detecting if two files are strictly identical, what to compare?

Question

I'd like to know how to compare two files to determine if it is exactly the same one. I know how to compare filename, date of creation/modification and even hash if required.

However I don't know how to compare meta data on the file (I actually don't know how it is stored) : security configuration, compatibility settings, potential antivirus timestamp and so on.

my final goal is to deep compare two file systems on separate computers

thanks steve

[edit] in order to clarify I reformulate the title of the question

*Strictly* identical, as in stored in the same location on the same disk? — fredley, Feb 01 '11 at 14:48
@fredley, same position in the filesystem, same attributes, same ... everything — Steve B, Feb 01 '11 at 14:53

score 2 · Accepted Answer · answered Feb 01 '11 at 14:56

What constitutes a file? On modern filesystem (say NTFS) you have

file atttibutes (times, FAT attributes)
unnamed file stream
zero or more alternate data streams (ADS)
Extended Attributes
NTFS security (it's stored in ADS yet we can identify it separately)

The rest (configuration, antivirus timestamp /what's this/ etc) is stored outside of the file and is not the file.

So you need to check the above mentioned bits of the file and compare them.

Different methods exist for reading different bits of information and you need to use them all to get all them together and compare them for different files.

score 1 · Answer 2 · answered Feb 01 '11 at 14:58

1

Just work through all the getters on System.IO.File.

GetAccessControl
GetAttributes
GetCreationTime
...
ReadAllBytes

If there's anything else that your definition of "same file" depends on (like the absolute path if on different machines), then get that as well, but you haven't made clear what that is.

answered Feb 01 '11 at 14:58

OrangeDog

36,653
12
122
207

The answer from "Eugene Mayevski 'EldoS" is what I was looking for. – Steve B Feb 01 '11 at 15:00

score 0 · Answer 3 · answered Feb 01 '11 at 14:52

0

You need to MD5, SHA hash both files and compare if both have the same sum.

Check the MD5CryptoServiceProvider and SHA512CryptoServiceProvider in System.Security.Cryptography.

It's something like this:

private string ComputeHashAsText(byte[] fileBytes)
{
    using (SHA512CryptoServiceProvider cryptoService = new SHA512CryptoServiceProvider())
    {
        return Encoding.ASCII.GetString(cryptoService.ComputeHash(fileBytes));
    }
}

public bool CompareFiles(string pathA, string pathB)
{
    string hashPathA = ComputeHashAsText(File.ReadAllBytes(pathA));
    string hashPathB = ComputeHashAsText(File.ReadAllBytes(pathB));

    return hashPathA == hashPathB;
}

In an actual solution you may want to compute the hash in chunks or something like this, because maybe files to compare are too big to read all bytes to memory and hash them.

answered Feb 01 '11 at 14:52

Matías Fidemraizer

63,804
18
124
206

my question is more "what to compare" than "how to compare". But thanks for the tip – Steve B Feb 01 '11 at 14:54
I've read "how to compare" (your question says "I don't know how to compare[...]") in your question! :D And I've answered that because is the best way of ensuring two files are the same one in different physical, network locations. I know you're asking about metadata, but, I repeat, compare two files because of metadata won't be necessarily a good way of knowing both are the same (what if two different stream have same metadata?). – Matías Fidemraizer Feb 01 '11 at 14:58
I want to compare both content (easy) AND metadata. My final goal is to detect full file system differences – Steve B Feb 02 '11 at 07:46
@Steve B - I see, understandable. – Matías Fidemraizer Feb 02 '11 at 07:54

detecting if two files are strictly identical, what to compare?

3 Answers3