How to keep the delimiter while using RegEx?

Question

I did a question about punctuation and regex, but it was confusing.

Supossing I have this text:

String text = "wor.d1, :word2. wo,rd3? word4!";

I'm doing this:

String parts[] = text.split(" ");

And I have this:

wor.d1, | :word2. | wor,d3? | word4!;

What I need to do to have this? (Keep the the symbols at the borders, but only that I specify: .,!?:, not all).

wor,d1 | , | : | word2 | . | wor,d3 | ? | word4 | !

UPDATE

I'm getting some good results with these regex, but it's giving an empty char before all splits on punctuation at start of a word.

There is a way to not have this empty char at the start?

Is this regex is good, or there is a more simple way?

public static final String PUNCTUATION_SEPARATOR =
        "("
        + "("
        + "(?=^[\"'!?.,;:(){}\\[\\]]+)"
        + "|"
        + "(?<=^[\"'!?.,;:(){}\\[\\]]+)"
        + ")"
        + "|"
        + "("
        + "(?=[\"'!?.,;:(){}\\[\\]]+($|\n))"
        + "|"
        + "(?<=[\"'!?.,;:(){}\\[\\]]+($|\n))"
        + ")"
        + ")";

See this question: http://stackoverflow.com/questions/275768/is-there-a-way-to-split-strings-with-string-split-and-include-the-delimiters — JJ., Aug 19 '11 at 21:05

m_vitaly · Answer 1 · 2011-08-19T21:19:59.640

2

Are you sure you want to use regex ? There's a faster implementation for splitting by single char: StringTokenizer. And it that can return the delimiters.

String str= "word1, word2. word3? word4!";
String delim = ",.!?";
StringTokenizer st = new StringTokenizer(str, delim, true);
while (st.hasMoreTokens()) {
  String token = st.nextToken();
  ... // token will be: "word1", ",", " word2", ".", etc...
}

edited Aug 19 '11 at 21:19

answered Aug 19 '11 at 21:14

m_vitaly

11,856
5
47
63

This works, but needs to be regex or something more complex, because I only want to spit on the borders (start and end) and not in the middle. – Renato Dinhani Aug 19 '11 at 21:30
I mean if the symbol is in the middle of the String (a-b, 20.50), I don't want to splitted, only in the borders (test, [100, etc.). – Renato Dinhani Aug 19 '11 at 21:58

score 1 · Answer 2 · answered Aug 19 '11 at 21:19

For simple separators I recommend the StringTokenizer. But here's a solution using regex and another auxiliary separator:

String s  = "one,two, three   four ,  five";
s = s.replaceAll("([,\\s]+)", "#$1#");
Pattern p = Pattern.compile("#");
String[] result = p.split(s);

score 1 · Answer 3 · answered Aug 19 '11 at 22:31

1

Here's a regex that I think will work:

/\s|(?=[\.,:?!](\W|$))|(?<=\W[\.:?!])/

answered Aug 19 '11 at 22:31

Jacob Eggers

9,062
2
25
43

score 0 · Answer 4 · answered Aug 19 '11 at 21:05

0

In my opinion you want this. First you explode your string and second step you use implode function.

answered Aug 19 '11 at 21:05

Yusuf ali

331
1
6
15

score 0 · Accepted Answer · answered Aug 23 '11 at 06:31

public static final String PUNCTUATION_SEPARATOR =
    "("
    + "("
    + "(?=^[\"'!?.,;:(){}\\[\\]-]+)"
    + "|"
    + "(?<=^[\"'!?.,;:(){}\\[\\]-]+)"
    + ")"
    + "|"
    + "("
    + "(?=[\"'!?.,;:(){}\\[\\]-]+($|\n))"
    + "|"
    + "(?<=[\"'!?.,;:(){}\\[\\]-]+($|\n))"
    + ")"
    + ")";

How to keep the delimiter while using RegEx?

UPDATE

5 Answers5

Linked