Just dump the data into a file and reload it after re-creating the tables. At most, you will have to edit a few SQL fields/values to comply with the new column/table names.
Doing it automated is *really* hard and I don't think it will be better (or more accurate) than just re-inserting the data via some db command. On 1/18/06, Luke Skibinski Holt <[EMAIL PROTECTED]> wrote: > > How probable would it be to implement a means of updating a database > schema when only minor changes occur? > > 'minor' has to be defined, but the removal of columns or entire models > could also be considered depending on their impact to the rest of the > app/project. With no regard for the actual data, models with no > relations could easily be dropped or have fields added/removed without > a sqlreset and then a clumsy find/replace on the data dump to reinsert > it back. > > I'm just looking for a quick way to update my models in the development > stage while keeping test data. Django is great for prototyping, and > despite my best laid plans (and usecases), things change and grow as I > discover new ways of doing them (as they rightfully should in a > prototype). > > Are there any such plans in the works? > > > Luke Skibinski Holt > >