Postgres UNACCENT for character with more than 1 diacritic

The UNACCENT function can strip diacritics off characters. However, in my case, it can only strip characters with 1 diacritic, for e.g.

• Thành
• Supermän
• äää

For characters with more than 1 diacritics, UNACCENT does nothing, for e.g.

• Hồ
• ầ
• phố

Is there a way to let Postgres strip the accents from these characters?

Thanks

PostgreSQL's unaccent module does not use Unicode normalization, but only a simple search-and-replace dictionary. The default dictionary, unaccent.rules, does not contain these Vietnamese characters, thus nothing is done.

You could create your own unaccent dictionary though. As explained in the documentation:

1. Create a text file vietnamese.rules with content like

   ầ  a
   Ầ  A
   ồ  o
   Ồ  O
   

2. Move vietnamese.rules into the folder $SHAREDIR/tsearch_data/ (usually /usr/share/postgresql/tsearch_data)

3. Run the function as

   SELECT unaccent('vietnamese', 'Hồ ầ phố');
   --              ^~~~~~~~~~~~~