StringUtils replace text in between two patterns

Question

Hi I found really useful the apache operator

StringUtils.substringBetween(fileContent, "<![CDATA[", "]]>")

to extract information inside

<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
<envelope>
    <xxxx>
        <yyyy>
            <![CDATA[

                    <?xml version="1.0" encoding="UTF-8" ?>
                    <Document >
                        <eee>
                            <tt>
                                <ss>zzzzzzz</ss>
                                <aa>2021-09-09T10:39:29.850Z</aa>
                                <aaaa>
                                    <Cd>cccc</Cd>
                                </aaaa>
                                <dd>ssss</dd>
                                <ff></ff>
                            </tt>
                        </eee>
                    </Document>
                ]]>
        </yyyy>
    </xxxx>
</envelope>

But now what I'm looking is another operator or regex that allow me to replace a dynamic xml

![CDATA["old_xml"]]

by another xml

![CDATA["new_xml"]]

Any idea idea how to accomplish this?

Regards.

This works great..... until you have an XML with two CDATA sections, one after the other. As has been discussed at [***great and passionate length***](https://stackoverflow.com/a/1732454/18157) on this site over the past decade, regex is categorically the WRONG tool for working with arbitrary XML, HTML, JSON, etc. You need a real parser for whatever flavor you're dealing with. _"Some people, when confronted with a problem, think "I know, I'll use regular expressions." Now they have two problems." -- [Jamie Zawinski](https://en.wikipedia.org/wiki/Jamie_Zawinski)_ — Jim Garrison, Oct 30 '21 at 00:09
100% agree, but I only extracting the "text" from CDATA, once that I extract the text I use DOM parser — paul, Oct 30 '21 at 09:22
If you insist on using regex, be prepared for it to break when you least expect it. Also prepare to be cursed by whoever has to maintain it. — Jim Garrison, Oct 31 '21 at 00:25

anubhava · Accepted Answer · 2021-10-31T03:33:54.393

1

Instead of StringUtils, you can use String#replaceAll method:

fileContent = fileContent
  .replaceAll("(?s)(<!\\[CDATA\\[).+?(]]>)", "$1foo$2");

Explanation:

(?s): Enable DOTALL mode so that . can match line breaks as well in .+?
(<!\\[CDATA\\[): Match opening <![CDATA[ substring and capture in group #1
.+?: Match 0 or more of any characters including line break
(]]>): Match closing ]]? substring and capture in group #2
$1foo$2: Replace with foo surrounded with back-references of capture group 1 and 2 on both sides

edited Oct 31 '21 at 03:33

answered Oct 29 '21 at 22:26

anubhava

761,203
64
569
643

@paul: Answer is updated. Had you provided this use-case data upfront we would have got to this answer in first go itself. – anubhava Oct 30 '21 at 04:59
`.replaceAll("(?s)(<!\\[CDATA\\[).+?(]]>)", "$1foo$2")` -- what is `(?s)`? Did you mean `(?:\\s)`? – Jim Garrison Oct 31 '21 at 00:21
@JimGarrison: `(?s)` is for enabling `DOTALL` mode so that `.` can match line breaks as well in `.+?` – anubhava Oct 31 '21 at 03:33

Arvind Kumar Avinash · Answer 2 · 2021-10-29T23:17:46.447

You can use the regex, (\<!\[CDATA\[).*?(\]\]>).

Demo:

public class Main {
    public static void main(String[] args) {
        String xml = """
                ...
                    <data><![CDATA[a < b]]></data>
                ...
                """;

        String replacement = "foo";

        xml = xml.replaceAll("(\\<!\\[CDATA\\[).*?(\\]\\]>)", "$1" + replacement + "$2");

        System.out.println(xml);
    }
}

Output:

...
    <data><![CDATA[foo]]></data>
...

Explanation of the regex:

( : Start of group#1
- \<!\[CDATA\[ : String <![CDATA[
) : End of group#1
.*? : Any character any number of times
( : Start of group#2
- \]\]>: String ]]>
) : End of group#2

StringUtils replace text in between two patterns

2 Answers2