apache/druid · error · ParseException

Could not convert value [%s] to long.

Error message

Could not convert value [%s] to long.

What it means

DimensionHandlerUtils.convertObjectToLong converts an ingested field value to a long. When the value is a String that cannot be parsed as a long (Long.parseLong fails), Druid throws a ParseException with this message. It indicates the input data does not match the declared long column type.

Source

Thrown at processing/src/main/java/org/apache/druid/segment/DimensionHandlerUtils.java:405

    } else if (valObj instanceof Boolean) {
      return Evals.asLong((Boolean) valObj);
    } else if (valObj instanceof String) {
      Long ret = DimensionHandlerUtils.getExactLongFromDecimalString((String) valObj);
      if (reportParseExceptions && ret == null) {
        final String message;
        if (objectKey != null) {
          message = StringUtils.nonStrictFormat(
              "Could not convert value [%s] to long for dimension [%s].",
              valObj,
              objectKey
          );
        } else {
          message = StringUtils.nonStrictFormat(
              "Could not convert value [%s] to long.",
              valObj
          );
        }
        throw new ParseException((String) valObj, message);
      }
      return ret;
    } else if (valObj instanceof List) {
      final String message;
      if (objectKey != null) {
        message = StringUtils.nonStrictFormat(
            "Could not ingest value [%s] as long for dimension [%s]. A long column cannot have multiple values in the same row.",
            valObj,
            objectKey
        );
      } else {
        message = StringUtils.nonStrictFormat(
            "Could not ingest value [%s] as long. A long column cannot have multiple values in the same row.",
            valObj
        );
      }
      throw new ParseException(
          valObj.getClass().toString(),

View on GitHub (pinned to 9b90983fd2)

Solutions

  1. Fix the input data so the value is an integer-parseable string (e.g. "1" not "1.0"), or drop/transform bad values.
  2. Use an ingest-time transform or parse expression (e.g. REGEXP_EXTRACT, CAST) to normalize values before the long conversion.
  3. Change the column type to string or double in the ingestion spec if values genuinely aren't integers.
  4. Handle ParseException in your ingestion error handler / tune maxParseExceptions so the job doesn't halt unexpectedly.

Example fix

// before: ingesting "1.5" into a long column
// after: add a transform to round/parse safely
"transforms": [{"type": "expression", "name": "myLongCol", "expression": "CAST(CAST(myCol AS DOUBLE) AS LONG)"}]
Defensive patterns

Strategy: try-catch

Validate before calling

boolean isParsableLong(Object v) {
  if (v == null) return true;
  if (v instanceof Number) return true;
  if (v instanceof String) {
    try { Long.parseLong((String) v); return true; } catch (NumberFormatException e) { return false; }
  }
  return false;
}

Type guard

Long asLong(Object v) {
  if (v instanceof Number) return ((Number) v).longValue();
  if (v instanceof String) { try { return Long.parseLong((String) v); } catch (NumberFormatException e) { return null; } }
  return null;
}

Try / catch

try {
  return DimensionHandlerUtils.convertObjectToLong(value, fieldName, objectKey);
} catch (ParseException e) {
  log.warn("Skipping non-long value [%s] for field [%s]", value, fieldName);
  return null; // or route row to a dead-letter channel
}

Prevention

When it happens

Trigger: Ingesting a row where a dimension/column declared as long receives a string like "abc" or "12.5" that Long.parseLong cannot parse; convertObjectToLong called on a non-numeric string value.

Common situations: Input format mismatch (e.g. CSV field "1,000" or "1.0" mapped to a long column); ETL producing locale-formatted numbers; schema drift where a column was re-declared as long but data still contains text.

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

AI-assisted analysis of apache/druid@9b90983fd2 (2026-09-07). Data as JSON: /api/errors/d6b9f040a917d239. Report an issue: GitHub.