Which collation to use so that `ş` and `s` are treated as unique values?

Question

The issue is that ş and s are interpreted by MySQL as identical values.

I'm new to MySQL, so I have no idea which collations would view them as unique.

The collations that I've tried using which don't work are:

1. utf8_general_ci
2. utf8_unicode_520_ci
3. utf8mb4_unicode_ci
4. utf8mb4_unicode_520_ci

Does anybody know which collation to use?

P.S. I also really need the collation to interpret emojis and other non-Latin characters, and, to my knowledge of MySQL and collations, the only collation able to do this is unicode?

Outdate

those two characters are considered equivalent, at least by Unicode's collation standard. unicode.org/reports/tr15 – Marathi

@Marathi i know that because that's the very issue i'm facing right now lol. thanks nonetheless. – Outdate

I'm saying there wouldn't be a unicode conforming collation that would treat those characters as unique. – Marathi

@Marathi ahh i see what youre saying now. my buddy just actually said unicode_bin will work since it doesn't strip accents – Outdate

That makes sense, since unicode_bin only treats characters as code points. But note that that isn't what collation actually is. That the mostly common mistake of how collation works. unicode.org/faq/collation.html – Marathi

Answer 1

utf8_turkish_ci and utf8_romanian_ci -- as shown in http://mysql.rjweb.org/utf8_collations.html

(Plus, of course, utf8_bin.)

For your added question: You are looking for a "character set" (not a "collation") that can represent Emoji and other non-Latin characters -- UTF-8 is the one to use. In MySQL, it is utf8mb4. The "collations" that are associated with that are named utf8mb4_.... Collations control ordering and equality, as indicated in the first part of your question about s and ş.

MySQL's CHARACTER SET utf8 is a subset of utf8mb4. Either can handle all the "letters" in the world. But only utf8mb4 can handle Emoji and some Chinese characters.

Oliviero