python用唯一数字替换列表中的元素

Question

I'm stuck with some problem for what I couldn't find a solution yet.

I have a list which looks like his and has actually 60000 elements:

lst = [[0, [-24.75, -24.75, -25.0], 0.00001, 2],
[10, [-26, -26, -26], 0.00011, 4], 
[0, [-3, -200000, -25.0], 0.000009, 42], 
[0, [-4, -4.7, -5], 0.00801, 7], 
[1, [-3, -200000, -25.0], 0.00089, 8], 
[2, [-3, -200000, -25.0], 0.000899, 18]]

Elements [lst[i][1] for i in range(len(lst))] are supposed to be Cartesian coordinates and some of them occur more than once.

I would like to give every coordinate a unique number, so that the list becomes:

lst = [[0, 0, 0.00001, 2],
[10, 1, 0.00011, 4], 
[0, 2, 0.000009, 42], 
[0, 3, 0.00801, 7], 
[1, 2, 0.00089, 8], 
[2, 2, 0.000899, 18]]

This means, if a coordinate [-3, -200000, -25.0] was once replaced with a number (here 2 ), all duplicates of this coordinate must also be replaced by the same number.

It is also okay if elements in the original list lst will be rearranged, main point is that I need to replace all of the coordinate triples with numbers.

Thank you in advance.

Answer 1

We can first construct a defaultdict :

from collections import defaultdict
from itertools import count

dispatcher = defaultdict(lambda c=count(0): next(c))

And now we can simply replace with a lookup. Now since the elements are not hashable, we can not simply add them to the dictionary. But we can convert them into a tuple.

So now we can use:

for lsti in lst:
    lsti[1] = dispatcher[tuple(lsti[1])]

For the given input this generates:

>>> lst
[[0, 0, 1e-05, 2], [10, 1, 0.00011, 4], [0, 2, 9e-06, 42], [0, 3, 0.00801, 7], [1, 2, 0.00089, 8], [2, 2, 0.000899, 18]]

Answer 2

Make a dictionary of values you've already seen. For each item in the list, if its coordinates are in that dictionary, use its saved value. Otherwise, insert that key into the dictionary and update the running count:

coordmap = {}
count = 0
for item in lst:
    coords = tuple(item[1])
    try:
        new_value = coordmap[coords]
    except KeyError:
        new_value = coordmap[coords] = count
        count += 1
    item[1] = new_value

Answer 3

Tuples are hashable. You can use this to maintain a dict of already visited coordinates quite easily.

import random  
seen = { }
i = 0
for l in lst:
    coord = tuple(l[1])
    if coord not in seen:
        seen[coord] = i
        i += 1

    l[1] = seen[coord]

Contents:

[[0, 0, 1e-05, 2],
 [10, 1, 0.00011, 4],
 [0, 2, 9e-06, 42],
 [0, 3, 0.00801, 7],
 [1, 4, 0.00089, 8],
 [2, 5, 0.000899, 18]]