目录
案例介绍
示例:
自定义Source,模拟1000条订单数据,每条数据间隔1秒钟(订单编号ID、用户编号ID、商品编号ID、消费金额、消费时间 )
要求:
- 随机生成订单编号ID(UUID)
- 随机生成用户编号ID(user1-user10)
- 随机生成商品编号ID(goods1-goods20)
- 随机生成消费金额(100-1000)
- 消费时间为当前系统时间
统计:
- 每30秒钟,统计一次各用户的最大消费订单信息,将结果写入MySQL;
- 统计30秒内,各用户的消费总额和订单数量,该数据每10秒更新一次,将结果打印输出;
- 统计30秒内,各商品的销售数量,该数据每10秒更新一次, 将结果打印输出。
开发步骤
-
创建订单样例类;
-
获取流处理环境;
-
创建自定义数据源;
(1)循环1000次; (2)随机构建订单信息; (3)上下文收集数据; (4)每隔一秒执行一次循环;
-
处理数据;
-
打印数据;
-
写入MySQL;
-
执行任务。
具体代码
定义样例类
//样例类
case class
Order(itemId: String, userId: String, goodsId: String, price: Int, createTime: Long)
Environment和Source流程
//TODO:1.environment
val env: StreamExecutionEnvironment = StreamExecutionEnvironment.getExecutionEnvironment
env.setParallelism(1)
//TODO:2.source,生产随机订单数据
val sourceDStream: DataStream[Order] = env.addSource(new OrderSourceFunction)
自定义随机生成订单类OrderSourceFunction
//自定义随机生成订单类
class OrderSourceFunction extends RichSourceFunction[Order] {
//定义变量
private var count: Long = 0L
private var isRunning: Boolean = true
override def run(sourceContext: SourceFunction.SourceContext[Order]): Unit = {
// 使用while循环生成1000个订单(订单编号ID、用户编号ID、商品编号ID、消费金额、消费时间 )
while (isRunning && count < 1000) {
// 随机生成订单ID(UUID)
val itemId: String = UUID.randomUUID().toString
// 随机生成用户编号ID(user1-user10)
val userID: String = "user" + (Random.nextInt(10) + 1)
// 随机生成商品编号ID(goods1-goods20)
val goodsID: String = "goods" + (Random.nextInt(20) + 1)
// 随机生成消费金额(100~1000)
val price = Random.nextInt(900) + 101
// 消费时间为当前系统时间
val createTime: Long = System.currentTimeMillis()
// println(new SimpleDateFormat("yyyy-MM-dd HH:mm:ss").format(new Date(createTime)))
// 收集数据
sourceContext.collect(Order(itemId, userID, goodsID, price, createTime))
// 统计生成的订单数
count += 1
// 每隔1秒生成一个订单
TimeUnit.SECONDS.sleep(1)
// TimeUnit.MILLISECONDS.sleep(1)
}
}
override def cancel(): Unit = {
isRunning = false
}
}
Transformation流程,实现案例的需求
//TODO:3.transformation,对数据进行窗口统计,并将有些数据写入MySQl中
(1)每30秒钟,统计一次各用户的最大消费订单信息,将结果写入MySQL;
// Order(7485091d-341c-4498-b00b-4cfca479de79,user6,goods9,815,1634218271508)
val resultDStream01: DataStream[Order] = sourceDStream
.assignAscendingTimestamps(_.createTime)
.keyBy(_.userId)
.window(TumblingEventTimeWindows.of(Time.seconds(30)))
.maxBy("price")
(2)统计30秒内,各用户的消费总额和订单数量,该数据每10秒更新一次,将结果打印输出;
// (user9,3440,4)
val resultDStream02: DataStream[(String, Int, Int)] = sourceDStream
.assignAscendingTimestamps(_.createTime)
.map(Data => (Data.userId, Data.price, 1))
.keyBy(_._1)
.window(SlidingEventTimeWindows.of(Time.seconds(30), Time.seconds(19)))
.reduce((preData, curData) => (preData._1, preData._2 + curData._2, preData._3 + curData._3))
(3)统计30秒内,各商品的销售数量,该数据每10秒更新一次, 将结果打印输出。
// (goods10,2)
val resultDStream03: DataStream[(String, Int)] = sourceDStream
.assignAscendingTimestamps(_.createTime)
.map(Data => (Data.goodsId, 1))
.keyBy(_._1)
.window(SlidingEventTimeWindows.of(Time.seconds(30), Time.seconds(10)))
.reduce((preData, curData) => (preData._1, preData._2 + curData._2))
Sink与Execution流程
//TODO:4.sink
resultDStream01.print("用户的最大消费订单信息:")
resultDStream01.addSink(new MySQLSinkFunction)
resultDStream02.print("用户的消费总额和订单数量:")
resultDStream03.print("商品的销售数量:")
// sourceDStream.writeAsCsv("src\\main\\resources\\new_OrderSource.csv", WriteMode.OVERWRITE)
//TODO:5.execution
env.execute("OrderStream Job")
自定义MySQLSinkFunction类,将数据存储于MySQL中
//自定义MySQLSinkFunction类,将数据存储于MySQL中
class MySQLSinkFunction extends RichSinkFunction[Order] {
var connection: Connection = null
var ps: PreparedStatement = null
override def open(parameters: Configuration): Unit = {
//加载驱动,打开连接
Class.forName("com.mysql.jdbc.Driver")
connection = DriverManager.getConnection(
"jdbc:mysql://localhost:3306/数据库名?useSSL=false&characterEncoding=utf8",
"用户名",
"密码")
ps = connection.prepareStatement("insert into 表名(itemId,userId,goodsId,price,createTime) values (?,?,?,?,?)")
}
override def invoke(value: Order, context: SinkFunction.Context): Unit = {
//元组的主键,取系统此时的时间戳
ps.setString(1, value.itemId)
ps.setString(2, value.userId)
ps.setString(3, value.goodsId)
ps.setInt(4, value.price)
ps.setLong(5, value.createTime)
//执行sql语句
ps.executeUpdate()
}
override def close(): Unit = {
if (connection != null)
connection.close()
if (ps != null)
ps.close()
}
}
实现结果
在控制台输出的结果:
存储于MySQL中的数据:
(数据表的结构)
表中数据
(查询产生的数据的数目)
本案例的完整代码
import org.apache.flink.configuration.Configuration
import org.apache.flink.streaming.api.functions.sink.{RichSinkFunction, SinkFunction}
import org.apache.flink.streaming.api.functions.source.{RichSourceFunction, SourceFunction}
import org.apache.flink.streaming.api.scala.{DataStream, StreamExecutionEnvironment, _}
import org.apache.flink.streaming.api.windowing.assigners.{SlidingEventTimeWindows, TumblingEventTimeWindows}
import org.apache.flink.streaming.api.windowing.time.Time
import java.util.UUID
import java.util.concurrent.TimeUnit
import scala.util.Random
import java.sql.{Connection, DriverManager, PreparedStatement}
object OrderStream {
//样例类
case class
Order(itemId: String, userId: String, goodsId: String, price: Int, createTime: Long)
def main(args: Array[String]): Unit = {
//TODO:1.environment
val env: StreamExecutionEnvironment = StreamExecutionEnvironment.getExecutionEnvironment
env.setParallelism(1)
//TODO:2.source,生产随机订单数据
val sourceDStream: DataStream[Order] = env.addSource(new OrderSourceFunction)
//TODO:3.transformation,对数据进行窗口统计,并将有些数据写入MySQl中
(1)每30秒钟,统计一次各用户的最大消费订单信息,将结果写入MySQL;
// Order(7485091d-341c-4498-b00b-4cfca479de79,user6,goods9,815,1634218271508)
val resultDStream01: DataStream[Order] = sourceDStream
.assignAscendingTimestamps(_.createTime)
.keyBy(_.userId)
.window(TumblingEventTimeWindows.of(Time.seconds(30)))
.maxBy("price")
(2)统计30秒内,各用户的消费总额和订单数量,该数据每10秒更新一次,将结果打印输出;
// (user9,3440,4)
val resultDStream02: DataStream[(String, Int, Int)] = sourceDStream
.assignAscendingTimestamps(_.createTime)
.map(Data => (Data.userId, Data.price, 1))
.keyBy(_._1)
.window(SlidingEventTimeWindows.of(Time.seconds(30), Time.seconds(19)))
.reduce((preData, curData) => (preData._1, preData._2 + curData._2, preData._3 + curData._3))
(3)统计30秒内,各商品的销售数量,该数据每10秒更新一次, 将结果打印输出。
// (goods10,2)
val resultDStream03: DataStream[(String, Int)] = sourceDStream
.assignAscendingTimestamps(_.createTime)
.map(Data => (Data.goodsId, 1))
.keyBy(_._1)
.window(SlidingEventTimeWindows.of(Time.seconds(30), Time.seconds(10)))
.reduce((preData, curData) => (preData._1, preData._2 + curData._2))
//TODO:4.sink
resultDStream01.print("用户的最大消费订单信息:")
resultDStream01.addSink(new MySQLSinkFunction)
resultDStream02.print("用户的消费总额和订单数量:")
resultDStream03.print("商品的销售数量:")
// sourceDStream.writeAsCsv("src\\main\\resources\\new_OrderSource.csv", WriteMode.OVERWRITE)
//TODO:5.execution
env.execute("OrderStream Job")
}
//自定义随机生成订单类
class OrderSourceFunction extends RichSourceFunction[Order] {
//定义变量
private var count: Long = 0L
private var isRunning: Boolean = true
override def run(sourceContext: SourceFunction.SourceContext[Order]): Unit = {
// 使用while循环生成1000个订单(订单编号ID、用户编号ID、商品编号ID、消费金额、消费时间 )
while (isRunning && count < 1000) {
// 随机生成订单ID(UUID)
val itemId: String = UUID.randomUUID().toString
// 随机生成用户编号ID(user1-user10)
val userID: String = "user" + (Random.nextInt(10) + 1)
// 随机生成商品编号ID(goods1-goods20)
val goodsID: String = "goods" + (Random.nextInt(20) + 1)
// 随机生成消费金额(100~1000)
val price = Random.nextInt(900) + 101
// 消费时间为当前系统时间
val createTime: Long = System.currentTimeMillis()
// println(new SimpleDateFormat("yyyy-MM-dd HH:mm:ss").format(new Date(createTime)))
// 收集数据
sourceContext.collect(Order(itemId, userID, goodsID, price, createTime))
// 统计生成的订单数
count += 1
// 每隔1秒生成一个订单
TimeUnit.SECONDS.sleep(1)
// TimeUnit.MILLISECONDS.sleep(1)
}
}
override def cancel(): Unit = {
isRunning = false
}
}
//自定义MySQLSinkFunction类,将数据存储于MySQL中
class MySQLSinkFunction extends RichSinkFunction[Order] {
var connection: Connection = null
var ps: PreparedStatement = null
override def open(parameters: Configuration): Unit = {
//加载驱动,打开连接
Class.forName("com.mysql.jdbc.Driver")
connection = DriverManager.getConnection(
"jdbc:mysql://localhost:3306/数据库名?useSSL=false&characterEncoding=utf8",
"用户名",
"密码")
ps = connection.prepareStatement("insert into 表名(itemId,userId,goodsId,price,createTime) values (?,?,?,?,?)")
}
override def invoke(value: Order, context: SinkFunction.Context): Unit = {
//元组的主键,取系统此时的时间戳
ps.setString(1, value.itemId)
ps.setString(2, value.userId)
ps.setString(3, value.goodsId)
ps.setInt(4, value.price)
ps.setLong(5, value.createTime)
//执行sql语句
ps.executeUpdate()
}
override def close(): Unit = {
if (connection != null)
connection.close()
if (ps != null)
ps.close()
}
}
}
总结:本案例使用Stream API 和·Window API 对自定义数据流(循环生成1000条)进行处理并将数据存储于MySQL数据中!
后续会继续更新有关Flink Stream及Flink SQL的内容!
(注:第12次发文,如有错误和疑问,欢迎在评论区指出!)
——2021.10.14