stanfordnlp/CoreNLP · error · IllegalArgumentException

Cannot match against null elements

Error message

Cannot match against null elements

What it means

SequenceMatcher's constructor requires the element list to match against. Passing a null List makes matching impossible, so IllegalArgumentException 'Cannot match against null elements' is thrown immediately when constructing the matcher (e.g. via SequencePattern.getMatcher(pattern, null)).

Solutions

1. Check the list for null before calling getMatcher and use Collections.emptyList() if matching should simply fail
2. Ensure the upstream annotator (e.g. tokenize, ssplit) has run so the annotation list exists
3. Guard CoreMap.get(...) results before matching

Example fix

// before
TokenSequenceMatcher m = pattern.getMatcher(coreMap.get(CoreAnnotations.TokensAnnotation.class)); // may be null
// after
List<CoreLabel> tokens = coreMap.get(CoreAnnotations.TokensAnnotation.class);
TokenSequenceMatcher m = pattern.getMatcher(tokens != null ? tokens : Collections.emptyList());

List<CoreLabel> tokens = coreMap.get(CoreAnnotations.TokensAnnotation.class);
if (tokens == null) { /* skip matching or run tokenizer */ }

boolean hasTokens(CoreMap cm) {
  return cm != null && cm.get(CoreAnnotations.TokensAnnotation.class) != null;
}

try {
  matcher = pattern.getMatcher(elements);
} catch (IllegalArgumentException e) {
  if (e.getMessage().contains("null elements")) {
    matcher = pattern.getMatcher(Collections.emptyList());
  }
}

When it happens

Trigger: pattern.getMatcher(null); calling getMatcher with a variable that is null because an upstream tokenization/annotation step produced no list; matching against a null field fetched from a CoreMap.

Common situations: CoreMap annotation lookups returning null (missing tokens/words annotation) before matching; pipeline ordering mistakes where matching runs before annotation creation.

Related errors

• Annotation field cannot be null
• Must supply a target label to compute precision and recall…
• Invalid annotation key
• Unknown rule type:
• Error creating composite rule: no annotation field

AI-assisted analysis of stanfordnlp/CoreNLP@1b7edd19c4 (2026-09-10). Data as JSON: /api/errors/cba00b073d43fa74. Report an issue: GitHub.

Appendix: source

Thrown at src/edu/stanford/nlp/ling/tokensregex/SequenceMatcher.java:97

   */
  public enum FindType { FIND_NONOVERLAPPING, FIND_ALL }
  private FindType findType = FIND_NONOVERLAPPING;

  // For FIND_ALL
  private Iterator<Integer> curMatchIter = null;
  private MatchedStates<T> curMatchStates = null;
  private final Set<String> prevMatchedSignatures = new HashSet<>();

  // Branching limit for searching with back tracking. Higher value makes the search faster but uses more memory.
  private int branchLimit = 32;

  protected SequenceMatcher(SequencePattern<T> pattern, List<? extends T> elements) {
    this.pattern = pattern;
    // NOTE: It is important elements DO NOT change as we do matches
    // TODO: Should we just make a copy of the elements?
    this.elements = elements;
    if (elements == null) {
      throw new IllegalArgumentException("Cannot match against null elements");
    }
    this.regionEnd = elements.size();
    this.priority = pattern.priority;
    this.score = pattern.weight;
    this.varGroupBindings = pattern.varGroupBindings;
    matchedGroups = new MatchedGroup[pattern.totalGroups];
  }

  public void setBranchLimit(int blimit){
    this.branchLimit = blimit;
  }


  /**
   * Interface that specifies what to replace a matched pattern with.
   *
   * @param <T>
   */

View on GitHub (pinned to 1b7edd19c4)