Meagan Longoria doesn’t want to repeat good work:
I presented at SQL Saturday Pittshburgh this past weekend about populating your data warehouse with a metadata-driven, pattern-based approach. One of the benefits I mentioned is that it’s easy to employ this pattern for restartability.
For instance, let’s say I am loading data from 30 tables and 5 files into the staging area of my data mart or data warehouse, and one of table loads fails. I don’t want to reload the other tables I just loaded. I want to load the ones that have not been recently loaded. Or let’s say I have 5 dimensions and 4 facts, and I had a failure loading a fact table. I don’t want to reload my dimensions, and I only want to reload the failed facts. How do we accomplish this?
Read on to learn how.