python - Please explain why these two builtin functions behave different when passed in keyword arguments -

consider these different behaviour::

>> def minus(a, b): >>    return - b  >> minus(**dict(b=2, a=1)) -1  >> int(**dict(base=2, x='100')) 4  >> import operator >> operator.sub.__doc__ 'sub(a, b) -- same - b.' >> operator.sub(**dict(b=2, a=1)) typeerror: sub() takes no keyword arguments 

why operator.sub behave differently int(x, [base]) ?

it implementation detail. python c api retrieve arguments separates between positional , keyword arguments. positional arguments not have name internally.

the code used retrieve arguments of operator.add functions (and similar ones sub) this:

pyarg_unpacktuple(a,#op,2,2,&a1,&a2) 

as can see, not contain argument name. whole code related operator.add is:

#define spam2(op,aop) static pyobject *op(pyobject *s, pyobject *a) { \   pyobject *a1, *a2; \   if(! pyarg_unpacktuple(a,#op,2,2,&a1,&a2)) return null; \   return aop(a1,a2); }  spam2(op_add           , pynumber_add) 

#define spam2(op,altop,doc) {#op, op_##op, meth_varargs, pydoc_str(doc)}, \                            {#altop, op_##op, meth_varargs, pydoc_str(doc)}, spam2(add,__add__, "add(a, b) -- same + b.") 

as can see, place a , b used in docstring. method definition not use meth_keywords flag necessary method accept keyword arguments.

generally spoken, can safely assume python-based function know argument name always accept keyword arguments (of course nasty stuff *args unpacking creating function doc arguments normal) while c functions may or may not accept keyword arguments. chances functions more few arguments or optional arguments accept keyword arguments later/optional ones. pretty have test it.

you can find discussion supporting keyword arguments everywhere on python-ideas mailinglist. there statement guido van rossum (the benevolent dictator life aka creator of python) on it:

hm. think many (most?) 1-arg , selected 2-arg functions (and 3+-arg functions) reduce readability, example of ord(char=x) showed.

i see syntactic feature state argument cannot given keyword argument (just added syntax state must keyword).

one area think adding keyword args outright wrong: methods of built-in types or abcs , overridable. e.g. consider pop() method on dict. since argument name undocumented, if subclasses dict , overrides method, or if create mutable mapping class tries emulate dict using duck typing, doesn't matter argument name -- callers (expecting dict, dict subclass, or dict-like duck) using positional arguments in call. if document argument names pop(), , users started use these, dict sublcasses , ducks broken (except if luck happened pick same name).