The problem appears to be that metadata (and tables) explicitly specify the default schema. If I set the schema to None in my metadata it works fine.
On Thursday, 12 February 2015 14:19:36 UTC+13, tim.mi...@leapfrog3d.com wrote: > > Hi, > > My diff list is full of remove_fk and add_fk pairs. For each foreign key > alembic wants to: > - remove a foreign key one with table = <tablename> and > - add same foreign key with table = <schema>.<tablename> > > The same is true for unique constraints. > > Any ideas? > > Thanks > Tim > > On Tuesday, 10 February 2015 14:05:27 UTC+13, Michael Bayer wrote: >> >> >> >> tim.mi...@leapfrog3d.com wrote: >> >> > Hi all, >> > >> > I would like to have a post-deployment check that verifies the >> sqlalchemy metadata in the code matches what is in the database. >> > I was thinking that doing something like >> > >> > alembic revision --autogenerate --sql > changes.sql >> > >> > and then check changes.sql. However, alembic tells me that "Using >> --sql with --autogenerate does not make any sense", and indeed, if you hack >> the code in command.py to remove the check you are blessed with tracebacks >> that do not make any sense (to me anyway ;-) ). >> > >> > Can anyone suggest a way I might do this? >> >> You’re actually doing something that they are starting to do in >> Openstack, >> so luckily I can confirm this is feasible. >> >> You want to use the autogenerate API directly. Call upon compare_metadata >> documented at >> http://alembic.readthedocs.org/en/latest/api.html#autogeneration. >> Ideally, >> if “diff == []”, then there’s no changes. >> >> -- You received this message because you are subscribed to the Google Groups "sqlalchemy-alembic" group. To unsubscribe from this group and stop receiving emails from it, send an email to sqlalchemy-alembic+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.