Python 3.0, rich comparisons and sorting order

[Steven Bethard]

Alex Martelli <aleaxit <at> yahoo.com> writes:
> 
> I wrote:
>    ...
> > (2) "is there a good use case for wanting to make a mapping with keys that
> > have incompatible types?" (my question to you)
> > 
> > To some degree, (1) has already been answered to my satisfaction by Carlos
> > Ribeiro's spreadsheet example.  If you could give me a real world example 
> > of when you'd want to do (2), I might be more convinced...
> 
> Does my example of 'tuples as concrete representations of expressions'
> which I posted to another subthread of this thread help?

Yeah, thanks.  It's still definitely not the kind of computation that 99% of 
users are going to want to do, but if you wanted to memoize such tuples and 
you also expected hashing said tuples to be more expensive than cmp'ing them, 
I can see that the mapping you replace dict with would need to be able to sort 
disparate types.

'Course rather than wrap your expression tuples in a class to define the 
__cmp__ function, it seems like a more reasonable solution might be to make 
the arbitrary cmp decision in the mapping class (the one implemented as a 
BTree).  If __cmp__ starts raising TypeErrors, your code could do something 
like:

def insert(self, val):
    try:
        c = cmp(self.val, val)
    except TypeError:
        c = -1 # self.val and val were not of the same type
               # make some arbitrary decision here
    ... # insert according to cmp result

Does take more work than it would now -- forces the arbitrary decision onto 
the BTree writer, instead of doing it automatically in cmp...  How much more 
work really depends on how complicated your 'arbitrary' scheme is...

Steve