I currently have an R-based algorithm that sorts a data.table by date and then finds the most recent non-NA / non-null value. I've found some success with the following StackOverflow question to implement a backfilling algorithm for some relatively large datasets:I currently have an R-based algorithm that sort