在電商網站中,訂單的支付作爲直接與錢掛鉤的一環,在業務流程中非常重要。對於訂單而言,爲了正確控制業務流程,也爲了增加用戶的支付意願,網站一般會設置一個支付失效時間,超過一段時間沒支付的訂單就會被取消。另外,對於訂單的支付,還應該保證最終支付的正確性,可以通過第三方支付平臺的交易數據來做一個實時對賬
第一個實現的效果,實時獲取訂單數據,分析訂單的支付情況,分別實時統計支付成功的和15分鐘後支付超時的情況
新建一個maven項目,這是基礎依賴,如果之前引入了,就不用加了
<properties>
<maven.compiler.source>8</maven.compiler.source>
<maven.compiler.target>8</maven.compiler.target>
<flink.version>1.10.1</flink.version>
<scala.binary.version>2.12</scala.binary.version>
<kafka.version>2.2.0</kafka.version>
</properties>
<dependencies>
<dependency>
<groupId>org.apache.flink</groupId>
<artifactId>flink-scala_${scala.binary.version}</artifactId>
<version>${flink.version}</version>
</dependency>
<dependency>
<groupId>org.apache.flink</groupId>
<artifactId>flink-streaming-scala_${scala.binary.version}</artifactId>
<version>${flink.version}</version>
</dependency>
<dependency>
<groupId>org.apache.kafka</groupId>
<artifactId>kafka_${scala.binary.version}</artifactId>
<version>${kafka.version}</version>
</dependency>
<dependency>
<groupId>org.apache.flink</groupId>
<artifactId>flink-connector-kafka_${scala.binary.version}</artifactId>
<version>${flink.version}</version>
</dependency>
<dependency>
<groupId>cn.hutool</groupId>
<artifactId>hutool-all</artifactId>
<version>5.5.6</version>
</dependency>
<dependency>
<groupId>org.apache.flink</groupId>
<artifactId>flink-table-planner-blink_2.12</artifactId>
<version>1.10.1</version>
</dependency>
</dependencies>
這個場景需要用到cep,所以再加入cep依賴
<dependencies>
<dependency>
<groupId>org.apache.flink</groupId>
<artifactId>flink-cep-scala_${scala.binary.version}</artifactId>
<version>${flink.version}</version>
</dependency>
</dependencies>
準備數據源文件src/main/resources/OrderLog.csv:
1234,create,,1611047605
1235,create,,1611047606
1236,create,,1611047606
1234,pay,akdb3833,1611047616
把java目錄改爲scala,新建com.mafei.orderPayMonitor.OrderTimeoutMonitor.scala 的object
/*
*
* @author mafei
* @date 2021/1/31
*/
package com.mafei.orderPayMonitor
import org.apache.flink.cep.{PatternSelectFunction, PatternTimeoutFunction}
import org.apache.flink.cep.scala.CEP
import org.apache.flink.cep.scala.pattern.Pattern
import org.apache.flink.streaming.api.TimeCharacteristic
import org.apache.flink.streaming.api.scala.{OutputTag, StreamExecutionEnvironment, createTypeInformation}
import org.apache.flink.streaming.api.windowing.time.Time
import java.util
/**
* 定義輸入樣例類類型,
*
* @param orderId 訂單id
* @param eventType 事件類別: 創建訂單create還是支付訂單pay
* @param txId 支付流水號
* @param ts 時間
*/
case class OrderEvent(orderId: Long, eventType:String,txId: String, ts: Long)
/**
* 定義輸出樣例類類型,
*/
case class OrderResult(orderId: Long, resultMsg: String)
object OrderTimeoutMonitor {
def main(args: Array[String]): Unit = {
val env = StreamExecutionEnvironment.getExecutionEnvironment
env.setParallelism(1)
env.setStreamTimeCharacteristic(TimeCharacteristic.EventTime)
// 1、從文件中讀取數據
val resource = getClass.getResource("/OrderLog.csv")
val orderEvnetStream = env.readTextFile(resource.getPath)
.map(d=>{
val arr = d.split(",")
OrderEvent(arr(0).toLong,arr(1),arr(2), arr(3).toLong) //把數據讀出來轉換成想要的樣例類類型
}).assignAscendingTimestamps(_.ts * 1000L) //指定ts字段
.keyBy(_.orderId) //按照訂單id分組
/**
* 2、定義事件-匹配模式
* 定義15分鐘內能發現訂單創建和支付
*/
val orderPayPattern = Pattern
.begin[OrderEvent]("create").where(_.eventType == "create") //先出現一個訂單創建的事件
.followedBy("pay").where(_.eventType == "pay") //後邊再出來一個支付事件
.within(Time.minutes(15)) //定義在15分鐘以內,觸發這2個事件
// 3、將pattern應用到流裏面,進行模式檢測
val patternStream = CEP.pattern(orderEvnetStream, orderPayPattern)
//4、定義一個側輸出流標籤,用於處理超時事件
val orderTimeoutTag = new OutputTag[OrderResult]("orderTimeout")
// 5、調用select 方法,提取並處理匹配的成功字符事件以及超時事件
val resultStream = patternStream.select(
orderTimeoutTag,
new OrderTimeoutSelect(),
new OrderPaySelect()
)
resultStream.print("pay")
resultStream.getSideOutput(orderTimeoutTag).print()
env.execute(" order timeout monitor")
}
}
//獲取超時之後定義的事件還沒觸發的情況,也就是訂單支付超時了。
class OrderTimeoutSelect() extends PatternTimeoutFunction[OrderEvent, OrderResult]{
override def timeout(map: util.Map[String, util.List[OrderEvent]], l: Long): OrderResult = {
val timeoutOrderId = map.get("create").iterator().next().orderId
OrderResult(timeoutOrderId, "超時了。。。。超時時間:"+l)
}
}
class OrderPaySelect() extends PatternSelectFunction[OrderEvent, OrderResult]{
override def select(map: util.Map[String, util.List[OrderEvent]]): OrderResult = {
val orderTs = map.get("create").iterator().next().ts
val paydTs = map.get("pay").iterator().next().ts
val payedOrderId = map.get("pay").iterator().next().orderId
OrderResult(payedOrderId, "訂單支付成功,下單時間:"+orderTs+" 支付時間:"+paydTs)
}
}
代碼結構及運行效果
用ProcessFunction來實現上面的場景
csv還可以用上面的數據,新建一個scala的object src/main/scala/com/mafei/orderPayMonitor/OrderTimeoutMonitorWithProcessFunction.scala
/*
*
* @author mafei
* @date 2021/1/31
*/
package com.mafei.orderPayMonitor
import org.apache.flink.api.common.state.{ValueState, ValueStateDescriptor}
import org.apache.flink.streaming.api.TimeCharacteristic
import org.apache.flink.streaming.api.functions.KeyedProcessFunction
import org.apache.flink.streaming.api.scala.{OutputTag, StreamExecutionEnvironment, createTypeInformation}
import org.apache.flink.util.Collector
object OrderTimeoutMonitorWithProcessFunction {
def main(args: Array[String]): Unit = {
val env = StreamExecutionEnvironment.getExecutionEnvironment
env.setParallelism(1)
env.setStreamTimeCharacteristic(TimeCharacteristic.EventTime)
// 1、從文件中讀取數據
val resource = getClass.getResource("/OrderLog.csv")
val orderEventStream = env.readTextFile(resource.getPath)
.map(d=>{
val arr = d.split(",")
OrderEvent(arr(0).toLong,arr(1),arr(2), arr(3).toLong) //把數據讀出來轉換成想要的樣例類類型
}).assignAscendingTimestamps(_.ts * 1000L) //指定ts字段
.keyBy(_.orderId) //按照訂單id分組
val resultStream = orderEventStream
.process(new OrderPayMatchProcess())
resultStream.print("支付成功的: ")
resultStream.getSideOutput(new OutputTag[OrderResult]("timeout")).print("訂單超時事件")
env.execute("訂單支付監控with ProcessFunction")
}
}
class OrderPayMatchProcess() extends KeyedProcessFunction[Long, OrderEvent, OrderResult]{
// 先定義狀態標識,標識create、payed、是否已經出現,以及對應的時間戳
lazy val isCreateOrderState: ValueState[Boolean] = getRuntimeContext.getState(new ValueStateDescriptor[Boolean]("isCreateOrderState", classOf[Boolean]))
lazy val isPayedOrderState: ValueState[Boolean] = getRuntimeContext.getState(new ValueStateDescriptor[Boolean]("isPayedOrderState", classOf[Boolean]))
lazy val timerTsState : ValueState[Long] = getRuntimeContext.getState(new ValueStateDescriptor[Long]("timerTsState", classOf[Long]))
// 定義一個側輸出流,捕獲timeout的訂單信息
val orderTimeoutOutputTag = new OutputTag[OrderResult]("timeout")
override def onTimer(timestamp: Long, ctx: KeyedProcessFunction[Long, OrderEvent, OrderResult]#OnTimerContext, out: Collector[OrderResult]): Unit = {
//到這裏,肯定不會出現訂單創建和支付同時存在的情況,因爲會在processElement處理掉
//如果只有訂單創建
if (isCreateOrderState.value()){
ctx.output(orderTimeoutOutputTag,OrderResult(ctx.getCurrentKey,"訂單沒支付或超時"))
}else if(isPayedOrderState.value()){
ctx.output(orderTimeoutOutputTag, OrderResult(ctx.getCurrentKey,"只有支付,沒看到訂單提交"))
}
isCreateOrderState.clear()
isPayedOrderState.clear()
timerTsState.clear()
}
override def processElement(i: OrderEvent, context: KeyedProcessFunction[Long, OrderEvent, OrderResult]#Context, collector: Collector[OrderResult]): Unit = {
/**
* 判斷當前事件類型,是create還是pay
* 分幾種情況:
* 1、判斷create和pay都來了
* 要看有沒有超時,沒有超時就正常輸出
* 超時了輸出到側輸出流
* 2、create或者pay有一個沒來
* 註冊一個定時器等着,然後等定時器觸發後再輸出
*
*/
val isCreate = isCreateOrderState.value()
val isPayed = isPayedOrderState.value()
val timerTs = timerTsState.value()
// 1、create來了
if (i.eventType == "create"){
// 1.1 如果已經支付過了,那是正常支付完成,輸出匹配成功的結果
if (isPayed){
isCreateOrderState.clear()
isPayedOrderState.clear()
timerTsState.clear()
context.timerService().deleteEventTimeTimer(timerTs)
collector.collect(OrderResult(context.getCurrentKey,"支付成功"))
}else{ //如果沒有支付過,那註冊一個定時器,等待15分鐘後觸發
context.timerService().registerEventTimeTimer(i.ts)
timerTsState.update(i.ts * 1000L + 900*1000L)
isCreateOrderState.update(true)
}
}
else if(i.eventType == "pay"){ //如果當前事件是支付事件
if(isCreate){ //判讀訂單創建事件已經發生
if(i.ts * 1000L < timerTs){ // 創建訂單到支付的時間在超時時間內,代表正常支付
collector.collect(OrderResult(context.getCurrentKey,"支付成功"))
}else{
context.output(orderTimeoutOutputTag, OrderResult(context.getCurrentKey,"已經支付,但是沒有找到訂單超時了"))
}
isCreateOrderState.clear()
isPayedOrderState.clear()
timerTsState.clear()
context.timerService().deleteEventTimeTimer(timerTs)
}else{ //如果沒看到訂單創建的事件,那就註冊一個定時器等着
context.timerService().registerEventTimeTimer(i.ts)
isPayedOrderState.update(true)
timerTsState.update(i.ts)
}
}
}
}
代碼結構及運行效果
上面實現了監測用戶支付的情況,實際中還需要對支付後的賬單跟第三方支付平臺做一個實時對賬功能
會涉及到2條數據流(支付和賬單)的合流計算
這裏模擬賬單,所以需要準備一個數據ReceiptLog.csv
akdb3833,alipay,1611047619
akdb3832,wechat,1611049617
上代碼: src/main/scala/com/mafei/orderPayMonitor/TxMatch.scala
/*
*
* @author mafei
* @date 2021/1/31
*/
package com.mafei.orderPayMonitor
import com.mafei.orderPayMonitor.OrderTimeoutMonitor.getClass
import org.apache.flink.api.common.state.{ValueState, ValueStateDescriptor}
import org.apache.flink.streaming.api.TimeCharacteristic
import org.apache.flink.streaming.api.functions.co.CoProcessFunction
import org.apache.flink.streaming.api.scala.{OutputTag, StreamExecutionEnvironment, createTypeInformation}
import org.apache.flink.util.Collector
case class ReceiptEvent(orderId: String, payChannel:String, ts: Long)
object TxMatch {
def main(args: Array[String]): Unit = {
val env = StreamExecutionEnvironment.getExecutionEnvironment
env.setParallelism(1)
env.setStreamTimeCharacteristic(TimeCharacteristic.EventTime)
// 1、從訂單文件中讀取數據
val resource = getClass.getResource("/OrderLog.csv")
val orderEventStream = env.readTextFile(resource.getPath)
.map(d=>{
val arr = d.split(",")
OrderEvent(arr(0).toLong,arr(1),arr(2), arr(3).toLong) //把數據讀出來轉換成想要的樣例類類型
}).assignAscendingTimestamps(_.ts * 1000L) //指定ts字段
.filter(_.eventType=="pay")
.keyBy(_.txId) //按照交易id分組
// 2、從賬單中讀取數據
val receiptResource = getClass.getResource("/ReceiptLog.csv")
val receiptEventStream = env.readTextFile(receiptResource.getPath)
.map(d=>{
val arr = d.split(",")
ReceiptEvent(arr(0),arr(1),arr(2).toLong) //把數據讀出來轉換成想要的樣例類類型
}).assignAscendingTimestamps(_.ts * 1000L) //指定ts字段
.keyBy(_.orderId) //按照訂單id分組
// 3、合併兩條流,進行處理
val resultStream = orderEventStream.connect(receiptEventStream)
.process(new TxPayMatchResult())
resultStream.print("match: ")
resultStream.getSideOutput(new OutputTag[OrderEvent]("unmatched-pay")).print("unmatched-pay")
resultStream.getSideOutput(new OutputTag[ReceiptEvent]("receipt")).print("unmatched-receipt")
env.execute()
}
}
class TxPayMatchResult() extends CoProcessFunction[OrderEvent,ReceiptEvent,(OrderEvent,ReceiptEvent)]{
lazy val orderEventState: ValueState[OrderEvent] = getRuntimeContext.getState(new ValueStateDescriptor[OrderEvent]("orderEvent", classOf[OrderEvent]))
lazy val receiptEventState: ValueState[ReceiptEvent] = getRuntimeContext.getState(new ValueStateDescriptor[ReceiptEvent]("payEvent", classOf[ReceiptEvent]))
// 定義自定義側輸出流
val unmatchedOrderEventTag = new OutputTag[OrderEvent]("unmatched-pay")
val unmatchedReceiptEventTag = new OutputTag[ReceiptEvent]("receipt")
override def processElement1(in1: OrderEvent, context: CoProcessFunction[OrderEvent, ReceiptEvent, (OrderEvent, ReceiptEvent)]#Context, collector: Collector[(OrderEvent, ReceiptEvent)]): Unit = {
//判斷支付賬單來了
val receiptEvent = receiptEventState.value()
if(receiptEvent != null){
//如果賬單已經過來了,那直接輸出
collector.collect((in1,receiptEvent))
orderEventState.clear()
receiptEventState.clear()
}else{
//如果沒來,那就註冊一個定時器,等待10秒鐘
context.timerService().registerEventTimeTimer(in1.ts*1000L + 10000L)
orderEventState.update(in1)
}
}
override def processElement2(in2: ReceiptEvent, context: CoProcessFunction[OrderEvent, ReceiptEvent, (OrderEvent, ReceiptEvent)]#Context, collector: Collector[(OrderEvent, ReceiptEvent)]): Unit = {
//判斷支付事件來了
val orderEvent = orderEventState.value()
if(orderEvent != null){
//如果賬單已經過來了,那直接輸出
collector.collect((orderEvent,in2))
orderEventState.clear()
receiptEventState.clear()
}else{
//如果沒來,那就註冊一個定時器,等待2秒鐘
context.timerService().registerEventTimeTimer(in2.ts*1000L + 2000L)
receiptEventState.update(in2)
}
}
override def onTimer(timestamp: Long, ctx: CoProcessFunction[OrderEvent, ReceiptEvent, (OrderEvent, ReceiptEvent)]#OnTimerContext, out: Collector[(OrderEvent, ReceiptEvent)]): Unit = {
if(orderEventState.value() != null){
ctx.output(unmatchedOrderEventTag, orderEventState.value())
}
else if(receiptEventState.value() != null){
ctx.output(unmatchedReceiptEventTag, receiptEventState.value())
}
orderEventState.clear()
receiptEventState.clear()
}
}
第二種, 使用join來實現這個效果
關於join的文檔:https://ci.apache.org/projects/flink/flink-docs-stable/dev/stream/operators/joining.html
這種方式優點是跟方便了,做了一層封裝,缺點也很明顯如果要實現一些複雜情況如沒匹配中的也輸出之類的就不行了,具體看實際場景需要
/*
*
* @author mafei
* @date 2021/1/31
*/
package com.mafei.orderPayMonitor
import com.mafei.orderPayMonitor.TxMatch.getClass
import org.apache.flink.streaming.api.TimeCharacteristic
import org.apache.flink.streaming.api.functions.co.ProcessJoinFunction
import org.apache.flink.streaming.api.scala.{StreamExecutionEnvironment, createTypeInformation}
import org.apache.flink.streaming.api.windowing.time.Time
import org.apache.flink.util.Collector
object TxMatchWithJoin {
def main(args: Array[String]): Unit = {
val env = StreamExecutionEnvironment.getExecutionEnvironment
env.setParallelism(1)
env.setStreamTimeCharacteristic(TimeCharacteristic.EventTime)
// 1、從訂單文件中讀取數據
val resource = getClass.getResource("/OrderLog.csv")
val orderEventStream = env.readTextFile(resource.getPath)
.map(d=>{
val arr = d.split(",")
OrderEvent(arr(0).toLong,arr(1),arr(2), arr(3).toLong) //把數據讀出來轉換成想要的樣例類類型
}).assignAscendingTimestamps(_.ts * 1000L) //指定ts字段
.filter(_.eventType=="pay")
.keyBy(_.txId) //按照交易id分組
// 2、從賬單中讀取數據
val receiptResource = getClass.getResource("/ReceiptLog.csv")
val receiptEventStream = env.readTextFile(receiptResource.getPath)
.map(d=>{
val arr = d.split(",")
ReceiptEvent(arr(0),arr(1),arr(2).toLong) //把數據讀出來轉換成想要的樣例類類型
}).assignAscendingTimestamps(_.ts * 1000L) //指定ts字段
.keyBy(_.orderId) //按照訂單id分組
val resultStream = orderEventStream.intervalJoin(receiptEventStream)
.between(Time.seconds(-3), Time.seconds(5))
.process(new TxMatchWithJoinResult())
resultStream.print()
env.execute()
}
}
class TxMatchWithJoinResult() extends ProcessJoinFunction[OrderEvent, ReceiptEvent,(OrderEvent,ReceiptEvent)]{
override def processElement(in1: OrderEvent, in2: ReceiptEvent, context: ProcessJoinFunction[OrderEvent, ReceiptEvent, (OrderEvent, ReceiptEvent)]#Context, collector: Collector[(OrderEvent, ReceiptEvent)]): Unit = {
collector.collect((in1,in2))
}
}