如何避免使用正则表达式将字符串括在双引号中？

Question

I'm using regex in java to get all strings excluding the double quotes AND the strings inside the double quotes for this string:

"Lorem ipsum mauris "libero" non "pulvinar" suscipit, nis "aenean" 
curae odio lobortis "nulla" suspendisse"

I can get the strings enclosed in the double quotes using:

((\")(\S+)(\"))

Result:

[^((\")(\S+)(\"))]

which is the opposite of what I want
but when I try to negate the pattern:

 [^((\\")(\\S+)(\\"))] 

the strings not enclosed in double quotes don't get targeted:

What I want is this:

 "Lorem ","ipsum","mauris","non","suscipit",",","nis","curae", 
 "odio","lobortis", "suspendiss" 

Any help would be appreciated

Answer 1

Character classes only negate individual characters and can't negate the pattern like you tried.

You can use this regex which uses look arounds to reject strings that are surrounded by double quotes,

(?!<")\b\w+\b(?!")

Here word boundary \\b ensures, that partial word should not be detected as a match. For eg in word "libero" if we don't put \\b around the regex, then it may detect iber as a match from the middle of that word.

Demo

Java code for same would be,

String s = "Lorem ipsum mauris \"libero\" non \"pulvinar\" suscipit, nis \"aenean\" curae odio lobortis \"nulla\" suspendisse";
Pattern p = Pattern.compile("(?!<\")\\b\\w+\\b(?!\")");
Matcher m = p.matcher(s);

while (m.find()) {
    System.out.println(m.group());
}

Which prints,

Lorem
ipsum
mauris
non
suscipit
nis
curae
odio
lobortis
suspendisse

Edit:

I realized you also want to get comma , as a matched string, so in that case you can change the regex a bit to say it like this,

(?!<")\b\w+\b(?!")|,

Although, now I am guessing that you string might have other special characters too other than comma, and in that case you can use a character class like this [,.!;] instead of just comma. Also depending upon how you want to group those characters, whether continuous once together like ,;! (then use [,.!;]+ ) or each of those special characters individually, then just keep the character class.