On Jul 13, 2013, at 22:59, Michael Bayer <mike...@zzzcomputing.com> wrote: > The original intent of "default" for Column is that it acts exactly like > "server_default", except it runs on the Python side for the case that your > schema wasn't created with this default, or the functionality of your default > is not supported by the database. But otherwise, just like a > "server_default", the value is None on the Python side until the INSERT > occurs.
For what it's worth I forget that all the time and end up flushing objects out just to get the default values right. > Really, a better solution here would be an ORM-level "default", that's what > you've implemented anyway. In order to keep things pythonic I wonder if this should just be left as default arguments for the constructor. If everyone would do this: class Foo(Base): start_end = sa.Column(sa.Timestamp()) def __init__(self, start_end=datetime(2001, 2, 3), **kw): Base.__init__(self, start_end. **kw) This is standard python and easy to read. For association tables you already need to do this, either this way or by supplying a factory function, so the step to doing it for other objects is not that big. The downside is that the default value is now at a different place than the column definition, but it does make it very clear that this operates at the Python object level instead of the core SQL level. Wichert. -- You received this message because you are subscribed to the Google Groups "sqlalchemy" group. To unsubscribe from this group and stop receiving emails from it, send an email to sqlalchemy+unsubscr...@googlegroups.com. To post to this group, send email to sqlalchemy@googlegroups.com. Visit this group at http://groups.google.com/group/sqlalchemy. For more options, visit https://groups.google.com/groups/opt_out.