聊聊reactive streams的tranform操做

序

本文主要展現一下reactive streams的一些transform操做

mergeWith

@Test
    public void testMerge(){
        Flux<String> flux1 = Flux.interval(Duration.ofSeconds(1))
                .take(3)
                .map(e -> "[flux1]:"+e);

        Flux<String> mergeFlux = Flux.interval(Duration.ofSeconds(1))
                .delayElements(Duration.ofSeconds(1))
                .take(3)
                .map(e -> "[flux2]:"+e)
                .mergeWith(flux1);

        mergeFlux.subscribe(e -> {
            LOGGER.info("subscribe:{}",e);
        });

        mergeFlux.blockLast();
    }

輸出實例

21:18:07.583 [main] DEBUG reactor.util.Loggers$LoggerFactory - Using Slf4j logging framework
21:18:08.618 [parallel-2] INFO com.example.demo.TransformTest - subscribe:[flux1]:0
21:18:09.619 [parallel-2] INFO com.example.demo.TransformTest - subscribe:[flux1]:1
21:18:09.645 [parallel-6] INFO com.example.demo.TransformTest - subscribe:[flux2]:0
21:18:10.619 [parallel-2] INFO com.example.demo.TransformTest - subscribe:[flux1]:2
21:18:10.649 [parallel-8] INFO com.example.demo.TransformTest - subscribe:[flux2]:1
21:18:11.654 [parallel-2] INFO com.example.demo.TransformTest - subscribe:[flux2]:2

能夠發現，他們是交叉合併的。

concatWith

@Test
    public void testConcat(){
        Flux<String> flux1 = Flux.interval(Duration.ofSeconds(1))
                .take(3)
                .map(e -> "[flux1]:"+e);

        Flux<String> concatFlux = Flux.interval(Duration.ofSeconds(1))
                .delayElements(Duration.ofSeconds(1))
                .take(3)
                .map(e -> "[flux2]:"+e)
                .concatWith(flux1);
        concatFlux.subscribe(e -> {
            LOGGER.info("subscribe:{}",e);
        });
        concatFlux.blockLast();
    }

輸出

21:19:00.779 [main] DEBUG reactor.util.Loggers$LoggerFactory - Using Slf4j logging framework
21:19:02.832 [parallel-4] INFO com.example.demo.TransformTest - subscribe:[flux2]:0
21:19:03.836 [parallel-6] INFO com.example.demo.TransformTest - subscribe:[flux2]:1
21:19:04.840 [parallel-8] INFO com.example.demo.TransformTest - subscribe:[flux2]:2
21:19:05.845 [parallel-2] INFO com.example.demo.TransformTest - subscribe:[flux1]:0
21:19:06.845 [parallel-2] INFO com.example.demo.TransformTest - subscribe:[flux1]:1
21:19:07.844 [parallel-2] INFO com.example.demo.TransformTest - subscribe:[flux1]:2

能夠發現concatWith只是鏈接兩個flux的數據，並非按emit的順序交叉來

zipWith

@Test
    public void testZip(){
        List<String> firstList = Lists.newArrayList("a","b","c","d","e","a","b");
        List<String> secondList = Lists.newArrayList("1","2","3","4","5");
        Flux<Tuple2<String,String>> zipFlux =  Flux.fromIterable(firstList)
                .zipWith(Flux.fromIterable(secondList));
        zipFlux.subscribe(e -> {
            LOGGER.info("subscribe:{}",e);
        });
    }

輸出以下

21:20:59.506 [main] DEBUG reactor.util.Loggers$LoggerFactory - Using Slf4j logging framework
21:20:59.516 [main] INFO com.example.demo.TransformTest - subscribe:[a,1]
21:20:59.517 [main] INFO com.example.demo.TransformTest - subscribe:[b,2]
21:20:59.517 [main] INFO com.example.demo.TransformTest - subscribe:[c,3]
21:20:59.517 [main] INFO com.example.demo.TransformTest - subscribe:[d,4]
21:20:59.517 [main] INFO com.example.demo.TransformTest - subscribe:[e,5]

能夠發現flux1相比flux2多餘的數據沒有被zip

flatMap

@Test
    public void testFlatMap(){
        List<String> secondList = Lists.newArrayList("1","2","3","4","5");
        Flux<String> flatMapFlux = Flux.fromIterable(secondList)
                .flatMap((str) ->{
                    return Mono.just(str).repeat(2).map(String::toUpperCase).delayElements(Duration.ofMillis(1));
                });
        flatMapFlux.subscribe(e -> {
            LOGGER.info("subscribe:{}",e);
        });
        flatMapFlux.blockLast();

        Flux<String> mapFlux = Flux.fromIterable(secondList)
                .repeat(2)
                .map(String::toUpperCase);
        mapFlux.subscribe(e -> {
                    LOGGER.info("map subscribe:{}",e);
                });
        mapFlux.blockLast();
    }

輸出

21:33:46.904 [main] DEBUG reactor.util.Loggers$LoggerFactory - Using Slf4j logging framework
21:33:46.958 [parallel-1] INFO com.example.demo.TransformTest - subscribe:1
21:33:46.959 [parallel-1] INFO com.example.demo.TransformTest - subscribe:2
21:33:46.959 [parallel-1] INFO com.example.demo.TransformTest - subscribe:3
21:33:46.960 [parallel-1] INFO com.example.demo.TransformTest - subscribe:4
21:33:46.960 [parallel-1] INFO com.example.demo.TransformTest - subscribe:5
21:33:46.960 [parallel-1] INFO com.example.demo.TransformTest - subscribe:2
21:33:46.960 [parallel-7] INFO com.example.demo.TransformTest - subscribe:3
21:33:46.960 [parallel-8] INFO com.example.demo.TransformTest - subscribe:4
21:33:46.960 [parallel-1] INFO com.example.demo.TransformTest - subscribe:5
21:33:46.961 [parallel-6] INFO com.example.demo.TransformTest - subscribe:1
21:33:46.963 [main] INFO com.example.demo.TransformTest - map subscribe:1
21:33:46.963 [main] INFO com.example.demo.TransformTest - map subscribe:2
21:33:46.963 [main] INFO com.example.demo.TransformTest - map subscribe:3
21:33:46.963 [main] INFO com.example.demo.TransformTest - map subscribe:4
21:33:46.963 [main] INFO com.example.demo.TransformTest - map subscribe:5
21:33:46.963 [main] INFO com.example.demo.TransformTest - map subscribe:1
21:33:46.963 [main] INFO com.example.demo.TransformTest - map subscribe:2
21:33:46.963 [main] INFO com.example.demo.TransformTest - map subscribe:3
21:33:46.963 [main] INFO com.example.demo.TransformTest - map subscribe:4
21:33:46.963 [main] INFO com.example.demo.TransformTest - map subscribe:5

flatMap是異步的

reduce

@Test
    public void testReduce(){
        List<String> secondList = Lists.newArrayList("1","2","3","4","5");
        Mono<Integer> reduceMono = Flux.fromIterable(secondList)
                .flatMap(e -> Mono.just(e).map(item -> Integer.valueOf(item)))
                .reduce((total, e) -> total + e);
        reduceMono.subscribe(e -> {
            LOGGER.info("subscribe:{}",e);
        });
    }

輸出

21:36:29.978 [main] DEBUG reactor.util.Loggers$LoggerFactory - Using Slf4j logging framework
21:36:30.014 [main] INFO com.example.demo.TransformTest - subscribe:15

groupBy

@Test
    public void testGroup(){
        List<String> firstList = Lists.newArrayList("a","b","c","d","e","a","b");
        Flux<GroupedFlux<String, String>> groupFlux = Flux.fromIterable(firstList)
                .map(String::toUpperCase)
                .groupBy(key -> key);
        groupFlux.subscribe(e -> {
            LOGGER.info("subscribe:{}",e.collectList().subscribe(item -> {
                LOGGER.info("item:{}",item);
            }));
        });
    }

輸出

21:37:00.912 [main] DEBUG reactor.util.Loggers$LoggerFactory - Using Slf4j logging framework
21:37:00.949 [main] INFO com.example.demo.TransformTest - subscribe:reactor.core.publisher.LambdaMonoSubscriber@5faeada1
21:37:00.951 [main] INFO com.example.demo.TransformTest - subscribe:reactor.core.publisher.LambdaMonoSubscriber@1563da5
21:37:00.951 [main] INFO com.example.demo.TransformTest - subscribe:reactor.core.publisher.LambdaMonoSubscriber@2bbf4b8b
21:37:00.951 [main] INFO com.example.demo.TransformTest - subscribe:reactor.core.publisher.LambdaMonoSubscriber@30a3107a
21:37:00.951 [main] INFO com.example.demo.TransformTest - subscribe:reactor.core.publisher.LambdaMonoSubscriber@33c7e1bb
21:37:00.951 [main] INFO com.example.demo.TransformTest - item:[A, A]
21:37:00.952 [main] INFO com.example.demo.TransformTest - item:[B, B]
21:37:00.952 [main] INFO com.example.demo.TransformTest - item:[C]
21:37:00.952 [main] INFO com.example.demo.TransformTest - item:[D]
21:37:00.952 [main] INFO com.example.demo.TransformTest - item:[E]

first

@Test
    public void testFirst(){
        List<String> firstList = Lists.newArrayList("a","b","c","d","e","a","b");
        List<String> secondList = Lists.newArrayList("1","2","3","4","5");
        Flux<String> firstFlux = Flux.fromIterable(firstList)
                .delayElements(Duration.ofMillis(200));
        Flux<String> secondFlux = Flux.fromIterable(secondList)
                .take(2);

        Flux<String> result = Flux.first(firstFlux, secondFlux);
        result.subscribe(e -> {
            LOGGER.info("subscribe:{}",e);
        });
    }

toIterable

@Test
    public void testToIterable(){
        List<String> firstList = Lists.newArrayList("a","b","c","d","e","a","b");
        Iterable<String> itr = Flux.fromIterable(firstList)
                .map(String::toUpperCase)
                .toIterable();
        itr.forEach(e -> LOGGER.info(e));
    }

輸出

21:39:35.031 [main] DEBUG reactor.util.Loggers$LoggerFactory - Using Slf4j logging framework
21:39:35.045 [main] INFO com.example.demo.TransformTest - A
21:39:35.045 [main] INFO com.example.demo.TransformTest - B
21:39:35.045 [main] INFO com.example.demo.TransformTest - C
21:39:35.045 [main] INFO com.example.demo.TransformTest - D
21:39:35.045 [main] INFO com.example.demo.TransformTest - E
21:39:35.045 [main] INFO com.example.demo.TransformTest - A
21:39:35.045 [main] INFO com.example.demo.TransformTest - B

小結

reactive streams的操做至關於在jdk的streams的基礎上實現了reactive化，能夠參照着瞭解。

doc

• Reactor – How to Combine Publishers (Flux/Mono)