stanfordnlp/CoreNLP · error · java.io.IOException

Invalid number of fields, should be >=1 and <= 31

Error message

Invalid number of fields, should be >=1 and <= 31

What it means

edu.stanford.nlp.ie.pascal.Prior reads a model file whose first line lists up to 31 index fields; it throws this IOException when the header has fewer than 1 or more than 31 whitespace-separated fields. The field count determines the size of the joint probability matrix (2^n entries), so the parser enforces this bound to keep the matrix allocation sane.

Solutions

  1. Check the file's first line: it must contain between 1 and 31 whitespace-separated field names.
  2. Verify you are loading the correct Prior model file, not another format's file.
  3. Re-download or regenerate the model file if it is empty or truncated.
  4. If you genuinely need more than 31 fields, reduce the field set — the format cannot represent it.

Example fix

// before
Prior p = new Prior(new FileReader("empty_or_wrong.txt"));
// after
BufferedReader r = new BufferedReader(new FileReader("prior.model"));
String first = r.readLine();
int n = first == null ? 0 : first.trim().split("\\s+").length;
if (n < 1 || n > 31) throw new IOException("Bad Prior header, fields=" + n);
Prior p = new Prior(r);
Defensive patterns

Strategy: validation

Validate before calling

// java: check the Prior header before constructing
String header = reader.readLine();
int n = (header == null || header.trim().isEmpty()) ? 0 : header.trim().split("\\s+").length;
if (n < 1 || n > 31) {
  throw new IOException("Prior header has " + n + " fields; need 1-31");
}

Try / catch

try {
  Prior p = new Prior(reader);
} catch (IOException e) {
  if (e.getMessage().contains("Invalid number of fields")) {
    LOG.error("Wrong or corrupt Prior model file: " + e.getMessage());
    p = loadFallbackPrior();
  } else throw e;
}

Prevention

When it happens

Trigger: Constructing Prior from a file whose first line is blank (0 fields) or contains more than 31 whitespace-separated field names, e.g. a corrupted or wrong-format model file passed to the Prior constructor.

Common situations: Pointing Prior at the wrong model file (a different format's file that happens to open); a truncated or empty model file from a failed download; editing a model file and adding fields beyond the 31-field hard limit.

Understand the failure class

Background: Schema validation failed / invalid input schema: payload rejected because its shape doesn't match the expected schema — this error's family across 28 libraries.

Related errors

AI-assisted analysis of stanfordnlp/CoreNLP@1b7edd19c4 (2026-09-10). Data as JSON: /api/errors/1243dcfc049b1cc1. Report an issue: GitHub.

Appendix: source

Thrown at src/edu/stanford/nlp/ie/pascal/Prior.java:35

  private Map<String, Integer> fieldIndices;
  private String[] indexFields;

  // n-dimensional boolean matrix. There will be 2^n entries in the matrix.
  private double[] matrix;

  public Prior(BufferedReader reader) throws IOException {
    String line;
    line = reader.readLine();
    if (line == null) {
      throw new IOException();
    }
    indexFields = line.split("\\s+");
    fieldIndices = new HashMap<String, Integer>();
    for (int i = 0; i < indexFields.length; ++i) {
      fieldIndices.put(indexFields[i], Integer.valueOf(i));
    }
    if (indexFields.length < 1 || indexFields.length > 31) {
      throw new IOException("Invalid number of fields, should be >=1 and <= 31");
    }
    int matrixSize = 1 << indexFields.length;
    matrix = new double[matrixSize];
    int matrixIdx = 0;
    while (matrixIdx < matrix.length && (line = reader.readLine()) != null) {
      String[] tokens = line.split("\\s+");
      for (int t = 0; matrixIdx < matrix.length && t < tokens.length; ++t) {
        matrix[matrixIdx++] = Double.parseDouble(tokens[t]);
      }
    }
  }

  /**
   * {@code Map<String, boolean>}
   */
  public double get(Set presentFields) {
    int index = 0;
    for (String field : indexFields) {

View on GitHub (pinned to 1b7edd19c4)