Matching everything but words, numbers and spaces

This code will replace everything except for words, but how do I get it to also leave the numbers and spaces untouched? e.g. "I didn't see him u开发者_JAVA百科ntil 1." -> "I didnt see him until 1"

text = regex.sub("\P{alpha}+","",text)

Don’t use Python’s re library on Unicode. It works very poorly. Use Matthew Barnett’s regex library instead. It works much, much better.

It also runs on both Python 2 and 3 and on both narrow and wide builds, but for reasons largely unrelated to that particular library I strongly recommend that you run only a wide build of Python 3 and eschew all other combinations.

Python regexes don't support Unicode properties. You can try:

text = re.sub("[^a-zA-Z0-9 ]+","",text)

Instead. If you do have something like Ponyguruma installed, you can use:

text = re.sub("[\P{Alnum}\PZ]+","",text) # pZ is shorthand for p{Separator}