How to identify a zip file in java?

Question

I want to identify my archive whether it is zip or rar. But the problem I get runtime error before I can validate my file. I want to create custom notification:

public class ZipValidator {
  public void validate(Path pathToFile) throws IOException {
    try {
      ZipFile zipFile = new ZipFile(pathToFile.toFile());
      String zipname = zipFile.getName();
    } catch (InvalidZipException e) {
      throw new InvalidZipException("Not a zip file");
    }
  }
}

At the moment I have runtime error:

java.util.zip.ZipException: error in opening zip file

Binnie

Get the file name. split the file name with . and get extension. Check whether extension is rar or zip . – Pygmy

@Ravindrababu - and What is stoppning one from simply changing the extension of a non zip file to zip or rar? – Avoirdupois

The filename (also the extension) can be changed without changing the file's content. Nothing stops you from naming zip file wrong.rar. A proper test is necessary. – Cohleen

Answer 1

I'd suggest to open a plain InputStream an reading the first few bytes (magic bytes) and not to rely on the file extension as this can be easily spoofed. Also, you can omit the overhead creating and parsing the files.

For RAR the first bytes should be 52 61 72 21 1A 07.

For ZIP it should be one of:

• 50 4B 03 04
• 50 4B 05 06 (empty archive)
• 50 4B 07 08 (spanned archive).

Source: https://en.wikipedia.org/wiki/List_of_file_signatures

Another point, just looked at your code:

Why do you catch die InvalidZipException, throw it away and construct a new one? This way you lose all the information from the original exception, making it hard to debug and understand what exactly went wrong. Either don't catch it at all or, if you have to wrap it, do it right:

} catch (InvalidZipException e) {
  throw new InvalidZipException("Not a zip file", e);
}

Depoliti

Answer 2

Merging the answers of nanda & bratkartoffel.

private static boolean isArchive(File f) {
    int fileSignature = 0;
    try (RandomAccessFile raf = new RandomAccessFile(f, "r")) {
        fileSignature = raf.readInt();
    } catch (IOException e) {
        // handle if you like
    }
    return fileSignature == 0x504B0304 || fileSignature == 0x504B0506 || fileSignature == 0x504B0708;
}

Cursory

Better use InputStream as it's handling a more general input, no? – Footlights

This does not work for me. It evaluates a .docx file to true – Efren

@ThanasisNamikazee a docx file is, in fact, a zip file with a changed file extension – Cursory

@FabianBraun its a fact I did not know about. So in reality this code works as expected. – Efren

This code does not handle self-extracting (SFX) archives. – Hotpress

@SergiuszWolicki well those would be executables so that's fair enough – Sounder

Answer 3

Exception is thrown in line

ZipFile zipFile = new ZipFile(pathToFile.toFile());

That's because if a non-ZipFile is given as parameter for the ZipFileconstructor the ZipException is thrown. So you have to check before generating a new ZipFile Object if your file path points to a correct ZipFile. One solution might be to check the extension of the file path like so

 PathMatcher matcher = FileSystems.getDefault().getPathMatcher("glob:*.zip");
 boolean extensionCorrect = matcher.matches(path); 

Taught

This does not work. For example I am using a .docx file and does not throw exception – Efren

The name is just a name and .zip maybe taken as a hint to ZIP content but can be misleading. – Cohleen

Answer 4

RandomAccessFile raf = new RandomAccessFile(f, "r");

long n = raf.readInt();

raf.close();

if (n == 0x504B0304)

    System.out.println("Should be a zip file");

else

    System.out.println("Not a zip file");

You can see it in the following link. http://www.coderanch.com/t/381509/java/java/check-file-zip-file-java

Upstart

Partly correct, there can be 3 different signatures, not just one. – Depoliti

Answer 5

Apache Tika was created to extract metadata from files, but one of its side benefits is it will determine a file's media-type by either its magic bytes and/or the files extension.

public String detect(InputStream stream, String name) throws IOException

Detects the media type of the given document. The type detection is based on the content of the given document stream and the name of the document.

Tantalizing

Answer 6

I resolved it using Apache Tika (here for gzip):

final MimeTypes mimeTypes = MimeTypes.getDefaultMimeTypes(ClassLoader.getSystemClassLoader());
boolean isCompressed = mimeTypes.forName("application/gzip").matches(
        FileUtils.readFileToByteArray(FileUtils.getFile(filepath))
);

Remark: the library use the magic bytes, so it should be possible to read only some start bytes of the file instead of using FileUtils.getFile().

Deprecatory