prestodb/presto · error · PrestoException

DIVISION_BY_ZERO

DIVISION_BY_ZERO

Error message

interval_day_to_second division by zero: %s ms / %s

What it means

Thrown by IntervalDayTimeOperators.divideByDouble when dividing an INTERVAL DAY TO SECOND by a DOUBLE that equals 0. Division by zero yields infinity, which cannot be represented as an interval millisecond long, so Presto fails fast with DIVISION_BY_ZERO instead of producing a garbage value.

Source

Thrown at presto-main-base/src/main/java/com/facebook/presto/type/IntervalDayTimeOperators.java:139

    @SqlType(StandardTypes.INTERVAL_DAY_TO_SECOND)
    public static long doubleMultiply(@SqlType(StandardTypes.DOUBLE) double left, @SqlType(StandardTypes.INTERVAL_DAY_TO_SECOND) long right)
    {
        try {
            return Math.addExact(
                Math.multiplyExact((long) left, right),
                (long) ((left - (long) left) * right));
        }
        catch (ArithmeticException e) {
            throw new PrestoException(NUMERIC_VALUE_OUT_OF_RANGE, format("interval_day_to_second multiply overflow: %s * %s ms", left, right), e);
        }
    }

    @ScalarOperator(DIVIDE)
    @SqlType(StandardTypes.INTERVAL_DAY_TO_SECOND)
    public static long divideByDouble(@SqlType(StandardTypes.INTERVAL_DAY_TO_SECOND) long left, @SqlType(StandardTypes.DOUBLE) double right)
    {
        if (right == 0) {
            throw new PrestoException(DIVISION_BY_ZERO, format("interval_day_to_second division by zero: %s ms / %s", left, right));
        }

        try {
            return multiplyByDouble(left, 1.0 / right);
        }
        catch (PrestoException e) {
            throw new PrestoException(NUMERIC_VALUE_OUT_OF_RANGE, format("interval_day_to_second division overflow: %s ms / %s", left, right));
        }
    }

    @ScalarOperator(NEGATION)
    @SqlType(StandardTypes.INTERVAL_DAY_TO_SECOND)
    public static long negate(@SqlType(StandardTypes.INTERVAL_DAY_TO_SECOND) long value)
    {
        try {
            return Math.negateExact(value);
        }
        catch (ArithmeticException e) {

View on GitHub (pinned to 55bb57d202)

Solutions

1. Guard the denominator: NULLIF(denominator, 0.0) or a CASE WHEN right = 0 THEN ... expression
2. Use try(INTERVAL '1' DAY / denominator) to get NULL instead of a failure
3. Verify upstream why the denominator is zero (missing data vs legitimate value)
4. Catch PrestoException with code DIVISION_BY_ZERO if handling at the engine level

Example fix

// before
SELECT INTERVAL '30' DAY / zero_ratio;
// after
SELECT INTERVAL '30' DAY / NULLIF(zero_ratio, 0.0); -- NULL instead of error, or try(... / ...)

-- SQL: guard the denominator before dividing
SELECT CASE WHEN denominator = 0.0 THEN NULL ELSE INTERVAL '30' DAY / denominator END
-- or
SELECT INTERVAL '30' DAY / NULLIF(denominator, 0.0)

public static boolean isSafeIntervalDivisor(double d) {
    return d != 0.0;
}

try {
    long r = IntervalDayTimeOperators.divideByDouble(intervalMs, divisor);
} catch (PrestoException e) {
    if (e.getErrorCode().getCode() == StandardErrorCode.DIVISION_BY_ZERO.getCode()) {
        // return NULL or a sentinel for zero denominators
    } else {
        throw e;
    }
}

When it happens

Trigger: DIVIDE operator with INTERVAL DAY TO SECOND left and a DOUBLE right of exactly 0.0 (or -0.0), e.g. SELECT INTERVAL '1' DAY / 0.0, or an interval divided by a computed double expression that evaluated to zero.

Common situations: Dividing by a ratio or count column that is 0 (empty cohorts, zero totals); data-driven denominators from joins where no rows matched.

Related errors

• DIVISION_BY_ZERO
• DIVISION_BY_ZERO
• INVALID_TABLE_PROPERTY
• ACCUMULO_TABLE_EXISTS
• NOT_SUPPORTED

AI-assisted analysis of prestodb/presto@55bb57d202 (2026-09-04). Data as JSON: /api/errors/46663cfb5666f418. Report an issue: GitHub.