I am designing a spatial database and I am including a variety of already existent datasources (ex: Lakes, rivers, road network, contour lines, cities, etc). Those sources can be updated every once in a while.

Except from these sources, the database is normalized in fifth normal form (5NF) or Boyce-Codd normal form (BCNF).

Should I also normalize those external datasources?

My concern is that it may be too much trouble for the benefits. I am thinking about the moment when I will update the sources. It would be much more trouble than to just "replace" the existing table.

  • You probably won't use your data in normal form, your tools will expect relatively flat, denormalized inputs.
  • You probably won't update your data in normal form, you'll receive flat, denormalized bulk updated.
  • You won't receive updates in the sort of granular, just-change-one-thing form that normalized databases excel at.

So, no, don't bother.

| improve this answer | |

Not the answer you're looking for? Browse other questions tagged or ask your own question.