I am currently in the process of refactoring my database schema.
As part of that I am changing the primary key on one of my tables.
The first thing I did was place a unique contraint on the column I want
to be the new pk, then for each table with a foriegn key I added a new
colum, populated that column, then replaced the foriegn key.
Then I dropped the old primary key from the original table, then created
the new primary key.
The question is, I would like to drop the now redundant unique contraint, but all of the foriegn keys are now dependent
onthe unique constraint, not the primary key.
I guess the big question is, is this even important? Is there a
performance penalty here?
I'm doing the work on my development machine, not my production server
so I have the option of trying things another way.