[sqlalchemy] Re: last_inserted_ids()

I am using 0.3.6 and reading the 0.3 doc. The problem is that actually I need it the way it works for me now, i.e. returning the auto_increment fields and not my primary key fields. Will upgrading sqlalchemy break this (undocumented) feature? On Jul 4, 12:58 am, Michael Bayer [EMAIL PROTECTED]

[sqlalchemy] Re: Creating where clauses programatically

I must be missing something Doesn't that evaluate the == expressions and result in an expression like and_(True, True, False, ...) ? On Jul 4, 1:04 am, Michael Bayer [EMAIL PROTECTED] wrote: On Jul 3, 4:06 pm, mc [EMAIL PROTECTED] wrote: I managed to avoid this issues because for select and

[sqlalchemy] Re: Creating where clauses programatically

SQLAlchemy Column objects override the '==' operator (and many others as well) so that python expressions get converted to appropriate SQL expressions. It's mentioned (briefly) here: http://www.sqlalchemy.org/docs/sqlconstruction.html#sql_whereclause Hope that helps, Simon -Original

[sqlalchemy] Re: SA trying to update read-only views in Postgres

Clear()ing the session before the mandatory commit resolved the issue (more a workaround than a solution :-). Andreas --On 6. Juli 2007 12:25:25 +0200 Andreas Jung [EMAIL PROTECTED] wrote: I have a self-referential mapper where a view is involved within a primaryjoin of a relation.

[sqlalchemy] Re: sqlite func,datetime

thanks, but that didn't work either. In the end i changed the func.db call to a datetime.datetime.now() python call which seems to work. I also noticed that i had to change all my date inserts to datetime objects as well (previously i was using postgres and inserted dates as date strings i.e.

[sqlalchemy] Re: last_inserted_ids()

On Jul 6, 2007, at 4:33 AM, mc wrote: I am using 0.3.6 and reading the 0.3 doc. The problem is that actually I need it the way it works for me now, i.e. returning the auto_increment fields and not my primary key fields. Will upgrading sqlalchemy break this (undocumented) feature? as

[sqlalchemy] Re: SA trying to update read-only views in Postgres

On Jul 6, 2007, at 6:25 AM, Andreas Jung wrote: I have a self-referential mapper where a view is involved within a primaryjoin of a relation. Traversing a tree represented through the mapper generates a bunch of UPDATE operations on the view (no idea why so far) although only SELECT

[sqlalchemy] How many objects are in a session

I noticed that Session has no method len. Wouldn't it be useful to add one, like this: --- orm/session.py 2007-07-06 14:38:22.0 +0200 +++ orm/session.py.orig 2007-07-06 14:43:48.0 +0200 @@ -642,6 +642,9 @@ def __iter__(self): return iter(list(self.uow.new) +

[sqlalchemy] Re: SA trying to update read-only views in Postgres

On Jul 6, 2007, at 6:25 AM, Andreas Jung wrote: I have a self-referential mapper where a view is involved within a primaryjoin of a relation. Traversing a tree represented through the mapper generates a bunch of UPDATE operations on the view (no idea why so far) although only SELECT

[sqlalchemy] Re: SA trying to update read-only views in Postgres

--On 6. Juli 2007 11:15:38 -0400 Michael Bayer [EMAIL PROTECTED] wrote: On Jul 6, 2007, at 6:25 AM, Andreas Jung wrote: I have a self-referential mapper where a view is involved within a primaryjoin of a relation. Traversing a tree represented through the mapper generates a bunch of

[sqlalchemy] Re: last_inserted_ids()

On Jul 6, 4:33 am, mc [EMAIL PROTECTED] wrote: I am using 0.3.6 and reading the 0.3 doc. The problem is that actually I need it the way it works for me now, i.e. returning the auto_increment fields and not my primary key fields. Will upgrading sqlalchemy break this (undocumented)

[sqlalchemy] Re: sqlite func,datetime

On Jul 6, 7:06 am, Nick [EMAIL PROTECTED] wrote: thanks, but that didn't work either. In the end i changed the func.db call to a datetime.datetime.now() python call which seems to work. I also noticed that i had to change all my date inserts to datetime objects as well (previously i was

[sqlalchemy] new Engine document

I just wrote this but haven't had time to carefully proof it...if people have corrections / comments let me know. the intro and all about executing are rewritten: http://www.sqlalchemy.org/docs_new/dbengine.html as well as some changes to metadata, no more BoundMetaData/ DynamicMetaData:

[sqlalchemy] Re: whats going to break in 0.4

-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On Jul 5, 2007, at 6:00 PM, Michael Bayer wrote: On Jul 5, 5:46 pm, Barry Warsaw [EMAIL PROTECTED] wrote: What I mean by that is that Mailman has at least three separate related collections of data: mailing lists, users, and messages. It should

[sqlalchemy] No update in cascade

I have a strange situation. If someone _really_ needs to know why I could explain. When an object gets deleted by SQLAlchemy, it's related objects are getting updated setting their foreign keys referring to the deleted items to NULL. But what if I don't want that to happen? Viewonly is not an

[sqlalchemy] Re: No update in cascade

On Jul 6, 2007, at 5:50 PM, Koen Bok wrote: I have a strange situation. If someone _really_ needs to know why I could explain. When an object gets deleted by SQLAlchemy, it's related objects are getting updated setting their foreign keys referring to the deleted items to NULL. But what