Does anybody know how to define a custom way of sorting, rather than the default English alphabet order? If there is a course/video on here about this, please pount me to it, as no joy so far with searching.

Geir Arne Hjelle RP Team on March 3, 2023

Hi Irene, you can define your custom sort order by using the key argument. This is covered in an earlier video in this course. You can also read about it at realpython.com/python-sort/#sorted-with-a-key-argument

Additionally, the following tutorial focuses on sorting dictionaries but does provide several examples of using key that’ll work on any iterable: realpython.com/sort-python-dictionary/#using-the-key-parameter-and-lambda-functions

Bartosz Zaczyński RP Team on March 3, 2023

@Irene Galstian Just like you, I’ve been bitten by the fact that Python’s built-in sorted() function doesn’t support Unicode characters well. However, you can achieve the expected result in a few ways.

First, you can set the locale to the desired language and use the following transformation function as the sorting key:

>>> import locale

>>> names = [
...     "Łukasz",
...     "Józef",
...     "Małgorzata",
...     "Paweł",
...     "Radosław",
...     "Mirosław",
...     "Stanisław",
...     "Żaneta",
...     "Świętosław",
... ]

>>> locale.setlocale(locale.LC_ALL, "pl_PL.UTF-8")
'pl_PL.UTF-8'
>>> for name in sorted(names, key=locale.strxfrm):
...     print(name)
...
Józef
Łukasz
Małgorzata
Mirosław
Paweł
Radosław
Stanisław
Świętosław
Żaneta

Alternatively, you can use a third-party library like PyICU, which is a Python binding to IBM’s ICU library, or a pure-Python implementation of the Unicode Collation Algorithm called PyUCA:

>>> import pyuca as pyuca

>>> names = [
...     "Łukasz",
...     "Józef",
...     "Małgorzata",
...     "Paweł",
...     "Radosław",
...     "Mirosław",
...     "Stanisław",
...     "Żaneta",
...     "Świętosław",
... ]

>>> collator = pyuca.Collator()
>>> for name in sorted(names, key=collator.sort_key):
...     print(name)
...
Józef
Łukasz
Małgorzata
Mirosław
Paweł
Radosław
Stanisław
Świętosław
Żaneta

If you’d like to sort more complex objects, for example, people by their first name and last name, then you can make a compound key by wrapping the collator instances in a tuple:

>>> from typing import NamedTuple
>>> import pyuca as pyuca

>>> class Person(NamedTuple):
...     first_name: str
...     last_name: str
...     
...     def __str__(self):
...         return f"{self.first_name} {self.last_name}"
... 
>>> collator = pyuca.Collator()

>>> def compound_key(person):
...     return (
...         collator.sort_key(person.first_name),
...         collator.sort_key(person.last_name),
...     )

>>> people = [
...     Person("Łukasz", "Nowak"),
...     Person("Józef", "Kowalski"),
...     Person("Małgorzata", "Wiśniewska"),
...     Person("Paweł", "Wójcik"),
...     Person("Radosław", "Kowalczyk"),
...     Person("Mirosław", "Krawczyk"),
...     Person("Stanisław", "Mazur"),
...     Person("Żaneta", "Jabłońska"),
...     Person("Świętosław", "Łuczak"),
... ]

>>> for person in sorted(people, key=compound_key):
...     print(person)
...
Józef Kowalski
Łukasz Nowak
Małgorzata Wiśniewska
Mirosław Krawczyk
Paweł Wójcik
Radosław Kowalczyk
Stanisław Mazur
Świętosław Łuczak
Żaneta Jabłońska

img on March 3, 2023

Bartosz, I can’t thank you enough! My task is a concordance of Vedic Sanskrit words, which should follow the order of Sanskrit alphabet: aāiīuū etc. Your example gives me hope.

Bartosz Zaczyński RP Team on March 3, 2023

@img Glad you found it helpful! Let us know if you encounter any other challenges.

Become a Member to join the conversation.