What is the correct way to define __hash__?

[Peng Yu]

Hi,

I'm wondering what is the general way to define __hash__. I could add
up all the members. But I am wondering if this would cause a
performance issue for certain classes.

Regards,
Peng


#!/usr/bin/env python

class A:
def __init__(self, a, b) :
self._a = a
self._b = b

def __str__(self):
return 'A(%s, %s)' %(self._a, self._b)

__repr__ = __str__

def __cmp__(self, other):
if self._a < other._a:
return -1
elif self._a > other._a:
return 1
elif self._b < other._b:
return -1
elif self._b > other._b:
return 1
else:
return 0

def __hash__(self):
return self._a + self._b

if __name__ == '__main__':

x = A(1, 1)

aset = set()
aset.add(x)
print aset

 

[Steven D'Aprano]

def __cmp__(self, other):
if self._a < other._a:
return -1
elif self._a > other._a:
return 1
elif self._b < other._b:
return -1
elif self._b > other._b:
return 1
else:
return 0

This can be simplified to:

return cmp((self._a, self._b), (other._a, other._b))

 

[Chris Rebert]

This can be simplified to:

return cmp((self._a, self._b), (other._a, other._b))

Assuming you're not using Python 3.x that is.

Cheers,
Chris

 

[greg]

Assuming you're not using Python 3.x that is.

If you're using Python 3, you won't be writing a __cmp__
method in the first place.