Separating __cmp__ from < and > ?

[Bjorn Pettersen]

Tim Peters wrote:
> 
> [Huaiyu Zhu]
> > In many situations it is conceptually possible to have a!=b without
> > either a<b or a>b.  Mathematically, this is the case for objects
> > that live in a partially ordered space.
> > ...
> 
> As you've discovered, since all comparisons in Python today funnel thru a
> single cmp function, you can't use the builtin operators to achieve this.
> 
> You can define your own (e.g.) lt gt eq ne etc methods, but can't map them
> to infix < > == != etc.
> 
> Allowing redefinition of individual comparison operators is one of the goals
> of David Ascher's "Rich Comparisons" proposal, although the *real* thrust of
> that was to allow NumPy to do stuff like return a matrix of boolean values
> when doing elementwise matrix compares.
> 
> The PEP for Rich Comparisons is at
> 
>     http://python.sourceforge.net/peps/pep-0207.html
> 
> although it's currently somewhat less than fully fleshed out <ahem>.

I ran into the same limitation when I wanted to create an abstraction
level on top of SQL, ie. I wanted to be able to say something like:

  t = Table("mytable")
  sel = Selector()
  sel << t['col1']           #  select mytable.col1
                             #  from mytable
  sel.where( t['col1'] > 2 ) #  where mytable.col1 > 2

the functional alternative was just too verbose, so I shelved the
idea...

-- bjorn