使用joda.DateTime将案例类的DStream转换为Spark DataFrame

I want to save a DStream into HDFS using parquet format. The problem is that my case class use joda.DateTime while Spark SQL doesn't support this. For example:I want to save a DStream into HDFS using parque