Operator overloading for part of an object

[David Humphrey]

"Scherer, Bill" <Bill.Scherer at VerizonWireless.com> wrote in message
news:mailman.990633334.870.python-list at python.org...
> On Wed, 23 May 2001, David Humphrey wrote:
>
> > I've been using operator overloading in both C++ and Python for some
time
> > now and have run into a situation I never thought about before.  How do
you
> > overload an operator that operates on some of the data for pair of
instances
> > and return only that modified data to the resulting instance?
> >
> > Here's a simple example.  Assume class C stores a name and a float and
> > defines the __add__ operator.  I want to be able to write expressions
like:
> >
> > c = a + b
> >
> > so that the float portion of instance c is the sum of the float portions
of
> > instances a and b.  Here's some code that specifies all this more
exactly:
> >
> > #==========================================
> > import sys
> >
> > class C:
> >    def __init__(self,aname,somedata=0.0):
> >       self.name = aname
> >       self.data = somedata
> >    def __add__(self,b):
> >       return self.data + b.data
> >    def report(self):
> >       print 'name = %s, data = %d\n' % (self.name,self.data)
>
> I think you need __add__ to return an instance of C:
>
>   def __add__(self, other):
>     return C(self.name, self.data + other.data)
>
> In your example c is reassigned to a float after the add, so it's no
> longer an instance of C.  With __add__ returning an instance of C, c in
> you example should work as you require (all untested, of course...)
>
Close, but not quite.  With the code suggested above, C will receive
self.name ('instance a' in the example) which will overwrite the original
name of c which is 'instance c'.  The problem here is that __add__ can't
know anything about the resulting object.  In C++, I think I can work around
this problem with a copy constructor, but it seems that Python doesn't offer
such facilities.  Therefore, I guess I'll have to do something like

c.datacopy(a+b)

Also, see Alex Martelli's reply to my original message.

> > #
> > # main code
> > #
> > a = C('instance a',1.0)
> > a.report()
> > b = C('instance b',2.0)
> > b.report()
> > c = C('instance c')
> > c.report()
> >
> > c = a + b
> > c.report()
> >
> > #==========================================
> >
> > In this case, I want c.report() to tell me
> >
> > name = c, data = 3.0
> >
> > but, instead, I get
> >
> > $ python ./test.py
> > name = instance a, data = 1
> >
> > name = instance b, data = 2
> >
> > name = instance c, data = 3
> >
> > Traceback (most recent call last):
> >   File "./test.py", line 26, in ?
> >     c.report()
> > AttributeError: 'float' object has no attribute 'report'
> >
> >
> > OK, I understand why it behaves this way, but I don't see how to
preserve
> > c.name while updating c.data with the computed value.  I konw I need to
> > overload the '=' operator, but don't see facilities in python to do
that.
> >
> > Thanks in advance for your advice.
> >
> > David L. Humphrey
> > Manager, Software Development
> > Bell Geospace, Inc
> >
> >
> >
>
> William K. Scherer
> Sr. Member of Applications Staff - Verizon Wireless
> Bill.Scherer_at_VerizonWireless.com
>
>
Regards,
DLH