从效果中分派多个动作:不同rxjs运算符之间的差异 [英] Dispatch multiple actions from effects: difference between different rxjs operators
问题描述
http请求成功后，我需要从ngrx效果中调度多个操作.
I need to dispatch multiple actions from an ngrx effect once an http request is successful.
似乎有几种方法行得通，有些则行不通，我不明白有什么区别.
There are several ways that seem to work and some that don't, I can't understand what the difference is.
@Effect()
loadUser = this.actions
.pipe(
ofType(campaignActions.type.LOAD_CAMPAIGN),
map((action: userActions.LoadUserAction) => action.payload),
switchMap((payload: { id: number }) =>
this.s.getUser(payload.id)
.pipe(
switchMap((data) => {
return from([
new userActions.LoadUserSuccessAction(),
new userActions.SomethingElseSuccessAction()
]);
}),
catchError(() => of(new userActions.LoadUserFailureAction()))
)
)
);
首先，我可以使用switchMap或mergeMap.我相信不同之处在于，如果多次触发此效果，那么如果我使用switchMap，而没有使用mergeMap，则所有正在进行的请求都将被取消.
First of all, I can use switchMap or mergeMap. I believe the difference is that if this effect is triggered multiple times, any ongoing requests will be cancelled if I use switchMap and not if I use mergeMap.
在分派多个动作方面，以下所有工作:
In terms of dispatching multiple actions, all the following works:
- 返回[action1，action2]
- 从([action1，action2])返回
- (动作1，动作2)的返回
以下内容无效:
- (([action1，action2])的返回->效果"UserEffects.loadUser"调度了无效的动作:[对象对象]，[对象对象]
前三个选项之间有什么区别? http请求后，mergeMap和switchMap怎么样?为什么最后一个选项不起作用?
What is the difference between all the first 3 options? What about mergeMap vs switchMap after the http request? Why doesn't the last option work?
推荐答案
这个问题可以分为几个部分:
This question can be separated into few sections:
+)mergeMap和switchMap之间的区别:
+) Difference between mergeMap and switchMap:
-
mergeMap = map + mergeAll.它将数据流映射到可观察流，每次内部可观察对象发出一个值，外部观察对象将收集所有发出的值并将其合并到新流中.
mergeMap = map + mergeAll. It maps a data stream to observable stream, and everytime the inner observable emits a value, the outer will collect and merge all emitted value into a new stream.
switchMap ，非常相似，但是正如您所知切换"，即，如果下一个内部可观测对象发出新值，则前一个被取消(一次一个内部可观测对象).
switchMap, very similar, but as you can tell it "switches", i.e. if the next inner observable emits a new value, the previous one gets canceled (one inner observable at a time).
+)mergeMap和switchMap之间的相似之处:它们都可以将任何类似于流的对象转换为可观察对象.
+) Similarity between mergeMap and switchMap: They can both convert any stream-like object to an observable.
+)您的代码之间的区别:
+) Difference between your code:
-
return [action1，action2]确实确实使switchMap(()=> ...)类似于from([action1，action2])，即从数组中创建一个可观察对象.
return [action1, action2] does indeed make switchMap(() => ...) similar to from([action1, action2]), i.e. create an observable from an array.
从([action1，action2])返回，这是一个内部可观察到的东西，发出2个动作，而switchMap将这些动作合并回到其流中.
return from([action1, action2]), this is an inner observable that emits 2 actions, and switchMap merges those actions back to its stream.
(action1，action2)的返回= from([[action1，action2]))
return of(action1, action2) = from([action1, action2])
返回([action1，action2])会创建一个可观察的对象，该对象发出一个值，该值是一个数组，您不能调用store.dispatch(array_of_actions)
return of([action1, action2]), however, creates an observable that emits a single value, which is an array, and you cannot call store.dispatch(array_of_actions)
这篇关于从效果中分派多个动作:不同rxjs运算符之间的差异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！
