Commented by: @pavel-zotov

It seems that when we define PK/UK-field based on ISO8859-1 charset and assign to this field collation CI_AI than we have NO ability to insert into this field values described in the starting post.

For example, this script (encoded in ISO8859-1) will fail:

set names iso8859_1;
create database '. . .' default character set iso8859_1;
create collation coll_fr for iso8859_1 from external ('FR_FR') case insensitive accent insensitive;
recreate table test_es(s varchar(10) character set iso8859_1 constraint test_es_pk primary key using index test_es_pk collate coll_es);
insert into test_es values('ñ');
insert into test_es values('.ñ');
insert into test_es values('..ñ');
insert into test_es values('ñ.');
insert into test_es values('ñ..');

But it works OK if encoding is UTF8.

Please see attached .zip -- there are two pair of files:
1) `c1178-iso.sql` and `c1178-iso.log` -- script encoded in ISO8859-1 and its log;
2) `c1178-utf`.sql `c1178-utf.log` -- similar, encoded in UTF8.

Is it correct behaviour that tables in ISO script will contain only those values:

select * from test_es;
S

ñ
n --- BTW! Why this was allowed to be inserted if collation is CI_AI and we can see spanish letter 'n' with tilda above this ??

select * from test_fr;
S

à

select * from test_pt;
S

ê