Why we should use RxJs of() function?

问题:

in service section of angular.io tutorial for angular2 I hit a method named of.for example :

getHeroes(): Observable<Hero[]> {
  return of(HEROES);
}

or in below sample:

getHero(id: number): Observable<Hero> {
  // Todo: send the message _after_ fetching the hero
  this.messageService.add(`HeroService: fetched hero id=${id}`);
  return of(HEROES.find(hero => hero.id === id));
}

in angular.io Just explained

used RxJS of() to return an Observable of mock heroes (Observable).

and It was not explained why we should use of function and exactly what does it do and what are its benefits ?

回答1:

The reason why they're using of() is because it's very easy to use it instead of a real HTTP call.

In a real application you would implement getHeroes() like this for example:

getHeroes(): Observable<Hero[]> {
  return this.http.get(`/heroes`);
}

But since you just want to use a mocked response without creating any real backend you can use of() to return a fake response:

const HEROES = [{...}, {...}];

getHeroes(): Observable<Hero[]> {
  return of(HEROES);
}

The rest of your application is going to work the same because of() is an Observable and you can later subscribe or chain operators to it just like you were using this.http.get(...).

The only thing that of() does is that it emits its parameters as single emissions immediately on subscription and then sends the complete notification.

回答2:

Observable.of() is useful for maintaining the Observable data type before implementing an asynchronous interaction (for example, an http request to an API).

As Brandon Miller suggests, Observable.of() returns an Observable which immediately emits whatever values are supplied to of() as parameters, then completes.

This is better than returning static values, as it allows you to write subscribers that can handle the Observable type (which works both synchronously and asynchronously), even before implementing your async process.

//this function works synchronously AND asynchronously
getHeroes(): Observable<Hero[]> { 
  return Observable.of(HEROES)
  //-OR-
  return this.http.get('my.api.com/heroes')
  .map(res => res.json());
}

//it can be subscribed to in both cases
getHeroes().subscribe(heroes => {
  console.log(heroes); // '[hero1,hero2,hero3...]'
}

//DON'T DO THIS
getHeroesBad(): Array<Hero> {
  return HEROES                             //Works synchronously
  //-OR-
  return this.http.get('my.api.com/heroes') //TypeError, requires refactor
}

回答3:

In the first example, I assume that the HEROES variable is an array of objects, in which every object corresponds to interface Hero (or is the instance of class Hero). When we call this function somewhere in the code, it will act like this:

heroService.getHeroes()
  .subscribe((heroes: Hero[]) => {
    console.log(heroes) 
    // will output array of heroes: [{...}, {...}, ...]
  })

It means, that when we use Observable's of creation method with array as argument, on subscription it will emit the whole array, not its' elements one-by-one

In the second example, when we subscribe to Observable, that was returned from getHero() method, it emits only one hero, whose id corresponds to given id.

Basically, when you create Observable with of method and supply arguments to it, on subscription it emits these arguments one-by-one

Here's a good reference