Christoph Zwerschke wrote:
> Janzert schrieb:
>> 2. Why not something like the following (untested code)?
>> @property
>> def permissions(self):
>> p_q = session.query(Permission).join(['tg_group', 'tg_user'])
>> p_q.filter_by(user_id = self.user_id)
>> return p_q.all()
>
> I would write something like that in a long statement, without the
> arbitrary p_q variable and avoid using hard coded table names.
>
> The other problem is that this code doesn't work at all.
>
Like I said it was untested. :/
> Also, the method must return a set, not a list with duplicates.
>
> There are basically two ways to collect the permissions via SQL.
>
> Either you make the two joins, creating all possible ways a user can
> have a certain permission, and then remove the duplicates, like that:
>
> def permissions(self):
> return set(Permission.query.join(Permission.groups,
> Group.users).filter_by(user_id=self.user_id).all())
>
> Or you create the query using exist clauses, so that the query does not
> produce any duplicates in the first place:
>
> def permissions(self):
> return set(Permission.query.filter(Permission.groups.any(
> Group.users.any(User.user_id==self.user_id))).all())
>
> The problem with this latter solution only works with SA >= 0.4 (and the
> former solution even only with SA >= 0.5).
>
> (Also, both solutions do a last join with the user table that is
> actually unnecessary. I currently see no elegant way to avoid this. It
> does not harm much, though.)
>
> -- Christoph
>
My understanding and testing seem to show that an inner join doesn't
return duplicates. Wrapping the result in a set doesn't seem necessary.
After playing around with SA 0.3.10, 0.4.6 and 0.5.0beta1 and taking
into account your comments I came up with this.
def permissions(self):
return Permission.query.join(["groups",
"users"]).filter_by(user_id=self.user_id).all()
permissions = property(permissions)
I tested by setting up a user belonging to two groups with each of them
containing two permissions. One of the permissions was unique to the
group and one was the same in both groups.
I put a diff from a TG 1.0.5 quickstarted project to what I used for
testing at http://paste.turbogears.org/paste/3056 and the full project
at http://janzert.com/files/jointest.zip
If this seems the right way to go I can create a trac issue and try to
work up a real patch.
Janzert
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"TurboGears" 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/turbogears?hl=en
-~----------~----~----~----~------~----~------~--~---
