Duplicate collation weights in SYSTEM collation tables

Some characters which have separate weights in a non-Unicode code page have equivalent weights in a Unicode code page.

For example, in code page 1046, both character X'A1' (arabic letter alef with madda above after lam) and character X'DC' (arabic letter alef with madda above final form) are equivalent to the Unicode character U+FE82 (arabic letter alef with madda above final form).

If you have the same set of data in a non-Unicode database with SYSTEM collation and a Unicode database with an equivalent collation, the ordering of sorted data might be different for the two databases.

When a non-Unicode database is converted to Unicode, sometimes the several non-Unicode characters convert to the same Unicode character. While these characters had different weights originally, they have the same weight in Unicode.