OK I rearranged how PG does "checkfirst", to only process a particular named
ENUM type once within a particular create/drop sequence without needing to
check the DB, so the "checkfirst" flag can now be honored normally. This is in
rde4eb56fb6c8 and you can get this at
http://hg.sqlalchemy.org/
I can make a ticket for this since the Enum type is using the deprecated event
system and PG's event here doesn't have access to the "checkfirst" flag.
Ticket #2311. For now you'd have to not use the "native" enum type with PG,
or patch out that "checkfirst=True" inside of postgresql/base.p
If I do this:
from sqlalchemy import *
from StringIO import StringIO
buf = StringIO()
pg_engine = create_engine('sqlite://', strategy='mock',
executor=lambda s,p=';': buf.write(s.__str__() + p))
buf.truncate(0)
tables = [x[1] for x in sorted(db.metadata.tables.items(), key=lambda
x: x[0])]
for tab
I need print this sql code if these ENUM types and tables and indexes
exist.
On 28 окт, 18:28, lestat wrote:
> How I can get raw sql for this table?
>
> class Test(db.Model):
> __tablename__ = 'test'
>
> id = db.Column(db.Integer, primary_key=True)
> birthday = db.Column(db.Date)
>
