Hi there,
Hurrican19@stripped <mailto:Hurrican19@stripped> wrote on Thursday, January
06, 2005 7:19 PM:
> Thanks for the reply! I show the following information for my DB,
> and shows the same for both the 3.23 DB And the 4.18a DB
>
> Field ---- Type ---- Collation
> avatardata ---- mediumtext ---- latin1_swedish_ci
It'd be better to use a BLOB, I think, because this would avoid the
translation issues. Other alternatives worth pursuing:
* Add the "BINARY" flag to the field - "avatardata MEDIUMTEXT BINARY", for
example.
* Use a different collation, preferably one that's not case insensitive (the
'ci' part of the collation)
* Use "latin1_bin" on the new DB (and you might need to use
_latin1_bin"data" instead of just "data" when doing inserts)
http://dev.mysql.com/doc/mysql/en/Charset-literal.html
* Force the collation type, using the COLLATION keyword, as in
http://dev.mysql.com/doc/mysql/en/Charset-collate.html
> I pasted a data table from the bad avatar and the good avatar
> to a file differential program, there was no differential at all
> that the system found..
If you want to send me (directly, rather than via the list) one of the
records, as taken from both the bad and good versions, I could take a look
for you. The binary files differ, so I'm not too sure what data is being
used when you did the paste-and-compare.
> I'm not too sure where or what to do to change this
> information? Do you mean
> that I recompile MySQL using different ./configure commands?
Shouldn't need to recompile, as long as you have the right charactersets
available (which should be the default, I think).
cheers,
Tom
