Join split words and punctuation with punctuation in the right place

Question

So I tried using join() after splitting a string into words and punctuation but it joins the string with a space in between the word and punctuation.

b = ['Hello', ',', 'who', 'are', 'you', '?']
c = " ".join(b)

But that returns:
c = 'Hello , who are you ?'

and I want:
c = 'Hello, who are you?'

Filial

Does calling the method on an empty string work? – Finkle

@CBroe: No, because you then get 'Hello,whoareyou?' instead.. – Surveyor

Yeah, right. So a simple join won’t do, because there’s different “types” of data, words and punctuation. So I guess one would need a self-written method that first checks which of those it is, and then concatenates the parts with a space or no space accordingly. – Finkle

Answer 1

You could join on the punctuation first:

def join_punctuation(seq, characters='.,;?!'):
    characters = set(characters)
    seq = iter(seq)
    current = next(seq)

    for nxt in seq:
        if nxt in characters:
            current += nxt
        else:
            yield current
            current = nxt

    yield current

c = ' '.join(join_punctuation(b))

The join_punctuation generator yields strings with any following punctuation already joined on:

>>> b = ['Hello', ',', 'who', 'are', 'you', '?']
>>> list(join_punctuation(b))
['Hello,', 'who', 'are', 'you?']
>>> ' '.join(join_punctuation(b))
'Hello, who are you?'

Surveyor

Answer 2

Do this after you get the result, not full, but works...

c = re.sub(r' ([^A-Za-z0-9])', r'\1', c)

Output:

c = 'Hello , who are you ?'
>>> c = re.sub(r' ([^A-Za-z0-9])', r'\1', c)
>>> c
'Hello, who are you?'
>>> 

Eberta

Good one, except that re.sub(r' (?=\W)', '', c) would be simpler. – Gripper

Answer 3

Maybe something like:

>>> from string import punctuation
>>> punc = set(punctuation) # or whatever special chars you want
>>> b = ['Hello', ',', 'who', 'are', 'you', '?']
>>> ''.join(w if set(w) <= punc else ' '+w for w in b).lstrip()
'Hello, who are you?'

This adds a space before words in b which aren't made up entirely of punctuation.

Sanders

Answer 4

Based on the answer of Martijn Pieters♦, I made a little generalization for languages with punctuation signs that can also be at the start of a word.

from string import punctuation


def join_punctuation(
    seq,
    characters_after=punctuation,
    characters_before="¡¿"
):
    characters_after = set(characters_after)
    characters_before = set(characters_before)
    seq = iter(seq)
    current = next(seq)

    for nxt in seq:
        if current in characters_before:
            current += nxt
        elif nxt in characters_after:
            current += nxt
        else:
            yield current
            current = nxt

    yield current

It would work the same way:

>>> b = ["Hola", ",", "¿", "Qué", "tal", "?"]
>>> list(join_punctuation(b))
['Hola,', '¿Qué', 'tal?']
>>> " ".join(join_punctuation(b))
'Hola, ¿Qué tal?'

Samford

Answer 5

How abt

c = " ".join(b).replace(" ,", ",")

L

You need to do the same for the ? question mark, and there may be more types of punctuation. – Surveyor