Re: unicode match normal forms

Hi Gianni,

many thanks for your detailed response.
It turned out that my postgresql installation is too old for normalize,so I will probably
a) use an external script to normalize existing data
b) change application code to normalize data before inserting or searching

Regards
Wolfgang
>> On 17 May 2021 13:27:40 -0000
>> hamann(dot)w(at)t-online(dot)de wrote:
>> > in unicode letter ä exists in two versions - linux and windows use a
>> > composite whereas macos prefers the decomposed form. Is there any way
>> > to make a semi-exact match that accepts both variants?
>> >> You should probably normalise the strings in whatever application code
>> handles the inserting. NFC is the "usually sensible" normal form to
>> use.
>> >> If you can't change the application code, you may use a trigger and
>> apply the `normalize(text[,form])→text` function to the values
>> >> https://www.postgresql.org/docs/13/functions-string.html#id-1.5.8.10.5.2.2.7.1.1.2
>> >> something vaguely like (totally untested!)::
>> >> create function normalize_filename() returns trigger as $$
>> begin
>> new.filename := normalize(new.filename);
>> return new;
>> end;
>> $$ language plpgsql;
>> >> create trigger normalize_filename
>> before insert or update
>> on that_table
>> for each row
>> execute function normalize_filename();
>> >> -- >> Dakkar - <Mobilis in mobile>
>> GPG public key fingerprint = A071 E618 DD2C 5901 9574
>> 6FE2 40EA 9883 7519 3F88
>> key id = 0x75193F88
>> >> >>