REgex for non repeating alphabets comma seperated

Another idea with capturing and checking by use of a lookahead:

^(?:([A-Z])(?!.*?\1),?\b)+$

You can test here at regex101 if it meets your requirements.

If you don't want to match single characters, e.g. A, change the + quantifier to {2,}.

The statement of the question is incomplete in several respects. I have made the following assumptions:

1. Considering that D,D,A is incorrect I assume that a letter cannot be followed by a comma followed by the same letter.
2. The string may contain the same letter more than once as long as #1 is satisfied.
3. Considering that A,,B,C is incorrect I assume a comma cannot follow a comma.
4. Since the examples contain only capital letters I will assume that lower-case letters are not permitted (though one need only set the case-indifferent flag (i) to permit either case).

We observe that the requirements are satisfied if and only if the string begins with a capital letter and is followed by a sequence of comma-capital letter pairs, provided that no capital letter is followed by a comma followed by the same letter. We therefore can attempt to match the following regular expression.

^(?:([A-Z]),(?!\1))*[A-Z]$

Demo

The elements of the expression are as follows.

^          # match beginning of string
(?:        # begin a non-capture group
  ([A-Z])  # match a capital letter and save to capture group 1
  ,        # match a comma
  (?!\1)   # use negative lookahead to assert next character is not equal
           # to the content of capture group 1
)*         # end non-capture group and execute it zero or more times
[A-Z]      # match a capital letter
$          # match end of string

Here is a big ugly regex solution:

var inputs = ['A,B', 'D,D,D', ',B,C', 'B,C,', 'A,,B'];
for (var i=0; i < inputs.length; ++i) {
    if (/^(?!.*?([^,]+).*,\1(?:,|$))[^,]+(?:,[^,]+)*$/.test(inputs[i])) {
        console.log(inputs[i] + " => VALID");
    }
    else {
        console.log(inputs[i] + " => INVALID");
    }
}

The regex has two parts to it. It uses a negative lookahead to assert that no two CSV entries ever repeat in the input. Then, it uses a straightforward pattern to match any proper CSV delimited input. Here is an explanation:

^                               from the start of the input
    (?!.*?([^,]+).*,\1(?:,|$))  assert that no CSV element ever repeats
    [^,]+                       then match a CSV element
    (?:,[^,]+)*                 followed by comma and another element, 0 or more times
$                               end of the input

This one could suit your needs:

^(?!,)(?!.*,,)(?!.*(\b[A-Z]+\b).*\1)[A-Z,]+(?<!,)$

• ^: the start of the string
  • (?!,): should not be directly followed by a comma
  • (?!.*,,): should not be followed by two commas
  • (?!.*(\b[A-Z]+\b).*\1): should not be followed by a value found twice
• [A-Z,]+: should contain letters and commas only
• $: the end of the string
  • (?<!,): should not be directly preceded by a comma

See https://regex101.com/r/1kGVSB/1