stanfordnlp/CoreNLP · error · IllegalArgumentException

Unknown output format

Error message

Unknown output format 

What it means

In outputTree, after printing trees the code switches on the user-selected output format (e.g. root-only, scores); any value outside the known enum of formats falls into the default branch and throws IllegalArgumentException("Unknown output format " + output). The library validates the format only at output time, not at argument-parse time.

Solutions

1. Use one of the accepted -output values (e.g. c, root, or scores).
2. Check spelling/case and re-run with -help to list supported formats.
3. In wrapper scripts, validate/interpolate the output value before invoking the pipeline.
4. If a format you need is genuinely missing, extend the switch in outputTree with a new case instead of passing an ad-hoc string.

Example fix

// before
java edu.stanford.nlp.sentiment.SentimentPipeline -model model.ser.gz -output tree -file in.txt
// after
java edu.stanford.nlp.sentiment.SentimentPipeline -model model.ser.gz -output scores -file in.txt

Set<String> validOutputs = new HashSet<>(Arrays.asList("c", "root", "scores"));
if (!validOutputs.contains(outputArg)) {
  throw new IllegalArgumentException("output must be one of " + validOutputs);
}

try {
  SentimentPipeline.main(args);
} catch (IllegalArgumentException e) {
  if (e.getMessage().startsWith("Unknown output format")) {
    System.err.println("Use -output with values: c, root, scores");
  } else throw e;
}

When it happens

Trigger: Passing -output <value> to SentimentPipeline main with a value not among the recognized formats (c, root, scores), or invoking outputTree programmatically with an unrecognized Output output value.

Common situations: Typos like -output Scores or -output tree; copying a CLI invocation from docs of a different version where the flag values changed; wrapping the class in a script that interpolates an unset variable, producing an empty format string.

Understand the failure class

Background: Invalid enum value errors: "Unknown type", "Invalid scope", "must be one of" — when a string is not on the library's allowed list — this error's family across 23 libraries.

Related errors

• Unknown format
• Unknown argument
• Please only specify one of -file, -fileList or -stdin
• You probably cannot read the serialized output, so printing…
• Unknown argument

AI-assisted analysis of stanfordnlp/CoreNLP@1b7edd19c4 (2026-09-10). Data as JSON: /api/errors/fdc3fef5b2c1c45d. Report an issue: GitHub.

Appendix: source

Thrown at src/edu/stanford/nlp/sentiment/SentimentPipeline.java:180

        Tree copy = tree.deepCopy();
        setIndexLabels(copy, 0);
        out.println(copy);
        outputTreeVectors(out, tree, 0);
        break;
      }
      case ROOT: {
        out.println("  " + sentence.get(SentimentCoreAnnotations.SentimentClass.class));
        break;
      }
      case PROBABILITIES: {
        Tree copy = tree.deepCopy();
        setIndexLabels(copy, 0);
        out.println(copy);
        outputTreeScores(out, tree, 0);
        break;
      }
      default:
        throw new IllegalArgumentException("Unknown output format " + output);
      }
    }
  }

  private static final String DEFAULT_TLPP_CLASS = "edu.stanford.nlp.parser.lexparser.EnglishTreebankParserParams";

  private static void help() {
    log.info("Known command line arguments:");
    log.info("  -sentimentModel <model>: Which model to use");
    log.info("  -parserModel <model>: Which parser to use");
    log.info("  -file <filename>: Which file to process");
    log.info("  -fileList <file>,<file>,...: Comma separated list of files to process.  Output goes to file.out");
    log.info("  -stdin: Process stdin instead of a file");
    log.info("  -input <format>: Which format to input, TEXT or TREES.  Will not process stdin as trees.  If trees are not already binarized, they will be binarized with -tlppClass's headfinder, which means they must have labels in that treebank's tagset.");
    log.info("  -output <format>: Which format to output, PENNTREES, VECTORS, PROBABILITIES, or ROOT.  Multiple formats can be specified as a comma separated list.");
    log.info("  -filterUnknown: remove unknown trees from the input.  Only applies to TREES input, in which case the trees must be binarized with sentiment labels");
    log.info("  -tlppClass: a class to use for building the binarizer if using non-binarized TREES as input.  Defaults to " + DEFAULT_TLPP_CLASS);
  }

View on GitHub (pinned to 1b7edd19c4)