Parsing a UTF-8 encoded XML Blob object

Question

Hi,
I am having a really strange problem, I am fetching a database BLOB object containing the XMLs and then parsing the XMLs. The XMLs are having some UTF-8 Encoded characters and when I am reading the XML from the BLOB, these characters lose their encoding, I had tried doing several things, but there is no means I am able to retain their UTF encoding. The characters causing real problem are mainly double qoutes, inverted commas, and apostrophe. I am attaching the piece of code below and you can see certain things I had ended up doing. What else can I try, I am using JAXP parser but I dont think that changing the parser may help because, here I am storing the XML file as I get from the database and on the very first stage it gets corrupted and I have to retain the UTF encoding. I tried to get the encoding info from the xml and it tells me cp1252 encoding, where did this come into picture and I couldn't try it retaining back to UTF -8
Here in the temp.xml itself gets corrupted. I had spend some 3 days on this issue. Help needed!!!
ResultSet rs = null;
    Statement stmt = null;
    Connection connection = null;
    InputStream inputStream = null;
    long cifElementId = -1;
    //Blob xmlData = null;
    BLOB xmlData=null;
    String xmlText = null;
    RubricBean rubricBean = null;
    ArrayList arrayBean = new ArrayList();
      rs = stmt.executeQuery(strQuery);
     // Iterate till result set has data
      while (rs.next()) {
        rubricBean = new RubricBean();
        cifElementId = rs.getLong("CIF_ELEMENT_ID");
                // get xml data which is in Blob format
        xmlData = (oracle.sql.BLOB)rs.getBlob("XML");
        // Read Input stream from blob data
         inputStream =(InputStream)xmlData.getBinaryStream(); 
        // Reading the inputstream of data into an array of bytes.
        byte[] bytes = new byte[(int)xmlData.length()];
         inputStream.read(bytes);  
       // Get the String object from byte array
         xmlText = new String(bytes);
       // xmlText=new String(szTemp.getBytes("UTF-8"));
        //xmlText = convertToUTF(xmlText);
        File file = new File("C:\\temp.xml");
        file.createNewFile();
        // Write to temp file
        java.io.BufferedWriter out = new java.io.BufferedWriter(new java.io.FileWriter(file));
        out.write(xmlText);
        out.close();

Answer 1

What the code you posted is doing:
// Read Input stream from blob data
inputStream =(InputStream)xmlData.getBinaryStream();Here you have a stream containing binary octets which encode some text in UTF-8.
// Reading the inputstream of data into an
into an array of bytes.
byte[] bytes = new byte[(int)xmlData.length()];
inputStream.read(bytes);Here you are reading between zero and xmlData.length() octets into a byte array. read(bytes[]) returns the number of bytes read, which may be less than the size of the array, and you don't check it.
xmlText = new String(bytes);Here you are creating a string with the data in the byte array, using the platform's default character encoding.
Since you mention cp1252, I'm guessing your platform is windows
// xmlText=new new String(szTemp.getBytes("UTF-8"));I don't know what szTemp is, but xmlText = new String(bytes, "UTF-8"); would create a string from the UTF-8 encoded characters; but you don't need to create a string here anyway.
//xmlText = convertToUTF(xmlText);
File file = new File("C:\\temp.xml");
file.createNewFile();
// Write to temp file
java.io.BufferedWriter out = new java.io.BufferedWriter(new java.io.FileWriter(file));This creates a Writer to write to the file using the platform's default character encoding, ie cp1252.
out.write(xmlText);This writes the string to out using cp1252.
So you have created a string treating UTF-8 as cp1252, then written that string to a file as cp1252, which is to be read as UTF-8. So it gets mis-decoded twice.
As the data is already UTF-8 encoded, and you want the output, just write the binary data to the output file without trying to convert it to a string and then back again:// not tested, as I don't have your Oracle classes
final InputStream inputStream = new BufferedInputStream((InputStream)xmlData.getBinaryStream());
final int length = xmlData.length();
final int BUFFER_SIZE = 1024;                  // these two can be
final byte[] buffer = new byte[BUFFER_SIZE];   // allocated outside the method
final OutputStream out = new BufferedOutputStream(new FileOutputStream(file));
for (int count = 0; count < length; ) {
   final int bytesRead = inputStream.read(buffer, 0, Math.min(BUFFER_SIZE, (length - count));
   out.write(buffer, 0, bytesRead);
   count += bytesRead;
}Pete