如何在 Java 中创建一个新的 zip 文件并向其中添加一个大目录？

Question

我正在尝试将文件目录添加到 zip。该目录大约有 150 个文件。在几个 5-75 个文件中，我不断崩溃并显示错误消息 "The process cannot access the file because it is being used by another process."

我尝试了延迟，这可能会有所帮助，但肯定不能解决错误。

使用以下代码： Is it possible to create a NEW zip file using the java FileSystem?

final File folder = new File("C:/myDir/img");
for (final File fileEntry : folder.listFiles()) {
    if (fileEntry.isDirectory()) {
        continue;
    }
    else {
        String filename = fileEntry.getName();
        String toBeAddedName = "C:/myDir/img/" + filename;
        Path toBeAdded = FileSystems.getDefault().getPath(toBeAddedName).toAbsolutePath();
        createZip(zipLocation, toBeAdded, "./" + filename);
        System.out.println("Added file " + ++count);
        //Delay because 'file in use' bug
        try { Thread.sleep(1000); } //1secs
        catch (InterruptedException e) {}
    }
}

public static void createZip(Path zipLocation, Path toBeAdded, String internalPath) throws Throwable {
    Map<String, String> env = new HashMap<String, String>();
    //Check if file exists.
    env.put("create", String.valueOf(Files.notExists(zipLocation)));
    //Use a zip filesystem URI
    URI fileUri = zipLocation.toUri();  //Here
    URI zipUri = new URI("jar:" + fileUri.getScheme(), fileUri.getPath(), null);
    System.out.println(zipUri);
    //URI uri = URI.create("jar:file:"+zipLocation);    //Here creates the zip
    //Try with resource
    try (FileSystem zipfs = FileSystems.newFileSystem(zipUri, env)) {
        //Create internal path in the zipfs
        Path internalTargetPath = zipfs.getPath(internalPath);
        //Create parent dir
        Files.createDirectories(internalTargetPath.getParent());
        //Copy a file into the zip file
        Files.copy(toBeAdded, internalTargetPath, StandardCopyOption.REPLACE_EXISTING);
    }
}

Answer 1

我不能保证这是您的问题的原因，但是您的代码以一种奇怪的或至少低效的方式将文件压缩到 ZIP 文件中。具体来说，您要为每个要压缩的文件打开一个新的FileSystem。我假设您这样做是因为您链接到的问答就是这样做的。但是，that answer 仅压缩 一个 文件，而您想同时压缩 多个 文件。在压缩目录的整个过程中，您应该保持 FileSystem 处于打开状态。

public static void compress(Path directory, int depth, Path zipArchiveFile) throws IOException {
    var uri = URI.create("jar:" + zipArchiveFile.toUri());
    var env = Map.of("create", Boolean.toString(Files.notExists(zipArchiveFile, NOFOLLOW_LINKS)));

    try (var fs = FileSystems.newFileSystem(uri, env)) {
        Files.walkFileTree(directory, Set.of(), depth, new SimpleFileVisitor<>() {

            private final Path archiveRoot = fs.getRootDirectories().iterator().next();

            @Override
            public FileVisitResult preVisitDirectory(Path dir, BasicFileAttributes attrs) throws IOException {
                // Don't include the directory itself
                if (!directory.equals(dir)) {
                    Files.createDirectory(resolveDestination(dir));
                }
                return FileVisitResult.CONTINUE;
            }

            @Override
            public FileVisitResult visitFile(Path file, BasicFileAttributes attrs) throws IOException {
                Files.copy(file, resolveDestination(file), REPLACE_EXISTING);
                return FileVisitResult.CONTINUE;
            }

            private Path resolveDestination(Path path) {
                /*
                 * Use Path#resolve(String) instead of Path#resolve(Path). I couldn't find where the
                 * documentation mentions this, but at least three implementations will throw a 
                 * ProviderMismatchException if #resolve(Path) is invoked with a Path argument that 
                 * belongs to a different provider (i.e. if the implementation types don't match).
                 *
                 * Note: Those three implementations, at least in OpenJDK 12.0.1, are the JRT, ZIP/JAR,
                 * and Windows file system providers (I don't have access to Linux's or Mac's provider
                 * source currently).
                 */
                return archiveRoot.resolve(directory.relativize(path).toString());
            }

        });
    }
}

注意：depth 参数的使用方式与maxDepth 在Files#walkFileTree 中的使用方式完全相同。

注意：如果您只关心目录本身中的文件（即不想递归遍历文件树），那么您可以使用Files#list(Path)。完成后不要忘记关闭Stream。

您可能一遍又一遍地打开和关闭FileSystem 导致您的问题，在这种情况下，上述应该可以解决问题。