What causes "' ' is not a directory" in pentaho/pentaho-kettle?

Print/resolve the actual dirname at runtime and verify it exists and is a directory (ls ).

pentaho/pentaho-kettle · error · KettleJobException

' ' is not a directory

Error message

'<dirname>' is not a directory

What it means

DummyJob.getDir wraps a java.io.File for the given dirname and throws KettleJobException when File.isDirectory() returns false — i.e. the path does not exist or exists but is a regular file. It is a fail-fast argument validator used by srcDir before scanning a directory of files.

Solutions

Print/resolve the actual dirname at runtime and verify it exists and is a directory (ls <dirname>).
Fix the job entry configuration to use an absolute path or correct variable expression.
If the path can legitimately be a file or missing, check new File(dirname).isDirectory() before calling and handle it.
Ensure the path is valid on the node that executes the job (agents/cluster nodes may not share the same filesystem).

Example fix

// before
File dir = job.getDir( "/data/input" ); // throws if missing
// after
File f = new File( "/data/input" );
if ( !f.isDirectory() ) {
  throw new KettleJobException( "Input directory missing: " + f.getAbsolutePath() );
}
File dir = job.getDir( f.getAbsolutePath() );

Defensive patterns

Strategy: validation

Validate before calling

// guard before calling getDir/srcDir
File f = new File( dirname );
if ( !f.exists() ) {
  throw new IllegalArgumentException( "Directory does not exist: " + f.getAbsolutePath() );
}
if ( !f.isDirectory() ) {
  throw new IllegalArgumentException( "Path is a file, not a directory: " + f.getAbsolutePath() );
}

Try / catch

try {
  File dir = job.getDir( dirname );
  job.srcDir( dirname, outDir );
} catch ( KettleJobException e ) {
  if ( e.getMessage() != null && e.getMessage().endsWith( "is not a directory" ) ) {
    logError( "Check path config: " + e.getMessage() );
    // skip or substitute a default directory
  } else {
    throw e;
  }
}

Prevention

Use absolute paths in job entry configuration, or verified variable substitutions
Validate directory existence on the executing node, not just the design machine
Guard with new File(path).isDirectory() before invoking directory-scanning methods
Watch for working-directory differences that break relative paths

When it happens

Trigger: Calling getDir (or srcDir, which calls it) with a dirname that is a file path, a typo'd/non-existent path, an empty string, or a path that exists relative to a different working directory than expected.

Common situations: Hard-coded job-entry directory settings pointing to a machine-local path that doesn't exist on the execution node; variable substitution (${Internal.Job.Filename.Directory}) resolving unexpectedly; path moved/renamed between job design and run; running the job from a different working directory so relative paths break.

Related errors

AI-assisted analysis of pentaho/pentaho-kettle@f3058517a1 (2026-09-13). Data as JSON: /api/errors/6bf59fd3b90d62b5. Report an issue: GitHub.

Appendix: source

Thrown at plugins/dummy/core/src/main/java/org/pentaho/di/pdi/jobentry/dummy/DummyJob.java:76

        return false;
      }
    };
    long files = 0;
    File[] allFiles = srcDir.listFiles( regexFiler );
    File outDir = new File( _targetDir );
    outDir.mkdirs();
    for ( int i = 0; i < allFiles.length; i++ ) {
      File cFile = allFiles[i];
      log.logDetailed( toString(), "processing file '" + cFile + "'" );
      processFile( cFile, outDir );
    }
    return files;
  }

  public File getDir( String dirname ) throws KettleJobException {
    File fl = new File( dirname );
    if ( !fl.isDirectory() ) {
      throw new KettleJobException( "'" + dirname + "' is not a directory" );
    }
    return fl;
  }

  public void processFile( File fl, File outDir ) throws FileNotFoundException {
    // do something with the file here

  }
}

View on GitHub (pinned to f3058517a1)