Ytukyg

This regex: b([A-z*]+)-(?=[A-z*]+b)

with this replacement: $1 

Applied on:

Jean-Pierre bought "blue-green-red" product-2345 and other blue-red stuff.

Gives me:

Jean Pierre bought "blue green red" product-2345 and other blue red stuff.

While I want:

Jean Pierre bought "blue-green-red" product-2345 and other blue red stuff.

https://regex101.com/r/SJzAaP/1

EDIT:

I am using Clojure (Java)

EDIT 2:

yellow-black-white -> yellow black white

product_a-b -> product_a-b

EDIT 3: Accepted answer translated in Clojure

(clojure.string/replace

 "Jean-Pierre bought "blue-green-red" product-2345 and other blue-red-green stuff yellow-black-white product_a-b"

 #"("[^"]*")|b([a-zA-Z]+)-(?=[a-zA-Z]+b)"

 (fn [[s1 s2 s3]] (if s2 s1 (str s3 " "))))



;;=> "Jean Pierre bought "blue-green-red" product-2345 and other blue red green stuff yellow black white product_a-b"

In Java, you may use something like

String s = "Jean-Pierre bought "blue-green-red" product-2345 and other blue-red stuff. yellow-black-white. product_a-b";

StringBuffer result = new StringBuffer();

Matcher m = Pattern.compile("("[^"]*")|\b([a-zA-Z]+)-(?=[a-zA-Z]+\b)").matcher(s);

while (m.find()) {

    if (m.group(1) != null) {

        m.appendReplacement(result, m.group(0));

    } else {

        m.appendReplacement(result, m.group(2) + " ");

    }

}

m.appendTail(result);

System.out.println(result.toString());

// => Jean Pierre bought "blue-green-red" product-2345 and other blue red stuff. yellow black white. product_a-b

See the Java demo.

The regex is

("[^"]*")|b([a-zA-Z]+)-(?=[a-zA-Z]+b)

Details

• 
  ("[^"]*") - Group 1: ", 0+ chars other than " and "
  


• 
  | - or


• 
  b - word boundary
  -([a-zA-Z]+) - Group 2: 1+ letters (may be replaced with (p{L}+) to match any letter)


• 
  - - a hyphen


• 
  (?=[a-zA-Z]+b) - a positive lookahead that, immediately to the right of the current location, requires 1+ letters and a word boundary.

If Group 1 matches (if (m.group(1) != null)) you just paste the match back into the result. If not, paste back Group 2 value and a space.

Adding clojure code here from the question, too, for better visibility:

(def s "Jean-Pierre bought "blue-green-red" product-2345 and other blue-red stuff. yellow-black-white. product_a-b"



(defn append [[g1 g2 g3]] (if g2 g1 (str g3 " ")))



(clojure.string/replace s #"("[^"]*")|b([a-zA-Z]+)-(?=[a-zA-Z]+b)" append)



;;=> "Jean Pierre bought "blue-green-red" product-2345 and other blue red stuff. yellow black white. product_a-b"

This should work if you don't need to handle too complex cases:

(?: |^)w+(-)(?![0-9])w+

This matches any instance of word(hyphen)word that has a space at the beginning or is the beginning of the line (so, the stuff in the quotes will not match because there would be a quote before it, not a space or the beginning of the line).

Let me know if this doesn't work for you. Live demo.

Try this one

(".*?")|((?<group>b([A-z*]+))-)

with substitution

${group} $1

You can test it here