一、实现UDF函数——更细粒度的控制流
1.1 函数类(Function Classes)
Flink暴露了所有udf函数的接口(实现方式为接口或者抽象类)。例如MapFunction, FilterFunction, ProcessFunction等等。
下面例子实现了FilterFunction接口:
class FilterFilter extends FilterFunction[String] {
override def filter(value: String): Boolean = {
value.contains("flink")
}
}
val flinkTweets = tweets.filter(new FlinkFilter)
还可以将函数实现成匿名类
val flinkTweets = tweets.filter(
new RichFilterFunction[String] {
override def filter(value: String): Boolean = {
value.contains("flink")
}
}
)
我们filter的字符串"flink"还可以当作参数传进去。
val tweets: DataStream[String] = ...
val flinkTweets = tweets.filter(new KeywordFilter("flink"))
class KeywordFilter(keyWord: String) extends FilterFunction[String] {
override def filter(value: String): Boolean = {
value.contains(keyWord)
}
}
1.2 匿名函数(Lambda Functions)
val tweets: DataStream[String] = ...
val flinkTweets = tweets.filter(_.contains("flink"))
1.3 富函数(Rich Functions)
“富函数”是DataStream API提供的一个函数类的接口,所有Flink函数类都有其Rich版本。它与常规函数的不同在于,可以获取运行环境的上下文,并拥有一些生命周期方法,所以可以实现更复杂的功能。
- RichMapFunction
- RichFlatMapFunction
- RichFilterFunction
- …
Rich Function有一个生命周期的概念。典型的生命周期方法有:
open()
方法是rich function的初始化方法,当一个算子例如map或者filter被调用之前open()会被调用。close()
方法是生命周期中的最后一个调用的方法,做一些清理工作。getRuntimeContext()
方法提供了函数的RuntimeContext的一些信息,例如函数执行的并行度,任务的名字,以及state状态
class MyFlatMap extends RichFlatMapFunction[Int, (Int, Int)] {
var subTaskIndex = 0
override def open(configuration: Configuration): Unit = {
subTaskIndex = getRuntimeContext.getIndexOfThisSubtask
// 以下可以做一些初始化工作,例如建立一个和HDFS的连接
}
override def flatMap(in: Int, out: Collector[(Int, Int)]): Unit = {
if (in % 2 == subTaskIndex) {
out.collect((subTaskIndex, in))
}
}
override def close(): Unit = {
// 以下做一些清理工作,例如断开和HDFS的连接。
}
}
二、Sink定义
Flink没有类似于spark中foreach方法,让用户进行迭代的操作。虽有对外的输出操作都要利用Sink完成。最后通过类似如下方式完成整个任务最终输出操作。
stream.addSink(new MySink(xxxx))
官方提供了一部分的框架的sink。除此以外,需要用户自定义实现sink。
2.1 Kafka
pom.xml文件定义
<!-- https://mvnrepository.com/artifact/org.apache.flink/flink-connector-kafka-0.11 -->
<dependency>
<groupId>org.apache.flink</groupId>
<artifactId>flink-connector-kafka-0.11_2.11</artifactId>
<version>1.10.0</version>
</dependency>
主函数中添加sink:
val union = high.union(low).map(_.temperature.toString)
union.addSink(new FlinkKafkaProducer011[String]("localhost:9092", "test", new SimpleStringSchema()))
2.2 Redis
pom.xml文件定义
<!-- https://mvnrepository.com/artifact/org.apache.bahir/flink-connector-redis -->
<dependency>
<groupId>org.apache.bahir</groupId>
<artifactId>flink-connector-redis_2.11</artifactId>
<version>1.0</version>
</dependency>
定义一个redis的mapper类,用于定义保存到redis时调用的命令:
class MyRedisMapper extends RedisMapper[SensorReading]{
override def getCommandDescription: RedisCommandDescription = {
new RedisCommandDescription(RedisCommand.HSET, "sensor_temperature")
}
override def getValueFromData(t: SensorReading): String = t.temperature.toString
override def getKeyFromData(t: SensorReading): String = t.id
}
在主函数中调用:
val conf = new FlinkJedisPoolConfig.Builder().setHost("localhost").setPort(6379).build()
dataStream.addSink( new RedisSink[SensorReading](conf, new MyRedisMapper) )
2.3 Elasticsearch
pom.xml
<dependency>
<groupId>org.apache.flink</groupId>
<artifactId>flink-connector-elasticsearch6_2.11</artifactId>
<version>1.10.0</version>
</dependency>
在主函数中调用:
val httpHosts = new util.ArrayList[HttpHost]()
httpHosts.add(new HttpHost("localhost", 9200))
val esSinkBuilder = new ElasticsearchSink.Builder[SensorReading]( httpHosts, new ElasticsearchSinkFunction[SensorReading] {
override def process(t: SensorReading, runtimeContext: RuntimeContext, requestIndexer: RequestIndexer): Unit = {
println("saving data: " + t)
val json = new util.HashMap[String, String]()
json.put("data", t.toString)
val indexRequest = Requests.indexRequest().index("sensor").`type`("readingData").source(json)
requestIndexer.add(indexRequest)
println("saved successfully")
}
} )
dataStream.addSink( esSinkBuilder.build() )
5.6.4 JDBC 自定义sink
<!-- https://mvnrepository.com/artifact/mysql/mysql-connector-java -->
<dependency>
<groupId>mysql</groupId>
<artifactId>mysql-connector-java</artifactId>
<version>5.1.44</version>
</dependency>
添加MyJdbcSink
class MyJdbcSink() extends RichSinkFunction[SensorReading]{
var conn: Connection = _
var insertStmt: PreparedStatement = _
var updateStmt: PreparedStatement = _
// open 主要是创建连接
override def open(parameters: Configuration): Unit = {
super.open(parameters)
conn = DriverManager.getConnection("jdbc:mysql://localhost:3306/test", "root", "123456")
insertStmt = conn.prepareStatement("INSERT INTO temperatures (sensor, temp) VALUES (?, ?)")
updateStmt = conn.prepareStatement("UPDATE temperatures SET temp = ? WHERE sensor = ?")
}
// 调用连接,执行sql
override def invoke(value: SensorReading, context: SinkFunction.Context[_]): Unit = {
updateStmt.setDouble(1, value.temperature)
updateStmt.setString(2, value.id)
updateStmt.execute()
if (updateStmt.getUpdateCount == 0) {
insertStmt.setString(1, value.id)
insertStmt.setDouble(2, value.temperature)
insertStmt.execute()
}
}
override def close(): Unit = {
insertStmt.close()
updateStmt.close()
conn.close()
}
}
在main方法中增加,把明细保存到mysql中
dataStream.addSink(new MyJdbcSink())