reactivex / rxjavamath Goto Github PK

I have a case where I'm using window to capture some data. There are times when the window comes in empty. However, by the logic of non-empty case, I have the following code:

sourceObservable.window(driverObservable).switchMap(window -> {
    Observable<Float> value = window.map(data -> data.value);

    Observable<Float> min = MathObservable.min(value);
    Observable<Float> max = MathObservable.max(value);

    return Observable.zip(min, max, Pair::new);
})...

In the above code, things work fine if window has any data. However, in time when the window is empty, map will ignore the operation and complete. However, min and max will, upon seeing complete without data, will throw an error. What I really want is for min and max to also complete without raising an error if the window is empty.

Overall, I need the whole operation not to produce a value if the window is empty. The only way I could quick-fix is via this very bad logic:

sourceObservable.window(driverObservable)
                .flatMap(Observable::toList)
                .filter(list -> list.size() > 0)
                .map(Observable::from)
.                switchMap(window -> {
                    Observable<Float> value = window.map(data -> data.value);

                    Observable<Float> min = MathObservable.min(value);
                    Observable<Float> max = MathObservable.max(value);

                    return Observable.zip(min, max, Pair::new);
})...

This works appropriately but it's an extra hit on logic that I'd hope I didn't need to do. Any suggestions? Maybe I missed something in the documentation?

I've noticed this project doesn't use buildViaTravis.sh so I guess the Github release screen doesn't trigger anything. In addition, it seems this secure credentials is different for each project.

@abersnaze do you have the means to generate this data?

@abersnaze could you switch the default branch over to 1.x?

I think that the current API is not very user friendly. Using Transformers would be a better approach. What do you think?
http://blog.danlew.net/2015/03/02/dont-break-the-chain/