Android RxJava使用介绍（4） RxJava的操作符

本篇文章继续介绍以下类型的操作符

• Combining Observables(Observable的组合操作符)
• Error Handling Operators(Observable的错误处理操作符)

Combining Observables(Observable的组合操作符)

combineLatest操作符

combineLatest操作符把两个Observable产生的结果进行合并，合并的结果组成一个新的Observable。这两个Observable中任意一个Observable产生的结果，都和另一个Observable最后产生的结果，按照一定的规则进行合并。流程图如下： 

调用例子如下:

//产生0,5,10,15,20数列 Observable<Long> observable1 = Observable.timer(0, 1000, TimeUnit.MILLISECONDS)
                .map(new Func1<Long, Long>() { @Override public Long call(Long aLong) { return aLong * 5;
                    }
                }).take(5); //产生0,10,20,30,40数列 Observable<Long> observable2 = Observable.timer(500, 1000, TimeUnit.MILLISECONDS)
                .map(new Func1<Long, Long>() { @Override public Long call(Long aLong) { return aLong * 10;
                    }
                }).take(5);


        Observable.combineLatest(observable1, observable2, new Func2<Long, Long, Long>() { @Override public Long call(Long aLong, Long aLong2) { return aLong+aLong2;
            }
        }).subscribe(new Subscriber<Long>() { @Override public void onCompleted() {
                System.out.println("Sequence complete.");
            } @Override public void onError(Throwable e) {
                System.err.println("Error: " + e.getMessage());
            } @Override public void onNext(Long aLong) {
                System.out.println("Next: " + aLong);
            }
        });

运行结果如下： 
Next: 0 
Next: 5 
Next: 15 
Next: 20 
Next: 30 
Next: 35 
Next: 45 
Next: 50 
Next: 60 
Sequence complete.

join操作符

join操作符把类似于combineLatest操作符，也是两个Observable产生的结果进行合并，合并的结果组成一个新的Observable，但是join操作符可以控制每个Observable产生结果的生命周期，在每个结果的生命周期内，可以与另一个Observable产生的结果按照一定的规则进行合并，流程图如下： 

join方法的用法如下： 
observableA.join(observableB, 
observableA产生结果生命周期控制函数， 
observableB产生结果生命周期控制函数， 
observableA产生的结果与observableB产生的结果的合并规则）

调用例子如下：

//产生0,5,10,15,20数列 Observable<Long> observable1 = Observable.timer(0, 1000, TimeUnit.MILLISECONDS)
                .map(new Func1<Long, Long>() { @Override public Long call(Long aLong) { return aLong * 5;
                    }
                }).take(5); //产生0,10,20,30,40数列 Observable<Long> observable2 = Observable.timer(500, 1000, TimeUnit.MILLISECONDS)
                .map(new Func1<Long, Long>() { @Override public Long call(Long aLong) { return aLong * 10;
                    }
                }).take(5);

        observable1.join(observable2, new Func1<Long, Observable<Long>>() { @Override public Observable<Long> call(Long aLong) { //使Observable延迟600毫秒执行 return Observable.just(aLong).delay(600, TimeUnit.MILLISECONDS);
            }
        }, new Func1<Long, Observable<Long>>() { @Override public Observable<Long> call(Long aLong) { //使Observable延迟600毫秒执行 return Observable.just(aLong).delay(600, TimeUnit.MILLISECONDS);
            }
        }, new Func2<Long, Long, Long>() { @Override public Long call(Long aLong, Long aLong2) { return aLong + aLong2;
            }
        }).subscribe(new Subscriber<Long>() { @Override public void onCompleted() {
                System.out.println("Sequence complete.");
            } @Override public void onError(Throwable e) {
                System.err.println("Error: " + e.getMessage());
            } @Override public void onNext(Long aLong) {
                System.out.println("Next: " + aLong);
            }
        });

运行结果如下： 
Next: 0 
Next: 5 
Next: 15 
Next: 20 
Next: 30 
Next: 35 
Next: 45 
Next: 50 
Next: 60 
Sequence complete.

groupJoin操作符

groupJoin操作符非常类似于join操作符，区别在于join操作符中第四个参数的传入函数不一致，其流程图如下： 

调用例子如下：

//产生0,5,10,15,20数列 Observable<Long> observable1 = Observable.timer(0, 1000, TimeUnit.MILLISECONDS)
                .map(new Func1<Long, Long>() { @Override public Long call(Long aLong) { return aLong * 5;
                    }
                }).take(5); //产生0,10,20,30,40数列 Observable<Long> observable2 = Observable.timer(500, 1000, TimeUnit.MILLISECONDS)
                .map(new Func1<Long, Long>() { @Override public Long call(Long aLong) { return aLong * 10;
                    }
                }).take(5);

        observable1.groupJoin(observable2, new Func1<Long, Observable<Long>>() { @Override public Observable<Long> call(Long aLong) { return Observable.just(aLong).delay(1600, TimeUnit.MILLISECONDS);
            }
        }, new Func1<Long, Observable<Long>>() { @Override public Observable<Long> call(Long aLong) { return Observable.just(aLong).delay(600, TimeUnit.MILLISECONDS);
            }
        }, new Func2<Long, Observable<Long>, Observable<Long>>() { @Override public Observable<Long> call(Long aLong, Observable<Long> observable) { return observable.map(new Func1<Long, Long>() { @Override public Long call(Long aLong2) { return aLong + aLong2;
                    }
                });
            }
        }).subscribe(new Subscriber<Observable<Long>>() { @Override public void onCompleted() {
                System.out.println("Sequence complete.");
            } @Override public void onError(Throwable e) {
                System.err.println("Error: " + e.getMessage());
            } @Override public void onNext(Observable<Long> observable) {
                observable.subscribe(new Subscriber<Long>() { @Override public void onCompleted() {

                    } @Override public void onError(Throwable e) {

                    } @Override public void onNext(Long aLong) {
                        System.out.println("Next: " + aLong);
                    }
                });
            }
        });

运行结果如下： 
Next: 0 
Next: 5 
Next: 10 
Next: 15 
Next: 20 
Next: 25 
Next: 30 
Next: 35 
Next: 40 
Next: 45 
Next: 50 
Next: 60 
Next: 55 
Sequence complete.

merge操作符

merge操作符是按照两个Observable提交结果的时间顺序，对Observable进行合并，如ObservableA每隔500毫秒产生数据为0,5,10,15,20；而ObservableB每隔500毫秒产生数据0,10,20,30,40，其中第一个数据延迟500毫秒产生，最后合并结果为：0,0,5,10,10,20,15,30,20,40;其流程图如下： 

调用例子如下：

//产生0,5,10,15,20数列 Observable<Long> observable1 = Observable.timer(0, 1000, TimeUnit.MILLISECONDS)
                .map(new Func1<Long, Long>() { @Override public Long call(Long aLong) { return aLong * 5;
                    }
                }).take(5); //产生0,10,20,30,40数列 Observable<Long> observable2 = Observable.timer(500, 1000, TimeUnit.MILLISECONDS)
                .map(new Func1<Long, Long>() { @Override public Long call(Long aLong) { return aLong * 10;
                    }
                }).take(5);

        Observable.merge(observable1, observable2)
                .subscribe(new Subscriber<Long>() { @Override public void onCompleted() {
                        System.out.println("Sequence complete.");
                    } @Override public void onError(Throwable e) {
                        System.err.println("Error: " + e.getMessage());
                    } @Override public void onNext(Long aLong) {
                        System.out.println("Next:" + aLong);
                    }
                });

运行结果如下： 
Next:0 
Next:0 
Next:5 
Next:10 
Next:10 
Next:20 
Next:15 
Next:30 
Next:20 
Next:40 
Sequence complete.

mergeDelayError操作符

从merge操作符的流程图可以看出，一旦合并的某一个Observable中出现错误，就会马上停止合并，并对订阅者回调执行onError方法，而mergeDelayError操作符会把错误放到所有结果都合并完成之后才执行，其流程图如下： 

调用例子如下：

//产生0,5,10数列,最后会产生一个错误 Observable<Long> errorObservable = Observable.error(new Exception("this is end!"));
        Observable < Long > observable1 = Observable.timer(0, 1000, TimeUnit.MILLISECONDS)
                .map(new Func1<Long, Long>() { @Override public Long call(Long aLong) { return aLong * 5;
                    }
                }).take(3).mergeWith(errorObservable.delay(3500, TimeUnit.MILLISECONDS)); //产生0,10,20,30,40数列 Observable<Long> observable2 = Observable.timer(500, 1000, TimeUnit.MILLISECONDS)
                .map(new Func1<Long, Long>() { @Override public Long call(Long aLong) { return aLong * 10;
                    }
                }).take(5);

        Observable.mergeDelayError(observable1, observable2)
                .subscribe(new Subscriber<Long>() { @Override public void onCompleted() {
                        System.out.println("Sequence complete.");
                    } @Override public void onError(Throwable e) {
                        System.err.println("Error: " + e.getMessage());
                    } @Override public void onNext(Long aLong) {
                        System.out.println("Next:" + aLong);
                    }
                });

运行结果如下： 
Next:0 
Next:0 
Next:5 
Next:10 
Next:10 
Next:20 
Next:30 
Next:40 
Error: this is end!

startWith操作符

startWith操作符是在源Observable提交结果之前，插入指定的某些数据，其流程图如下： 

调用例子如下：

Observable.just(10,20,30).startWith(2, 3, 4).subscribe(new Subscriber<Integer>() {
            @Override public void onCompleted() {
                System.out.println("Sequence complete.");
            }

            @Override public void onError(Throwable e) {
                System.err.println("Error: " + e.getMessage());
            }

            @Override public void onNext(Integer value) {
                System.out.println("Next:" + value);
            }
        });

运行结果如下： 
Next:2 
Next:3 
Next:4 
Next:10 
Next:20 
Next:30 
Sequence complete.

switchOnNext操作符

switchOnNext操作符是把一组Observable转换成一个Observable，转换规则为：对于这组Observable中的每一个Observable所产生的结果，如果在同一个时间内存在两个或多个Observable提交的结果，只取最后一个Observable提交的结果给订阅者，其流程图如下： 

调用例子如下：

//每隔500毫秒产生一个observable Observable<Observable<Long>> observable = Observable.timer(0, 500, TimeUnit.MILLISECONDS).map(new Func1<Long, Observable<Long>>() { @Override public Observable<Long> call(Long aLong) { //每隔200毫秒产生一组数据（0,10,20,30,40) return Observable.timer(0, 200, TimeUnit.MILLISECONDS).map(new Func1<Long, Long>() { @Override public Long call(Long aLong) { return aLong * 10;
                    }
                }).take(5);
            }
        }).take(2);

        Observable.switchOnNext(observable).subscribe(new Subscriber<Long>() { @Override public void onCompleted() {
                System.out.println("Sequence complete.");
            } @Override public void onError(Throwable e) {
                System.err.println("Error: " + e.getMessage());
            } @Override public void onNext(Long aLong) {
                System.out.println("Next:" + aLong);
            }
        });

运行结果如下： 
Next:0 
Next:10 
Next:20 
Next:0 
Next:10 
Next:20 
Next:30 
Next:40 
Sequence complete.

zip操作符

zip操作符是把两个observable提交的结果，严格按照顺序进行合并，其流程图如下： 

调用例子如下：

Observable<Integer> observable1 = Observable.just(10,20,30);
        Observable<Integer> observable2 = Observable.just(4, 8, 12, 16);
        Observable.zip(observable1, observable2, new Func2<Integer, Integer, Integer>() {
            @Override public Integer call(Integer integer, Integer integer2) { return integer + integer2;
            }
        }).subscribe(new Subscriber<Integer>() {
            @Override public void onCompleted() {
                System.out.println("Sequence complete.");
            }

            @Override public void onError(Throwable e) {
                System.err.println("Error: " + e.getMessage());
            }

            @Override public void onNext(Integer value) {
                System.out.println("Next:" + value);
            }
        });

运行结果如下： 
Next:14 
Next:28 
Next:42 
Sequence complete.

Error Handling Operators(Observable的错误处理操作符)

onErrorReturn操作符

onErrorReturn操作符是在Observable发生错误或异常的时候（即将回调oError方法时），拦截错误并执行指定的逻辑，返回一个跟源Observable相同类型的结果，最后回调订阅者的onComplete方法，其流程图如下： 

调用例子如下：

Observable<Integer> observable = Observable.create(new Observable.OnSubscribe<Integer>() { @Override public void call(Subscriber<? super Integer> subscriber) { if (subscriber.isUnsubscribed()) return; //循环输出数字 try { for (int i = 0; i < 10; i++) { if (i == 4) { throw new Exception("this is number 4 error！");
                        }
                        subscriber.onNext(i);
                    }
                    subscriber.onCompleted();
                } catch (Exception e) {
                    subscriber.onError(e);
                }
            }
        });

        observable.onErrorReturn(new Func1<Throwable, Integer>() { @Override public Integer call(Throwable throwable) { return 1004;
            }
        }).subscribe(new Subscriber<Integer>() { @Override public void onCompleted() {
                System.out.println("Sequence complete.");
            } @Override public void onError(Throwable e) {
                System.err.println("Error: " + e.getMessage());
            } @Override public void onNext(Integer value) {
                System.out.println("Next:" + value);
            }
        });

运行结果如下： 
Next:0 
Next:1 
Next:2 
Next:3 
Next:1004 
Sequence complete.

onErrorResumeNext操作符

onErrorResumeNext操作符跟onErrorReturn类似，只不过onErrorReturn只能在错误或异常发生时只返回一个和源Observable相同类型的结果，而onErrorResumeNext操作符是在错误或异常发生时返回一个Observable，也就是说可以返回多个和源Observable相同类型的结果，其流程图如下： 

调用例子如下：

Observable<Integer> observable = Observable.create(new Observable.OnSubscribe<Integer>() { @Override public void call(Subscriber<? super Integer> subscriber) { if (subscriber.isUnsubscribed()) return; //循环输出数字 try { for (int i = 0; i < 10; i++) { if (i == 4) { throw new Exception("this is number 4 error！");
                        }
                        subscriber.onNext(i);
                    }
                    subscriber.onCompleted();
                } catch (Exception e) {
                    subscriber.onError(e);
                }
            }
        });

        observable.onErrorResumeNext(new Func1<Throwable, Observable<? extends Integer>>() { @Override public Observable<? extends Integer> call(Throwable throwable) { return Observable.just(100,101, 102);
            }
        }).subscribe(new Subscriber<Integer>() { @Override public void onCompleted() {
                System.out.println("Sequence complete.");
            } @Override public void onError(Throwable e) {
                System.err.println("Error: " + e.getMessage());
            } @Override public void onNext(Integer value) {
                System.out.println("Next:" + value);
            }
        });

运行结果如下： 
Next:0 
Next:1 
Next:2 
Next:3 
Next:100 
Next:101 
Next:102 
Sequence complete.

onExceptionResumeNext操作符

onExceptionResumeNext操作符和onErrorResumeNext操作符类似，不同的地方在于onErrorResumeNext操作符是当Observable发生错误或异常时触发，而onExceptionResumeNext是当Observable发生异常时才触发。

这里要普及一个概念就是，java的异常分为错误（error）和异常（exception）两种，它们都是继承于Throwable类。

错误（error）一般是比较严重的系统问题，比如我们经常遇到的OutOfMemoryError、StackOverflowError等都是错误。错误一般继承于Error类，而Error类又继承于Throwable类，如果需要捕获错误，需要使用try..catch(Error e)或者try..catch(Throwable e)句式。使用try..catch(Exception e)句式无法捕获错误

异常（Exception）也是继承于Throwable类，一般是根据实际处理业务抛出的异常，分为运行时异常（RuntimeException）和普通异常。普通异常直接继承于Exception类，如果方法内部没有通过try..catch句式进行处理，必须通过throws关键字把异常抛出外部进行处理（即checked异常）；而运行时异常继承于RuntimeException类，如果方法内部没有通过try..catch句式进行处理，不需要显式通过throws关键字抛出外部，如IndexOutOfBoundsException、NullPointerException、ClassCastException等都是运行时异常，当然RuntimeException也是继承于Exception类，因此是可以通过try..catch(Exception e)句式进行捕获处理的。 
onExceptionResumeNext流程图如下： 

调用例子如下：

Observable<Integer> observable = Observable.create(new Observable.OnSubscribe<Integer>() { @Override public void call(Subscriber<? super Integer> subscriber) { if (subscriber.isUnsubscribed()) return; //循环输出数字 try { for (int i = 0; i < 10; i++) { if (i == 4) { throw new Exception("this is number 4 error！");
                        }
                        subscriber.onNext(i);
                    }
                    subscriber.onCompleted();
                } catch (Throwable e) {
                    subscriber.onError(e);
                }
            }
        });

        observable.onExceptionResumeNext(Observable.just(100, 101, 102)).subscribe(new Subscriber<Integer>() { @Override public void onCompleted() {
                System.out.println("Sequence complete.");
            } @Override public void onError(Throwable e) {
                System.err.println("Error: " + e.getMessage());
            } @Override public void onNext(Integer value) {
                System.out.println("Next:" + value);
            }
        });

运行结果如下： 
Next:0 
Next:1 
Next:2 
Next:3 
Next:100 
Next:101 
Next:102 
Sequence complete.

retry操作符

retry操作符是当Observable发生错误或者异常时，重新尝试执行Observable的逻辑，如果经过n次重新尝试执行后仍然出现错误或者异常，则最后回调执行onError方法；当然如果源Observable没有错误或者异常出现，则按照正常流程执行。其流程图如下： 

调用例子如下：

Observable<Integer> observable = Observable.create(new Observable.OnSubscribe<Integer>() { @Override public void call(Subscriber<? super Integer> subscriber) { if (subscriber.isUnsubscribed()) return; //循环输出数字 try { for (int i = 0; i < 10; i++) { if (i == 4) { throw new Exception("this is number 4 error！");
                        }
                        subscriber.onNext(i);
                    }
                    subscriber.onCompleted();
                } catch (Throwable e) {
                    subscriber.onError(e);
                }
            }
        });

        observable.retry(2).subscribe(new Subscriber<Integer>() { @Override public void onCompleted() {
                System.out.println("Sequence complete.");
            } @Override public void onError(Throwable e) {
                System.err.println("Error: " + e.getMessage());
            } @Override public void onNext(Integer value) {
                System.out.println("Next:" + value);
            }
        });

运行结果如下： 
Next:0 
Next:1 
Next:2 
Next:3

Next:0 
Next:1 
Next:2 
Next:3

Next:0 
Next:1 
Next:2 
Next:3 
Error: this is number 4 error！

retryWhen操作符

retryWhen操作符类似于retry操作符，都是在源observable出现错误或者异常时，重新尝试执行源observable的逻辑，不同在于retryWhen操作符是在源Observable出现错误或者异常时，通过回调第二个Observable来判断是否重新尝试执行源Observable的逻辑，如果第二个Observable没有错误或者异常出现，则就会重新尝试执行源Observable的逻辑，否则就会直接回调执行订阅者的onError方法。其流程图如下： 

调用例子如下：

Observable<Integer> observable = Observable.create(new Observable.OnSubscribe<Integer>() { @Override public void call(Subscriber<? super Integer> subscriber) {
                System.out.println("subscribing");
                subscriber.onError(new RuntimeException("always fails"));
            }
        });

        observable.retryWhen(new Func1<Observable<? extends Throwable>, Observable<?>>() { @Override public Observable<?> call(Observable<? extends Throwable> observable) { return observable.zipWith(Observable.range(1, 3), new Func2<Throwable, Integer, Integer>() { @Override public Integer call(Throwable throwable, Integer integer) { return integer;
                    }
                }).flatMap(new Func1<Integer, Observable<?>>() { @Override public Observable<?> call(Integer integer) {
                        System.out.println("delay retry by " + integer + " second(s)"); //每一秒中执行一次 return Observable.timer(integer, TimeUnit.SECONDS);
                    }
                });
            }
        }).subscribe(new Subscriber<Integer>() { @Override public void onCompleted() {
                System.out.println("Sequence complete.");
            } @Override public void onError(Throwable e) {
                System.err.println("Error: " + e.getMessage());
            } @Override public void onNext(Integer value) {
                System.out.println("Next:" + value);
            }
        });

运行结果如下： 
subscribing 
delay retry by 1 second(s) 
subscribing 
delay retry by 2 second(s) 
subscribing 
delay retry by 3 second(s) 
subscribing 
Sequence complete.

好了，先介绍这么多，下回继续介绍其他的操作符，敬请期待！