XX and YY are both of the same direction <symbol 'ONETOMANY> error
Andreas Jung
Hash: SHA1
I am getting the following error after upgrading from post-0.5.2 version
(SVN) to 0.5.3:
======================================================================
ERROR: testToolLookupVisitedBy (tool.tests.test_tool_lookup.TestToolLookup)
- ----------------------------------------------------------------------
Traceback (most recent call last):
File "/HRS2/local/lib/python2.4/unittest.py", line 260, in run
testMethod()
File
"/local/HRS2/Devel/junga/tb-dev/parts/modules-svn/toolbox/tool/tests/test_tool_lookup.py",
line 47, in testToolLookupVisitedBy
rows = TL.toolsVisitedBy('ajung')
File
"/local/HRS2/Devel/junga/tb-dev/parts/modules-svn/toolbox/tool/__init__.py",
line 70, in toolsVisitedBy
visited_tools = session.query(VT).filter_by(benutzer=username.lower())
File
"/local/HRS2/Devel/junga/tb-dev/eggs/SQLAlchemy-0.5.3-py2.4.egg/sqlalchemy/orm/session.py",
line 895, in query
return self._query_cls(entities, self, **kwargs)
File
"/local/HRS2/Devel/junga/tb-dev/eggs/SQLAlchemy-0.5.3-py2.4.egg/sqlalchemy/orm/query.py",
line 91, in __init__
self._set_entities(entities)
File
"/local/HRS2/Devel/junga/tb-dev/eggs/SQLAlchemy-0.5.3-py2.4.egg/sqlalchemy/orm/query.py",
line 100, in _set_entities
self.__setup_aliasizers(self._entities)
File
"/local/HRS2/Devel/junga/tb-dev/eggs/SQLAlchemy-0.5.3-py2.4.egg/sqlalchemy/orm/query.py",
line 114, in __setup_aliasizers
mapper, selectable, is_aliased_class = _entity_info(entity)
File
"/local/HRS2/Devel/junga/tb-dev/eggs/SQLAlchemy-0.5.3-py2.4.egg/sqlalchemy/orm/util.py",
line 492, in _entity_info
mapper = class_mapper(entity, compile)
File
"/local/HRS2/Devel/junga/tb-dev/eggs/SQLAlchemy-0.5.3-py2.4.egg/sqlalchemy/orm/util.py",
line 567, in class_mapper
mapper = mapper.compile()
File
"/local/HRS2/Devel/junga/tb-dev/eggs/SQLAlchemy-0.5.3-py2.4.egg/sqlalchemy/orm/mapper.py",
line 653, in compile
raise sa_exc.InvalidRequestError("One or more mappers failed to
compile. Exception was probably "
InvalidRequestError: One or more mappers failed to compile. Exception
was probably suppressed within a hasattr() call. Message was: One or
more mappers failed to compile. Exception was probably suppressed
within a hasattr() call. Message was: One or more mappers failed to
compile. Exception was probably suppressed within a hasattr() call.
Message was: One or more mappers failed to compile. Exception was
probably suppressed within a hasattr() call. Message was:
Hierarchies.subscribed_by and back-reference Hierarchies.subscriber are
both of the same direction <symbol 'ONETOMANY>. Did you mean to set
remote_side on the many-to-one side ?
The related code is:
class Hierarchies(Base, AsDictMixin):
__tablename__ = 'hierarchies'
__table_args__ = ( { 'autoload' : True, })
__mapper_args__ = ({'extension' : HierachiesDeletionLogger()})
id = Column(Integer, Sequence('hierarchies_seq'), primary_key=True)
parent_id = Column(Integer, ForeignKey('hierarchies.id'))
hierarchyshare_id = Column(Integer, ForeignKey('hierarchies.id'))
pos = Column(Integer)
Hierarchies.subscribed_by = relation('Hierarchies',
primaryjoin=Hierarchies.hierarchyshare_id==Hierarchies.id,
backref=backref("subscriber",
remote_side=Hierarchies.hierarchyshare_id),
remote_side=Hierarchies.hierarchyshare_id,
uselist=True,
)
Anything I am missing or something that changed over the last two weeks
at this point?
Andreas
- --
ZOPYX Ltd. & Co. KG - Charlottenstr. 37/1 - 72070 Tübingen - Germany
Web: www.zopyx.com - Email: in...@zopyx.com - Phone +49 - 7071 - 793376
Registergericht: Amtsgericht Stuttgart, Handelsregister A 381535
Geschäftsführer/Gesellschafter: ZOPYX Limited, Birmingham, UK
- ------------------------------------------------------------------------
E-Publishing, Python, Zope & Plone development, Consulting
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (Darwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
iEYEARECAAYFAknUzXIACgkQCJIWIbr9KYwdZgCfVfo9ZN2bNPM4iaxZoFXdcuuE
yPoAoMaqN2Wr219oL+kviY7dtotIqh/M
=RG8E
-----END PGP SIGNATURE-----
Adrian
works fine with 0.5.2 on another with 0.5.3 I get the error you
mentioned.
> Geschäftsführer/Gesellschafter: ZOPYX Limited, Birmingham, UK
> - ------------------------------------------------------------------------
> E-Publishing, Python, Zope & Plone development, Consulting
>
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.9 (Darwin)
> iEYEARECAAYFAknUzXIACgkQCJIWIbr9KYwdZgCfVfo9ZN2bNPM4iaxZoFXdcuuE
> yPoAoMaqN2Wr219oL+kviY7dtotIqh/M
> =RG8E
> -----END PGP SIGNATURE-----
>
> < 1KViewDownload
Michael Bayer
On Apr 2, 10:36 am, Andreas Jung <li...@zopyx.com> wrote:
> The related code is:
>
> class Hierarchies(Base, AsDictMixin):
> __tablename__ = 'hierarchies'
> __table_args__ = ( { 'autoload' : True, })
> __mapper_args__ = ({'extension' : HierachiesDeletionLogger()})
>
> id = Column(Integer, Sequence('hierarchies_seq'), primary_key=True)
> parent_id = Column(Integer, ForeignKey('hierarchies.id'))
> hierarchyshare_id = Column(Integer, ForeignKey('hierarchies.id'))
> pos = Column(Integer)
>
> Hierarchies.subscribed_by = relation('Hierarchies',
>
> primaryjoin=Hierarchies.hierarchyshare_id==Hierarchies.id,
> backref=backref("subscriber",
> remote_side=Hierarchies.hierarchyshare_id),
> remote_side=Hierarchies.hierarchyshare_id,
> uselist=True,
> )
backwards reference, indicating they are both one-to-many from
hierarchies.id to hiearchies.parent_id. The remote_side in the many-
to-one backref should point to Hierarchies.id.
>
> Anything I am missing or something that changed over the last two weeks
> at this point?
>
> Andreas
>
> - --
> ZOPYX Ltd. & Co. KG - Charlottenstr. 37/1 - 72070 Tübingen - Germany
> Geschäftsführer/Gesellschafter: ZOPYX Limited, Birmingham, UK
> - ------------------------------------------------------------------------
> E-Publishing, Python, Zope & Plone development, Consulting
>
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.9 (Darwin)
> iEYEARECAAYFAknUzXIACgkQCJIWIbr9KYwdZgCfVfo9ZN2bNPM4iaxZoFXdcuuE
> yPoAoMaqN2Wr219oL+kviY7dtotIqh/M
> =RG8E
> -----END PGP SIGNATURE-----
>
> < 1KViewDownload
Adrian
perfectly fine in 0.5.2:
mapper(Contact, CONTACTS)
mapper(Atom, ATOMS,
properties = {
'Contacts': relation(Contact,
primaryjoin = Contact.atom_id==ATOMS.c.id,
foreign_keys = [Contact.atom_id], lazy=True,
uselist=True, viewonly=True,
backref = backref('Atom', primaryjoin =
Contact.atom_id==ATOMS.c.id,
foreign_keys =
[ATOMS.c.id],
lazy=True, uselist=False,
viewonly=True)
)
})
In: atom = session.query(Atom).filter(Atom.id==57).one()
In: atom.Contacts
Out: [<Contact(57, 9)>]
In 0.5.3 however, the backref() will throw the following error:
ArgumentError: Atom.Contacts and back-reference Contact.Atom are both
remote_side on the many-to-one side ?
Michael Bayer
in this case you have set the foreign_keys collection backwards on the
backref. the foreign key is always contact.atom_id. its also not needed
if you have ForeignKey objects on your table metadata.