Re: alembic upgrade not idempotent?

[Mike Bayer]

On 04/18/2016 12:04 PM, bertr...@cryptosense.com wrote:
Sorry for the confusion, I meant "alembic upgrade head" indeed.  As you
confirmed the expected behavior, I did a bit of debugging to find out
where I might have done a mistake, and I found the problem.

I wrote a test revision that looked like the following:

|
defupgrade():
     bind =op.get_bind()
     db =Session(bind)
     db.execute('alter table cats rename to dogs')
|

With level = DEBUG for [logger_sqlalchemy], I get:

|
alembic upgrade head
INFO [sqlalchemy.engine.base.Engine]SELECT CAST('test plain returns'AS
VARCHAR(60))AS anon_1
INFO [sqlalchemy.engine.base.Engine]()
INFO [sqlalchemy.engine.base.Engine]SELECT CAST('test unicode returns'AS
VARCHAR(60))AS anon_1
INFO [sqlalchemy.engine.base.Engine]()
INFO [alembic.runtime.migration]Contextimpl SQLiteImpl.
INFO [alembic.runtime.migration]Willassume non-transactional DDL.
INFO [sqlalchemy.engine.base.Engine]PRAGMA table_info("alembic_version")
INFO [sqlalchemy.engine.base.Engine]()
INFO [sqlalchemy.engine.base.Engine]PRAGMA table_info("alembic_version")
INFO [sqlalchemy.engine.base.Engine]()
INFO [sqlalchemy.engine.base.Engine]
CREATE TABLE alembic_version (
         version_num VARCHAR(32)NOT NULL
)


INFO [sqlalchemy.engine.base.Engine]()
INFO [sqlalchemy.engine.base.Engine]COMMIT
INFO [alembic.runtime.migration]Runningupgrade ->80def0e180bc,Renamecats
to dogs
INFO [sqlalchemy.engine.base.Engine]BEGIN(implicit)
INFO [sqlalchemy.engine.base.Engine]alter table cats rename to dogs
INFO [sqlalchemy.engine.base.Engine]()
INFO [sqlalchemy.engine.base.Engine]INSERT INTO alembic_version
(version_num)VALUES ('80def0e180bc')
INFO [sqlalchemy.engine.base.Engine]()
INFO [sqlalchemy.engine.base.Engine]ROLLBACK
|

As you can see, my omission of db.commit() caused alembic's version
upgrade to be rolled back.  However, the table has really been renamed
from "cats" to "dogs".  Therefore, a second run of alembic upgrade head
will fail in this case.

in keeping with the tradition of "long-lived bugs that remain invisible 
for years and all the sudden get reported twice in two days", this has 
just been added yesterday in 
https://bitbucket.org/zzzeek/alembic/issues/369/orm-session-creates-a-subtransaction-on#comment-27063967 
.






--
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
<mailto:sqlalchemy-alembic+unsubscr...@googlegroups.com>.
For more options, visit https://groups.google.com/d/optout.

--
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.