Detecting character encoding from BLOB stream... (PLSQL)

Question

I'am looking for a procedure/function which can return me the character encoding of a "text/xml/csv/slk" file stored in BLOB..
For example...
I have 4 files in different encodings (UTF8, Utf8BOM, ISO8859_2, Windows1252)...
With java I'can simply detect the character encoding with JuniversalCharDet (http://code.google.com/p/juniversalchardet/)...
thank you

Answer 1

Solved...
On my local PC I have installed Java 1.5.0_00 (because on DB is 1.5.0_10)...
With Jdeveloper I have recompiled source code from:
http://juniversalchardet.googlecode.com/svn/trunk/src/org/mozilla/universalchardet
http://code.google.com/p/juniversalchardet/
After that I have made a JAR file and uploaded it with loadjava to my database...
C:\>loadjava -grant r_inis_prod -force -schema insurance2 -verbose -thin -user username/password@ip:port:sid chardet.jarAfter that I have done a java procedure and PLSQL wrapper example below:
   public static String verifyEncoding(BLOB p_blob) {
       if (p_blob == null) return "-1";
       try
        InputStream is = new BufferedInputStream(p_blob.getBinaryStream());
        UniversalDetector detector = new UniversalDetector(null);
        byte[] buf = new byte[p_blob.getChunkSize()];
        int nread;
        while ((nread = is.read(buf)) > 0 && !detector.isDone()) {
            detector.handleData(buf, 0, nread);
        detector.dataEnd();
        is.close();
       return detector.getDetectedCharset();
       catch(Exception ex) {
           return "-2";
   }as you can see I used -2 for exception and -1 if input blob is null.
then i have made a PLSQL procedure:
function f_preveri_encoding(p_blob in blob) return varchar2 is
language Java name 'Zip.Zip.verifyEncoding(oracle.sql.BLOB) return java.lang.String';After that I have uploaded 2 different txt files in my blob field.. (first one is encoded with UTF-8, second one with WINDOWS-1252)..
example how to call:
declare
   l_blob blob;
   l_encoding varchar2(100);
begin
select vsebina into l_blob from dok_vsebina_dokumenta_blob where id = 401587359 ;
l_encoding := zip_util.f_preveri_encoding(l_blob);
if l_encoding = 'UTF-8' then
   dbms_output.put_line('file is encoded with UTF-8');
elsif l_encoding = 'WINDOWS-1252' then
   dbms_output.put_line('file is encoded with WINDOWS-1252');
else
    dbms_output.put_line('other enc...');
end if;
end;Now I can get encoding from blob and convert it to database encoding and store datas in CLOB field..
Here you have a chardet.jar file if you need this functionality..
https://docs.google.com/open?id=0B6Z9wNTXyUEeVEk3VGh2cDRYTzg
Edited by: peterv6i.blogspot.com on Nov 29, 2012 1:34 PM
Edited by: peterv6i.blogspot.com on Nov 29, 2012 1:34 PM
Edited by: peterv6i.blogspot.com on Nov 29, 2012 1:38 PM