apache/flink · error · NumberFormatException

Orphaned minus sign.

Error message

Orphaned minus sign.

What it means

ShortParser.parseField throws NumberFormatException("Orphaned minus sign.") when the field starts with '-' but has no digits after it — either the minus is the entire field or it is immediately followed by the delimiter.

Source

Thrown at flink-core/src/main/java/org/apache/flink/types/parser/ShortParser.java:136

     * @param delimiter The delimiter that terminates the field.
     * @return The parsed value.
     * @throws NumberFormatException Thrown when the value cannot be parsed because the text
     *     represents not a correct number.
     */
    public static final short parseField(byte[] bytes, int startPos, int length, char delimiter) {
        long val = 0;
        boolean neg = false;

        if (bytes[startPos] == delimiter) {
            throw new NumberFormatException("Empty field.");
        }

        if (bytes[startPos] == '-') {
            neg = true;
            startPos++;
            length--;
            if (length == 0 || bytes[startPos] == delimiter) {
                throw new NumberFormatException("Orphaned minus sign.");
            }
        }

        for (; length > 0; startPos++, length--) {
            if (bytes[startPos] == delimiter) {
                return (short) (neg ? -val : val);
            }
            if (bytes[startPos] < 48 || bytes[startPos] > 57) {
                throw new NumberFormatException("Invalid character.");
            }
            val *= 10;
            val += bytes[startPos] - 48;

            if (val > OVERFLOW_BOUND && (!neg || val > UNDERFLOW_BOUND)) {
                throw new NumberFormatException("Value overflow/underflow");
            }
        }

View on GitHub (pinned to 2f3c205e92)

Solutions

1. Replace bare '-' placeholders in the source data with a proper value or an explicit null literal
2. Configure the CSV connector's null-literal handling if '-' denotes missing
3. Validate/clean fields before they reach a SMALLINT-typed parser

Example fix

// before
-- CSV: 5,-,7 with schema (a SMALLINT, b SMALLINT, c SMALLINT)

// after
-- CSV: 5,\N,7 with null-literal configured, or 5,0,7

if (field.equals("-") || field.endsWith("-") && field.length() == 1) {
    /* placeholder: map to null */
}

catch (NumberFormatException e) { log.info("placeholder '-' in short field: {}", rawLine); }

When it happens

Trigger: A CSV short field containing just '-' (e.g. '5,-,7'), or '-' immediately before the delimiter after the sign-consumption step. The parser advances startPos past '-' and then finds length==0 or the delimiter byte.

Common situations: Data quality issues where '-' is used as a placeholder for missing/unknown values; truncation of negative numbers during export; user-typed placeholders in generated test files.

Related errors

• Empty field.
• Invalid character.
• Value overflow/underflow
• Row too short: {}
• Line could not be parsed: '{}' ParserError {} Expect field

AI-assisted analysis of apache/flink@2f3c205e92 (2026-08-14). Data as JSON: /api/errors/efb2f49686d1cff0. Report an issue: GitHub.