> We are thinking (and thinking and thinking and thinking...) about
> converting our database to UTF8. The thing is that we have some
> columns that we would like to keep as ANSI columns, the reasons for that are:

It is no problem to mix various types of character sets in one
database.

For the "blob" columns you should probably use NONE as the character
set so these won't get altered by a transliteration or string
conversion.

You should do some tests though to see what will happen when you use
UTF8 as the client connection character set.


Regards

Stefan