Difference between catch and onErrorResumeNext

Question

In RxJS, there seems to be very little difference between an Observable instance's catch method and onErrorResumeNext method, besides the fact that onErrorResumeNext concatenates the original Observable with the Observable parameters whether an error happens or not.

If that's the case, isn't the naming a bit confusing? Because in case there is an error , onErrorResumeNext works exactly the same way than catch does:

var testObservable = Rx.Observable.return(1).concat(Rx.Observable.throw("Error"))

// Both onError and onCatch will emit the same result: 1, 2
var onError = testObservable.onErrorResumeNext(Rx.Observable.return(2));
var onCatch = testObservable.catch(Rx.Observable.return(2));

Is there a strong reason for not always using catch?

Solution

They are distinct.

As you've noted, both operators handle failure similarly; however, they differ in their handling of completion.

OnErrorResumeNext is simply a specialization with the following semantics:

"My query concatenates two observables. If the first observable fails, then resume with the next observable anyway."

Catch is far more general:

"My query avoids failure. If the observable fails, then continue with another observable."

If you're avoiding failure on the boundary of concatenation, then use OnErrorResumeNext; otherwise, to avoid failures in general use Catch.

Comments

[Brandon]

The name most likely comes from the Visual Basic On Error Resume Next statement, which has semantics similar to this operator.

[Vít Kotačka]

Your reply makes sense, thanks. I guess that the confusion is about the naming because, as you point out, the main task of onErrorResumeNext is to concatenate, but this is not reflected at all in the name.

[Dave Sexton]

@GiantSquid Well, I think concatenation is implied by the phrase "resume next", but I can see how that could be unclear at first.