Question

Does anyone know why these two return different ratios.

>>> import difflib
>>> difflib.SequenceMatcher(None, '10101789', '11426089').ratio()
0.5
>>> difflib.SequenceMatcher(None, '11426089', '10101789').ratio()
0.625

Answer 1

This gives some ideas of how matching works.

>>> import difflib
>>> 
>>> def print_matches(a, b):
...     s =  difflib.SequenceMatcher(None, a, b)
...     for block in s.get_matching_blocks():
...         print "a[%d] and b[%d] match for %d elements" % block
...     print s.ratio()
... 
>>> print_matches('01017', '14260')
a[0] and b[4] match for 1 elements
a[5] and b[5] match for 0 elements
0.2
>>> print_matches('14260', '01017')
a[0] and b[1] match for 1 elements
a[4] and b[2] match for 1 elements
a[5] and b[5] match for 0 elements
0.4

It looks as if it matches as much as it can on the first sequence against the second and continues from the matches. In this case (‘01017’, ‘14260’), the righthand match is on the 0, the last character, so no further matches on the right are possible. In this case (‘14260’, ‘01017’), the 1s match and the 0 still is available to match on the right, so two matches are found.

I think the matching algorithm is commutative against sorted sequences.