Question

As I’m French, I’m trying to make a little function that can add the good definite article before a country name. I have no problem except for the few countries that start with a diacritic. Here’s my code :

#!/usr/bin/env python
# -*- coding: utf-8 -*-
def article(nomPays):
    voyelles = ['A','E','É','I','O','U','Y']
    if nomPays == 'Mexique':
        return 'du'
    elif nomPays[0] in voyelles:
        return 'de l\''
    elif nomPays[-1] == 'e':#signe négatif pour compter à partir de la dernière lettre
        return 'de la'
    else:
        return 'du'

print article('Érythrée')

If I enter Allemagne instead of Érythrée, the behaviour is correct : it returns ‘de l”. But Érythrée returns ‘de la’. It means my function doesn’t recognize the character É as part of the voyelles list.

Can anyone explain me why and how I can resolve this?

Answer 1

The problem is that you’re using str in Python 2, where str is a sequence of bytes and so nomPays[0] will give the first byte of the string, not the first character. In single-byte encodings this isn’t a problem, but with multi-byte encodings like UTF-8 the first byte of “Érythrée” is a lead byte and not the whole character “É”.

You need to change to use unicode to grab the first character:

firstChar = unicode(nomPays, 'UTF-8')[0].encode('UTF-8')

Actually, it’d probably be easier to use startswith:

if any(nomPays.startswith(voyelle) for voyelle in voyelles):

Alternatively you could use unicode throughout your application, or switch to Python 3, where all this is handled much better.