Jython Data Extraction problem

[Steve Williams]

The data is an 8-byte 2s complement binary integer stored in a MSSQL 
2005 CHAR column.  (COBOL did that, not me). I'm using zxJDBC to read 
the data and Jython to process.
I could extract the integer if it wasn't returned in the resultset as 
unicode.  Things like ord(char) and struct.unpack('>B',char) to get at 
the bits don't seem to work.  struct.unpack('>g',string) is not available.
I've tried fooling with charset in the url and the connect statement and 
tried encode, but I have no experience here.
Is there anyway to get Jython/zxJDBC to stop converting the string to 
unicode?
Any advice is welcome.