CODEWITHSUNDEEP
python
thecoparyew
thecoparyew

Reputation: 735

Map arrays to integers in python

I have an array [[0, 1, 1], [1, 0, 0], [1, 1, 0], [0, 1, 1], [1, 0, 0]] and I want to get this [[0], [1], [2], [0], [1]] - so I want to specify all the same vectors with unique number.

I have an idea how to do it, but it is really ugly, pretty slow and not pythonic at all. How could I do it more pythonic and nicer?

def get_multiclass_list_nodes(y_):
    _class = 0
    prev_list = list()
    new_y = list()
    for el in y_:
        el = el.astype('int32').tolist()
        if el in prev_list:
            index = -1
            for i, val in enumerate(prev_list):
                if val == el:
                    index = i + 1
                    break
            new_y.append([index])
        else:
            _class += 1
            prev_list.append(el)
            new_y.append([_class])
    return np.array(new_y)

Upvotes: 0

Views: 52

Answers (2)

Cyphase
Cyphase

Reputation: 12022

This should work for you:

def map_lists_to_ints(list_of_lists):  # Name this better
    tuples = (tuple(l) for l in list_of_lists)

    seen = {}
    next_int = 0
    for tup in tuples:
        if tup not in seen:
            seen[tup] = next_int
            next_int += 1

        yield seen[tup]

list_of_lists = [[0, 1, 1], [1, 0, 0], [1, 1, 0], [0, 1, 1], [1, 0, 0]]

result = list(map_lists_to_ints(list_of_lists))

print(result)

listified_result = [[x] for x in result]

print(listified_result)

Output:

[0, 1, 2, 0, 1]
[[0], [1], [2], [0], [1]]

Upvotes: 2

Paulo Almeida
Paulo Almeida

Reputation: 8071

Maybe this is abusing numpy a bit, but an easy way:

a = np.array([[0, 1, 1], [1, 0, 0], [1, 1, 0], [0, 1, 1], [1, 0, 0]])
lst = [str(x) for x in a]
result = np.unique(lst, return_inverse=True)[1]

Perhaps you'll want to tweak the final result, but it's trivial from there.

Upvotes: 0

Related Questions