Binary file read/write Question

[santosh kulkarni]

I am using the following code to read in a PDF file and writing the contents out to a different file. I am expecting that I should be able to
read the second file using the Acrobat reader. I can but the contents are blank! The original file is a proper PDF with text and images.Am I missing something obvious?
Everything is on the same system(same charset etc.)

  
public static void main(String[] args) {
// TODO Auto-generated method stub
byte[] bytes=new byte[2048];
StringBuffer stringbuffer= new StringBuffer();
 try {
            RandomAccessFile file= new RandomAccessFile(new File("c://development//two.pdf"), "rw");
            while((file.read(bytes))!=-1)
            {
            stringbuffer.append(new String(bytes));
            }
            RandomAccessFile file1= new RandomAccessFile(new File("c://development//twoAgain.pdf"), "rw");
            file1.write(new String(stringbuffer).getBytes());
            file1.close();
        } catch (FileNotFoundException e1) {
            // TODO Auto-generated catch block
            e1.printStackTrace();
        } catch (IOException e) {
            // TODO Auto-generated catch block
            e.printStackTrace();
        }
}
 

[Paul Clapham]

Converting the (binary) bytes to String data and back to bytes again assumes that the bytes represent text. If they don't, then changes may occur that damage the document (as in this case).

So don't use String or StringBuffer. Just read the bytes from one file and write them to the other file.

Personally I wouldn't use RandomAccessFiles for this task. I would use a BufferedInputStream to read into a buffer of something like 2048 bytes and a BufferedOutputStream to write that buffer into. Keep doing that (in a loop) until you have copied the entire file.

[santosh kulkarni]

Thanks for the prompt reply Paul. your comments make sense.
Allow me to elaborate though.
I simplified the code before posting.
What I want to do is Base64 encode the pdf, send it over a web service and
decode and be able to view/print it on the other side.
I was able to do this when I started from a already encoded pdf as a
(java.lang.)String and got the bytes and wrote a file and poof it came up.
I'll keep trying, may be store it in arraylist or something
S

[Santhosh Kumar]

Hi,

If you implement the steps you mentioned as they should, I feel that solution would work smooth. The important thing to keep in mind is that to read the input pdf and write the output pdf in binary format (you should use streams instead of Reader or Writers). I tried it out with sample pdf (around 61kb which had some images too) and the below code worked fine in creating a duplicate pdf.

package com;
 
import java.io.*;
import sun.misc.*;
 
public class PdfCopy {
    private static int BUFFER_SIZE = 500 * 1024;
    private static String inFile = "C:/pdfin.pdf";
    private static String outFile = "C:/pdfout.pdf";
 
    public static void main(String[] args) throws IOException {
        
        //Pdf if binary file so read in binary format
        ByteArrayOutputStream bao = new ByteArrayOutputStream (BUFFER_SIZE);
        byte[] buffer = new byte[BUFFER_SIZE];
        BufferedInputStream bi = new BufferedInputStream (new FileInputStream (inFile));
        while (true) {
            int bytesRead = bi.read(buffer);
            if (bytesRead == -1) {
                break;
            } else {
                bao.write(buffer, 0, bytesRead);
            }
        }
        bi.close();
        
        //Base64 encode it
        String base64EncodedPdf = new BASE64Encoder().encodeBuffer(bao.toByteArray());
        
        //Here you can send the string over wire
        
        //Base64 decode it
        byte[] newPdfBytes = new BASE64Decoder().decodeBuffer(base64EncodedPdf);
        
        //Write new pdfbytes to the outpuf stream
        BufferedOutputStream bo = new BufferedOutputStream (new FileOutputStream (outFile));
        bo.write(newPdfBytes);
        bo.close();
    }
    
}
 

[santosh kulkarni]

Thanks Santhosh, I was trying to avoid pre-determined buffer size. I realized, in my case and most of the cases 512K(or some other simlar number) ought to be more than enough.
Santosh