使用从0开始的唯一项替换列表中的项

Question

Let's say I have a list like this:

Y=[1018, 1018, 1011, 1012, 1013, 1014, 1019, 1019, 1017]

What's the most pythonic way to replace each number with

• the lowest unused integer (>=0), if the number has not been seen before
• the same integer that has been used to replace the number otherwise

So that the list becomes:

Y=[0, 0, 1, 2, 3, 4, 5, 5, 6]

It's not important that first element is 0, but there must be a unique maximal matching (= assignment) between the two lists of numbers, ie also this is a good solution:

Y=[3, 3, 4, 0, 2, 5, 6, 6, 1]

EDIT: what I tried is a for loop using find, my solution is very ugly, I know there is better way to do it, it's not relevant how bad I did it :D

Answer 1

The first idea that comes to mind is to convert the values to a set() and enumerate() them, store the pairs in a dict, and use a mapping list comprehension to create the new list:

>>> Y=[1018, 1018, 1011, 1012, 1013, 1014, 1019, 1019, 1017]
>>> mapping={v:k for k,v in enumerate(set(Y))}
>>> Y1=[mapping[y] for y in Y]
>>> Y1
[5, 5, 0, 1, 2, 3, 6, 6, 4]

Answer 2

You can also use a defaultdict with itertools.count , eg:

from collections import defaultdict
from itertools import count

dd = defaultdict(lambda c=count(): next(c))

Y=[1018, 1018, 1011, 1012, 1013, 1014, 1019, 1019, 1017]
mapped = [dd[el] for el in Y]
# [0, 0, 1, 2, 3, 4, 5, 5, 6]

How this works is that a defaultdict will return the value for an existing key but where that key doesn't exist, it will assign the key to a default value - in this case that value is the next number in sequence.

Answer 3

Here's what I typically use. The core logic is basically the same as what @JonClements wrote.

#!/usr/bin/env python3

# mypy static typing annotations
from typing import Dict, Generic, List, TypeVar

T = TypeVar('T')

class Interner(Generic[T]):
    def __init__(self):
        self._values = [] # type: List[T]
        self._keys = {} # type: Dict[T, int]
    def intern(self, val: T) -> int:
        idx = self._keys.setdefault(val, len(self._keys))
        if idx == len(self._values):
            self._values.append(val)
        return idx
    def unintern(self, idx: int) -> T:
        return self._values[idx]
    pass

def main():
    data = [1018, 1018, 1011, 1012, 1013, 1014, 1019, 1019, 1017]
    pool = Interner() # type: Interner[int]
    return [pool.intern(i) for i in data]

if __name__ == '__main__':
    print(main())