JDBC throws SQLException: failUTFConversion

Question

hi,
i'm running: oracle 8.1.5 (server with an instance running utf8), classes111b.zip (jdbc - thin driver), nlscharset11.zip (to support utf8), orion (client - java application server/web server)
on the server, i've
$ echo $NLS_LANG
AMERICAN_AMERICA.UTF8
and
select * from v$nls_parameters where parameter='NLS_CHARACTERSET'
returns utf8
on the client, which is the web server, i get:
$ echo $NLS_LANG
AMERICAN_AMERICA.UTF8
when i do getString() on a JSP page, i got the SQLException:
java.sql.SQLException: Fail to convert between UTF8 and UCS2: failUTFConversion at oracle.jdbc.dbaccess.DBError.check_error(DBError.java) at oracle.jdbc.oracore.JavaConversion.CHARBytesToJavaChars(JavaConversion.java) at oracle.jdbc.oracore.JavaConversion.CharBytesToString(JavaConversion.java) at oracle.jdbc.driver.OracleStatement.getStringValue(OracleStatement.java) at oracle.jdbc.driver.OracleResultSet.getString(OracleResultSet.java) at oracle.jdbc.driver.OracleResultSet.getString(OracleResultSet.java)
i don't have any problem if the database contains only english, i.e. 1 byte utf8... however, i got that exception when the database contains chinese character
what should i do to make things work?
besides, i don't know where the web container called the code in nlscharset11.zip (although i included it in my classpath)
thx

Answer 1

hi, thx
for a)
i'm using orion (www.orionserver.com) as my client
according to their instruction, "Also remember to put the driver (zip/jar) for your database to the /lib dir in Orion." i put both the classes111b.zip and nlscharset11.zip into the /usr/local/orion/lib directory (where my ORION_HOME=/usr/local/orion) and have the CLASSPATH=/usr/java/jdk1.3/lib/tools.jar:/usr/share/java/ant.jar:/usr/share/java/jaxp.jar:/usr/share/java/parser.jar:/usr/local/orion/orion.jar:/usr/local/orion/lib/classes111b.zip :/usr/local/orion/lib/nlscharset11.zip
i also tried to add all classes in nlscharset11.zip to classes111b.zip, so i'm sure that the web server can access to the required classes. in addition, i attempted to move the modified classes111b.zip to another directory, and in that case, the web server gives the following error
# java -jar orion.jar
Error initializing server: DriverManagerDataSource driver 'oracle.jdbc.driver.OracleDriver' not found
for b)
how can i check whether the dumped bytes are correct UTF-8 bytes?
i have the following data in my utf8 database
New York
a$i
$i%;
and this is what i got from select dump(col, 16) from tab1
Typ=1 Len=8: 4e,65,77,20,59,6f,72,6b
Typ=1 Len=3: 61,a4,e9
Typ=1 Len=4: a4,e9,a5,bb
where a4,e9 and a5,bb are chinese char $i and %; respectively
is this correct? seems that the database stores a chinese char as 2 HEX codes... what does Typ means?
any suggestion (software) that i can use to enter unicode char?
thx