stanfordnlp/CoreNLP · error · RuntimeException

Got unsplittable line: "

Error message

Got unsplittable line: "

What it means

ClassicCounter.valueOfIgnoreComments parses a TSV file where each non-comment line must contain exactly two tab-separated fields (key<TAB>count). If a non-comment line does not split into exactly 2 fields, a RuntimeException with the offending line is thrown. Comment lines (starting with '#') are skipped, but only after the split check.

Solutions

1. Inspect the reported line and fix the delimiter to a tab character
2. Pre-filter lines: skip blank lines and lines starting with '#' before calling the parser
3. Regenerate the input file with the matching saveCounter/toString output format

Example fix

// before
List<String> lines = Arrays.asList(content.split("\n"));
ClassicCounter<String> c = ClassicCounter.valueOfIgnoreComments(lines);
// after
List<String> lines = Arrays.stream(content.split("\n"))
    .map(l -> l.replace("\r", "").trim())
    .filter(l -> !l.isEmpty() && !l.startsWith("#"))
    .collect(Collectors.toList());
ClassicCounter<String> c = ClassicCounter.valueOfIgnoreComments(lines);

for (String line : lines) {
  String t = line.replace("\r", "");
  if (t.isEmpty() || t.startsWith("#")) continue;
  if (t.split("\t", -1).length != 2)
    throw new IllegalArgumentException("bad TSV line: " + t);
}

try {
  c = ClassicCounter.valueOfIgnoreComments(lines);
} catch (RuntimeException e) {
  log.error("Malformed counter file: {}", e.getMessage());
  throw new IOException("counter file format invalid", e);
}

When it happens

Trigger: Calling ClassicCounter.valueOfIgnoreComments on a file/lines that contain a line without exactly one tab character; lines using spaces instead of tabs; trailing blank lines or header rows.

Common situations: Reading a counts file edited by hand or exported from Excel with space separators; files with CRLF where fields.length checks fail unexpectedly; empty or malformed last line.

Understand the failure class

Background: "Invalid ... format", "must be in format X", "does not look like a ..." — invalid argument format errors across CLI tools and libraries — this error's family across 17 libraries.

Related errors

• Expected left paren!
• Expected right paren!
• Bad data format:
• Cannot find matching labelled span for
• Error extracting labelled spans for column :

AI-assisted analysis of stanfordnlp/CoreNLP@1b7edd19c4 (2026-09-10). Data as JSON: /api/errors/05ce133306d0d617. Report an issue: GitHub.

Appendix: source

Thrown at src/edu/stanford/nlp/stats/ClassicCounter.java:569

   * <blockquote>
   * StringKey\tdoubleValue\n
   * </blockquote>
   *
   * @param s String representation of a Counter, where entries are one per
   *     line such that each line is either a comment (begins with #)
   *     or key \t value
   * @return The Counter with String keys
   */
  public static ClassicCounter<String> valueOfIgnoreComments(String s) {
      ClassicCounter<String> result = new ClassicCounter<>();
      String[] lines = s.split("\n");
      for (String line : lines) {
        String[] fields = line.split("\t");
        if (fields.length != 2) {
          if (line.startsWith("#")) {
            continue;
          } else {
            throw new RuntimeException("Got unsplittable line: \"" + line + '\"');
          }
        }
        result.setCount(fields[0], Double.parseDouble(fields[1]));
      }
      return result;
    }


  /**
   * Converts from the format printed by the toString method back into
   * a Counter&lt;String&gt;.  The toString() doesn't escape, so this only
   * works providing the keys of the Counter do not have commas or equals signs
   * in them.
   *
   * @param s A String representation of a Counter
   * @return The Counter
   */
  public static ClassicCounter<String> fromString(String s) {

View on GitHub (pinned to 1b7edd19c4)