I have input search use preg_replace, however I wish my search input accept others language
Keep - Chinese, Japaneses, German.. etc.
Remove - symbol character like @#$%^*() those
This one only keep english preg_replace("/[^a-zA-Z0-9]+/", "", $search);
any way to set up for multi language?
Though for java, a concise overview here.
You can use the so called Posix notation:
[^\p{Alnum}\p{M}] The first is the alphanumeric group, and the second the combining diacritical marks: the accents. The latter should not be forgotten because one can write ĉ as one Unicode point c-circumflex, but also as 'c' followed by a combining circumflex ^ (zero width, here represented by the normal circumflex). In some languages there are more than one marks to a base letter.
Correction:
[^\p{L}\p{N}\p{M}] 
Here's the PHP unicode regex reference. The plus + isn't necessary as PHP will loop through the string. The \s will match all whitespace characters.
preg_replace("![^\p{L}\p{N}\s]!", "", $search); If you want to match only the space character itself you would add it in the brackets as a literal:
preg_replace("![^\p{L}\p{N} ]!", "", $search); Update Added bit about spaces per comment request
\s.
æis as much of a unicode character as^. Your best bet it to make a blacklist, and simply replace with that. (preg_replace("/[!\"#¤%&\(\)=/]+", "", $search);).