There are two files supplied, representative of what they'd download from the website and import. Each has multiple entity records in them: lowest-level, then a parent, and then a parent of that.
We know that outside the spec, there are going to be other tables -- tables that are part of the whole solution -- but we're not supposed to focus on them.
We were told that we should throw out the data before each import. But looking to the future, aren't there going to be relationships to those other tables. What happens to any relationship established to those tables that we're not considering?
Will this make a mess for any further expansion? Should we instead do a careful update as part of the import?
My approach would be not to throw away any data, since we're already set up not to step on existing data.
(dictated to Mark Baum by J Fletcher)