you're the best sir! worked for my test case and for my live code. I
like the slightly-more-intuitive "selectobj.c.key" rather than
"selectobj.c.columnname" usage too.
On 3/5/07, Michael Bayer <[EMAIL PROTECTED]> wrote:
>
> no problem, that test uncovered a place I could ratchet things down
> some more so everybody wins. you want rev 2383.
>
> On Mar 5, 2007, at 3:09 PM, Pete Taylor wrote:
>
> > yessir, when using selectables i always just use
> > selectobj.c.colname...
> >
> > i've attached a quick test case, and it throws the same error that my
> > live code does, so it at least proves i can break things the same way
> > twice :D i wouldn't be surprised if it's something easy that i'm just
> > doing wrong, but here it is all the same.
> >
> > the test case relation for lineItems is the most bare-bones version of
> > the one's i've tried, but this is the first one I tried... it merely
> > substitutes foreign_keys for where, pre-2309, had foreignkey.
> >
> > Thanks for looking!
> > Pete
> >
> > On 3/5/07, Michael Bayer <[EMAIL PROTECTED]> wrote:
> >>
> >> please forward an runnable example test case, preferably a single
> >> file with <200 lines if possible.
> >>
> >> you're also making usage of Column instances directly off the
> >> selectable youre mapping, right ?
> >>
> >>
> >> On Mar 5, 2007, at 1:55 PM, Pete Taylor wrote:
> >>
> >>>
> >>> in regard to post-2309 revisions and the foreignkey keyword
> >>> deprecation...
> >>>
> >>> I know this conversation seems to have terminated as of 2/21, but
> >>> it's
> >>> the only thread that was tracing through a problem i'm having. I'm
> >>> certain my problem is just one of my own lack of understanding on
> >>> how
> >>> to use foreign_keys now, but I'll lay it out there...
> >>>
> >>> I have a legacy table that, in effect, maps to two classes. A
> >>> container-style class and a line items class, which for convenience
> >>> we'll just call Container and LineItem. A Container has_many
> >>> LineItem, as you might imagine. The issue, however, is that the
> >>> Container class is mapped against a select statement from that
> >>> table,
> >>> not the table directly, while the LineItem class is mapped
> >>> against the
> >>> table, largely because there's an incrementing LineItem id that's a
> >>> primary key on the table, but has no meaning to the Container.
> >>>
> >>> before 2309, I was able to map the lineItem relationship back to the
> >>> container class using an explicit primary join and a foreignkey
> >>> statement.
> >>>
> >>> I've read through the self-referential example rather a lot, and the
> >>> two mailing list threads, and have tried all the combinations and
> >>> permutations of primaryjoin, foreign_keys, and remote_side that I
> >>> can
> >>> think of (obviously not all of them or it would have worked ;) ),
> >>> from
> >>> the mappers and relationships on both classes, and I can't seem to
> >>> figure out the right way to make it work. I can send along a
> >>> simplified mockup of the relationships and what I've got
> >>> currently if
> >>> anyone has time to take a look, but I may in fact just be heading
> >>> the
> >>> wrong direction...
> >>>
> >>> Any help would be greatly appreciated!
> >>> Pete Taylor
> >>>
> >>> --
> >>> "All guilt is relative, loyalty counts, and never let your
> >>> conscience
> >>> be your guide."
> >>> - Lucas Buck, American Gothic
> >>>
> >>>>
> >>
> >>
> >>>
> >>
> >
> >
> > --
> > "All guilt is relative, loyalty counts, and never let your conscience
> > be your guide."
> > - Lucas Buck, American Gothic
> >
> > >
> > #!/usr/bin/env python
> >
> > import datetime
> > from sqlalchemy import *
> >
> > db = create_engine('sqlite:///:memory:')
> > metadata = BoundMetaData(db)
> >
> > class Container(object): pass
> > class LineItem(object): pass
> >
> > items = Table('items', metadata,
> > Column('item_policy_num', String(10), primary_key=True,
> > key='policyNum'),
> > Column('item_policy_eff_date', DateTime, primary_key=True,
> > key='policyEffDate'),
> > Column('item_type', String(20), primary_key=True, key='type'),
> > Column('item_id', Integer, primary_key=True, key='id'),
> > )
> >
> >
> > container_select = select(
> > [items.c.policyNum, items.c.policyEffDate, items.c.type],
> > distinct=True,
> > ).alias('container_select')
> >
> > LineItem.mapper = mapper(LineItem, items)
> >
> > Container.mapper = mapper(Container, container_select, order_by=asc
> > (container_select.c.item_type), properties=dict(
> > lineItems = relation(LineItem, lazy=False, cascade='all, delete-
> > orphan', order_by=asc(items.c.type),
> > primaryjoin=and_(
> > container_select.c.item_policy_num==items.c.policyNum,
> >
> > container_select.c.item_policy_eff_date==items.c.policyEffDate,
> > container_select.c.item_type==items.c.type
> > ),
> > foreign_keys=[
> > items.c.policyNum,
> > items.c.policyEffDate,
> > items.c.type,
> > ],
> > )
> > ))
> >
> > def main():
> > metadata.create_all()
> > session = create_session()
> > con = Container()
> > con.policyNum = "9999999999"
> > con.policyEffDate = datetime.date.today()
> > con.type = "TESTER"
> > session.save(con)
> > for i in range(0, 10):
> > li = LineItem()
> > li.id = i
> > con.lineItems.append(li)
> > session.save(li)
> > session.flush()
> > session.clear()
> > con = session.query(Container).selectfirst()
> > print "DEBUG: con has %s line items" % len(con.lineItems)
> >
> > if __name__ == "__main__":
> > main()
>
>
> >
>
--
"All guilt is relative, loyalty counts, and never let your conscience
be your guide."
- Lucas Buck, American Gothic
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"sqlalchemy" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at
http://groups.google.com/group/sqlalchemy?hl=en
-~----------~----~----~----~------~----~------~--~---
