HELP!!! string logical sorting

[kucingkembar]

sorry for my bad english,
i have this problem :

thelist = ["t9","t8","t11","t10"]
print(thelist)
thelist.sort()
print(thelist)

Output:
['t9', 't8', 't10', 't11']
['t10', 't11', 't8', 't9']

expected output is :

Output:
['t8', 't9', 't10', 't11']

i google for [python string "logical" "sorting"] but found nothing
please help me!!!

[deanhystad]

Natural sorting. When sorting strings, t11 logically comes before t8, but it doesn’t look natural

[Pedroski55]

Try this:

thelist = ["t9","t8","t11","t10"]
# not quite good
sorted_list = sorted(thelist, key=lambda x: len(x))
# ok
sorted_list = sorted(thelist, key=lambda x: int(x[1:]))

[kucingkembar]

thank you for both answers,
the real problem is the above is a simplified example,
the real case is used for list files in folders that may contain hundreds or thousands of files

[DeaD_EyE]

Fix for your example.

def by_number(text):
    return int(text[1:])


thelist = ["t9","t8","t11","t10"]
print(thelist)

# using the key function by_number, which returns int
# the int is then used for comparison

thelist.sort(key=by_number)
print(thelist)

If you work with filenames, it's similar. Then you have to know the structure of the name, decompose it until only the number is left and then converting it to an int. If your filenames have a ISO8601 prefix, then you can parse it with datetime. datetime objects are sortable.

Example with ISO8601 prefix and pathlib:

from datetime import date as Date
from pathlib import Path


def by_date(path):
    date_str, _ = path.name.split("_", maxsplit=1)
    return Date.fromisoformat(date_str)


def walk(root, pattern):
    for path in Path(root).glob(pattern):
        yield path


def main():
    root = "."

    pattern = "????-??-??_*.*"
    # or more explicit glob pattern: pattern = "[0-2][0-0][0-9][0-9]-[0-1][0-9]-[0-3][0-9]_*.*"
    # 2022-10-12_bla1.txt matches the glob pattern

    for path in sorted(walk(root, pattern), key=by_date):
        print(path)


if __name__ == "__main__":
    main()

[Pedroski55]

@kucingkembar

You may paste an example list with file names if you wish!

[kucingkembar]

this project is translating raw manga/manhua/comics from non-alphabet words to English,
the data obtained using selenium to [save page as - web page complete],
so the image names may be random,
but if you read them using image-viewer, they have right order,
like these :
-001.jpg, 002.jpg
-1.jpg, 10.jpg
-6183416182372182_86981783_1.webp ,6183416182372182_86981783_2.webp
-1722013291_90649092356080.jpg, 1722013293_31683094936372.jpg

[Gribouillis]

I would use lists as sorting keys, where the integers have been converted to Python's int type

examples = [
    ["002.jpg", "001.jpg"],
    ["10.jpg", "1.jpg"],
    ["6183416182372182_86981783_2.webp", "6183416182372182_86981783_1.webp"],
    ["1722013293_31683094936372.jpg", "1722013291_90649092356080.jpg"],
]

import re


def numed(s):
    L = re.split(r"(\d+)", s)
    for i in range(1, len(L), 2):
        L[i] = int(L[i])
    return L


print("Sorting keys:")
for names in examples:
    print([numed(x) for x in names])

print("Sorted names")
for names in examples:
    print("unsorted:", names)
    print("sorted  :", sorted(names, key=numed))

Output:
Sorting keys:
[['', 2, '.jpg'], ['', 1, '.jpg']]
[['', 10, '.jpg'], ['', 1, '.jpg']]
[['', 6183416182372182, '_', 86981783, '_', 2, '.webp'], ['', 6183416182372182, '_', 86981783, '_', 1, '.webp']]
[['', 1722013293, '_', 31683094936372, '.jpg'], ['', 1722013291, '_', 90649092356080, '.jpg']]
Sorted names
unsorted: ['002.jpg', '001.jpg']
sorted  : ['001.jpg', '002.jpg']
unsorted: ['10.jpg', '1.jpg']
sorted  : ['1.jpg', '10.jpg']
unsorted: ['6183416182372182_86981783_2.webp', '6183416182372182_86981783_1.webp']
sorted  : ['6183416182372182_86981783_1.webp', '6183416182372182_86981783_2.webp']
unsorted: ['1722013293_31683094936372.jpg', '1722013291_90649092356080.jpg']
sorted  : ['1722013291_90649092356080.jpg', '1722013293_31683094936372.jpg']

There are also solutions in Pypi such as natsort, but I don't know these modules. Install with care.

[kucingkembar]

@Gribouillis
your code is working flawlessly,
but i don't understand this part:

key=numed

how the sort only pick [int] data only, and ignore the [str]

[Gribouillis]

how the sort only pick [int] data only, and ignore the [str]

The numed function takes a string and returns a list where the (non negative) integers have been converted as python int
for example

print(numed('foo10bar035spam'))  # -> prints ['foo', 10, 'bar', 35, 'spam']
print(numed('foo2bar035spam'))    # -> ['foo', 2, 'bar', 35, 'spam']

The sorted function compares these lists instead of comparing the original strings, which produces the correct result. Lists are compared with lexicographic order

>>> ['foo', 2, 'bar', 35, 'spam'] < ['foo', 10, 'bar', 35, 'spam']
True

Strings are not ignored, but 'foo' and 'foo' compare equal.