Bug #41106 | utf8_lithuanian_ci has errors | ||
---|---|---|---|
Submitted: | 28 Nov 2008 16:15 | Modified: | 29 Nov 2008 10:41 |
Reporter: | Steponas K. | Email Updates: | |
Status: | Not a Bug | Impact on me: | |
Category: | MySQL Server: Charsets | Severity: | S3 (Non-critical) |
Version: | 5.0 + | OS: | Any (Different symbols are treated as equal) |
Assigned to: | CPU Architecture: | Any | |
Tags: | collation error |
[28 Nov 2008 16:15]
Steponas K.
[28 Nov 2008 20:14]
Domas Mituzas
According to official language bodies in Lithuania (Standards institute & National language comission), i has same sorting weight as y.
[28 Nov 2008 20:19]
Domas Mituzas
I should probably be more verbose. According to Lithuanian standards, these letters should be treated equally when sorting: a and ą e, ę and ė i, į and y u, ų and ū these ones are treated as separate characters: c and č, s and š, z and ž Other letters (such as q, x and w) do not interfer with lithuanian language ordering and may follow standard latin order.
[29 Nov 2008 10:41]
Steponas K.
I uderstand the problem with the sorting order. If it's a lithuanian standart then the enquiry must go to the lithuaninan institutions. But what to do with unique indexes? There are only such solutions as writing *workaraunds* that inplement unique index functionalinty without actually adding an unique index on that column. The more logic way would be to use a non-lithuanian collation. Because the comparison operators (=, LIKE) work not as I and most lithuanian programmers would expect them to work. Well but if it is "according to national rules" then I can't do anything except to accept it as it is. Even if it does not match my "natural order" algorythm. Thank you, good luck.