When the data source has a large amount of duplicate data for certain fields and these duplicate records are not needed for subsequent operations, we want to filter the data at the source of the real-time task