how to guarantee the index is always following the file's original order no matter what. Currently, I'm using val df = spark.read.options(Map("header"-> "true", "inferSchema" -> "true")).csv("filePath").withColumn("index", monotonically_increasing_id()) .

Thanks!