Just tried running the code you posted, but in my case it keeps using the default query_class (Query) instead of the "MyPrefulteredQuery". Not sure if there was a breaking change on SA since you posted this.
On Tuesday, July 22, 2014 at 11:35:49 AM UTC-7, Meg Mitchell wrote: > > I had the same problem and decided to look into implementing a custom > relationship like you suggested. I'll post what I have here in case anyone > else has the same problem. :) > > The first thing I noticed was the query_class attribute on the > RelationshipProperty class. If you don't want to implement a new > relationship, you can just do this, which is a little cleaner than > respecifying the primaryjoin: > > children = relationship(Child, query_class=MyPrefilteredQuery) > > > (Where MyPrefilteredQuery is the PreFilteredQuery class you implemented as > per > https://bitbucket.org/zzzeek/sqlalchemy/wiki/UsageRecipes/PreFilteredQuery > .) > > Making your own relationship is pretty straightforward from there, though: > > > > from sqlalchemy.util.langhelpers import public_factory > from sqlalchemy.orm.relationships import RelationshipProperty > > class MyPrefilteredRelationship(RelationshipProperty): > def __init__(self, *args, **kwargs): > kwargs['query_class'] = MyPrefilteredQuery > super(MyPrefilteredRelationship, self).__init__(*args, **kwargs) > > filtered_relationship = public_factory(MyPrefilteredRelationship, > ".mylib.db") > > > Then you can use it like so: > > from mylib.db import filtered_relationship > > class Parent(Base): > __table__ = parent_table > children = filtered_relationship(Child) > > > On Sunday, February 20, 2011 8:19:58 PM UTC-8, robert wrote: >> >> On Feb 20, 8:59 pm, Michael Bayer <mike...@zzzcomputing.com> wrote: >> > the cleanest and most foolproof way is to join on relationships that >> have the condition: >> > >> > related_non_deleted = relationship(Related, >> primaryjoin=and_(Related.deleted==False, Related.foo_id==id), >> viewonly=True) >> >> I agree that the explicit approach is the most foolproof. The main >> reason I was trying to avoid it is that, in my case, the condition is >> ubiquitous. Thus all of my relationships that look like this >> >> children = relationship(Child) >> >> would have to change to >> >> children = relationship(Child, primaryjoin=and_(Child.deleted==False, >> Child.parent_id==id) >> >> Perhaps I could write a custom relationship() to simplify things. >> I'll give that a try. I'll also have to >> add .filter(foo.deleted==False) to all top level queries. Maybe a >> custom Session that always answers a query with that filter already >> added would be cleaner. There may ultimately be many queries and many >> relationships in this app, and I'm trying to avoid having to remember >> (or worse, have someone else remember) to always include the deleted >> condition. >> >> Thanks for your help. I've got a bit of digging to do now. > > -- 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 https://groups.google.com/group/sqlalchemy. For more options, visit https://groups.google.com/d/optout.
