On Wed, Sep 21, 2011 at 12:30 PM, Ben Pfaff <b...@nicira.com> wrote:
> On Wed, Sep 21, 2011 at 11:41:58AM -0700, Reid Price wrote:
> > On Wed, Sep 21, 2011 at 10:58 AM, Ben Pfaff <b...@nicira.com> wrote:
> >
> > > I wish there was a simple way to just test for "iterable" and
> > > "mapping" types (is there?). Then we could get rid of a lot of ugly
> > > tests for specific types, in favor of more general tests.
> > >
> >
> > The standard way to determine iterables is with
> >
> > hasattr(obj, '__iter__')
> >
> > It's not the prettiest (and note behavior with strings/unicode), but it
> > generally behaves as expected.
> > There isn't a great way to determine whether something is a 'map' in
> > general, as {}[key] and [][index]
> > both use the __getitem__ method. In practice, it is fairly rare to see
> > non-dict maps.
>
> The Python reference manual says:
> It is also recommended that mappings provide the methods
> keys(), values(), items(), has_key(), get(), clear(),
> setdefault(), iterkeys(), itervalues(), iteritems(), pop(),
> popitem(), copy(), and update() behaving similar to those for
> Python's standard dictionary objects.
>
> That seems to distinguish in practice:
> >>> hasattr({}, "iterkeys")
> True
> >>> hasattr([], "iterkeys")
> False
> Any comment on whether how (un)wise it might be to use this to
> distinguish maps from simple sequences?
>
I think it seems reasonably safe to rely upon one (or several) attributes
to distinguish maps from sequences. I would recommend using keys
and/or values. The others might suffice, but are mostly too specific
(iter*),
too general (pop, copy, clear, items), or possibly deprecated (has_key).
-Reid
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
