>>>>> "Piet" == Piet Wesselman <pwessel@stripped> writes:
Piet> I just upgraded to 3.23.3. In my data several languages are used, so the new
Piet> feature (described in D.1.5 Changes in release 3.23.0):
Piet> Changed LIKE character comparison to behave as =; This
Piet> means that 'e' LIKE 'é' is now true.
Piet> was one of the main reasons for the upgrade. To my surprise however this
Piet> does work on characters like é, è an ê, but not on the
> (German a.o.) Umlaute
Piet> like ä, ö and ü. Even worse, LIKE does not map them to themselves
> any more,
Piet> so a word like 'Bürger' cannot be found in any way.
Piet> Is there a list of characters that are equivalents? Can MySql be patched to
Piet> enlarge the list of matched characters? Maybe this is something to include
Piet> in future versions...
Piet> Thanks for any help.
Piet> Piet Wesselman
The problem is that different people have different ideas how accent
characters should be mapped. For example in Swedish 'e' and 'é' are
usually thought to be the same character (and sometimes even ü and y)
while 'a' and 'ä' has nothing to do with each other. The current german
character map was done by a German and I am not the right person to
Anyway, you can find the mapping at strings/ctype-german1.c and
client/ctype-german1.c. The MySQL manual includes information how you
can change this (You only have to change the map 'sort_order_german')