MBCS data in stepexecution in spring batch database - character-encoding

Question

i am using spring batch with Jdbc and Postgres DB. all the job and step execution data gets saved in spring batch created tables in PostgreS DB.
i am using it to save some step context data, which gets saved in batch_step_execution_context table in the column SERIALIZED_CONTEXT.
the data i am saving has some MBCS characters.
but i see that when writing data to the table and reading from it its using ISO-8859-1 charset.
hence my mbcs characters though serialized by Xstream default serializer gets stored as garbage.
any way to workaround this, so i can retrieve and save data as MBCS.
please find the code snippet from JDBCExecutionContextDao.
private String serializeContext(ExecutionContext ctx) {
Map<String, Object> m = new HashMap<String, Object>();
for (Entry<String, Object> me : ctx.entrySet()) {
m.put(me.getKey(), me.getValue());
}
ByteArrayOutputStream out = new ByteArrayOutputStream();
String results = "";
try {
serializer.serialize(m, out);
results = new String(out.toByteArray(), "ISO-8859-1");
}
catch (IOException ioe) {
throw new IllegalArgumentException("Could not serialize the execution context", ioe);
}
return results;
}
private class ExecutionContextRowMapper implements RowMapper<ExecutionContext> {
#Override
public ExecutionContext mapRow(ResultSet rs, int i) throws SQLException {
ExecutionContext executionContext = new ExecutionContext();
String serializedContext = rs.getString("SERIALIZED_CONTEXT");
if (serializedContext == null) {
serializedContext = rs.getString("SHORT_CONTEXT");
}
Map<String, Object> map;
try {
ByteArrayInputStream in = new ByteArrayInputStream(serializedContext.getBytes("ISO-8859-1"));
map = serializer.deserialize(in);
}
catch (IOException ioe) {
throw new IllegalArgumentException("Unable to deserialize the execution context", ioe);
}
for (Map.Entry<String, Object> entry : map.entrySet()) {
executionContext.put(entry.getKey(), entry.getValue());
}
return executionContext;
}
}
i expect to store and retrieve mbcs data.

Answer 1

hence my mbcs characters though serialized by Xstream default serializer gets stored as garbage.
The XStreamExecutionContextStringSerializer is deprecated, we recommend using the Jackson2ExecutionContextStringSerializer instead.
That said, if you want to use Multi-Byte characters for the execution context's data, you need to increase the size of the columns in the meta-data schema. This is explained in the International and Multi-byte Characters section of the reference documentation.