Ravindra BagaleCourses & study guides Track your progress

5. Python for Pipelines (pandas to files/DB)

5.3 Idempotent Runs Idea

Why. Re-running a job twice should not double the warehouse rows.

What. Idempotent ≈ same input re-run gives the same safe result (overwrite partition, merge on keys, truncate+load carefully).

How. Prefer “replace yesterday’s partition” over blind append.