additional difference algorithm suggeestion

[mgcdanny]

Hi,

This dictdiffer library rocks! Thank you all.

This might be another nice difference algorithm per #42. Basically similar to the symmetric difference operator from python's Set data structure.

Setup:

a = {'numbers' : [1, 2, 3]}
b = {'numbers': [0, 1, 2, 3]}

result = dd.diff(a, b)
for r in result:
    print(r)

Current output:

('change', ['numbers', 0], (1, 0))
('change', ['numbers', 1], (2, 1))
('change', ['numbers', 2], (3, 2))
('add', 'numbers', [(3, 3)])

Proposed output:

('symmetric_diff', ['numbers'], (, 0))

So (,0) would represent the addition of 0 to the second dict and that everything in the first dict is contained in the second. I think this would be like

(set(a) - set(b), set(b)-set(a))

Or at least some way of getting at the fact that the value 0 was added to the second list and not the value 3.

Cheers!

[jirikuncar]

@mgcdanny what do you think about support for sets in #45?

a = {'numbers' : {1, 2, 3}}
b = {'numbers': {0, 1, 2, 3}}
assert list(diff(a, b)) == [('add', ['numbers'], (0, {0}))]

[mgcdanny]

Yeah, that seems to works really nicely. Good idea! Thanks

[lnielsen]

+1

[tiborsimko]

Closing the issue now that PR #45 was merged.