One of our legacy databases stores binary data in a MySQL CHAR(32) Latin1 field. Converting this field to binary would be the ideal solution but unfortunately isn't easily feasible.
The problem lies in the fact that the MySQL JDBC driver converts MySQL's latin1 charset to Windows-1252 whereas our legacy system stores these bytes using IANA Latin1 (or ISO-8859-1) charset and these implementations do not have a 1:1 mapping.
Additionally, it seems to me that SnapshotReader altogether ignores charset mapping when consuming the ResultSet.
My proposal is for me to submit a PR which exposes a new configuration parameter:
charset.overrides which is a comma delimited list of MySQL_CHARSET:JAVA_CHARSET which can be used to override the default JDBC driver mapping.