Regex for removing part of a line if it is preceded by some word in Java

Question

There's a properties language bundle file:

label.username=Username:
label.tooltip_html=Please enter your username.</center></html>
label.password=Password:
label.tooltip_html=Please enter your password.</center></html>

How to match all lines that have both "_html" and "</center></html>" in that order and replace them with the same line except the ending "</center></html>". For example, line:

label.tooltip_html=Please enter your username.</center></html>

should become:

label.tooltip_html=Please enter your username.

Note: I would like to do this replacement using an IDE (IntelliJ IDEA, Eclipse, NetBeans...)

Answer 1

if (line.contains("_html=")) {
   line = line.replace("</center></html>", "");
}

No regExp needed here ;) (edit) as long as all lines of the property file are well formed.

Answer 2

/^(.*)<\/center><\/html>/ 

finds you the

label.tooltip_html=Please enter your username. 

part. then you can just put the string together correctly.

Answer 3

Try something like this:

Pattern p = Pattern.compile(".*(_html).*</center></html>");
Matcher m = p.matcher(input_line); // get a matcher object
String output = input_line;
if (m.matches()) {
  String output = input_line.replace("</center></html>", "");   
}

Answer 4

String s = "label.tooltip_html=Please enter your password.</center></html>";
Pattern p = Pattern.compile("(_html.*)</center></html>");
Matcher m = p.matcher(s);
System.out.println(m.replaceAll("$1"));

Answer 5

Since you clarified that this regex is to be used in the IDE, I tested this in Eclipse and it works:

FIND:
(_html.*)</center></html>

REPLACE WITH:
$1

Make sure you turn on the Regular expressions switch in the Find/Replace dialog. This will match any string that contains _html.* (where the .* greedily matches any string not containing newlines), followed by </center></html>. It uses (…) brackets to capture what was matched into group 1, and $1 in the replacement substitutes in what group 1 captured.

This effectively removes </center></html> if that string is preceded by _html in that line.

If there can be multiple </center></html> in a line, and they are all to be removed if there's a _html_ to their left, then the regex will be more complicated, but it can be done in one regex with \G continuing anchor if absolutely need be.

---

Variations

Speaking more generally, you can also match things like this:

(delete)this part only(please)

This now creates 2 capturing groups. You can match strings with this pattern and replace with $1$2, and it will effectively delete this part only, but only if it's preceded by delete and followed by please. These subpatterns can be more complicated, of course.