Collations for migrations

In an Oracle application, the following collation parameters have been used with Unicode (4 bytes):

• case-sensitive,
• language-specific sorting,
• no padding spaces, i.e. spaces count as normally differentiating part of strings.
  The PRIMARY and UNIQUE constraints, i.e. Data Definition, are thus very selective and allow more individual different strings than the "_ci" collations of MariaDB. Such a difference in uniqueness would result in erroneously rejected data in MariaDB.

A case-sensitive, language-specific and no-padding collation like e.g. "utf8mb4_<language>_nopad_cs" has been missing in MariaDB.

The collation supporting this detailed level of uniqueness is utf8mb4_nopad_bin at the moment. However, its sorting order is not identical to language-specific collations.

Hence, it is recommended to consider developing a set of case-sensitive, language-specific and no-padding collations in MariaDB.