Regular Expression to find a pangram

Question

Awhile back I encountered a simple programming puzzle to determine whether a string was a pangram, or a perfect pangram. (Although it's worth noting that this particular challenge gave a different meaning than the standard definition of a "perfect pangram," to mean any sentence where every letter appears n times for some positive n). And it was a fun challenge. But recently it got me wondering whether it would be possible to solve this challenge purely with regular expressions. If it is, I expect it will be a pretty gnarly regex. But I'm not even sure it is possible.

Still, if anyone is able to do it, it would be the folks here at Code Golf. So here's the challenge:

Write a regular expression which can determine whether or not a given string is a pangram. You can assume that all non-alpha characters have already been stripped out (or don't, if you really want to challenge yourself), and that everything has been transformed into lowercase.

Bonus challenge: write another regular expression to determine whether the given string is a "perfect pangram" by the definition given above.

Examples of an ordinary pangram:

thequickbrownfoxjumpsoverthelazydog

Example of perfect pangrams:

abcdefghijklmnopqrstuvwxyz

jocknymphswaqfdrugvexblitz

squdgykilpjobzarfnthcwmvexhmfjordwaltzcinqbuskpyxveg

I'm not really sure if this is even possible just with regex alone. If you find that it is, kudos!

Answer 1

Assuming you're using a flavor of regex that include look-aheads, detecting a pangram is straightforward, but tedious. Using Python syntax:

r"(?=.*a)(?=.*b)(?=.*c)(?=.*d)(?=.*e)(?=.*f)(?=.*g)(?=.*h)(?=.*i)(?=.*j)(?=.*k)(?=.*l)(?=.*m)(?=.*n)(?=.*o)(?=.*p)(?=.*q)(?=.*r)(?=.*s)(?=.*t)(?=.*u)(?=.*v)(?=.*w)(?=.*x)(?=.*y)(?=.*z).*"

https://regex101.com/r/Z7NjEQ/2

The (?=.*a) block tests if what follows after that point matches .*a; another way to say that is, is this expression followed by any characters and then an a, or just, does this include an a? Then repeat that check for every letter of the alphabet.

Without lookaheads, it's still possible, but even more tedious, because you have to account for every possible ordering of the alphabet. If you were limited to the letters abc, here's what it might look like:

r".*(a.*b.*c|a.*c.*b|b.*a.*c|b.*c.*a|c.*a.*b|c.*b.*a).*"

https://regex101.com/r/OnXQe7/1/

This checks for an a, any set of characters, a b, any set of characters, then a c, OR an a, any set of characters, a c, any set of characters, then a b, OR...you get the idea.

You could use the same approach to check for the standard meaning of perfect pangram (every character exactly once), but not the modified version used here.

Answer 2

Regexp: 45 43 20 19 18 bytes

(?!.*(.).*\1).{26}

Assuming the input is [a-z] (lowercase, only alphabets a to z, spaces etc removed)

A byte saved, thanks to Neil