处理方法,对大家解决问题具有一定的参考价值,需要的朋友们下面随着小编来一起学习吧!
问题描述
我在Postgres数据库中有3个表,正在使用R2dbc以关系方式查询和连接它们。
我有3个实体类(可能不应该是数据类,但不应该影响示例)
@Entity
@Table(name = "parent", schema = "public", catalog = "Test")
data class MyParentObject(
@GeneratedValue(strategy = GenerationType.IDENTITY)
@Id
@org.springframework.data.annotation.Id
@Column(name = "id")
var id: Int = 0,
@Transient
var childData: List<MyChildObject>? = null
)
@Entity
@Table(name = "child", schema = "public", catalog = "Test")
data class MyChildObject(
@GeneratedValue(strategy = GenerationType.IDENTITY)
@Id
@org.springframework.data.annotation.Id
@Column(name = "id")
var id: Int = 0,
@Column(name = "parent_id")
var parentId: Int? = null
@Transient
var grandchildData: List<MyGrandchildObject>? = null
)
@Entity
@Table(name = "grandchild", schema = "public", catalog = "Test")
data class MyGrandchildObject(
@GeneratedValue(strategy = GenerationType.IDENTITY)
@Id
@org.springframework.data.annotation.Id
@Column(name = "id")
var id: Int = 0
@Column(name = "child_id")
var childId: Int? = null
)
父母是一对多对孩子,这是孙子对一对多。父项id和子项id的作用类似于fkey。
我有一个RestController,它可以通过这些方法返回填充了所有子数据的所有父数据
fun viewAllParents(): Mono<MutableList<MyParentObject>> =
parentRepository.findAll()
.flatMap { Mono.just(it).addChildData(it.id) }
.collectList()
fun Mono<MyParentObject>.addChildData(id: Int): Mono<MyParentObject> =
this.zipWith(childRepository.getAllByParentIdEquals(id).collectList())
.map {
it.t1.childData = it.t2
it.t1
}
我还有另一个RestController,它可以通过这些方法返回包含所有孙子数据的所有ChildData(与上面基本相同)
fun viewAllChildren(): Mono<MutableList<MyChildObject>> =
childRepository.findAll()
.flatMap { Mono.just(it).addGrandchildData(it.id) }
.collectList()
fun Mono<MyChildObject>.addGrandchildData(id: Int): Mono<MyChildObject> =
this.zipWith(childOfChildRepository.getAllByChildIdEquals(id).collectList())
.map {
it.t1.childOfChildData = it.t2
it.t1
}
我不能做的,也是我的问题是,我如何让viewAllParents()
也填充孙子数据。我是否需要将var grandchildData: List<MyGrandchildObject>
转换为Flux,并用孙子存储库中的新Flux将其压缩?还是我看错了?
如有任何提示,我们将不胜感激。
推荐答案
我真的很喜欢使用反应器获取分层数据的挑战。 我不知道Kotlin,但我曾尝试用Java重现这个问题。我无法创建具有父-子-孙层次结构的PostgreSQL表,但我尝试通过Web客户端模拟类似的东西(基本上逻辑将保持不变)。这是我的代码,这是我尝试执行的操作,并能够获得您想要的结果:https://github.com/harryalto/reactive-springwebflux
解决方案的关键在于处理程序代码,我使用该代码基于孩子列表构建一个子流,并使用该子流将所有子流联系在一起
public Flux<Parent> getFamiliesHierarchy() { return getAllParents() .flatMap(parent -> getAllChildsList(parent.getId()) .flatMap(childList -> getChildsWithKids(childList)) .map(childList -> parent.toBuilder().children(childList).build() ) ); }
下面是完整的代码
@Component
@Slf4j
public class FamilyHandler {
@Autowired
private WebClient webClient;
public Flux<Parent> getAllParents() {
return webClient
.get()
.uri("parents")
.retrieve()
.bodyToFlux(Parent.class);
}
public Mono<List<Child>> getAllChildsList(final Integer parentId) {
ParameterizedTypeReference<List<Child>> childList =
new ParameterizedTypeReference<List<Child>>() {};
return webClient
.get()
.uri("childs?parentId=" + parentId)
.retrieve()
.bodyToMono(childList);
}
public Flux<GrandChild> getGrandKids(final Integer childId) {
return webClient
.get()
.uri("grandKids?childId=" + childId)
.retrieve()
.bodyToFlux(GrandChild.class);
}
public Mono<List<GrandChild>> getGrandKidsList(final Integer childId) {
ParameterizedTypeReference<List<GrandChild>> grandKidsList =
new ParameterizedTypeReference<List<GrandChild>>() {};
return webClient
.get()
.uri("grandKids?childId=" + childId)
.retrieve()
.bodyToMono(grandKidsList);
}
private Mono<List<Child>> getChildsWithKids(final List<Child> childList) {
return Flux.fromIterable(childList).flatMap(child ->
Mono.zip(Mono.just(child), getGrandKidsList(child.getId()))
.map(tuple2 -> tuple2.getT1().toBuilder().grandChildren(tuple2.getT2()).build())
).collectList();
}
public Flux<Parent> getFamiliesHierarchy() {
return getAllParents()
.flatMap(parent ->
getAllChildsList(parent.getId())
.flatMap(childList -> getChildsWithKids(childList))
.map(childList -> parent.toBuilder().children(childList).build()
)
);
}
}`
我使用json-server模拟服务器
下面是我的db.json文件
{
"parents":[
{
"id": 1,
"name" : "Parent1",
"path":"1"
},
{
"id": 2,
"name" : "Parent2",
"path":"2"
}
],
"childs":[
{
"id": 1,
"parentId": 1,
"name": "child1Parent1",
"path":"1.1"
},
{
"id":2,
"parentId": 1,
"projectName": "child2Parent1",
"path":"1.2"
},
{
"id":3,
"parentId": 2,
"projectName": "child1Parent2",
"path":"2.1"
},
{
"id":4,
"parentId": 2,
"projectName": "child2Parent2",
"path":"2.2"
}
],
"grandKids":[
{
"id":1,
"childId": 2,
"projectName": "grandKid1child2Parent1",
"path":"1.2.1"
},
{
"id":3,
"childId": 2,
"projectName": "grandKid1child2Parent1",
"path":"1.2.2"
},
{
"id":2,
"childId": 4,
"projectName": "grandKid1child1Parent2",
"path":"2.2.1"
},
{
"id":4,
"childId": 4,
"projectName": "grandKid1child1Parent2",
"path":"2.2.2"
},
{
"id":5,
"childId": 3,
"projectName": "grandKid1child1Parent2",
"path":"2.1.1"
}
]
}
这是我的控制器代码
@RestController
@Slf4j
public class FamilyController {
@Autowired
private FamilyHandler familyHandler;
@GetMapping(FAMILIES_ENDPOINT_V1)
public Flux<Parent> viewAllParents() {
return familyHandler.getFamiliesHierarchy();
}
}
我们可以轻松地为r2DBC存储库更新代码。
--更新-- 我能够创建样例数据,并使用R2DBC创建了等效项 以下是指向gist
的链接这篇关于如何使用Reator和R2dbc压缩嵌套列表的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持IT屋!
查看全文