跳至正文

🌊 Flow

📖 什么是Flow?​

Flow是一个异步地按顺序返回多个值的数据流。可以把它想象成suspend函数的列表版本!

💡 基本概念​

单个值 vs 多个值​

// suspend - 一个值
suspend fun fetchData(): String {
    delay(1000)
    return "数据"
}

// Flow - 多个值
fun fetchDataStream(): Flow<String> = flow {
    emit("数据1")
    delay(500)
    emit("数据2")
    delay(500)
    emit("数据3")
}

fun main() = runBlocking {
    // 单个值
    println(fetchData())

    // 多个值
    fetchDataStream().collect { value ->
        println(value)
    }
}

Flow创建​

import kotlinx.coroutines.flow.*

fun main() = runBlocking {
    // 1. flow构建器
    val flow1 = flow {
        emit(1)
        emit(2)
        emit(3)
    }

    // 2. flowOf
    val flow2 = flowOf(1, 2, 3)

    // 3. asFlow
    val flow3 = listOf(1, 2, 3).asFlow()

    flow1.collect { println(it) }
}

🎯 实战示例​

实时数据​

fun tickerFlow(): Flow<Int> = flow {
    var counter = 0
    while (counter < 5) {
        emit(counter++)
        delay(1000)
    }
}

fun main() = runBlocking {
    tickerFlow().collect { value ->
        println("$value 秒经过")
    }
}

数据库变化监测​

data class User(val id: Int, val name: String)

fun observeUsers(): Flow<List<User>> = flow {
    repeat(3) { i ->
        delay(1000)
        val users = listOf(
            User(i * 2, "用户${i * 2}"),
            User(i * 2 + 1, "用户${i * 2 + 1}")
        )
        emit(users)
    }
}

fun main() = runBlocking {
    observeUsers().collect { users ->
        println("更新的用户:")
        users.forEach { println("  - ${it.name}") }
    }
}

传感器数据​

fun sensorData(): Flow<Double> = flow {
    repeat(10) {
        val reading = (20..30).random() + Math.random()
        emit(reading)
        delay(500)
    }
}

fun main() = runBlocking {
    sensorData().collect { temperature ->
        println("温度: ${"%.1f".format(temperature)}°C")
    }
}

🔧 Flow运算符​

map - 转换​

fun main() = runBlocking {
    flowOf(1, 2, 3, 4, 5)
        .map { it * it }  // 平方
        .collect { println(it) }
    // 1, 4, 9, 16, 25
}

filter - 过滤​

fun main() = runBlocking {
    flowOf(1, 2, 3, 4, 5)
        .filter { it % 2 == 0 }  // 仅偶数
        .collect { println(it) }
    // 2, 4
}

transform - 复杂转换​

fun main() = runBlocking {
    flowOf(1, 2, 3)
        .transform { value ->
            emit("转换: $value")
            emit("平方: ${value * value}")
        }
        .collect { println(it) }
}

take - 数量限制​

fun main() = runBlocking {
    flowOf(1, 2, 3, 4, 5)
        .take(3)  // 仅前3个
        .collect { println(it) }
    // 1, 2, 3
}

🎨 高级运算符​

buffer - 缓冲​

fun main() = runBlocking {
    val time = measureTimeMillis {
        flow {
            repeat(3) {
                delay(100)
                emit(it)
            }
        }
            .buffer()  // 通过缓冲提升性能
            .collect {
                delay(300)
                println(it)
            }
    }
    println("耗时: ${time}ms")
}

conflate - 仅最新值​

fun main() = runBlocking {
    flow {
        repeat(5) {
            delay(100)
            emit(it)
        }
    }
        .conflate()  // 如果收集者较慢则跳过中间值
        .collect {
            delay(300)
            println("收集: $it")
        }
}

zip - 结合​

fun main() = runBlocking {
    val numbers = flowOf(1, 2, 3)
    val strings = flowOf("一", "二", "三")

    numbers.zip(strings) { num, str ->
        "$num - $str"
    }.collect { println(it) }
    // 1 - 一
    // 2 - 二
    // 3 - 三
}

combine - 组合​

fun main() = runBlocking {
    val flow1 = flow {
        emit("A")
        delay(500)
        emit("B")
    }

    val flow2 = flow {
        emit(1)
        delay(200)
        emit(2)
        delay(200)
        emit(3)
    }

    flow1.combine(flow2) { a, b ->
        "$a$b"
    }.collect { println(it) }
}

🔥 实用模式​

搜索功能​

fun searchQuery(query: String): Flow<List<String>> = flow {
    delay(300)  // 防抖
    val results = listOf(
        "${query}1",
        "${query}2",
        "${query}3"
    ).filter { it.contains(query) }
    emit(results)
}

fun main() = runBlocking {
    flowOf("kot", "kotlin", "kotli")
        .flatMapLatest { query ->
            searchQuery(query)
        }
        .collect { results ->
            println("搜索结果: $results")
        }
}

分页​

fun loadPage(page: Int): Flow<List<String>> = flow {
    delay(500)
    val items = List(10) { "项目${page * 10 + it}" }
    emit(items)
}

fun main() = runBlocking {
    flowOf(1, 2, 3)
        .flatMapConcat { page ->
            loadPage(page)
        }
        .collect { items ->
            println("加载页面: $items")
        }
}

错误处理​

fun riskyFlow(): Flow<Int> = flow {
    emit(1)
    emit(2)
    throw Exception("发生错误!")
    emit(3)
}

fun main() = runBlocking {
    riskyFlow()
        .catch { e ->
            println("捕获错误: ${e.message}")
            emit(-1)  // 默认值
        }
        .collect { println("值: $it") }
}

重试​

var attemptCount = 0

fun unstableFlow(): Flow<String> = flow {
    attemptCount++
    if (attemptCount < 3) {
        throw Exception("临时错误")
    }
    emit("成功!")
}

fun main() = runBlocking {
    unstableFlow()
        .retry(3) { cause ->
            println("重试... (${cause.message})")
            delay(1000)
            true
        }
        .collect { println(it) }
}

🛠️ StateFlow和SharedFlow​

StateFlow - 状态管理​

class Counter {
    private val _count = MutableStateFlow(0)
    val count: StateFlow<Int> = _count

    fun increment() {
        _count.value++
    }
}

fun main() = runBlocking {
    val counter = Counter()

    launch {
        counter.count.collect { value ->
            println("计数: $value")
        }
    }

    delay(100)
    counter.increment()
    delay(100)
    counter.increment()
    delay(100)
}

SharedFlow - 事件​

class EventBus {
    private val _events = MutableSharedFlow<String>()
    val events: SharedFlow<String> = _events

    suspend fun emit(event: String) {
        _events.emit(event)
    }
}

fun main() = runBlocking {
    val bus = EventBus()

    // 订阅者1
    launch {
        bus.events.collect { event ->
            println("订阅者1: $event")
        }
    }

    // 订阅者2
    launch {
        bus.events.collect { event ->
            println("订阅者2: $event")
        }
    }

    delay(100)
    bus.emit("事件1")
    delay(100)
    bus.emit("事件2")
    delay(100)
}

🤔 常见问题​

Q1. Flow何时执行?​

A: 调用collect时执行(冷流)!

fun main() = runBlocking {
    val flow = flow {
        println("Flow 开始!")
        emit(1)
    }

    println("Flow 创建")
    delay(1000)

    println("调用 collect")
    flow.collect { println(it) }
}
// Flow 创建
// (等待1秒)
// 调用 collect
// Flow 开始!
// 1

Q2. List和Flow的区别是?​

A: List是立即的,Flow是渐进的!

fun main() = runBlocking {
    // List - 所有数据在内存中
    val list = listOf(1, 2, 3)

    // Flow - 需要时才生成
    val flow = flow {
        repeat(3) {
            delay(1000)
            emit(it)
        }
    }
}

Q3. 多次收集Flow会怎样?​

A: 每次都会重新执行!

fun main() = runBlocking {
    val flow = flow {
        println("执行!")
        emit(1)
    }

    flow.collect { println("第一次收集: $it") }
    flow.collect { println("第二次收集: $it") }
}
// 执行!
// 第一次收集: 1
// 执行!
// 第二次收集: 1

🎬 总结​

用Flow实现响应式编程!

核心要点:
✅ 异步数据流
✅ 冷流(收集时执行)
✅ map、filter等运算符
✅ 用StateFlow管理状态
✅ 用SharedFlow传递事件

下一步: 在Context和Dispatcher中了解协程执行环境!