Flink實時寫入MongoDB

通過flink的RichSinkFunction，實現連接MongoDB，實時寫入數據（也可以自定義一個類繼承RichSinkFunction）

此處需注意，由於RichSinkFunction是序列化對象，此時可以使用 @transient (private) lazy來表示不需序列化，否則可能會報異常。（其中@trainsient可以避免overhead，lazy可以第一次被調用時正確地初始化以避免NPE）。

代碼如下：

    streamData.addSink(new RichSinkFunction[String] {
      lazy val mongoClient = new MongoClient(new ServerAddress("host", port))
      override def invoke(value: String): Unit = {
        if (mongoClient != null) {
          val data = DataUtils.MapLoader(value)
          val db = mongoClient.getDatabase("db")
          val collection = db.getCollection("collection")
          val list = new util.ArrayList[Document]()
          val doc = new Document()
          val date = new DateTime().getMillis
          doc.append("createtime", date)
          doc.append("updatetime", date)
          data.foreach(t => doc.append(t._1, t._2))
          list.add(doc)
          collection.insertMany(list)
        }
      }
    })

MongoDB鑑權：

    val serverAddress = new ServerAddress("host", port)
    val credential: util.ArrayList[MongoCredential] = new util.ArrayList[MongoCredential]
    //MongoCredential.createScramSha1Credential()三個參數分別爲 用戶名 數據庫名稱 密碼
    val mongoCredential1: MongoCredential = MongoCredential.createScramSha1Credential("", "", "")
    credential.add(mongoCredential1)
    val mongoClient = new MongoClient(ServerAddress addr, List<MongoCredential> credentialsList)

 pom文件：

<dependency>
   <groupId>org.mongodb</groupId>
   <artifactId>mongo-java-driver</artifactId>
   <version>3.10.1</version>
</dependency>