如何在python中替换字符串中的非字母和数字字符

Question

I understand that to replace non-alphanumeric characters in a string a code would be as follows:

words = re.sub("[^\w]", " ",  str).split()

However, ^\\w replaces non-alphanumeric characters. I want to replace both non-alphabetic and numeric chars in a string like:

"baa!!!!! baa sheep23? baa baa"

and I want it to have an outcome like this:

 "baa baa sheep baa baa"

If I do words = re.sub("[^\\w\\d]", " ", str).split() , I get an outcome with numeric characters, like 'sheep23' . I guess that it could be because "^" affects \\d as well and it counts as if I want the non-numeric characters removed. How do I do this right?

Answer 1

Use str.translate :

>>> from string import punctuation, digits
>>> s = "baa!!!!! baa sheep23? baa baa"
>>> s.translate(None, punctuation+digits)
'baa baa sheep baa baa'

Answer 2

No need to do regex here, just a simple comprehension will work:

>>> import string
>>> word = "baa!!!!! baa sheep23? baa baa"
>>> "".join([l for l in word if l in string.ascii_letters+string.whitespace])
'baa baa sheep baa baa'

Answer 3

Try this regex:

[^a-zA-Z]

This matches anything that is not a letter.

Or this if you want to keep spaces:

[^a-zA-Z\\s]

Answer 4

What about this regex?

[^\w]|\d

EDIT:

As @Avinash said this not removes _ . If you want to remove also _ use:

[^\w]|[\d_]

and if you also want to replace multiple spaces with a single one use:

([^\w]|[\d_])+

Here's your example with an addition of underscores:

In [1]: import re

In [2]: s = "baa!!!!! baa sheep23? baa baa___"

In [3]: re.sub("([^\w]|[\d_])+", " ",  s)
Out[3]: 'baa baa sheep baa baa '

In [4]: re.sub("([^\w]|[\d_])+", " ",  s).split()
Out[4]: ['baa', 'baa', 'sheep', 'baa', 'baa']

Answer 5

Through re.sub function,

>>> s = "baa!!!!! baa sheep23? baa baa"
>>> m = re.sub(r'[^A-Za-z ]', "", s)
>>> m
'baa baa sheep baa baa'

Answer 6

Instead of replacing every non-letter with a space then split you can do it all in one go:

>>> re.split("[^a-zA-Z]+", "baa!!!!! baa sheep23? baa baa")
['baa', 'baa', 'sheep', 'baa', 'baa']

[^\\w] is equivalent to [^a-zA-Z0-9_] (modulo language settings), you need to keep in your character class only what you want - and [^a-zA-Z] obviously includes spaces.