Character Encodings when using Streams for Text

Question

hi there,
i have a table with a key column (ordinary varchar) and a text column (memo field in ms access, connection is done via the jdbc2odbc bridge). if i read the stuff with ResultSet.getAsciiStream() everything is fine, i.e., printing the resulting strings looks ok. then i use this method to read it and write it into a second table with the same structure but using PreparedStatement.setCharacterStream(). now i read from the second table using ResultSet.getCharacterStream() and print the strings. but now they seem garbled. something with the text conversion must have gone wrong. i never set any character encoding so the encodings should all be default.
any ideas?
robert

Answer 1

Can you confirm that the data has been stored to the
second table correctly?
Ie, is it getCharacterStream or setCharacterStream
that's causing the problems?that's difficult to answer: if i use the access interface to look into both tables they look alike. especially the target table looks ok.
but if i read the data with getCharacterStream() i get something that looks as if this had happened: during storing with setCharacterStream() for every character there seem to be two bytes stored in the table. after reading the output looks like every second character was ok and every other character seems garbled. so it looks like on reading not two bytes in the database made a character but only one so every second byte becomes crap.
unfortunately there is neither method in Reader that allows for setting the encoding nor is there a variant of getCharacterStream() that accepts an encoding parameter. (the same holds for setCharacterStream(); plus, if i create a StringReader there is no way to set the encoding.)
do you have still any ideas?
robert