ansaurus

Question

Is it possible to solve this problem with Regular Expression?

Answer 1

+2 A:

Maybe like this? (Thanks to KennyTM for final optimizations.)

.{10}[^.?!]*[.?!]+

Thomas 2010-10-02 10:38:25

No need to escape these characters inside character classes. And this regex will always match the entire string.

Tim Pietzcker 2010-10-02 10:41:34

You're right, there was a problem. The escaping doesn't hurt, I believe.

Thomas 2010-10-02 10:51:23

I also try this: prog = re.compile('/([^\.\?\!]{10,}[^\.\?\!]*?[\.\?\!]+)+/')result = prog.split(test). But the result was whole sentence from the example.

Ilija 2010-10-02 11:02:35

@Ilija: I think you want to use prog.findAll() instead of prog.split(). I've updated the regex for use in Python.

Thomas 2010-10-02 11:11:07

Yes, this will work. Thanks a lot.

Ilija 2010-10-02 11:17:38

You could use `{10}` as a short cut of `{10,10}`. Also, I think `[^.?!]*` don't need to be lazy any more.

KennyTM 2010-10-02 11:35:31

@KennyTM: Yes to the shortcut. The expression evolved over time, so this small improvement was overlooked. ;) The [^.?!]*? is absolutely necessary (unless I'm missing something) to switch to non-greedy matching. Otherwise it would just match all of: 'Hello, this is part 1. And this should be part 2.', instead of matching only part 1 and part 2 in distinct matches. It depends on Python's default setting for greediness, of course.

Thomas 2010-10-02 11:41:44

@Thomas: No it won't match all of them. Notice that it is `[^.?!]*` not `.*`.

KennyTM 2010-10-02 11:43:40

@KennyTM: Haha, darn, got lost in my own expression. You're absolutely right. With that change, it should be as concise as it gets, I hope.

Thomas 2010-10-02 11:46:47

Answer 2

+2 A:

.{10,}?[.!?]+\s*

should work. It will also keep repeated punctuation characters together, so it splits Do you want lime??? Yes. I love when I drink tequila. into Do you want lime??? and Yes. I love when I drink tequila.

However, it doesn't take quoted speech into account and will break Peter said "Hi! How about dinner tonight?" and left. into Peter said "Hi!, How about dinner tonight? and " and left.

Could that be a problem that needs to be taken into account?

Tim Pietzcker 2010-10-02 10:39:58

No, that doesn't need to be taken into account? Now, I'm testing,

Ilija 2010-10-02 10:49:15

How would it split: Hello ?? !! ... My name is not important.So I'm not saying it ...

Thomas 2010-10-02 10:53:20

I'm trying this in Python prog = re.compile('.{10,}?[.!?]+\s*')result = prog.split(test) but it gives me emtpy list

Ilija 2010-10-02 11:00:28

@Thomas: Then it should probably be .{10,}?[.!?][\s.!?]*

ninjalj 2010-10-02 11:33:14

ansaurus

tags:

views:

answers:

Is it possible to solve this problem with Regular Expression?

related questions