apache/hadoop · error · PathIsNotDirectoryException

Is not a directory

Error message

Is not a directory

What it means

PathIsNotDirectoryException ('Is not a directory') thrown by Rmdir.processPath (Delete.java:208) when 'hadoop fs -rmdir' targets an existing path that is a regular file. rmdir is the POSIX-style 'remove empty directory only' command, so a file argument is rejected before listStatus/delete are attempted.

Source

Thrown at hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/fs/shell/Delete.java:208

      "[--ignore-fail-on-non-empty] <dir> ...";
    public static final String DESCRIPTION =
      "Removes the directory entry specified by each directory argument, " +
      "provided it is empty.\n"; 
    
    private boolean ignoreNonEmpty = false;
    
    @Override
    protected void processOptions(LinkedList<String> args) throws IOException {
      CommandFormat cf = new CommandFormat(
          1, Integer.MAX_VALUE, "-ignore-fail-on-non-empty");
      cf.parse(args);
      ignoreNonEmpty = cf.getOpt("-ignore-fail-on-non-empty");
    }

    @Override
    protected void processPath(PathData item) throws IOException {
      if (!item.stat.isDirectory()) {
        throw new PathIsNotDirectoryException(item.toString());
      }      
      if (item.fs.listStatus(item.path).length == 0) {
        if (!item.fs.delete(item.path, false)) {
          throw new PathIOException(item.toString());
        }
      } else if (!ignoreNonEmpty) {
        throw new PathIsNotEmptyDirectoryException(item.toString());
      }
    }
  }

  // delete files from the trash that are older
  // than the retention threshold.
  static class Expunge extends FsCommand {
    private static final String OPTION_FILESYSTEM = "fs";

    public static final String NAME = "expunge";
    public static final String USAGE =

View on GitHub (pinned to 2add963021)

Solutions

1. Use 'hadoop fs -rm /data/file.txt' for files
2. Guard in scripts: 'hadoop fs -test -d /path && hadoop fs -rmdir /path || hadoop fs -rm /path'
3. Use 'hadoop fs -rm -r' when you want both files and directories removed uniformly

Example fix

# before
hadoop fs -rmdir /job/_SUCCESS    # Is not a directory (_SUCCESS is a file)

# after
hadoop fs -rm /job/_SUCCESS

if (!fs.getFileStatus(path).isDirectory()) {
  throw new PathIsNotDirectoryException(path.toString());
}

static boolean isEmptyableDirectory(FileSystem fs, Path p) throws IOException {
  return fs.getFileStatus(p).isDirectory();
}

try {
  fs.delete(path, false);
} catch (PathIsNotDirectoryException e) {
  // it is a file: plain rm/delete is the correct command
}

When it happens

Trigger: 'hadoop fs -rmdir /data/file.txt'; scripts that assume a path is a directory (e.g. a _SUCCESS-style marker that is actually a file); parameterized cleanup where $DIR sometimes names a file.

Common situations: Marker files confused with directories; porting 'rm -d' semantics; cleanup jobs iterating over a mix of files and dirs with a single command.

Related errors

• Directory is not empty
• Is not a directory
• Illegal option {}
• Not enough arguments: expected {} but got {}
• Too many arguments: expected {} but got {}

AI-assisted analysis of apache/hadoop@2add963021 (2026-08-22). Data as JSON: /api/errors/efba0496f1a8a52e. Report an issue: GitHub.