Improve a regex to include coupled-dashes matching

I have this regex: [1-9]([A-Z&&[^SLOIBZ]][0-9A-Z&&[^SLOIBZ]]\d){2}[A-Z&&[^SLOIBZ]]{2}\d{2}, which I got from this site and that works as described here.

The problem is that this regex does not support dashes, and I want to modify it accordingly. As per now, this regex matches "1EG4TE5MK72" but not "1EG4-TE5-MK72"

I improved the original regex to: [1-9]([A-Z&&[^SLOIBZ]][0-9A-Z&&[^SLOIBZ]]\\d(-)?){2}[A-Z&&[^SLOIBZ]]{2}\\d{2} (added (-)?) and now both strings are matched. But also "1EG4TE5-MK72" gets matched, whereas it should be invalid (not matched).

In other words: only 0 or 2 dashes are allowed.

Can someone help me with that?

You can unroll it, and use a backreference to check correspondence with previous existence of dash (or not).

[1-9][A-Z&&[^SLOIBZ]][0-9A-Z&&[^SLOIBZ]]\d(-?)[A-Z&&[^SLOIBZ]][0-9A-Z&&[^SLOIBZ]]\d[A-Z&&[^SLOIBZ]]{2}\d{2}

Regex Explanation:

• [1-9]: number in range 1-9
• [A-Z&&[^SLOIBZ]][0-9A-Z&&[^SLOIBZ]]\d: your first three characters
• (-?): optional dash capturing group
• [A-Z&&[^SLOIBZ]][0-9A-Z&&[^SLOIBZ]]\d: your middle three characters
• \1: backreference to the capturing group
• [A-Z&&[^SLOIBZ]]{2}\d{2}: your last three characters

Check the demo here.

If you're not looking for a regular-expression pattern, you can try the following method.

This follows the pattern, as specified in the PDF file provided.

Pos.    1   2   3   4   5   6   7   8   9  10  11
Type    C   A  AN   N   A  AN   N   A   A   N   N

C  – Numeric 1 thru 9
N  – Numeric 0 thru 9 
AN – Either A or N 
A  – Alphabetic Character (A...Z); Excluding (S, L, O, I, B, Z)

Additionally, the following conditions are evaluated.

- The MBI has 11 characters.
- The MBI’s 2nd, 5th, 8th, and 9th characters are always letters.
- Characters 1, 4, 7, 10, and 11 are always numbers.
- The 3rd and 6th characters are letters or numbers.
- We don’t use dashes in the MBI.

I was unsure how to handle a lowercase argument—I settled with an uppercase requirement.

The dash separator is optional, otherwise it must contain 2, and position can be arbitrary.

I did not run any extensive tests, although "1EG4-TE5-MK73" will return true, and "ABCD-EFG-HIJK" will return false.

The logic of this evaluation works in a procedural manner—there will be an early return of false if a transition is not met—ultimately leading to a return of true.

/** @param string case-sensitive; A-Z */
boolean match(String string) {
    if (string.length() < 11 || string.length() > 13) return false;
    if (string.contains("-")) {
        if (string.split("-").length != 3) return false;
        string = string.replace("-", "");
    }
    String value;
    int position = 1;
    for (char character : string.toCharArray()) {
        value = String.valueOf(character);
        switch (position) {
            case 2, 5, 8, 9 -> {
                if (!value.matches("[A-Z&&[^SLOIBZ]]")) return false;
            }
            case 1, 4, 7, 10, 11 -> {
                if (position == 1 && !value.matches("[1-9]")) return false;
                else if (!value.matches("\\d")) return false;
            }
            case 3, 6 -> {
                if (!value.matches("[A-Z\\d&&[^SLOIBZ]]")) return false;
            }
        }
        position++;
    }
    return true;
}