Java Regular Expression - Nested Object?

vvvvvvvvvvvvvvvv

You can't use repetition effectively with capturing groups.  The simplest Java style RE I can see that works here just to get the details is:

    String regex =

        "\\s*properties=\\{" +  // leading section for "properties={"

        "\\s*(\\w+=\\{[^}]*\\})," + // accept "keyword={anything},"

        "\\s*(\\w+=\\{[^}]*\\})," + //  [repeat 4 times]

        "\\s*(\\w+=\\{[^}]*\\})," +

        "\\s*(\\w+=\\{[^}]*\\})," +

        "\\s*\\}" // trailing section to accept "}"

That doesn't attempt to recognize the individual attributes for each property, and will only match exactly 4 properties.  It needs quite a bit of work to be useful.  Note that each \ escape in the regex needs to be doubled to \\ in the string literal.  The \s* accepts 0 or more spaces, tabs or newlines.  The \\w+ accepts 1 or more "word" characters (letters or digits, basically) that make up the property name.  The [^}]* pattern between \{ and \} matches 0 or more non-right-brace characters.

You can test that with something like:

    Pattern pat = Pattern(regex);

    Matcher matcher = pat.matcher(input_string);

    if (matcher.matches()) {

        MatchResult m = matcher.toMatchResult();

        for (int g=1; g<=m.groupCount(); ++g) {

            System.out.println("Group " + g + " : " + m.group(g));

        }

    }

    else System.out.println("No match.");

You'll need input_string to be a pasted copy of the string in your example.

A regex isn't very good for parsing expressions with arbitrary repetition; and is extremely NOT good at parsing expressions with arbitrary nesting.  With only two levels, though, you could take each m.group(g) group string and use a different regex to parse out the individual attribute names and values for just one property.

I think a top-down parser is about as easy to write, a whole lot easier to read and debug, and is *certainly* a lot more flexible.