Compare 2 strings without considering accents in Python [duplicate]

Question

This question already has answers here:

What is the best way to remove accents (normalize) in a Python unicode string? (13 answers)

Closed 5 years ago.

I would like to compare 2 strings and have True if the strings are identical, without considering the accents.

Example : I would like the following code to print 'Bonjour'

if 'séquoia' in 'Mon sequoia est vert':
 print 'Bonjour'

• tripleee: Convert to fully decomposed normal form, remove accents, compare.
• Basj: Linked: stackoverflow.com/questions/517923/…

Answer 1

You should use unidecode function from Unidecode package:

from unidecode import unidecode
if unidecode(u'séquoia') in 'Mon sequoia est vert':
 print 'Bonjour'

Answer 2

You should take a look at Unidecode. With the module and this method, you can get a string without accent and then make your comparaison:

def remove_accents(data):
 return ''.join(x for x in unicodedata.normalize('NFKD', data) if x in string.ascii_letters).lower()
if remove_accents('séquoia') in 'Mon sequoia est vert':
 # Do something
 pass

Reference from stackoverflow

• Benjamin Lowry: This would not work if the word was "séQuoIa" since the remove_accents method makes all of the characters lowercase.

Answer 3

(sorry, late to the party!!)

How about instead, doing this:

>>> unicodedata.normalize('NFKD', 'î ï í ī į ì').encode('ASCII', 'ignore').decode('ascii')
'i i i i i i'

No need to loop over anything. @Maxime Lorant answer is very inefficient.

>>> import timeit
>>> code = """
import string, unicodedata
def remove_accents(data):
 return ''.join(x for x in unicodedata.normalize('NFKD', data) if x in string.ascii_letters).lower()
"""
>>> timeit.timeit("remove_accents('séquoia')", setup=code)
3.6028339862823486
>>> timeit.timeit("unicodedata.normalize('NFKD', 'séquoia').encode('ASCII', 'ignore')", setup='import unicodedata')
0.7447490692138672

Hint: less is better

Also, I'm sure the package unidecode @Seur suggested has other advantages, but it is still very slow compared to the native option that requires no 3rd party libraries.

>>> timeit.timeit("unicodedata.normalize('NFKD', 'séquoia').encode('ASCII', 'ignore')", setup="import unicodedata")
0.7662729263305664
>>> timeit.timeit("unidecode.unidecode('séquoia')", setup="import unidecode")
7.489392042160034

Hint: less is better

Putting it all together:

clean_text = unicodedata.normalize('NFKD', 'séquoia').encode('ASCII', 'ignore').decode('ascii')
if clean_text in 'Mon sequoia est vert':
 ...

• Arpad Horvath -- Слава Україні: I tried using Python 3.12. If we need to encode with ASCII at all we need also to revert it to be able to use the in operator. Bytes can't be at the left side of in. unicodedata.normalize('NFKD', 'î ï í ī į ì Í').encode('ASCII', 'ignore').decode('ASCII') I think the answer above is for Python 2.x
• Javier Buzzi: @ArpadHorvath--СлаваУкраїні It is for py 2.x, this is over 4 years old :P notice the u in front of the strings