How to split a string by any characters but letters? In other words, I want only words from the text, nothing else.
s="(This# is an5example!)" what_i_want=['This', 'is', 'an', 'example'] 
1 Answer
Please keep the content from @Andj below into account. I've copied the relevant part below:
"français".isalpha()will returnTruewhile"franc\u0327ais".isalpha()will returnFalse. For many languages for longer strings it will always returnFalse. Python's definition of Alphabetic differs from Unicode's definition.
Original answer:
You can iterate over the characters in the string and test if they are alphabetical characters. Add some logic so you don't put empty strings in the result list and you're done:
s = "(This# is an5example!)" word = "" word_list = [] for character in s: if character.isalpha(): word += character elif len(word) > 0: word_list.append(word) word = "" print(word_list) output
['This', 'is', 'an', 'example'] 4 Comments
user19273
Of course, but is this the most pythonic way?
Edo Akse
Not sure, but no imports and a single pass over the data
Andj
@EdoAkse
str.isalpha() is fine for English, but can be dangerous for many other languages. "français".isalpha() will return True while "franc\u0327ais".isalpha() will return False. For many languages for longer strings it will always return False. Python's definition of Alphabetic differs from Unicode's definition.Edo Akse
That I did not know, I'll update the answer to reflect this.
join. Simplyre.findall("[a-zA-Z]+", s)''.join([c if c.isalpha() else '\n' for c in s]).split()[A-Za-z], in Posix notation as a Unicode set[[:alpha:]&[:ASCII:]]. Although technically, Letters could be anything in[\p{L}]or maybe more importantly[\p{Alphabetic}], although begs the question do you include ideographs, etc in the mix? Although in strictest sense it would be[\p{L}]. But then, I tend to expect a high level of imprecision in terminology from Python developers, since Python itself tends towards the same imprecisions.