来自RDD火花流的Scala提取物CassandraRow

Question

我有以下代码： -

val kafkaStream = KafkaUtils.createDirectStream[String, String, StringDecoder, StringDecoder](ssc, kafkaParams, topics)
val records = kafkaStream.map(_._2)

records.foreachRDD { rdd =>
  {
    rdd.collect().foreach(a =>
      {
        implicit val formats = DefaultFormats
        import sqlContext.implicits._
        val jValue = parse(a)
        val record = jValue.extract[historyevent]
        println("JSON String" + a.toString())

        val history_gpsdt = record.gpsdt //"2017-04-12 00:25:10"
        val history_latitude = record.latitude //6.854678
        val history_longitude = record.longitude //78.583751

        val rss = sc.cassandraTable("db", "table").select("imei", "date", "gpsdt").where("imei=? and date=?", record.imei, record.date)
// some more set of statements
      }
   }
 }

所以，这里我从Kafka Streaming获取了一组JSON字符串，我从中逐一提取每一行，然后对数据执行一些逻辑。因此，我只是想了解我是使用最佳方式还是我们可以更优化它。请建议。谢谢，