How to Think, in RxJava, Before Reacting

This talk was given at IndicThreads Conference, 3-4 June 2016, Pune, India (http://pune16.indicthreads.com/).

RxJava is a library that helps in writing code using reactive paradigm. When writing programs using RxJava, the hardest part is thinking in Reactive, which requires one to let go of old imperative and stateful habits of programming, and start thinking in a different paradigm. At my work in ThoughtWorks I have been using RxJava for more that six months in production applications. My talk will be based on the knowledge that I gained and lessons that I learnt from my practical experience of RxJava.

My goals from this talk are:
- To demystify the reactive paradigms concepts for the attendees and
- Provide enough information about the mechanics of RxJava so that attendees can straight away start experimenting with the library

The topics that I will be talking about are:
– Introduction to Reactive Programming
– Creating Observable (the source)
– Creating Observer (the consumer)
– Subscribing (connecting consumer to source)
– Operators (for Transforming, Filtering and Combining)
– Building Resilience (Error handling)
– Using Concurrency (Schedulers)
– Unit Testing Utilities

Transcript

1. Praveer Gupta How to think in RxJava before reacting
2. Reactive Programming “a programming paradigm oriented around data flows and the propagation of change” – Wikipedia
3. Reactive Programming 0 A B 1 fx = A1 + B1 C 2 3 2 5
4. • Java VM implementation of ReactiveX • An API for asynchronous programming with observable streams What is Rx in RxJava? RxJs Rx.NET RxScala RxClojure RxSwift Rx.rb RxPY & Others
5. • Client Side Applications of RxJava Thread switching Event handling
6. • Client Side • Server Side Applications of RxJava Internal Processing Service A Service B Service C Request Total Request Processing Time Service Boundary Parallel Sequential Response Combine
7. RxJava Observable Subscriber observeOn subscribeOn fromCallable timer interval map flatMap timestamp filter distinctUntilChanged just error onError onNext onCompleted take buffer zip combineLatest concat merge Scheduler io computation newThread Observer Pattern Reactive Functional Asynchronous Resiliency TestSubscriber TestScheduler
8. RxJava fromCallable timer interval just error Observable Subscriber onError onNext onCompleted Observer Pattern Reactive map flatMap timestamp filter distinctUntilChanged take buffer zip combineLatest concat merge Functional observeOn subscribeOn Scheduler io computation newThread Asynchronous TestSubscriber TestScheduler Resiliency retry timeout onErrorResumeNext Observable Subscriber onError onNext onCompleted Observer Pattern Reactive Concept fromCallable timer interval just error Create map flatMap timestamp filter distinctUntilChanged take buffer zip combineLatest concat merge Functional Operate observeOn subscribeOn Scheduler io computation newThread Asynchronous Resiliency retry timeout onErrorResumeNext TestSubscriber TestScheduler Concurrent Handle errors Test How to think in RxJava
9. How to think in RxJava • Introduction
10. Introduction • Observer & Iterator Pattern
11. Subscriber (Consumer) Observable (Source) 2 onNext( ) 2 2 1 1 onNext( ) 1 Source Consumer subscribe Produces data/event asynchronously Consumes data/event as they arrive Pushes data/event Observer & Iterator Pattern
12. onCompleted() Observable Subscriber (Source) (Consumer) unsubscribe Pushes completion event Observer & Iterator Pattern Complete
13. onError( ) X X X Observable Subscriber (Source) (Consumer) unsubscribe Pushes error event Observer & Iterator Pattern Error
14. onNext( ) 1 Observable Subscriber onError( ) X onCompleted() Pushing Streams of Data/Events Observer & Iterator Pattern
15. Introduction • Observer & Iterator Pattern • RxJava in Action
16. Subscriber<Integer> subscriber = Subscribers.create(
 n -> System.out.println(n),
 e -> System.out.print("Error"),
 () -> System.out.println("Completed")
 ); Observable<Integer> observable = Observable.just(1, 2, 3); observable.subscribe(subscriber); Subscriber<Integer> subscriber = Subscribers.create(
 n -> System.out.println(n),
 e -> System.out.print("Error"),
 () -> System.out.println("Completed")
 ); Subscriber<Integer> subscriber = Subscribers.create(
 n -> System.out.println(n),
 e -> System.out.println("Error"),
 () -> System.out.println("Completed")
 ); Subscriber<Integer> subscriber = Subscribers.create(
 n -> System.out.println(n),
 e -> System.out.println("Error"),
 () -> System.out.println("Completed")
 ); Subscriber<Integer> subscriber = Subscribers.create(
 n -> System.out.println(n),
 e -> System.out.println("Error"),
 () -> System.out.println("Completed")
 ); Observable<Integer> observable = Observable.just(1, 2, 3); Subscriber<Integer> subscriber = Subscribers.create(
 n -> System.out.println(n),
 e -> System.out.println("Error"),
 () -> System.out.println("Completed")
 ); Observable<Integer> observable = Observable.just(1, 2, 3); RxJava in Action Source Consumer onNext onError onCompleted connect
17. Observable.just(1, 2, 3).subscribe(
 n -> System.out.println(n),
 e -> System.out.println("Error"),
 () -> System.out.println("Completed")
 ); 1 Output: 2 3 Completed RxJava in Action
18. Introduction • Observer & Iterator Pattern • RxJava in Action • Creating Observables
19. Creating Observables • fromCallable
20. Observable.fromCallable(new Callable<String>() {
 @Override
 public String call() throws Exception {
 return "result";
 }
 }).subscribe( n -> System.out.println(n), e -> System.out.println(“Error"), () -> System.out.println("Completed") ); fromCallable Observable.fromCallable(new Callable<String>() {
 @Override
 public String call() throws Exception {
 return "result";
 }
 }).subscribe( n -> System.out.println(n), e -> System.out.println(“Error"), () -> System.out.println("Completed") );
21. Observable.fromCallable(() -> "result"); .subscribe( s -> System.out.println(s), e -> System.out.println("Error"), () -> System.out.println("Completed") ); fromCallable Output: result Completed Observable.fromCallable(() -> "result"); .subscribe( s -> System.out.println(s), e -> System.out.println("Error"), () -> System.out.println("Completed") );
22. Creating Observables • fromCallable
23. Introduction • Observer Pattern • RxJava in Action • Creating Observables
24. How to think in RxJava • Introduction • Operators
25. Subscriber Operators Observable Operator 1 Operator 2 Operator 3 functional
26. Operators • Filtering
27. Filtering Timeline Events/Data Completion Event Operator Timeline RxMarbles
28. Observable.interval(1, TimeUnit.SECONDS) .take(2)
 .subscribe( n -> System.out.println(n), e -> System.out.println("Error"), () -> System.out.println("Completed") ); Output: 0 1 Completed Filtering Observable.interval(1, TimeUnit.SECONDS) .take(2)
 .subscribe( n -> System.out.println(n), e -> System.out.println("Error"), () -> System.out.println("Completed") ); Observable.interval(1, TimeUnit.SECONDS) .take(2)
 .subscribe( n -> System.out.println(n), e -> System.out.println("Error"), () -> System.out.println("Completed") );
29. Operators • Filtering • Transforming
30. Transforming Observable Interleaved Completion of source Completion of flatMapped Observable
31. Observable.just("A", "B")
 .flatMap(e -> Observable.just(1, 2).map(n -> e + n))
 .subscribe(
 n -> System.out.println(n),
 e -> System.out.println("Error"),
 () -> System.out.println("Completed")
 ); Output: A1 A2 B1 B2 Completed Transforming Observable.just("A", "B")
 .flatMap(e -> Observable.just(1, 2).map(n -> e + n))
 .subscribe(
 n -> System.out.println(n),
 e -> System.out.println("Error"),
 () -> System.out.println("Completed")
 ); Observable.just("A", "B")
 .flatMap(e -> Observable.just(1, 2).map(n -> e + n))
 .subscribe(
 n -> System.out.println(n),
 e -> System.out.println("Error"),
 () -> System.out.println("Completed")
 );
32. Operators • Filtering • Transforming • Combining
33. Combining Emitted only when each source has emitted
34. Observable.zip(
 Observable.just("A", "B", "C"),
 Observable.interval(1, TimeUnit.SECONDS),
 (a, n) -> a + n
 ).subscribe(
 n -> System.out.println(n),
 e -> System.out.println("Error"),
 () -> System.out.println("Completed")
 ); Output: A0 B1 C2 Completed Combining Observable.zip(
 Observable.just("A", "B", "C"),
 Observable.interval(1, TimeUnit.SECONDS),
 (a, n) -> a + n
 ).subscribe(
 n -> System.out.println(n),
 e -> System.out.println("Error"),
 () -> System.out.println("Completed")
 );
35. Operators • Filtering • Transforming • Combining
36. How to think in RxJava • Introduction • Operators • Multithreading
37. Multithreading • subscribeOn • observeOn
38. Multithreading Observable
 .fromCallable(() -> {
 printThreadNameAnd("Observable");
 return 1;
 })
 .filter(n -> {
 printThreadNameAnd("Operator 1");
 return n == 1;
 })
 .map(n -> {
 printThreadNameAnd("Operator 2");
 return "A" + n;
 })
 .subscribe(n -> printThreadNameAnd(“Subscriber")); printThreadNameAnd("Done subscribing"); Observable Operator 1 Operator 2 Subscriber Observable
 .fromCallable(() -> {
 printThreadNameAnd("Observable");
 return 1;
 })
 .filter(n -> {
 printThreadNameAnd("Operator 1");
 return n == 1;
 })
 .map(n -> {
 printThreadNameAnd("Operator 2");
 return "A" + n;
 })
 .subscribe(n -> printThreadNameAnd(“Subscriber")); printThreadNameAnd("Done subscribing");
39. Observable
 .fromCallable(() -> {
 printThreadNameAnd("Observable");
 return 1;
 })
 .filter(n -> {
 printThreadNameAnd("Operator 1");
 return n == 1;
 })
 .map(n -> {
 printThreadNameAnd("Operator 2");
 return "A" + n;
 })
 .subscribe(n -> printThreadNameAnd(“Subscriber")); printThreadNameAnd("Done subscribing"); Observable
 .fromCallable(() -> {
 printThreadNameAnd("Observable");
 return 1;
 })
 .filter(n -> {
 printThreadNameAnd("Operator 1");
 return n == 1;
 })
 .map(n -> {
 printThreadNameAnd("Operator 2");
 return "A" + n;
 })
 .subscribe(n -> printThreadNameAnd(“Subscriber")); printThreadNameAnd("Done subscribing"); Observable
 .fromCallable(() -> {
 printThreadNameAnd("Observable");
 return 1;
 })
 .filter(n -> {
 printThreadNameAnd("Operator 1");
 return n == 1;
 })
 .map(n -> {
 printThreadNameAnd("Operator 2");
 return "A" + n;
 })
 .subscribe(n -> printThreadNameAnd(“Subscriber")); printThreadNameAnd("Done subscribing"); Observable
 .fromCallable(() -> {
 printThreadNameAnd("Observable");
 return 1;
 })
 .filter(n -> {
 printThreadNameAnd("Operator 1");
 return n == 1;
 })
 .map(n -> {
 printThreadNameAnd("Operator 2");
 return "A" + n;
 })
 .subscribe(n -> printThreadNameAnd(“Subscriber")); printThreadNameAnd("Done subscribing"); Observable
 .fromCallable(() -> {
 printThreadNameAnd("Observable");
 return 1;
 })
 .filter(n -> {
 printThreadNameAnd("Operator 1");
 return n == 1;
 })
 .map(n -> {
 printThreadNameAnd("Operator 2");
 return "A" + n;
 })
 .subscribe(n -> printThreadNameAnd(“Subscriber")); printThreadNameAnd("Done subscribing"); Observable
 .fromCallable(() -> {
 printThreadNameAnd("Observable");
 return 1;
 })
 .filter(n -> {
 printThreadNameAnd("Operator 1");
 return n == 1;
 })
 .map(n -> {
 printThreadNameAnd("Operator 2");
 return "A" + n;
 })
 .subscribe(n -> printThreadNameAnd(“Subscriber")); printThreadNameAnd("Done subscribing"); Observable
 .fromCallable(() -> {
 printThreadNameAnd("Observable");
 return 1;
 })
 .filter(n -> {
 printThreadNameAnd("Operator 1");
 return n == 1;
 })
 .map(n -> {
 printThreadNameAnd("Operator 2");
 return "A" + n;
 })
 .subscribe(n -> printThreadNameAnd(“Subscriber")); printThreadNameAnd("Done subscribing"); Output: main: Observable main: Operator 1 main: Operator 2 main: Subscriber main: Done subscribing RxJava is blocking by default Multithreading
40. Observable
 .fromCallable(() -> {
 printThreadNameAnd("Observable");
 return 1;
 })
 .subscribeOn(Schedulers.io())
 .filter(n -> {
 printThreadNameAnd("Operator 1");
 return n == 1;
 })
 .map(n -> {
 printThreadNameAnd("Operator 2");
 return "A" + n;
 })
 .subscribe(n -> printThreadNameAnd("Subscriber")); printThreadNameAnd("Done subscribing"); Observable
 .fromCallable(() -> {
 printThreadNameAnd("Observable");
 return 1;
 })
 .subscribeOn(Schedulers.io())
 .filter(n -> {
 printThreadNameAnd("Operator 1");
 return n == 1;
 })
 .map(n -> {
 printThreadNameAnd("Operator 2");
 return "A" + n;
 })
 .subscribe(n -> printThreadNameAnd("Subscriber")); printThreadNameAnd("Done subscribing"); Output: main: Done subscribing RxCachedThreadScheduler-1: Observable RxCachedThreadScheduler-1: Operator 1 RxCachedThreadScheduler-1: Operator 2 RxCachedThreadScheduler-1: Subscriber Observable
 .fromCallable(() -> {
 printThreadNameAnd("Observable");
 return 1;
 })
 .subscribeOn(Schedulers.io())
 .filter(n -> {
 printThreadNameAnd("Operator 1");
 return n == 1;
 })
 .map(n -> {
 printThreadNameAnd("Operator 2");
 return "A" + n;
 })
 .subscribe(n -> printThreadNameAnd("Subscriber")); printThreadNameAnd("Done subscribing"); Observable
 .fromCallable(() -> {
 printThreadNameAnd("Observable");
 return 1;
 })
 .subscribeOn(Schedulers.io())
 .filter(n -> {
 printThreadNameAnd("Operator 1");
 return n == 1;
 })
 .map(n -> {
 printThreadNameAnd("Operator 2");
 return "A" + n;
 })
 .subscribe(n -> printThreadNameAnd("Subscriber")); printThreadNameAnd("Done subscribing"); Observable
 .fromCallable(() -> {
 printThreadNameAnd("Observable");
 return 1;
 })
 .subscribeOn(Schedulers.io())
 .filter(n -> {
 printThreadNameAnd("Operator 1");
 return n == 1;
 })
 .map(n -> {
 printThreadNameAnd("Operator 2");
 return "A" + n;
 })
 .subscribe(n -> printThreadNameAnd("Subscriber")); printThreadNameAnd("Done subscribing"); Observable
 .fromCallable(() -> {
 printThreadNameAnd("Observable");
 return 1;
 })
 .subscribeOn(Schedulers.io())
 .filter(n -> {
 printThreadNameAnd("Operator 1");
 return n == 1;
 })
 .map(n -> {
 printThreadNameAnd("Operator 2");
 return "A" + n;
 })
 .subscribe(n -> printThreadNameAnd("Subscriber")); printThreadNameAnd("Done subscribing"); Observable
 .fromCallable(() -> {
 printThreadNameAnd("Observable");
 return 1;
 })
 .subscribeOn(Schedulers.io())
 .filter(n -> {
 printThreadNameAnd("Operator 1");
 return n == 1;
 })
 .map(n -> {
 printThreadNameAnd("Operator 2");
 return "A" + n;
 })
 .subscribe(n -> printThreadNameAnd("Subscriber")); printThreadNameAnd("Done subscribing"); Observable
 .fromCallable(() -> {
 printThreadNameAnd("Observable");
 return 1;
 })
 .subscribeOn(Schedulers.io())
 .filter(n -> {
 printThreadNameAnd("Operator 1");
 return n == 1;
 })
 .map(n -> {
 printThreadNameAnd("Operator 2");
 return "A" + n;
 })
 .subscribe(n -> printThreadNameAnd("Subscriber")); printThreadNameAnd("Done subscribing"); Multithreading
41. Observable
 .fromCallable(() -> {
 printThreadNameAnd("Observable");
 return 1;
 })
 .subscribeOn(Schedulers.io())
 .filter(n -> {
 printThreadNameAnd("Operator 1");
 return n == 1;
 }) .observeOn(Schedulers.newThread())
 .map(n -> {
 printThreadNameAnd("Operator 2");
 return "A" + n;
 })
 .subscribe(n -> printThreadNameAnd("Subscriber")); printThreadNameAnd("Done subscribing"); Output: main: Done subscribing RxCachedThreadScheduler-1: Observable RxCachedThreadScheduler-1: Operator 1 RxNewThreadScheduler-1: Operator 2 RxNewThreadScheduler-1: Subscriber Observable
 .fromCallable(() -> {
 printThreadNameAnd("Observable");
 return 1;
 })
 .subscribeOn(Schedulers.io())
 .filter(n -> {
 printThreadNameAnd("Operator 1");
 return n == 1;
 }) .observeOn(Schedulers.newThread())
 .map(n -> {
 printThreadNameAnd("Operator 2");
 return "A" + n;
 })
 .subscribe(n -> printThreadNameAnd("Subscriber")); printThreadNameAnd("Done subscribing"); Observable
 .fromCallable(() -> {
 printThreadNameAnd("Observable");
 return 1;
 })
 .subscribeOn(Schedulers.io())
 .filter(n -> {
 printThreadNameAnd("Operator 1");
 return n == 1;
 }) .observeOn(Schedulers.newThread())
 .map(n -> {
 printThreadNameAnd("Operator 2");
 return "A" + n;
 })
 .subscribe(n -> printThreadNameAnd("Subscriber")); printThreadNameAnd("Done subscribing"); Observable
 .fromCallable(() -> {
 printThreadNameAnd("Observable");
 return 1;
 })
 .subscribeOn(Schedulers.io())
 .filter(n -> {
 printThreadNameAnd("Operator 1");
 return n == 1;
 }) .observeOn(Schedulers.newThread())
 .map(n -> {
 printThreadNameAnd("Operator 2");
 return "A" + n;
 })
 .subscribe(n -> printThreadNameAnd("Subscriber")); printThreadNameAnd("Done subscribing"); Observable
 .fromCallable(() -> {
 printThreadNameAnd("Observable");
 return 1;
 })
 .subscribeOn(Schedulers.io())
 .filter(n -> {
 printThreadNameAnd("Operator 1");
 return n == 1;
 }) .observeOn(Schedulers.newThread())
 .map(n -> {
 printThreadNameAnd("Operator 2");
 return "A" + n;
 })
 .subscribe(n -> printThreadNameAnd("Subscriber")); printThreadNameAnd("Done subscribing"); Observable
 .fromCallable(() -> {
 printThreadNameAnd("Observable");
 return 1;
 })
 .subscribeOn(Schedulers.io())
 .filter(n -> {
 printThreadNameAnd("Operator 1");
 return n == 1;
 }) .observeOn(Schedulers.newThread())
 .map(n -> {
 printThreadNameAnd("Operator 2");
 return "A" + n;
 })
 .subscribe(n -> printThreadNameAnd("Subscriber")); printThreadNameAnd("Done subscribing"); Observable
 .fromCallable(() -> {
 printThreadNameAnd("Observable");
 return 1;
 })
 .subscribeOn(Schedulers.io())
 .filter(n -> {
 printThreadNameAnd("Operator 1");
 return n == 1;
 }) .observeOn(Schedulers.newThread())
 .map(n -> {
 printThreadNameAnd("Operator 2");
 return "A" + n;
 })
 .subscribe(n -> printThreadNameAnd("Subscriber")); printThreadNameAnd("Done subscribing"); Observable
 .fromCallable(() -> {
 printThreadNameAnd("Observable");
 return 1;
 })
 .subscribeOn(Schedulers.io())
 .filter(n -> {
 printThreadNameAnd("Operator 1");
 return n == 1;
 }) .observeOn(Schedulers.newThread())
 .map(n -> {
 printThreadNameAnd("Operator 2");
 return "A" + n;
 })
 .subscribe(n -> printThreadNameAnd("Subscriber")); printThreadNameAnd("Done subscribing"); Multithreading
42. Multithreading • subscribeOn • observeOn • Scheduler
43. Scheduler • Schedulers.newThread() • Schedulers.io() • Schedulers.computation() • Schedulers.from(java.util.concurrent.Executor executor)
44. Observable.interval(1, TimeUnit.SECONDS)
 .take(3)
 .subscribe(
 n -> printThreadNameAnd("Subscriber"),
 e -> System.out.println("Error"),
 () -> System.out.println("Completed")
 ); Output: RxComputationThreadPool-1: Subscriber RxComputationThreadPool-1: Subscriber RxComputationThreadPool-1: Subscriber Completed Scheduler Scheduler: interval operates by default on the computation Scheduler.
45. Multithreading • subscribeOn • observeOn • Scheduler
46. How to think in RxJava • Introduction • Operators • Multithreading • Resiliency
47. Resiliency • Default error handling
48. Default error handling Observable.fromCallable(() -> { throw new IOException(); })
 .subscribe(
 n -> System.out.println(n),
 throwable -> printErrorClassName(throwable),
 () -> System.out.println("Completed")
 ); Output: Error: IOException Observable.fromCallable(() -> { throw new IOException(); })
 .subscribe(
 n -> System.out.println(n),
 throwable -> printErrorClassName(throwable),
 () -> System.out.println("Completed")
 ); Observable.fromCallable(() -> { throw new IOException(); })
 .subscribe(
 n -> System.out.println(n),
 throwable -> printErrorClassName(throwable),
 () -> System.out.println("Completed")
 ); Observable throws Exception handled in onError
49. Default error handling Observable.just("A")
 .map(text -> Integer.parseInt(text))
 .subscribe(
 n -> System.out.println(n),
 throwable -> printErrorClassName(throwable),
 () -> System.out.println("Completed")
 ); Output: Error: NumberFormatException Observable.just("A")
 .map(text -> Integer.parseInt(text))
 .subscribe(
 n -> System.out.println(n),
 throwable -> printErrorClassName(throwable),
 () -> System.out.println("Completed")
 ); Observable.just("A")
 .map(text -> Integer.parseInt(text))
 .subscribe(
 n -> System.out.println(n),
 throwable -> printErrorClassName(throwable),
 () -> System.out.println("Completed")
 ); Operator throws Exception handled in onError
50. Resiliency • Default error handling • Errors not handled
51. Errors not handled Observable.just("A")
 .map(text -> Integer.parseInt(text))
 .subscribe(
 n -> System.out.println(n),
 throwable -> { throw new RuntimeException(); },
 () -> System.out.println("Completed")
 ); Output: Exception in thread "main" rx.exceptions.OnErrorFailedException: Error occurred when trying to propagate error to Observer.onError Observable.just("A")
 .map(text -> Integer.parseInt(text))
 .subscribe(
 n -> System.out.println(n),
 throwable -> { throw new RuntimeException(); },
 () -> System.out.println("Completed")
 ); Observable.just("A")
 .map(text -> Integer.parseInt(text))
 .subscribe(
 n -> System.out.println(n),
 throwable -> { throw new RuntimeException(); },
 () -> System.out.println("Completed")
 ); onError throws Exception
52. Errors not handled Observable.just("A")
 .map(text -> Integer.parseInt(text))
 .subscribe(
 n -> System.out.println(n)
 ); Output: Exception in thread "main" rx.exceptions.OnErrorNotImplementedException Observable.just("A")
 .map(text -> Integer.parseInt(text))
 .subscribe(
 n -> System.out.println(n)
 ); onError function missing
53. Resiliency • Default error handling • Errors not handled • Recovery Mechanisms
54. Recovery Mechanisms • Closing the stream gracefully
55. Closing the stream gracefully Observable.error(new IOException())
 .onErrorReturn(throwable -> -1)
 .subscribe(
 n -> System.out.println(n),
 throwable -> printErrorClassName(throwable),
 () -> System.out.println("Completed")
 ); Output: -1 Completed Observable.error(new IOException())
 .onErrorReturn(throwable -> -1)
 .subscribe(
 n -> System.out.println(n),
 throwable -> printErrorClassName(throwable),
 () -> System.out.println("Completed")
 ); Observable.error(new IOException())
 .onErrorReturn(throwable -> -1)
 .subscribe(
 n -> System.out.println(n),
 throwable -> printErrorClassName(throwable),
 () -> System.out.println("Completed")
 ); handling error by retuning a value indicating error
56. Recovery Mechanisms • Closing the stream gracefully • Using a backup
57. Using a backup Observable<Integer> aService = Observable .error(new IOException());
 Observable<Integer> backupService = Observable.just(1); 
 aService
 .onErrorResumeNext(backupService)
 .subscribe(
 n -> System.out.println(n),
 throwable -> printErrorClassName(throwable),
 () -> System.out.println("Completed")
 ); Output: 1 Completed Observable<Integer> aService = Observable .error(new IOException());
 Observable<Integer> backupService = Observable.just(1); 
 aService
 .onErrorResumeNext(backupService)
 .subscribe(
 n -> System.out.println(n),
 throwable -> printErrorClassName(throwable),
 () -> System.out.println("Completed")
 ); Observable<Integer> aService = Observable .error(new IOException());
 Observable<Integer> backupService = Observable.just(1); 
 aService
 .onErrorResumeNext(backupService)
 .subscribe(
 n -> System.out.println(n),
 throwable -> printErrorClassName(throwable),
 () -> System.out.println("Completed")
 ); resuming with another service
58. Recovery Mechanisms • Closing the stream gracefully • Using a backup • Retrying
59. Retrying Observable.error(new IOException())
 .retryWhen(errorStream ->
 Observable.range(1, 3)
 .zipWith(errorStream, (n, e) -> n)
 .doOnNext(n -> System.out.println("Retry: " + n))
 ).subscribe(
 n -> System.out.println(n),
 throwable -> printErrorClassName(throwable),
 () -> System.out.println("Completed")
 ); Output: Retry: 1 Retry: 2 Retry: 3 Completed Observable.error(new IOException())
 .retryWhen(errorStream ->
 Observable.range(1, 3)
 .zipWith(errorStream, (n, e) -> n)
 .doOnNext(n -> System.out.println("Retry: " + n))
 ).subscribe(
 n -> System.out.println(n),
 throwable -> printErrorClassName(throwable),
 () -> System.out.println("Completed")
 ); Observable.error(new IOException())
 .retryWhen(errorStream ->
 Observable.range(1, 3)
 .zipWith(errorStream, (n, e) -> n)
 .doOnNext(n -> System.out.println("Retry: " + n))
 ).subscribe(
 n -> System.out.println(n),
 throwable -> printErrorClassName(throwable),
 () -> System.out.println("Completed")
 ); conditional retry logic
60. Recovery Mechanisms • Closing the stream gracefully • Using a backup • Retrying
61. Resiliency • Points of failure • Errors not handled • Recovery Mechanisms
62. How to think in RxJava • Introduction • Operators • Multithreading • Resiliency • Testing Utilities
63. Testing Utilities • TestSubscriber
64. TestSubscriber TestSubscriber<Long> testSubscriber = TestSubscriber.create(); Observable.interval(1, TimeUnit.SECONDS).take(2)
 .subscribe(testSubscriber); 
 testSubscriber.awaitTerminalEvent();
 testSubscriber.assertCompleted(); testSubscriber.assertNoErrors(); testSubscriber.assertUnsubscribed(); testSubscriber.assertValueCount(2);
 testSubscriber.assertValues(0L, 1L);
65. Testing Utilities • TestSubscriber • TestScheduler
66. TestScheduler TestScheduler testScheduler = Schedulers.test(); TestSubscriber<Long> testSubscriber = TestSubscriber.create(); 
 Observable.interval(1, TimeUnit.SECONDS, testScheduler) .take(2).subscribe(testSubscriber); 
 testScheduler.advanceTimeTo(1500, TimeUnit.MILLISECONDS);
 testSubscriber.assertValueCount(1);
 testSubscriber.assertNotCompleted(); 
 testScheduler.advanceTimeBy(1000, TimeUnit.MILLISECONDS);
 testSubscriber.assertValueCount(2);
 testSubscriber.assertCompleted();
67. Testing Utilities • TestSubscriber • TestScheduler
68. How to think in RxJava • Introduction • Operators • Multithreading • Resiliency • Testing Utilities
69. How to think in RxJava fromCallable timer interval just error Observable Subscriber onError onNext onCompleted Observer Pattern Reactive map flatMap timestamp filter distinctUntilChanged take buffer zip combineLatest concat merge Functional observeOn subscribeOn Scheduler io computation newThread Asynchronous TestSubscriber TestScheduler Resiliency retry timeout onErrorResumeNext Observable Subscriber onError onNext onCompleted Observer Pattern Reactive Concept fromCallable timer interval just error Create map flatMap timestamp filter distinctUntilChanged take buffer zip combineLatest concat merge Functional Operate observeOn subscribeOn Scheduler io computation newThread Asynchronous Resiliency retry timeout onErrorResumeNext TestSubscriber TestScheduler Concurrent Handle errors Test
70. http://reactivex.io/ https://github.com/ReactiveX/RxJava
71. Questions ? http://praveer09.github.io @praveergupta praveer