Your question is Backfilling Missing Pipeline Data. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
You're working on a data pipeline and discover that some historical data is missing because of an upstream outage, bad deployment, or delayed source delivery. You need a safe way to restore the missing records without corrupting downstream tables or double counting data.
What strategies would you use to backfill missing data in a data pipeline?