Pre-requisite: The schema of MongoDB documents is unknown. It has null values and missing values: For instance CCNO in the following documents (1) has value (2) has null value (3) is missing.
I am fetching them in spark using MongoSpark.load(SparkSession, ReadConfig), but it is replacing missing values with nulls in the dataframe.
Please suggest a way to differentiate between the manually inserted nulls and spark inserted nulls(for missing values).
