C#.net identify zip file

Viewed 23808

I am currently using the SharpZip api to handle my zip file entries. It works splendid for zipping and unzipping. Though, I am having trouble identifying if a file is a zip or not. I need to know if there is a way to detect if a file stream can be decompressed. Originally I used

FileStream lFileStreamIn = File.OpenRead(mSourceFile);
lZipFile = new ZipFile(lFileStreamIn);
ZipInputStream lZipStreamTester = new ZipInputStream(lFileStreamIn, mBufferSize);// not working
lZipStreamTester.Read(lBuffer, 0, 0);
if (lZipStreamTester.CanDecompressEntry)
{

The LZipStreamTester becomes null every time and the if statement fails. I tried it with/without a buffer. Can anybody give any insight as to why? I am aware that i can check for file extension. I need something that is more definitive than that. I am also aware that zip has a magic #(PK something), but it isn't a guarantee that it will always be there because it isn't a requirement of the format.

Also i read about .net 4.5 having native zip support so my project may migrate to that instead of sharpzip but I still need didn't see a method/param similar to CanDecompressEntry here: http://msdn.microsoft.com/en-us/library/3z72378a%28v=vs.110%29

My last resort will be to use a try catch and attempt an unzip on the file.

6 Answers

I used https://en.wikipedia.org/wiki/List_of_file_signatures, just adding an extra byte on for my zip files, to differentiate between my zip files and Word documents (these share the first four bytes).

Here is my code:

public class ZipFileUtilities
{
    private static readonly byte[] ZipBytes1 = { 0x50, 0x4b, 0x03, 0x04, 0x0a };
    private static readonly byte[] GzipBytes = { 0x1f, 0x8b };
    private static readonly byte[] TarBytes = { 0x1f, 0x9d };
    private static readonly byte[] LzhBytes = { 0x1f, 0xa0 };
    private static readonly byte[] Bzip2Bytes = { 0x42, 0x5a, 0x68 };
    private static readonly byte[] LzipBytes = { 0x4c, 0x5a, 0x49, 0x50 };
    private static readonly byte[] ZipBytes2 = { 0x50, 0x4b, 0x05, 0x06 };
    private static readonly byte[] ZipBytes3 = { 0x50, 0x4b, 0x07, 0x08 };

    public static byte[] GetFirstBytes(string filepath, int length)
    {
        using (var sr = new StreamReader(filepath))
        {
            sr.BaseStream.Seek(0, 0);
            var bytes = new byte[length];
            sr.BaseStream.Read(bytes, 0, length);

            return bytes;
        }
    }

    public static bool IsZipFile(string filepath)
    {
        return IsCompressedData(GetFirstBytes(filepath, 5));
    }

    public static bool IsCompressedData(byte[] data)
    {
        foreach (var headerBytes in new[] { ZipBytes1, ZipBytes2, ZipBytes3, GzipBytes, TarBytes, LzhBytes, Bzip2Bytes, LzipBytes })
        {
            if (HeaderBytesMatch(headerBytes, data))
                return true;
        }

        return false;
    }

    private static bool HeaderBytesMatch(byte[] headerBytes, byte[] dataBytes)
    {
        if (dataBytes.Length < headerBytes.Length)
            throw new ArgumentOutOfRangeException(nameof(dataBytes), 
                $"Passed databytes length ({dataBytes.Length}) is shorter than the headerbytes ({headerBytes.Length})");

        for (var i = 0; i < headerBytes.Length; i++)
        {
            if (headerBytes[i] == dataBytes[i]) continue;

            return false;
        }

        return true;
    }

 }

There may be better ways to code this particularly the byte compare, but as its a variable length byte compare (depending on the signature being checked), I felt at least this code is readable - to me at least.

Thanks to dkackman and Kiquenet for answers above. For completeness, the below code uses the signature to identify compressed (zip) files. You then have the added complexity that the newer MS Office file formats will also return match this signature lookup (your .docx and .xlsx files etc). As remarked upon elsewhere, these are indeed compressed archives, you can rename the files with a .zip extension and have a look at the XML inside.

Below code, first does a check for ZIP (compressed) using the signatures used above, and we then have a subsequent check for the MS Office packages. Note that to use the System.IO.Packaging.Package you need a project reference to "WindowsBase" (that is a .NET assembly reference).

    private const string SignatureZip = "50-4B-03-04";
    private const string SignatureGzip = "1F-8B-08";

    public static bool IsZip(this Stream stream)
    {
        if (stream.Position > 0)
        {
            stream.Seek(0, SeekOrigin.Begin);
        }

        bool isZip = CheckSignature(stream, 4, SignatureZip);
        bool isGzip = CheckSignature(stream, 3, SignatureGzip);

        bool isSomeKindOfZip = isZip || isGzip;

        if (isSomeKindOfZip && stream.IsPackage()) //Signature matches ZIP, but it's package format (docx etc).
        {
            return false;
        }

        return isSomeKindOfZip;
    }

    /// <summary>
    /// MS .docx, .xslx and other extensions are (correctly) identified as zip files using signature lookup.
    /// This tests if System.IO.Packaging is able to open, and if package has parts, this is not a zip file.
    /// </summary>
    /// <param name="stream"></param>
    /// <returns></returns>
    private static bool IsPackage(this Stream stream)
    {
        Package package = Package.Open(stream, FileMode.Open, FileAccess.Read);
        return package.GetParts().Any();
    }
Related