fast way to filter a set?

[Mike C. Fletcher]

fortepianissimo wrote:

>But this created two lists (keys and dummy)? Compared to the filter()
>method, which created a list and a new set, maybe your suggestion is
>still a bit faster...
>  
>
Generators will probably be the easiest way to avoid the list-creation 
overhead...

 >>> from __future__ import generators
 >>> def fget( s, function ):
...     for item in s:
...         if function(item):
...             yield item
...            
 >>> r = Set.Set( range(20))
 >>> Set.Set( fget(r, lambda x: x%2))
[19, 1, 3, 17, 5, 7, 9, 11, 13, 15]
 >>>

HTH,
Mike

>--- Skip Montanaro <skip at pobox.com> wrote:
>  
>
>>    fortepianissimo> I know I can do things like
>>    fortepianissimo> s=Set(range(1,11))
>>    fortepianissimo> s=Set(filter(lambda x:x%2==0,s))
>>
>>    fortepianissimo> But this seems a bit slow since filter returns a
>>list
>>    fortepianissimo> which then must be converted back to a set. Any
>>tips?
>>    
>>
...
_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/