我知道这有点晚了，但是，我创建了一个解决方案，适用于任何java版本。

    public static String getRealtivePath(File root, File file) 
    {
        String path = file.getPath();
        String rootPath = root.getPath();
        boolean plus1 = path.contains(File.separator);
        return path.substring(path.indexOf(rootPath) + rootPath.length() + (plus1 ? 1 : 0));
    }

我假设你有fromPath(一个文件夹的绝对路径)，和toPath(一个文件夹/文件的绝对路径)，你正在寻找一个路径，代表文件/文件夹在toPath作为一个相对路径从fromPath(你当前的工作目录是fromPath)，然后像这样的工作:

public static String getRelativePath(String fromPath, String toPath) {

  // This weirdness is because a separator of '/' messes with String.split()
  String regexCharacter = File.separator;
  if (File.separatorChar == '\\') {
    regexCharacter = "\\\\";
  }

  String[] fromSplit = fromPath.split(regexCharacter);
  String[] toSplit = toPath.split(regexCharacter);

  // Find the common path
  int common = 0;
  while (fromSplit[common].equals(toSplit[common])) {
    common++;
  }

  StringBuffer result = new StringBuffer(".");

  // Work your way up the FROM path to common ground
  for (int i = common; i < fromSplit.length; i++) {
    result.append(File.separatorChar).append("..");
  }

  // Work your way down the TO path
  for (int i = common; i < toSplit.length; i++) {
    result.append(File.separatorChar).append(toSplit[i]);
  }

  return result.toString();
}

private String relative(String left, String right){
    String[] lefts = left.split("/");
    String[] rights = right.split("/");
    int min = Math.min(lefts.length, rights.length);
    int commonIdx = -1;
    for(int i = 0; i < min; i++){
        if(commonIdx < 0 && !lefts[i].equals(rights[i])){
            commonIdx = i - 1;
            break;
        }
    }
    if(commonIdx < 0){
        return null;
    }
    StringBuilder sb = new StringBuilder(Math.max(left.length(), right.length()));
    sb.append(left).append("/");
    for(int i = commonIdx + 1; i < lefts.length;i++){
        sb.append("../");
    }
    for(int i = commonIdx + 1; i < rights.length;i++){
        sb.append(rights[i]).append("/");
    }

    return sb.deleteCharAt(sb.length() -1).toString();
}

当使用java.net.URI.relativize时，你应该注意Java错误: JDK-6226081 (URI应该能够将路径与部分根相对化)

目前，URI的relativize()方法只在一个URI是另一个URI的前缀时才会相对化URI。

这本质上意味着java.net.URI.relativize将不会创建“…这是给你的。

这里有一个其他库免费的解决方案:

Path sourceFile = Paths.get("some/common/path/example/a/b/c/f1.txt");
Path targetFile = Paths.get("some/common/path/example/d/e/f2.txt"); 
Path relativePath = sourceFile.relativize(targetFile);
System.out.println(relativePath);

输出

..\..\..\..\d\e\f2.txt

[编辑]实际上它输出更多。因为源文件不是目录。对于我的情况，正确的解决方案是:

Path sourceFile = Paths.get(new File("some/common/path/example/a/b/c/f1.txt").parent());
Path targetFile = Paths.get("some/common/path/example/d/e/f2.txt"); 
Path relativePath = sourceFile.relativize(targetFile);
System.out.println(relativePath);

递归产生一个较小的解决方案。如果结果不可能(例如不同的Windows磁盘)或不切实际(根目录只是普通目录)，则抛出异常。

/**
 * Computes the path for a file relative to a given base, or fails if the only shared 
 * directory is the root and the absolute form is better.
 * 
 * @param base File that is the base for the result
 * @param name File to be "relativized"
 * @return the relative name
 * @throws IOException if files have no common sub-directories, i.e. at best share the
 *                     root prefix "/" or "C:\"
 */

public static String getRelativePath(File base, File name) throws IOException  {
    File parent = base.getParentFile();

    if (parent == null) {
        throw new IOException("No common directory");
    }

    String bpath = base.getCanonicalPath();
    String fpath = name.getCanonicalPath();

    if (fpath.startsWith(bpath)) {
        return fpath.substring(bpath.length() + 1);
    } else {
        return (".." + File.separator + getRelativePath(parent, name));
    }
}