Since all databases should be, the source code for our user is versioned using a control source. The database is updated using a series of SQL scripts generated by the Red Gate comparison tool, which essentially matches up-migration in the many database migration infrastructures that seem to have appeared recently.
But what's the point of βdownβ migration within this framework? Often the code for up-migration is extremely complex (usually a complex data migration as functions evolve), and I struggle to understand that I need to write everything in the reverse order for down. This, of course, is something that I have never felt. Am I missing something here ...?
source share