[Tutor] simple question about numeric types

[Kent Johnson]

Emanuele Rocca wrote:
> Hello list,
> I've got a question partially related to this thread.
> 
> * Gregor Lingl <glingl at aon.at>, [2006-05-03  0:24 +0200]:
>>  v=<some numeric value>
>>  if isinstance(v, int) or isinstance(v, float):
>>       <block>
> 
> I wonder which is the recommended way to check the type of a value.

Often the recommendation is just to assume that a value is the correct 
type and be ready to catch an exception if it is not. This allows 'duck 
typing' to work - your function (or whatever) will accept any value that 
implements the correct methods, whether it has a particular inheritance 
relationship with the expected type or not. For example you could make a 
function that accepted any mapping object whether it inherits from dict 
or not.
> 
> In other words, what should I choose between:
>     isinstance(v, int)

This is good because it will accept subclasses of int as well as plain 
ints. (Not that subclasses of int are that common, but other classes 
might be subclassed.)

>     type(v) is int
>     v.__class__ is int

I think these are equivalent, though I would prefer the first. They are 
the most restrictive choice because they require an exact class match.

Kent

> 
> There's simply more than one way to do it or some of them are
> discouraged?
> 
> I've been told on #python that .__class__ is probably the worst one, but
> without precise motivation, just as a personal opinion.
> 
> ciao,
>     ema
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Tutor maillist  -  Tutor at python.org
> http://mail.python.org/mailman/listinfo/tutor