automatic "lotjes trekken"

François Pinard pinard at iro.umontreal.ca
Sun Aug 25 06:01:50 CEST 2002


[Tim Peters]

> [François Pinard]

> > By the way, I do not remember any standard library helper to enumerate
> > all permutations (or combinations, or arrangements) of a given list
> > (or tuple).  Ideally one at a time, of course.  Would this be useful?

> [...]  I've likely got all the most useful algorithms coded up already
> (from permutations of n things taken k at a time, to partitions of an
> integer, in lexicographic or "Grey code" orderings).

Hello, Tim.  Re-reading this old email of yours.  Copy to Python list,
in case there is interest to some readers.

What do you call "partitions of an integer"?

How should I read the last "or" above?  Where the expression "Grey code"
comes from?  I presume there is no relation with "Gray code", because
lexicographic ordering is usually not compatible with these.

> There is a particular irritation inherent in your natural generalization
> from list to "list (or tuple)": you should really add "(or sequence)",
> including strings and user-defined sequence types.  A polymorphic
> generator has to be written very carefully to work with all those types!
> It also has to be written inefficiently, since it can't assume the base
> sequence type is mutable.  [...]

Indeed.  I never noticed that as a difficulty before, because the
application usually knows what to expect, and could convert the result of
a combinator received as a list, back to a tuple, or string, or whatever.
As you underline, this allows the combinator routine to be written more
efficiently, indeed.

> Let's make that concrete.  [...]  Another irritation is that the
> for/__getitem__ protocol isn't really suited to this kind of application
> [...]

I do not quote your examples here, as you published elegant solutions
based on generators since then.  That irritation is gone by now! :-)

-- 
François Pinard   http://www.iro.umontreal.ca/~pinard




More information about the Python-list mailing list