Python查找所有出现连字的单词并在位置处替换

Question

I have to replace all occurrences of patterns with hyphen like cccc-come or oh-oh-oh-oh , etc. with the last token ie come or oh in this example, where

• The number of character between hyphen is arbitrary, it can be one ore more characters
• the token to match is the last token in the hyphenation, hence come in cc-come .

• the input string may have one or more occurrences of it like the following sentences:

  cccc-come to home today cccc-come to me

  oh-oh-oh-oh it's a bad life oh-oh-oh-oh

• Need to find the start and end position of the matched token via finditer

   r = re.compile(pattern, flags=re.I | re.X | re.UNICODE) for m in r.finditer(text): word=m.group() characterOffsetBegin=m.start() characterOffsetEnd=m.end() # now replace and store indexes 

[UPDATE]

Assumed that those hyphenated words does not belong to a fixed dictionary, I'm adding this constraint to it:

• The number of character between hyphen must range from a minimum to a max, like {1,3} so that the capture group must match c-come , or cc-come , but not a hyphenated real word like fine-tuning or like inter-face , etc.

Answer 1

You can just use re.sub() to replace all without having to iterate over matched indices:

import re

s = 'c-c-c-c-come to home today c-c-c-c-come to me'

print(re.sub(r'(\w+(?:-))+(\w+)', '\\2', s))
# come to home today come to me

Answer 2

Here is one possible expression:

import re

text = ("c-c-c-c-come to home today c-c-c-c-come to me, "
        "oh-oh-oh-oh it's a bad life oh-oh-oh-oh")
pattern = r"(?<=-)\w+(?=[^-\w])"
r = re.compile(pattern, flags=re.I | re.X | re.UNICODE)
for m in r.finditer(text):
    word = m.group()
    characterOffsetBegin = m.start()
    print(word, characterOffsetBegin)

Output:

come 8
come 35
oh 56

Answer 3

An option using a capturing group and a backreference might be:

(?<!\S)(\w{2,3})(?:-\1)*-(\w+)(?!\S)

That will match:

• (?<!\\S) Negative lookbehind, assert what is on the left is not a non whitespace char
• (\\w{2,3}) Capture in group 1 two or three times a word char
• (?:-\\1)* Repeat 0+ times matching a hyphen followed by a backreference to what is matched in group 1
• -(\\w+) Match - followed by matching 1+ word chars in group 2
• (?!\\S) Negative lookahead, assert what is on the right is not a non whitespace char

In the replacement use the second capturing group \\\\2 or r'\\2

Regex demo | Python demo

For example

import re

text = "c-c-c-c-come oh-oh-oh-oh it's a bad life oh-oh-oh-oh"
pattern = r"(?<!\S)(\w{1,3})(?:-\1)*-(\w+)(?!\S)"
text = re.sub(pattern, r'\2', text)
print(text)

Result

come oh it's a bad life oh

Answer 4

It can be done without regular expressions. Code:

s = "c-c-c-c-come to home today c-c-c-c-come to me"
s = " ".join(w if "-" not in w else w[w.rindex('-') + 1:] for w in s.split(" "))

Output:

come to home today come to me