Getting MySQL compare Unicode Greek Extended characters correctly
posted: February 8th, 2007 · by: Sven
Lately I ran into an interesting issue with MySQL’s string comparsion that I haven’t seen before.
I’ve been setting up a simple vocabulary and grammar learning program for my spouse who’s started learning ancient greek a while ago. After she’s entered some testdata containing several funny looking ancient greek characters we saw that MySQL 4.1 seems to treat the following characters as equal when compared as VARCHAR:
| Char. | Unicode Codepos. |
UTF-8 | Name |
 |
U+03B7 | 206 183 | eta |
 |
U+1F75 | 225 189 181 | eta w/ oxia |
 |
U+1FC4 | 225 191 135 | eta w/ persispomeni and ypogegrammenti |
These characters are stored and retrieved correctly (which was a nice thing to watch, by the way). But when it comes to compare them to each other they are wrongly regarded the same character.
Read the rest of this entry