Tell me the regular, which should cut any characters, except for the comma (only Cyrillic and Latinnia should remain). Made outline, it seems to work, but I don’t know the comma how to leave.

$ Key_Words = MB_STRTOLOWER ($ Key_Words);
$ Key_Words = Preg_replace ("/ [^ a-za-i \ s] / iu", "", $ Key_Words);

Answer 1, Authority 100%

If you need to remain only letters (Cyrillic + Latina) and comma, the code should have a look:

preg_replace ("/ [^ a-yaya-z,] / iu", ", $ target);

UPD:

It is worth paying attention to that the regular expression above corresponds only to parts Cyrillic, namely, Russian alphabet . In general, the concept of Cyrillic is essentially wider, and includes the characters used not only in the modern Russian alphabet, but also in a number of other Slavic languages. Cyrillic Unicode characters can be seen, for example, here .

The same, concerns latice. Formally, in addition to the symbols of the English alphabet (/ [A-Z] / I ), there are many other characters, for example, ñ from the Spanish alphabet. The regular expression above, corresponds only to the symbols English alphabet, and not to all Latin symbols.

Upd2:

If you really want to leave the string characters that formally enter Latin or Cyrillic, then you can use the following code:

preg_replace ("/ [^, \ p {cyrillic} \ p {latin}] / ui", ", $ target);

This regular expression leaves very a large set of characters (for example, Cyrillic characters ). I do not think that it is worth using it at all, but for those who formally apply to the concepts of Latin and Cyrillic , it may be interesting.

Answer 2, Authority 36%

I think, if you use Unicode, then the classes of signs need to use from there:

$ Key_Words = preg_replace ("/ [^ \ p {l}, \ s] / u", "", $ Key_Words);

Here we leave \ p {l} – all letter characters, , – comma, \ s – Blind symbols (for Blank need to be replaced by Symbol of the Blank ).

Test example https://regex101.com/r/wf4gz2/1