On Feb 23, 2013, at 1:13 AM, Eric Rasmussen <ericrasmus...@gmail.com> wrote:
> But that's just a long-winded way to express a reduce operation*, so for your
> example you could also write:
>
> import sqlalchemy as sa
>
> criteria = (('male', 35), ('female', 35))
> Useraccount = model.Useraccount
> query = session.query(Useraccount)
> ands = [sa.and_(Useraccount.gender == gender, Useraccount.age == age) for
> gender, age in criteria]
> or_clauses = reduce(sa.or_, ands)
> query = query.filter(or_clauses)
>
this is fine but I'd make one offhand note that nesting the conjunctions, i.e.
or_(or_(or_(or_(x, y), z), q), b) has the effect of the compiler traversing it
in a deep recursion loop, if you're using a lot of values (like, hundreds). We
had a user getting recursion overflows due to this. I illustrated a compiler
plug in for this user that did a non-recursive "unwrap" of the nested structure
first before passing it down to the default compiler, which isn't built in
because it adds a good chunk of performance overhead to all conjunctions.
Ultimately he went with doing or_(*everything) instead of his original approach
of "x |= (y), x |= (q)" etc.
--
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 http://groups.google.com/group/sqlalchemy?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.
