|Perl: the Markov chain saw|
OK, from the DBAdmin I used to work with, here are the rules:
Really that last one is the real answer. Write your code to views, not tables.
If the tables are still in flux, what are you doing writing production level code? Worries about posistion are naive, views and clean selects fix that at the database level. Your code depends on the DB, if you can't trust the DB to behave, you are sunk, period.
The short answer to your problems is: select from views and with field names, don't try to fix your DB mistakes with code, forget what order your fields are in, finish the DB design before you write final code.
You might also try factoring out the DB access into a nice local module so you can move common actions into a central place where you can change them just once. That way, at least you can protect the rest of your code from the changes by interposing a code layer.
In reply to Re: Topics in Perl Programming: Table-Mutation Tolerant Database Fetches with DBI