Here it is:
commit 85368d25ed158c85bd19f4a63400884ab1cda26a
Author: Mike Bayer <m...>
Date: Sat Jun 8 18:54:14 2013 -0400
get nested joins to render on oracle 8
Sounds like the right commit notes. You still maintaining 0.8? Should
that change be patchable in 0.7?
On Wednesday, January 21, 2015 at 1:13:21 PM UTC-5, Michael Bayer wrote:
>
>
>
> Kent <[email protected] <javascript:>> wrote:
>
> > Mike,
> >
> > When using use_ansi=False for Oracle (8) in conjunction with
> joinedload-ing an inline view property, SQLAlchemy-0.8.7 renders an inner
> join instead of an outer join. This has been fixed in SQLAlchemy-0.9.0,
> but, as I'm not prepared for the migration yet, I was hoping and failing to
> find the bug ticket and hopefully a patch. Do you know when/where this was
> fixed and whether the fix would be patch'able in 0.7 or at least 0.8?
> >
> > The attached script runs on 0.9.0+ but the assertion fails on 0.8.7.
> >
> > The only difference in SQL output is the outer join "(+)":
> >
> > SELECT products.productid AS products_productid, anon_1.productid AS
> anon_1_productid, anon_1.siteid AS anon_1_siteid, anon_1.qty AS anon_1_qty
> > FROM products, (SELECT inventory.productid AS productid,
> inventory.siteid AS siteid, sum(inventory.qty) AS qty
> > FROM inventory GROUP BY inventory.productid, inventory.siteid) anon_1
> > WHERE anon_1.productid(+) = products.productid ORDER BY anon_1.siteid
> >
> > Interestingly, use-ansi=True correctly renders "LEFT OUTER JOIN" in
> 0.8.7 but it fails to render as an outer join with use-ansi=False.
> >
> > Thanks for you time and exceptional software,
> > Kent
> >
>
>
> no specific fixes are logged, however
> http://docs.sqlalchemy.org/en/rel_0_9/changelog/migration_09.html#many-join-and-left-outer-join-expressions-will-no-longer-be-wrapped-in-select-from-as-anon-1
>
> refers to a very large change in how the ORM decides to join things.
> That would cause some kinds of joinedloads to render differently , which
> would impact how (+) comes out as well, but i wouldn’t think it would have
> the effect that the missing (+) is the only change, it would be more than
> that.
>
> So I have no better idea than you, so the method I’d do is just to git
> bisect (http://git-scm.com/docs/git-bisect) through revisions until you
> find the fix. If it’s a big merge revision, I can look into it to find
> something specific, but if you can get me a rev ID that would be a good
> start.
>
>
>
--
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 [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/sqlalchemy.
For more options, visit https://groups.google.com/d/optout.
