Problem in setBinaryStream when using SequenceInputStream

[Renz Cruz]

I need help here is the java code

        static String url = "jdbc:oracle:thin:@xxxxxx.155:1521:orcl";
        static String username = "gfs_imp";
        static String password = "passw0rd";
        Class.forName("oracle.jdbc.driver.OracleDriver");
        Connection conn = DriverManager.getConnection(url, username, password);
        conn.setAutoCommit(false);

        String sql = "INSERT INTO GFS_BLOB (FILE_ID_HASH, CONTENT) VALUES (?, ?)";
        PreparedStatement stmt = conn.prepareStatement(sql);

        stmt.setString(1, "102");
        File file1 = new File("C:\\blobtest\\Input\\file1.jpg");
        File file2 = new File("C:\\blobtest\\Input\\file2.jpg");
        FileInputStream   fis = new FileInputStream(file1);
        FileInputStream   fis2 = new FileInputStream(file2);
        SequenceInputStream inStream;
        inStream = new SequenceInputStream(fis,fis2);

        stmt.setBinaryStream(2, inStream, (int) file1.length() + file2.length());
        stmt.execute();

        conn.commit();
        fis.close();
        conn.close();

I'm getting this error

Exception in thread "main" java.lang.AbstractMethodError: oracle.jdbc.driver.T4CPreparedStatement.setBinaryStream(ILjava/io/InputStream;J)V
at com.svi.blob.Main.saveBlob(Main.java:56)
at com.svi.blob.Main.main(Main.java:36)

Thanks,

[Rob Spoor]

Hi Rey, welcome to the Ranch!

Could you please UseCodeTags next time? I've added them for you this time, and look how much nicer it looks.

As for the problem, you're calling PreparedStatement.setBinaryStream(int, InputStream, long) because the casting to int only applies to file1.length(). By adding file2.length() you're turning the thing into a long again. The problem is that that method was added in Java 6.

There are two solutions:
1) apply the cast to the result of the addition: (int)(file1.length() + file2.length())
2) update your Oracle drivers to one that supports the features added in Java 6

Personally I'd go for the second, because then the code you compile will actually work.

[Renz Cruz]

Hi Rob thank you for that quick reply it solved my problem. I have another question now I can put many files in one blob in my db, can you please suggest a way how can i retrieve the blob and split it to get the original form of the files?


Thank you for your help.

[Rob Spoor]

That's not a good way to do it frankly. If you still want to do it you must specify where one file ends and another one starts. One way to do this is use a DataOutputStream (and DataInputStream for reading back). First write the size of the file, then write the file. Then write the size of the next file, write the next file, etc.

Reading is then just as easy: first read back the size, then read exactly that many bytes. Read the size again, read bytes again, etc. You may want to add a -1 after all files so you'll know you're done reading when the size is -1.

[Renz Cruz]

Are you saying that I need to rewrite the files? Read the files and then write it to a new file using DataOutputStream, write first the size of the file, then write the file?

Can you comment on my code I'm using .skip to get the file that I need I think its simillar to your solution.

public static void main(String[] args) throws Exception{
        // TODO code application logic here

      Main aa = new Main();
     aa.getBlob();
    // aa.saveBlob();
    }

    public void saveBlob() throws Exception{

        Class.forName("oracle.jdbc.driver.OracleDriver");
        Connection conn = DriverManager.getConnection(url, username, password);
        conn.setAutoCommit(false);

        String sql = "INSERT INTO GFS_BLOB (FILE_ID_HASH, CONTENT) VALUES (?, ?)";
        PreparedStatement stmt = conn.prepareStatement(sql);

        stmt.setString(1, "105");
        File file1 = new File("C:\\blobtest\\Input\\file1.jpg");
        File file2 = new File("C:\\blobtest\\Input\\file2.jpg");
        File file3 = new File("C:\\blobtest\\Input\\file3.pdf");
        FileInputStream   fis = new FileInputStream(file1);
        FileInputStream   fis2 = new FileInputStream(file2);
        FileInputStream   fis3 = new FileInputStream(file3);

        
        ArrayList<InputStream> inputStream = new ArrayList<InputStream>();
        inputStream.add(fis);
        inputStream.add(fis2);
        inputStream.add(fis3);

        
        Enumeration<InputStream> eInput = Collections.enumeration(inputStream);
        SequenceInputStream inStream = new SequenceInputStream(eInput);



        System.out.println(fis2.available());
        stmt.setBinaryStream(2, inStream, (int) (file1.length() + file2.length() + file3.length()));
        stmt.execute();

        conn.commit();
        fis.close();
        conn.close();


    }

    public void getBlob() throws Exception{
        Class.forName("oracle.jdbc.driver.OracleDriver");
        Connection conn = DriverManager.getConnection(url, username, password);
        conn.setAutoCommit(false);
        String sql = "SELECT FILE_ID_HASH, CONTENT FROM GFS_BLOB where FILE_ID_HASH = '105' ";
        PreparedStatement stmt = conn.prepareStatement(sql);
        ResultSet resultSet = stmt.executeQuery();
        while (resultSet.next()) {
            String fileID = resultSet.getString(1);
            File image = new File("C:\\blobtest\\Output\\"+fileID+"2.pdf");
            FileOutputStream fos = new FileOutputStream(image);

            byte[] buffer = new byte[1];
            InputStream is = resultSet.getBinaryStream(2);
         //  is.skip(3682);   gets the second file by skipping the first file
             is.skip(97195); // gets the third file by skipping the first and second file
           while (is.read(buffer) > 0) {
                 fos.write(buffer);
           }
          fos.close();

        }
    conn.close();

    }
   

Thank you.

[Rob Spoor]

How do you know where each file starts? You should add a few more InputStreams, for the lengths.

private InputStream toInputStream(long size)
{
    ByteArrayOutputStream baos = new ByteArrayOutputStream(8);
    new DataInputStream(baos).writeLong(size);
    return new ByteArrayInputStream(baos.toByteArray());
}

inputStream.add(toInputStream(file1.size());
inputStream.add(fis);   
inputStream.add(toInputStream(file2.size());
inputStream.add(fis2);   
inputStream.add(toInputStream(file3.size());
inputStream.add(fis3);   
inputStream.add(toInputStream(-1); // to indicate there are no more files

The reading:

InputStream is = resultSet.getBinaryStream(2);
DataInputStream dis = new DataInputStream(is);
long size = dis.readLong();
dis.skip(size);
// skipped past the first file
size = dis.readLong();
// now copy size bytes only to the file

The -1 is added so you know when to stop when iterating through all files:

long size = dis.readLong();
while (size != -1)
{
    // copy size bytes
    size = dis.readLong();
}

[Renz Cruz]

Hi there is some problem in the code

 
private InputStream toInputStream(long size)
{
    ByteArrayOutputStream baos = new ByteArrayOutputStream(8);
    new DataInputStream(baos).writeLong(size); //error on this line
    return new ByteArrayInputStream(baos.toByteArray());
}
   

and

 
InputStream is = resultSet.getBinaryStream(2);
DataInputStream dis = new DataInputStream(is);
long size = dis.readLong();
dis.skip(size);    //  it skips the whole file
// skipped past the first file
size = dis.readLong();
// now copy size bytes only to the file
   

Please make it clear because I'm new with java.

Thanks,

[Rob Spoor]

Renz Cruz wrote:

 
    new DataInputStream(baos).writeLong(size); //error on this line

Sorry, that should be DataOutputStream of course.

 
InputStream is = resultSet.getBinaryStream(2);
DataInputStream dis = new DataInputStream(is);
long size = dis.readLong();
dis.skip(size);    //  it skips the whole file
// skipped past the first file
size = dis.readLong();
// now copy size bytes only to the file
   

The way I've written the files to the database is as follows:
1) the file size in 4 bytes; DataOutputStream converted the long to bytes
2) the file contents
3) steps 1 and 2 again
4) steps 1 and 2 again
5) -1 in 4 bytes

Because I've used DataOutputStream to convert one long into 4 bytes I used DataInputStream to do the inverse - convert the first 4 bytes back into a long. This is the size of the first file (step 1).
Because you wanted the second file we can skip the entire first file. That's what the dis.skip(size) does* (step 2). You then read the size of the second file (step 3, part 1) and the file (step 3, part 2).


* Actually this code has a potential flaw. skip is allowed to actually skip less. skip should instead be looped until all is really skipped:

while (size > 0)
{
    size -= dis.skip(size);
}

This tries to skip the remaining size each time, subtracting the actual number of skipped bytes. So if size starts at 20 and the first time only 16 bytes are skipped, size becomes 4 and skip is attempted again.

[Renz Cruz]

Hi Rob thank you for your help I really appreciate it.

[Rob Spoor]

You're welcome.