4
\$\begingroup\$

We have two lists:

l=["a","b","c"]
s=[["a","b","c"],
["a","d","c"],
["a-B1","b","c"],
["a","e","c"],
["a_2","c"],
["a","d-2"],
["a-3","b","c-1-1","d"]]
print l 
print s

Now, I am try to see if each 2nd-level list of s has fuzzy match to any of items in list l:

matches=list() 
matchlist2=list()   
print s2
for i in range(len(s)):
    matches.append([])
    for j in range(len(s[i])):

        for x in l:
            if s[i][j].find(x)>=0:
                print s[i][j]
                matches[i].append(True)

                break
        else:
            matches[i].append(False)


    matchlist2.append(all(x for x in matches[i]))
print matches
print matchlist2

This gives me what was intended. But I am not happy with how it has so many loops. I am also working with pandas and if there is pandas solution that will be great to. In pandas, there are just two columns of two dataframes.

[[True, True, True], [True, False, True], [True, True, True], [True, False, True], [True, True], [True, False], [True, True, True, False]]

[True, False, True, False, True, False, False]
\$\endgroup\$
5
\$\begingroup\$

Extract this into functions

You are doing one well defined thing (finding fuzzy matches), so this should be at least one function, but I suggest more because you are returning two results.

Write automatic tests

Re-factoring is very hard without tests, re-running and checking the input manually is very time consuming, an automatic test runs in milliseconds.

Use the list comprehensions and zip

You are basically checking the single-dimensional list (xs) against each row of the two-dimensional list (matrix) and seeing which elements are contained and which are not. This can be expressed very easily in code.

def fuzzy_matches(xs, matrix):
    """
    >>> list(fuzzy_matches(["a","b","c"], [["a", "b43", "x"], ["w", "b", "5cfg"]]))
    [[True, True, False], [False, True, True]]
    """
    return ([i in j for i, j in zip(xs, line)] for line in matrix)
| improve this answer | |
\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.