使用Python替换或交换文件中的子串

Question

Suppose I have a line in an ASCII file of the following form:

{text1} {stringA} {text2} {stringB} {text3}

where {stringA} and {stringB} are substrings of interest. Let's call them "A" and "B" respectively. The strings {text1} , {text2} , and {text3} are strings of any length (possibly empty) that do not contain either A or B.

What I want to do in Python is simply swap A and B such that the line goes from

{text1} {stringA} {text2} {stringB} {text3}

to

{text1} {stringB} {text2} {stringA} {text3}

I'd appreciate any help here. I think that by getting help on this question, it will help me learn to better work with regular expressions in Python.

Note that {text1} , {text2} , and {text3} are unknown strings.

We know exactly the substrings A and B. We know that A precedes B in the line. However, we don't know what (if anything) is before/between/after them.

Examples (A=John, B=Tim):

(1) This:

"I told John to give the bag to Tim."

is changed to this:

"I told Tim to give the bag to John."

(2) This:

"John said hello to Tim."

is changed to this:

"Tim said hello to John."

(3) This:

"John!h9aghagTim"

is changed to this:

"Tim!h9aghagJohn"

Answer 1

>>> import re
>>> text = '{text1} {stringA} {text2} {stringB} {text3}'
>>> re.sub(r'(stringA)(.*)(stringB)', r'\3\2\1', text)
'{text1} {stringB} {text2} {stringA} {text3}'

Replace stringA and stringB with your substrings of interest, note that you may want to re.escape() them in case the substrings can have characters with a special meaning in regex.

Test cases:

>>> stringA = 'John'
>>> stringB = 'Tim'
>>> regex = re.compile(r'(%s)(.*)(%s)' % (stringA, stringB))
>>> regex.sub(r'\3\2\1', "I told John to give the bag to Tim.")
'I told Tim to give the bag to John.'
>>> regex.sub(r'\3\2\1', "John said hello to Tim.")
'Tim said hello to John.'
>>> regex.sub(r'\3\2\1', "John!h9aghagTim")
'Tim!h9aghagJohn'

Answer 2

The approach to go for is to use capturing groups so that you can refer them to later

result = re.sub(r"(\{text1\}) (\{stringA\}) (\{text2\}) (\{stringB\}) (\{text3\})", r"\1 \4 \3 \2 \5", subject)

The capture group is identified by the parenthesis () and you refer to them in python by \\x where x is the number of the capture group

Update 1

Your examples makes it more obvious what you want and how you currently think about regexes. Regexes match patterns of characters. You want to swap names (Tom,Tim,...) so we need to come up with a pattern to match a name which is only possible by complete enumeration. In my language there ar (I think) thousand of first names and some of them are also used to refer to objects and not person. To make that distinction you have to take context into account which a regex cannot. Let me know if this makes sense cause it's important if you want to go any further.

Update 2

I suspect your question is out of curiosity and not to solve a real life problem. But if we go along than this would get you far but it's not perfect and cannot be

regex

(.*)\b(John|Tim|Jo)\b(.*)\b(John|Tim|Jo)\b

replace with

\1\4\3\2

In python

result = re.sub(r"(?sm)(.*)\b(John|Tim|Jo)\b(.*)\b(John|Tim|Jo)\b", r"\1\4\3\2", subject)

Note the \\b in the regex which states that the match should happen at word boundaries. This prevents matches like Johndoe.

Also observe that the regex above will fail for the sentence

Tim bought some top level domains of Jordan that end with Jo from John