Is there a better way to compare dictionary values

You can use sets for this too

>>> a = {'x': 1, 'y': 2}
>>> b = {'y': 2, 'x': 1}
>>> set(a.iteritems())-set(b.iteritems())
set([])
>>> a['y']=3
>>> set(a.iteritems())-set(b.iteritems())
set([('y', 3)])
>>> set(b.iteritems())-set(a.iteritems())
set([('y', 2)])
>>> set(b.iteritems())^set(a.iteritems())
set([('y', 3), ('y', 2)])

Uhm, you are describing dict1 == dict2 ( check if boths dicts are equal )

But what your code does is all( dict1[k]==dict2[k] for k in dict1 ) ( check if all entries in dict1 are equal to those in dict2 )

If the dicts have identical sets of keys and you need all those prints for any value difference, there isn't much you can do; maybe something like:

diffkeys = [k for k in dict1 if dict1[k] != dict2[k]]
for k in diffkeys:
  print k, ':', dict1[k], '->', dict2[k]

pretty much equivalent to what you have, but you might get nicer presentation for example by sorting diffkeys before you loop on it.

If the true intent of the question is the comparison between dicts (rather than printing differences), the answer is

dict1 == dict2

This has been mentioned before, but I felt it was slightly drowning in other bits of information. It might appear superficial, but the value comparison of dicts has actually powerful semantics. It covers

• number of keys (if they don't match, the dicts are not equal)
• names of keys (if they don't match, they're not equal)
• value of each key (they have to be '==', too)

The last point again appears trivial, but is acutally interesting as it means that all of this applies recursively to nested dicts as well. E.g.

 m1 = {'f':True}
 m2 = {'f':True}
 m3 = {'a':1, 2:2, 3:m1}
 m4 = {'a':1, 2:2, 3:m2}
 m3 == m4  # True

Similar semantics exist for the comparison of lists. All of this makes it a no-brainer to e.g. compare deep Json structures, alone with a simple "==".