Retrieving file extension from a file when it doesn't exist in the filename (in C#)?

Question

I tried searching for this but my Google has failed me. I have a directory full of files that are just their GUID:

b3445ffb-55f4-4538-bc6f-13534fd549f6

I know they can be only a handful of file extensions (doc,docx,pdf,jpg), but obviously the file extension does not exist. I could write a script that simply tries to open the file using all the known file extensions, but that wouldn't be very efficient. Is there anyway to read the file and determine what the file should be?

OSX stores file type codes inside the file, and I was hoping that Windows stored similar metadata inside the file and the file extension was simply a historical artifact. Am I to be so lucky?

NB Since you might be wondering why I have a directory full of GUIDs, the database was suppose to keep track of the GUID and match it up with the filename and extension, but the table was dropped.

Answer 1

No, Windows does not provide this metadata in the file system. Because you only have a few types to worry about, it would not be hard to programmatically examine the file header and see what extension to rename the file to.

Answer 2

I would suggest opening the files in notepad and looking for some indicator of the file type.

Eg: PDF starts with %PDF etc

and look for those indicators and do some kind of process of elimination for images or rename to jpg anf try and open?

Also try and restore from a backup for more clues.

Answer 3

To find out the kind of content that is in a file you need to know the signature or "magic number" of the content you're looking for. Some extensions might not have this characteristic. You could use such a file signature table to create a class that recognizes some extensions.

Answer 4

These are the file headers that should be on every valid file for that specific type.

JPEG

0xFF 
0xD8 

PDF

0x25
0x50
0x44
0x46

DOC

0xD0 
0xCF 
0x11 
0xE0 
0xA1 
0xB1 
0x1A 
0xE1

DOCX

0x50 
0x4B 
0x03 
0x04

For fun I wrote a little app in F# that matches the files. (Gotta love pattern matching!)

module Program =

    let main () =

        let files = 
            seq {
                for path in System.IO.Directory.GetFiles(directory) do
                    use fs = System.IO.File.OpenRead(path)
                    let buffer = Array.zeroCreate 8
                    let read = fs.Read(buffer, 0, 8)
                    match buffer with
                    | [| 0xFFuy; 0xD8uy; _; _; _; _; _; _; |] -> 
                        yield (path, ".jpg")
                    | [| 0x25uy; 0x50uy; 0x44uy; 0x46uy; _; _; _; _; |] -> 
                        yield (path, ".pdf")
                    | [| 0x50uy; 0x4Buy; 0x03uy; 0x04uy; _; _; _; _; |] -> 
                        yield (path, ".docx")
                    | [| 0xD0uy; 0xCFuy; 0x11uy; 0xE0uy; 0xA1uy; 0xB1uy; 0x1Auy; 0xE1uy; |] -> 
                        yield (path, ".doc")
                    | _ -> 
                        yield (path, ".unk")
            }
            |> Seq.toArray

        System.Console.ReadKey true |> ignore

    main()

Answer 5

Each file will have a different format, so you may well be able to poke into them and figure out what the beginning of a PDF file looks like relative to a .doc. Docx is a zipped format (although I'm not sure it's compressed) so it will have common file names stored somewhere inside. JPGs are probably pretty specific.

How many files are you talking about?

Answer 6

If it is just a few file types I would try to open the file using automation. First, load it as a picture, if fail, try load it in a word object, if fail, try to load it in a excel object and so on. You shoul'nt need more than 20 lines of code to fix the most ordinary files extentions (docx,xls,pdf,jpg).

Here are an VB-exemple of what I mean. Just add reference to Microsoft.Office.Interop.Excel and Microsoft.Office.Interop.Word to the project first. Use an onlone-vb-c#-converter to get it in C# or rewrite it by yourselfe, its just an example.

Public Class Form1
    Private MyFolder As String = "C:\MyFolder\"
    Dim p As New PictureBox
    Dim w As New Microsoft.Office.Interop.Word.Application
    Dim x As New Microsoft.Office.Interop.Excel.Application

    Private Sub Form1_Load(ByVal sender As System.Object, ByVal e As System.EventArgs) Handles MyBase.Load

        For Each file In IO.Directory.GetFiles(MyFolder)
            ProcessFile(file)
        Next
    End Sub

    Sub ProcessFile(ByVal FileName As String)
        If TryJpeg(FileName) Then Exit Sub
        If TryWordDoc(FileName) Then Exit Sub
        If TryExcelDoc(FileName) Then Exit Sub    
    End Sub

    Function TryJpeg(ByVal Filename As String) As Boolean

        Try
            p.Image = Image.FromFile(MyFolder & Filename)
            'it worked, so we assume it is a picture, rename it to jpg.
            FileSystem.Rename(MyFolder & Filename, MyFolder & Filename & ".jpg")
            Return True
        Catch ex As Exception
            Return False
        End Try
    End Function

    Function TryWordDoc(ByVal Filename As String) As Boolean

        Try
            w.Documents.Open(MyFolder & Filename)
            'it worked, so we assume it is a word document, rename it to docx.
            FileSystem.Rename(MyFolder & Filename, MyFolder & Filename & ".docx")
            Return True
        Catch ex As Exception
            Return False
        End Try
    End Function

    Function TryExcelDoc(ByVal Filename As String) As Boolean
        Try
            x.Workbooks.Open(MyFolder & Filename)
            'it worked, so we assume it is a excel document, rename it to xlsx.
            FileSystem.Rename(MyFolder & Filename, MyFolder & Filename & ".xlsx")
            Return True
        Catch ex As Exception
            Return False
        End Try
    End Function

End Class

Answer 7

As has been said already, Windows relies only on the file extension to determine a file's type. You can however read the first few bytes of each file and look for a distinguishing signature for each type. From the list provided by Jordão, the ones you want are:

 .PDF  = 25 50 44 46 ("%PDF")
 .JPG  = FF D8 FF
 .DOC  = D0 CF 11 E0 A1 B1 1A E1 (same for .XLS, .PPT, etc)
 .DOCX = 50 4B 03 04 (same for .ZIP, .XLSX, .PPTX, .JAR, etc)

Note that the signature for DOC files is the same for other Office file formats before Office 2007 (Word, Excel, Powerpoint, etc). Also note that DOCX files are actually ZIP files with a different extension, so this signature is the same for ZIP files and other file formats based on ZIP (other Office 2007/2010 applications, Java JARs, etc).