Spark Structured streaming API支持的输出源有：Console、Memory、File和Foreach。其中Console在前两篇博文中已有详述，而Memory使用非常简单。本文着重介绍File和Foreach两种方式，并介绍如何在源码基本扩展新的输出方式。

1. File

　　Structured Streaming支持将数据以File形式保存起来，其中支持的文件格式有四种:json、text、csv和parquet。其使用方式也非常简单只需设置checkpointLocation和path即可。checkpointLocation是检查点保存的路径，而path是真实数据保存的路径。

如下所示的测试例子：



// Create DataFrame representing the stream of input lines from connection to host:port 
val lines = spark.readStream 
.format("socket") 
.option("host", host) 
.option("port", port) 
.load() 
 
// Split the lines into words 
val words = lines.as[String].flatMap(_.split(" ")) 
 
// Generate running word count 
val wordCounts = words.groupBy("value").count() 
 
// Start running the query that prints the running counts to the console 
val query = wordCounts.writeStream 
.format("json") 
.option("checkpointLocation","root/jar") 
.option("path","/root/jar") 
.start()



注意：

    File形式不能设置"compelete"模型，只能设置"Append"模型。由于Append模型不能有聚合操作，所以将数据保存到外部File时，不能有聚合操作。

2. Foreach

　　foreach输出方式只需要实现ForeachWriter抽象类，并实现三个方法，当Structured Streaming接收到数据就会执行其三个方法，如下的测试示例：



/* 
* Licensed to the Apache Software Foundation (ASF) under one or more 
* contributor license agreements. See the NOTICE file distributed with 
* this work for additional information regarding copyright ownership. 
* The ASF licenses this file to You under the Apache License, Version 2.0 
* (the "License"); you may not use this file except in compliance with 
* the License. You may obtain a copy of the License at 
* 
* http://www.apache.org/licenses/LICENSE-2.0 
* 
* Unless required by applicable law or agreed to in writing, software 
* distributed under the License is distributed on an "AS IS" BASIS, 
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 
* See the License for the specific language governing permissions and 
* limitations under the License. 
*/ 
 
// scalastyle:off println 
package org.apache.spark.examples.sql.streaming 
 
import org.apache.spark.sql.SparkSession 
 
/** 
* Counts words in UTF8 encoded, '\n' delimited text received from the network. 
* 
* Usage: StructuredNetworkWordCount <hostname> <port> 
* <hostname> and <port> describe the TCP server that Structured Streaming 
* would connect to receive data. 
* 
* To run this on your local machine, you need to first run a Netcat server 
* `$ nc -lk 9999` 
* and then run the example 
* `$ bin/run-example sql.streaming.StructuredNetworkWordCount 
* localhost 9999` 
*/ 
object StructuredNetworkWordCount { 
def main(args: Array[String]) { 
if (args.length < 2) { 
System.err.println("Usage: StructuredNetworkWordCount <hostname> <port>") 
System.exit(1) 
} 
 
val host = args(0) 
val port = args(1).toInt 
 
val spark = SparkSession 
.builder 
.appName("StructuredNetworkWordCount") 
.getOrCreate() 
 
import spark.implicits._ 
 
// Create DataFrame representing the stream of input lines from connection to host:port 
val lines = spark.readStream 
.format("socket") 
.option("host", host) 
.option("port", port) 
.load() 
 
// Start running the query that prints the running counts to the console 
val query = wordCounts.writeStream 
.outputMode("append") 
.foreach(new ForearchWriter[Row]{ 
override def open(partitionId:Long,version:Long):Boolean={ 
            println("open") 
            return true 
        } 
override def process(value:Row):Unit={
            val spark = SparkSession.builder.getOrCreate() 
            val seq = value.mkString.split(" ") 
            val row = Row.fromSeq(seq) 
            val rowRDD:RDD[Row] = sparkContext.getOrCreate().parallelize[Row](Seq(row)) 
             
            val userSchema = new StructType().add("name","String").add("age","String") 
            val peopleDF = spark.createDataFrame(rowRDD,userSchema) 
            peopleDF.createOrReplaceTempView(myTable) 
            spark.sql("select * from myTable").show() 
        } 
         
override def close(errorOrNull:Throwable):Unit={ 
            println("close") 
        } 
     }) 
.start() 
 
query.awaitTermination() 
} 
} 
// scalastyle:on println



 

　　上述程序是直接继承ForeachWriter类的接口，并实现了open()、process()、close()三个方法。若采用显示定义一个类来实现，需要注意Scala的泛型设计，如下所示：



class myForeachWriter[T<:Row](stream:CatalogTable) extends ForearchWriter[T]{ 
open(partionId:Long,version:Long):Boolean={ 
        println("open") 
        true 
    } 
     
process(value:T):Unit={ 
        println(value) 
    } 
     
close(errorOrNull:Throwable):Unit={ 
        println("close") 
    } 
}



 

3. 自定义

　　若上述Spark Structured Streaming API提供的数据输出源仍不能满足要求，那么还有一种方法可以使用：修改源码。

如下通过实现一种自定义的Console来介绍这种使用方式：

3.1 ConsoleSink

　　Spark有一个Sink接口，用户可以实现该接口的addBatch方法，其中的data参数是接收的数据，如下所示直接将其输出到控制台：



class ConsoleSink(streamName:String) extends Sink{ 
    override def addBatch(batchId:Long, data;DataFrame):Unit = { 
        data.show()         
    } 
}



 

3.2 DataStreamWriter

　　在用户自定义的输出形式时，并调用start()方法后，Spark框架会去调用DataStreamWriter类的start()方法。所以用户可以直接在该方法中添加自定义的输出方式，如我们向其传递上述创建的ConsoleSink类示例，如下所示：



def start():StreamingQuery={ 
    if(source == "memory"){ 
        ... 
    }else if(source=="foreach"){ 
        ... 
else if(source=="consoleSink"){ 
        val streamName:String = extraOption.get("streamName") mathc{ 
            case Some(str):str 
            case None=>throw new AnalysisException("streamName option must be specified for Sink") 
        } 
         
        val sink = new consoleSink(streamName) 
        df.sparkSession.sessionState.streamingQueryManager.startQuery( 
            extraOption.get("queryName"), 
            extraOption.get("checkpointLocation"), 
            df, 
            sink, 
            outputMode, 
            useTempCheckpointLocaltion = true, 
            recoverFromCheckpointLocation = false, 
            trigger = trigger 
        ) 
    }else{ 
        ... 
    } 
}



3.3 Structured Streaming

　　在前两部修改和实现完成后，用户就可以按正常的Structured Streaming API方式使用了，唯一不同的是在输出形式传递的参数是"consoleSink"字符串，如下所示：



def execute(stream:CatalogTable):Unit={ 
    val spark = SparkSession 
.builder 
.appName("StructuredNetworkWordCount") 
.getOrCreate() 
/**1. 获取数据对象DataFrame*/
    val lines = spark.readStream 
.format("socket") 
.option("host", "localhost") 
.option("port", 9999) 
.load() 
     
/**2. 启动Streaming开始接受数据源的信息*/ 
    val query:StreamingQuery = lines.writeStream 
                .outputMode("append") 
.format("consoleSink") 
                .option("streamName","myStream") 
                .start() 
                 
    query.awaitTermination() 
}



4. 参考文献

[1]. Structured Streaming Programming Guide.



[2]. Kafka Integration Guide .