Parse OLE Object - C#

[narinder2]

I have an Access database (images.mdb) that has 2 columns: one is the id of
the picture (an integer) and one (column named picture) is a field of type
OLE Object which contains an image stored as on OLE Object (it can store
jpg, bmp, gif, but I don't
know what image is stored inside).
I want to retrieve the picture stored in the database and identified by a
given id and display it in a web page (.aspx).
I write in Visual C# , but it does not matter, VB answers are just as
wellcome.
My problem is that this OLE Object field does not contain just the raw array
of bytes that form the image. So I can not just read the array of bytes and
output it to the browser.
no, the OLE Object contains some extra information about the type of the
file stored
(which would be good to know so I can know what kind of image it is). but I
don't know how to get this information.
I also don't know how to separate this information from the actual image.
Does anyone know how to solve this ?

[kenobewan]

I notice that you copied a post from 2005:

using System;
using System.Collections.Generic;
using System.ComponentModel;
using System.Data;
using System.Data.OleDb;
using System.Drawing;
using System.Drawing.Imaging;
using System.Text;
using System.IO;
using System.Windows.Forms;
namespace OleImages
{
    public partial class Form1 : Form
    {
        public Form1()
        {
            InitializeComponent();
        }
        private void Form1_Load(object sender, EventArgs e)
        {
            String strConn = @"Provider = Microsoft.Jet.OLEDB.4.0;Data Source = C:\Nwind.mdb;";
            OleDbConnection conn = new OleDbConnection(strConn);
            Byte[] byPicture;
            String strCmd = "Select Picture From Categories where CategoryID=1";
            OleDbCommand cmd = new OleDbCommand(strCmd, conn);
            try
            {
                conn.Open();
                byPicture = (Byte[]) cmd.ExecuteScalar();
                conn.Close();
                MemoryStream ms = new MemoryStream();
                Bitmap bm;
                ms.Write(byPicture, 78, byPicture.Length - 78);
                bm = new Bitmap(ms);
                pictureBox1.Image = bm;
                String strPath = Environment.GetFolderPath(Environment.SpecialFolder.Desktop) + "\\test.jpg";
                bm.Save(strPath, ImageFormat.Jpeg);
            }
            catch
            {
 
            }
        }
    }
}

[narinder2]

Hi,

Thanks for the reply !!!

Yes i copied that post because i was not able to find any solution on that post and currentlyi am also facing the same issue.

I have tried the solution provided by you but i get "Parameter is not valid error" while initializing the Bitmap. Moreover, the Ole Object field in my case can contain any file: word, excel, jpg, bmp, gif. So, really not sure whether the offset will be 78 in every case.

i am not getting any way forward... will appreciate any help on this.

[stucky]

I recently had to fight this battle, and found no answers online. I normally don't write code at the bit level, so excuse any novice mistakes, but lacking a library to deal with OLE Files from Access I had to parse it as well as I could.

In C# I retrieve my document from Access (or SQL Server if the Access Database has been upsized) as a byte array. Access encapsulates files with its own header, which isn't an OLE file structure. It's something different...

byte[] doc = ld.GetSupportingDocument(docID);
 
MemoryStream ms = new MemoryStream();
ms.Write(doc, 0, doc.Length);
int firstByte;
int secondByte;
ms.Seek(0, SeekOrigin.Begin);
firstByte = ms.ReadByte();
secondByte = ms.ReadByte();
 
if (firstByte != 0x15 && secondByte != 0x1C) {
    ErrorResponse("Stored object is not an Access File.");
    return;
}
 

The first two bytes are a signature, if they don't equal to 0x15 and 0x1C, it's not an access ole file. The next short is the end of the file type:

int fileTypeLoc = 20; // begin of the file type
short offset; // end of the file type
 
byte[] buffer = new byte[2];
ms.Read(buffer, 0, 2);
offset = BitConverter.ToInt16(buffer, 0);
 

Keeping track of how far I've read into the file, I store a portion of the bytes as a string, starting from 0x14 (the 20th byte) up to the offset I retrieved in the previous block.

long seekTotal = 0; 
seekTotal += offset;
 
string docType = String.Empty;
for (int i = fileTypeLoc; i < offset; i++) {
    docType += (char)doc[i];
}
 

The next bit is how I'm figuring out what type of file it is (so that when I serve it via HTTP, I can set the file name and content type properly. There's no real parsing going on in this block, with the exception of the Package type. A package can be anything, a zip file, a gif, a pdf, whatever. When you have a package, the original file name is stored in the access header, so I read in 256 bytes (an arbitrary number I selected based on trial and error), and pluck the original file extension from it. Because of my database I have no worries that it's anything but a pdf, but if you can't guaruntee that, you need to do a better job of parsing than I am.

bool packageIsPdf = false;
string ext = "dat";
string filename = "supporting-document";
string contentType = "application/octet-stream";
if (docType.Contains("Word.Document.8")) {
    ext = "doc";
    contentType = "application/ms-word";
} else if (docType.Contains("AcroExch.Document.7")) {
    contentType = "application/pdf";
    ext = "pdf";
} else if (docType.Contains("Package")) {
    // packages are generic and require more processing
    string packageBuffer = String.Empty;
    for (int i = 20; i < 256; i++) {
        packageBuffer += (char)doc[i];
    }
    if (packageBuffer.Contains(".pdf")) {
        contentType = "application/pdf";
        ext = "pdf";
        packageIsPdf = true;
    } else if (packageBuffer.Contains(".zip")) {
        contentType = "application/zip";
        ext = "zip";
    } else {
        ext = "dat";
    }
} else if (docType.Contains("Excel.Sheet.8")) {
    ext = "xls";
    contentType = "application/ms-excel";
} else if (docType.Contains("PowerPoint.Show.8")) {
    ext = "ppt";
    contentType = "application/ms-powerpoint";
} else if (docType.Contains("Word.Document.12")) {
    ext = "docx";
    contentType = "application/ms-word";
} else if (docType.Contains("PowerPoint.Show.12")) {
    ext = "pptx";
    contentType = "application/ms-powerpoint";
} else if (docType.Contains("Excel.Sheet.12")) {
    ext = "xlsx";
    contentType = "application/ms-excel";
}
 

Read 8 more bytes. These bytes should always be 01 05 00 00 02 00 00 00.

// magic eight bytes 01 05 00 00 02 00 00 00
ms.Seek(seekTotal, SeekOrigin.Begin);
buffer = new byte[8];
ms.Read(buffer, 0, 8);
seekTotal += 8;
 

Read the next long. Move to that location.

// Second offset to move to 
buffer = new byte[4];
ms.Read(buffer, 0, 4);
seekTotal += 4;
long offset2 = BitConverter.ToInt32(buffer, 0);
seekTotal += offset2;
ms.Seek(seekTotal, SeekOrigin.Begin);
 

Read 8 empty bytes.

// eight empty bytes
buffer = new byte[8];
ms.Read(buffer, 0, 8);
seekTotal += 8;
 

The next long will tell you how many bytes your encapsulated file is

// next n bytes are the length of the file
buffer = new byte[4];
ms.Read(buffer, 0, 4);
seekTotal += 4;
long fileByteLength = BitConverter.ToInt32(buffer, 0);
 

The next N bytes consist of your file. Create a new buffer of this length and read from your memory stream into the buffer.

// next N bytes are the file
byte[] data = new byte[fileByteLength];
 
// store file bytes in data buffer
ms.Read(data, 0, Convert.ToInt32(fileByteLength));
 

If your file is a PDF, you have another headache to deal with, OLE2 Compound Files. I deal with extracting the pdf from the OLE2 file in another method using the Gembox Compound File 1.1 library.

    if (contentType == "application/pdf" && !packageIsPdf) {
        data = GetPdfFromOle(data, Convert.ToInt32(seekTotal), Convert.ToInt32(fileByteLength));
    }
 

If everything went well, I can return the byte array to my Response object

    if (data == null) {
        Response.Write("Unable to retrieve file");
        Response.End();
    }
 
    string contentDisposition = String.Format("attachment; filename={0}.{1}", filename, ext);
    Response.ContentType = contentType;
    Response.AppendHeader("Content-Disposition", contentDisposition);
    Response.BinaryWrite(data);
    Response.End();
 

A code sample of extracting the pdf from an OLE2 Compound file using Gembox's library

private byte[] GetPdfFromOle(byte[] data, int offset, int length) {
    string tmpFileName = Path.GetTempFileName();
 
    FileStream fstmp = new FileStream(tmpFileName, FileMode.Create, FileAccess.Write, FileShare.None);
    fstmp.Write(data, 0, data.Length);
    fstmp.Close();
 
    byte[] pdfData = null;
    Ole2CompoundFile ole2file = new Ole2CompoundFile();
 
    try {
 
        ole2file.Load(tmpFileName, false);
        foreach (Ole2Stream entry in ole2file.Root) {
            log.Debug(entry.Name);
 
            if (entry.Name.ToLower() == "contents") {
                pdfData = entry.GetData();
                break;
            }
        }
    }catch (Exception ex) {
        ErrorResponse(ex.Message);
    } finally {
        File.Delete(tmpFileName);
        ole2file.Close();
    }
    return pdfData;
}
 

Hope this helps.