Spark SQL Structured Data Write atomic files

So. Basically this is my problem.

I've set up a Spark SQL Structured data

val hiveQ = filteredDF.writeStream
  .trigger(Trigger.ProcessingTime(5, TimeUnit.SECONDS))
  .outputMode("append")
  .format("csv")
  .option("header", "false")
  .option(
    "path",
    conf.getString("dest.hdfs.prefix") + msgType + "/" + tableName)
  .option("checkpointLocation",
          conf.getString("dest.hdfs.temp.prefix") + tableName)
  .start()

What I'm trying to achieve somehow is that file written in hdfs should be "atomic", I mean, since there is another job on server that read from same directory and move files from this folder to another, I encounter errors because Spark is appending data to same files.

Is there a way to have something like .tmp files and just when it finalise them, let change extensions to .csv?

Honestly I have no idea how to move, could be partitionBy somehow help me? adding like some currentTimestamp column and use it as partition?