stanfordnlp/CoreNLP · error · UnsupportedOperationException

Argument array lengths differ

Error message

Argument array lengths differ: <keys> vs. <values>

What it means

CoreLabel.initFromStrings throws UnsupportedOperationException when the keys and values string arrays passed to the constructor have different lengths. The label format requires a 1:1 pairing of annotation keys and values.

Solutions

1. Verify the keys and values arrays come from the same split and have equal length before constructing.
2. Normalize the input line's delimiter (e.g. split on \t consistently, trim trailing separators).
3. Pad or trim one array to match the other only if the format genuinely allows missing columns.
4. Log both Arrays.toString(keys) and Arrays.toString(values) to find which column is missing/extra.

Example fix

// before
String[] parts = line.split("\\t");
new CoreLabel(keys, parts); // lengths may differ
// after
String[] parts = line.split("\\t", -1);
if (keys.length != parts.length)
  throw new IllegalArgumentException("Expected " + keys.length + " columns, got " + parts.length + ": " + line);
new CoreLabel(keys, parts);

if (keys == null || values == null || keys.length != values.length) {
  throw new IllegalArgumentException("keys and values must be non-null and equal length");
}
new CoreLabel(keys, values);

try {
  return new CoreLabel(keys, values);
} catch (UnsupportedOperationException e) {
  if (e.getMessage().startsWith("Argument array lengths differ")) {
    throw new IllegalArgumentException("Malformed label record: " + e.getMessage(), e);
  }
  throw e;
}

When it happens

Trigger: Calling new CoreLabel(keys, values) (or the constructor variants that delegate to initFromStrings) with keys.length != values.length, e.g. malformed CoNLL/TSV line splits where a column was dropped or an extra tab added.

Common situations: Loading labels from CSV/TSV files with inconsistent column counts; splitting a line on a wrong delimiter so a value containing the delimiter shifts columns; constructing labels programmatically with partially filled arrays.

Understand the failure class

Background: "must be a positive integer", "cannot be empty", "invalid argument": how invalid-argument errors work across open-source libraries — this error's family across 33 libraries.

Related errors

• Unknown key
• CORE: CoreLabel.initFromStrings: Bad type for
• Bad data format:
• Cannot find matching labelled span for
• ERROR: typed SINGLETON feature.

AI-assisted analysis of stanfordnlp/CoreNLP@1b7edd19c4 (2026-09-10). Data as JSON: /api/errors/a59ccf58e82a5d93. Report an issue: GitHub.

Appendix: source

Thrown at src/edu/stanford/nlp/ling/CoreLabel.java:177

    cl.setValue(word);
    return cl;
  }

  /**
   * Class that all "generic" annotations extend.
   * This allows you to read in arbitrary values from a file as features, for example.
   */
  public interface GenericAnnotation<T> extends CoreAnnotation<T> {  }

  public static final Map<String, Class<? extends GenericAnnotation<String>>> genericKeys = Generics.newHashMap();

  public static final Map<Class<? extends GenericAnnotation<String>>, String> genericValues = Generics.newHashMap();


  @SuppressWarnings({"unchecked", "rawtypes"})
  private void initFromStrings(String[] keys, String[] values) {
    if (keys.length != values.length) {
      throw new UnsupportedOperationException("Argument array lengths differ: " +
              Arrays.toString(keys) + " vs. " + Arrays.toString(values));
    }
    for (int i = 0; i < keys.length; i++) {
      String key = keys[i];
      String value = values[i];
      Class coreKeyClass = AnnotationLookup.toCoreKey(key);

      //now work with the key we got above
      if (coreKeyClass == null) {
        if (key != null) {
          throw new UnsupportedOperationException("Unknown key " + key);
        }
      } else {
        try {
          Class<?> valueClass = AnnotationLookup.getValueType(coreKeyClass);
          if(valueClass.equals(String.class)) {
            this.set(coreKeyClass, values[i]);
          } else if(valueClass == Integer.class) {

View on GitHub (pinned to 1b7edd19c4)