Detecting mime-type of a MySQL BLOB in Java

Question

Yes I know that we shouldn't store images on a database, thanks.

That said, is there a way to detect in java the mime type of a BLOB stored in mysql?

It is only for images (.gif, .png, .jpeg, etc.) I don't need a common purpose tool.

Thanks a lot guys

Bonus points if the proposed solution does not involve 3rd party libs :)

Answer 1

I imagine that you can take a look at the headers. You need to read the data into a byte array and then examine the bytes to see if they match up to the headers for the different file-types.

For example, for a GIF, the first three bytes are "GIF" (47₁₆ 49₁₆ 46₁₆) followed by either "87a" (38₁₆ 37₁₆ 61₁₆) or "89a" (38₁₆ 39₁₆ 61₁₆).

So something like this should work (uses a FileInputStream for demonstration purposes, but you can get the binary data from a BLOB by using ResultSet#getBinaryStream(int) or ResultSet#getBinaryStream(String)):

import java.io.FileInputStream;
import java.io.FileOutputStream;
import java.io.IOException;
import java.util.Arrays;

public class IdentifyImage {
    public static void main(String[] args) throws IOException {
        FileInputStream in = null;

        try {
            in = new FileInputStream("sample.gif");

            //The following are in base 10
            byte[] gifHeader87a = {71, 73, 70, 56, 55, 97};
            byte[] gifHeader89a = {71, 73, 70, 56, 57, 97};

            byte[] bytes = new byte[6];
            in.read(bytes, 0, 6);

            if(Arrays.equals(gifHeader89a, bytes) || Arrays.equals(gifHeader87a, bytes)) {
               System.out.println("It's a GIF!");
            }

        } finally {
            if (in != null) {
                in.close();
            }
        }
    }
}

You just need to look up the bytes in the header for other file types (like JPG, PNG, etc.). Not sure if this is the best way. Someone may have a better solution.

As far as 3rd-party libraries, jMimeMagic is pretty good. This is probably the easiest solution :).

EDIT

I found this article here that lists header information for different file types. There is some sample code, but it's written in another language (Clarion).

Answer 2

Honestly, storing images in databases is a-ok if you remember that you can always employ caching strategies to avoid certain issues. And modern computers tend to be able to handle the load for most scenarios. Anyhow, a key thing to remember when storing the blob is to store the mimetype with it on the way in as it tends to be pretty easy to divine then.