为了账号安全,请及时绑定邮箱和手机立即绑定

Kotlin Coroutines Flow 系列(五) 其他的操作符

标签:
Kotlin

八. Flow 其他的操作符

8.1 Transform operators

transform

在使用 transform 操作符时,可以任意多次调用 emit ,这是 transform 跟 map 最大的区别:

fun main() = runBlocking {

    (1..5).asFlow()
        .transform {
            emit(it * 2)
            delay(100)
            emit(it * 4)
        }
        .collect { println(it) }
}

transform 也可以使用 emit 发射任意值:

fun main() = runBlocking {

    (1..5).asFlow()
        .transform {
            emit(it * 2)
            delay(100)
            emit("emit $it")
        }
        .collect { println(it) }
}

8.2 Size-limiting operators

take

take 操作符只取前几个 emit 发射的值。

fun main() = runBlocking {

    (1..5).asFlow()
        .take(2)
        .collect { println(it) }
}

8.3 Terminal flow operators

Kotlin Coroutines Flow 系列(一) Flow 基本使用 一文最后,我整理了 Flow 相关的 Terminal 操作符。本文介绍 reduce 和 fold 两个操作符。

reduce

类似于 Kotlin 集合中的 reduce 函数,能够对集合进行计算操作。

例如,对平方数列求和:

fun main() = runBlocking {

    val sum = (1..5).asFlow()
        .map { it * it }
        .reduce { a, b -> a + b }

    println(sum)
}

例如,计算阶乘:

fun main() = runBlocking {

    val sum = (1..5).asFlow().reduce { a, b -> a * b }

    println(sum)
}

fold

也类似于 Kotlin 集合中的 fold 函数,fold 也需要设置初始值。

fun main() = runBlocking {

    val sum = (1..5).asFlow()
        .map { it * it }
        .fold(0) { a, b -> a + b }

    println(sum)
}

在上述代码中,初始值为0就类似于使用 reduce 函数实现对平方数列求和。

而对于计算阶乘:

fun main() = runBlocking {

    val sum = (1..5).asFlow().fold(1) { a, b -> a * b }

    println(sum)
}

初始值为1就类似于使用 reduce 函数实现计算阶乘。

8.4 Composing flows operators

zip

zip 是可以将2个 flow 进行合并的操作符。

fun main() = runBlocking {

    val flowA = (1..5).asFlow()
    val flowB = flowOf("one", "two", "three","four","five")
    flowA.zip(flowB) { a, b -> "$a and $b" }
        .collect { println(it) }
}

执行结果:

1 and one
2 and two
3 and three
4 and four
5 and five

zip 操作符会把 flowA 中的一个 item 和 flowB 中对应的一个 item 进行合并。即使 flowB 中的每一个 item 都使用了 delay() 函数,在合并过程中也会等待 delay() 执行完后再进行合并。

fun main() = runBlocking {

    val flowA = (1..5).asFlow()
    val flowB = flowOf("one", "two", "three", "four", "five").onEach { delay(100) }

    val time = measureTimeMillis {
        flowA.zip(flowB) { a, b -> "$a and $b" }
            .collect { println(it) }
    }

    println("Cost $time ms")
}

执行结果:

1 and one
2 and two
3 and three
4 and four
5 and five
Cost 561 ms

如果 flowA 中 item 个数大于 flowB 中 item 个数:

fun main() = runBlocking {

    val flowA = (1..6).asFlow()
    val flowB = flowOf("one", "two", "three","four","five")
    flowA.zip(flowB) { a, b -> "$a and $b" }
        .collect { println(it) }
}

执行合并后新的 flow 的 item 个数 = 较小的 flow 的 item 个数。

执行结果:

1 and one
2 and two
3 and three
4 and four
5 and five

combine

combine 虽然也是合并,但是跟 zip 不太一样。

使用 combine 合并时,每次从 flowA 发出新的 item ,会将其与 flowB 的最新的 item 合并。

fun main() = runBlocking {

    val flowA = (1..5).asFlow().onEach { delay(100)  }
    val flowB = flowOf("one", "two", "three","four","five").onEach { delay(200)  }
    flowA.combine(flowB) { a, b -> "$a and $b" }
        .collect { println(it) }
}

执行结果:

1 and one
2 and one
3 and one
3 and two
4 and two
5 and two
5 and three
5 and four
5 and five

flattenMerge

其实,flattenMerge 不会组合多个 flow ,而是将它们作为单个流执行。

fun main() = runBlocking {

    val flowA = (1..5).asFlow()
    val flowB = flowOf("one", "two", "three","four","five")

    flowOf(flowA,flowB)
        .flattenConcat()
        .collect{ println(it) }
}

执行结果:

1
2
3
4
5
one
two
three
four
five

为了能更清楚地看到 flowA、flowB 作为单个流的执行,对他们稍作改动。

fun main() = runBlocking {

    val flowA = (1..5).asFlow().onEach { delay(100) }
    val flowB = flowOf("one", "two", "three","four","five").onEach { delay(200) }

    flowOf(flowA,flowB)
        .flattenMerge(2)
        .collect{ println(it) }
}

执行结果:

1
one
2
3
two
4
5
three
four
five

8.5 Flattening flows operators

flatMapConcat、flatMapMerge 类似于 RxJava 的 concatMap、flatMap 操作符。

flatMapConcat

flatMapConcat 由 map、flattenConcat 操作符实现。

@FlowPreview
public fun <T, R> Flow<T>.flatMapConcat(transform: suspend (value: T) -> Flow<R>): Flow<R> =
    map(transform).flattenConcat()

在调用 flatMapConcat 后,collect 函数在收集新值之前会等待 flatMapConcat 内部的 flow 完成。

fun currTime() = System.currentTimeMillis()

var start: Long = 0

fun main() = runBlocking {

    (1..5).asFlow()
        .onStart { start = currTime() }
        .onEach { delay(100) }
        .flatMapConcat {
            flow {
                emit("$it: First")
                delay(500)
                emit("$it: Second")
            }
        }
        .collect {
            println("$it at ${System.currentTimeMillis() - start} ms from start")
        }
}

执行结果:

1: First at 114 ms from start
1: Second at 619 ms from start
2: First at 719 ms from start
2: Second at 1224 ms from start
3: First at 1330 ms from start
3: Second at 1830 ms from start
4: First at 1932 ms from start
4: Second at 2433 ms from start
5: First at 2538 ms from start
5: Second at 3041 ms from start

flatMapMerge

flatMapMerge 由 map、flattenMerge 操作符实现。

@FlowPreview
public fun <T, R> Flow<T>.flatMapMerge(
    concurrency: Int = DEFAULT_CONCURRENCY,
    transform: suspend (value: T) -> Flow<R>
): Flow<R> = map(transform).flattenMerge(concurrency)

flatMapMerge 是顺序调用内部代码块,并且并行地执行 collect 函数。

fun currTime() = System.currentTimeMillis()

var start: Long = 0

fun main() = runBlocking {

    (1..5).asFlow()
        .onStart { start = currTime() }
        .onEach { delay(100) }
        .flatMapMerge {
            flow {
                emit("$it: First")
                delay(500)
                emit("$it: Second")
            }
        }
        .collect {
            println("$it at ${System.currentTimeMillis() - start} ms from start")
        }
}

执行结果:

1: First at 116 ms from start
2: First at 216 ms from start
3: First at 319 ms from start
4: First at 422 ms from start
5: First at 525 ms from start
1: Second at 618 ms from start
2: Second at 719 ms from start
3: Second at 822 ms from start
4: Second at 924 ms from start
5: Second at 1030 ms from start

flatMapMerge 操作符有一个参数 concurrency ,它默认使用DEFAULT_CONCURRENCY,如果想更直观地了解 flatMapMerge 的并行,可以对这个参数进行修改。例如改成2,就会发现不一样的执行结果。

flatMapLatest

当发射了新值之后,上个 flow 就会被取消。

fun currTime() = System.currentTimeMillis()

var start: Long = 0

fun main() = runBlocking {

    (1..5).asFlow()
        .onStart { start = currTime() }
        .onEach { delay(100) }
        .flatMapLatest {
            flow {
                emit("$it: First")
                delay(500)
                emit("$it: Second")
            }
        }
        .collect {
            println("$it at ${System.currentTimeMillis() - start} ms from start")
        }
}

执行结果:

1: First at 114 ms from start
2: First at 220 ms from start
3: First at 321 ms from start
4: First at 422 ms from start
5: First at 524 ms from start
5: Second at 1024 ms from start

九. Flow VS Reactive Streams

天生的多平台支持

由于 Kotlin 语言自身对多平台的支持,使得 Flow 也可以在多平台上使用。

互操作性

Flow 仍然属于响应式范畴。开发者通过 kotlinx-coroutines-reactive 模块中 Flow.asPublisher() 和 Publisher.asFlow() ,可以方便地将 Flow 跟 Reactive Streams 进行互操作。

该系列的相关文章:

点击查看更多内容
1人点赞

若觉得本文不错,就分享一下吧!

评论

作者其他优质文章

正在加载中
感谢您的支持,我会继续努力的~
扫码打赏,你说多少就多少
赞赏金额会直接到老师账户
支付方式
打开微信扫一扫,即可进行扫码打赏哦
今天注册有机会得

100积分直接送

付费专栏免费学

大额优惠券免费领

立即参与 放弃机会
意见反馈 帮助中心 APP下载
官方微信

举报

0/150
提交
取消