Testing for an empty dictionary in Python
steve at REMOVE-THIS-cybersource.com.au
Mon Mar 24 03:39:06 CET 2008
On Sun, 23 Mar 2008 18:56:51 -0700, John Machin wrote:
>> Python knows the truth value of built-in types like dicts without
>> actually converting them to bools, or for that matter calling __len__
>> or __nonzero__ on them.
> What the 2.5.1 interpreter does is call PyObject_IsTrue, which checks to
> see if the built_in or extension type is a mapping (not just a dict)
> with a length method and if so calls it; similarly for sequences:
> else if (v->ob_type->tp_as_mapping != NULL &&
> v->ob_type->tp_as_mapping->mp_length != NULL)
> res = (*v->ob_type->tp_as_mapping->mp_length)(v);
> else if (v->ob_type->tp_as_sequence != NULL &&
> v->ob_type->tp_as_sequence->sq_length != NULL)
> res = (*v->ob_type->tp_as_sequence->sq_length)(v);
> Was that what you meant by "without ... calling __len__"?
What I meant was that the interpreter *didn't* do was lookup the name
'__len__' via the normal method resolution procedure. There's no need to
search the dict's __dict__ attribute looking for an attribute named
__len__, or indeed any sort of search at all. It's a direct pointer
lookup to get to mp_length.
I didn't mean to imply that Python magically knew whether a dict was
empty without actually checking to see if it were empty. Apologies for
being less than clear.
Also, since I frequently criticize others for confusing implementation
details with Python the language, I should criticize myself as well. What
I described is specific to the CPython implementation -- there's no
requirement for other Python versions to do the same thing.
More information about the Python-list