Using python re need to match string that starts and ends with two possible patterns each

Question

The | symbol in regular expressions seems to divide the entire pattern, but I need to divide a smaller pattern... I want it to find a match that starts with either "Q: " or "A: ", and then ends before the next either "Q: " or "A: ". In between can be anything including newlines.

My attempt:

string = "Q: This is a question. \nQ: This is a 2nd question \non two lines. \n\nA: This is an answer. \nA: This is a 2nd answer \non two lines.\nQ: Here's another question. \nA: And another answer."

pattern = re.compile("(A: |Q: )[\w\W]*(A: |Q: |$)")

matches = pattern.finditer(string)
for match in matches:
    print('-', match.group(0))

The regex I am using is (A: |Q: )[\w\W]*(A: |Q: |$).

Here is the same string over multiple lines, just for reference:

Q: This is a question. 
Q: This is a 2nd question 
on two lines. 

A: This is an answer. 
A: This is a 2nd answer 
on two lines.
Q: Here's another question. 
A: And another answer.

So I was hoping the parenthesis would isolate the two possible patterns at the start and the three at the end, but instead it treats it like 4 separate patterns. Also it would include at the end the next A: or Q:, but hopefully you can see what I was going for. I was planning to just not use that group or something.

If it's helpful, this is for a simple study program that grabs the questions and answers from a text file to quiz the user. I was able to make it with the questions and answers being only one line each, but I'm having trouble getting an "A: " or "Q: " that has multiple lines.

Do you need to map each question to the right answer? Are they all following in set order? — Wiktor Stribiżew
– Wiktor Stribiżew, Commented Oct 28, 2021 at 14:39
@WiktorStribiżew Originally (when each Q: and A: was one line) I would go through and get all the Qs first into a list, then all the As. So the correct Qs and As would all have matching index numbers. — Mark
– Mark, Commented Oct 28, 2021 at 14:53
Then use two separate regexps and then zip the outputs. Use my regex I shared in the comments. I could provide an answer but I am on a mobile now. — Wiktor Stribiżew
– Wiktor Stribiżew, Commented Oct 28, 2021 at 15:01

Wizard.Ritvik · Accepted Answer · 2021-10-28 15:06:53Z

1

One approach could be to use a negative lookahead ?! to match a newline followed by an A: | Q: block, as follows:

^([AQ]):(?:.|\n(?![AQ]:))+

You can also try it out here on the Regex Demo.

Here's another approach suggested by @Wiktor that should be a little faster:

^[AQ]:.*(?:\n+(?![AQ]:).+)*

A slight modification where we match .* instead of like \n+ (but note that this also captures blank lines at the end):

^[AQ]:.*(?:\n(?![AQ]:).*)*

edited Oct 28, 2021 at 15:06

answered Oct 28, 2021 at 14:27

Wizard.Ritvik

12.1k5 gold badges38 silver badges76 bronze badges

Sign up to request clarification or add additional context in comments.

10 Comments

Wiktor Stribiżew Over a year ago

I think (?m)^[AQ]:.*(?:\n(?![AQ]:).+)* would be a much faster pattern for what you tried to achieve with yours.

Wiktor Stribiżew Over a year ago

You should never use an alternation of . and whitespace/line break patterns. There are too many issues related to that pattern. Simply use re.DOTALL to make . match line breaks.

Wiktor Stribiżew Over a year ago

Maybe. In my regex, to match empty lines, you need to replace .+ with .*.

Mark Over a year ago

@WiktorStribiżew rv.kvetch Great! You've both been very helpful because I would definitely like to save blank lines if possible. Trailing ones don't matter but there might be some in the middle of a Q/A

Wiktor Stribiżew Over a year ago

I have just published a YT video about the evil (?:\s|.)* pattern.

|

HelixAchaos · Accepted Answer · 2021-10-28 14:27:47Z

1

I suggest just using a for-loop for this as it's easier for me at least. To answer your question, why not just target until the period rather than the next A: | Q:? You'd probably have to use lookaheads otherwise.

(A: |Q: )[\s\S]*?\.

[\s\S] (Conventionally used to match every character though [\w\W] work as well)

*? is a lazy quantifier. It matches as few characters as it can. If we had just (A: |Q: )[\s\S]*?, then it'd only match the (A: |Q: ), but we have the ending \..

\. matches a literal period.

For the for-loop:

questions_and_answers = []
for line in string.splitlines():
    if line.startswith(("Q: ", "A: ")):
        questions_and_answers.append(line)
    else:
        questions_and_answers[-1] += line

# ['Q: This is a question. ', 'Q: This is a 2nd question on two lines. ', 'A: This is an answer. ', 'A: This is a 2nd answer on two lines.', "Q: Here's another question. ", 'A: And another answer.']```

answered Oct 28, 2021 at 14:27

HelixAchaos

1311 silver badge3 bronze badges

2 Comments

Mark Over a year ago

Unfortunately I can't use a period because the text might not include a period. But according to you and the other answer it seems like lookaheads is what I was looking for, so thank you.

Mark Over a year ago

Actually your alternative without re is very good! I'll give it a try. I want to upvote your answer but unfortunately I don't have enough points apparently.

Collectives™ on Stack Overflow

Using python re need to match string that starts and ends with two possible patterns each

2 Answers 2

10 Comments

2 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

10 Comments

2 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related