postgresql – Clean up bad data with REGEXP_REPLACE too aggressive with non-Latin characters

Try using REGEXP_REPLACE to clean a text field in a table.

Here is my request:

select the original title, REGEXP_REPLACE (title, & # 39;[^a-zA-Z0-9]+ & # 39;, & # 39; & # 39;) own
of tmp_table2
O txn_title as'%?% & # 39;

Here are my lines:

cleaning of origin
We? ? Lionel Casson and Lionel Casson
  ? ?? ? ? . ?
Ph phobias the child: dead end or passage? Phobias in the child stalemate or passage
- Snälla from! Kan of the sätta? : om vägledni i förskolan Sn lla du Kan of the s tta dig om v gledning i f rskolan

The first three lines above are suitable for my needs, but the 4th is not, I want to keep non-Latin characters (words with umlauts, etc.)

How can I keep these types of (non-Latin) characters by using a regexp in the replacement function above, while retaining the rest of the results, as shown in the "clean" column?

Thank you