Numeric functions - e6data Docs

ReferenceApplies to: Unified engine deployments

Numeric functions resolve an overload from the input types. Cast an input explicitly when an example needs floating-point behavior. Most required null arguments produce NULL. Invalid domains, overflow, or unsupported conversions can produce errors.

Magnitude and rounding

Function	Returns	Meaning
`abs(number)`	Numeric	Absolute value
`ceil(number)`, `floor(number)`	Numeric	Round toward positive or negative infinity
`round(number[, digits])`	Numeric	Round to a precision
`trunc(number[, digits])`	Numeric	Truncate fractional digits
`signum(number)`	Numeric	Sign of the argument

SELECT i_item_sk,
       ABS(CAST(i_current_price AS DOUBLE)) AS magnitude,
       CEIL(CAST(i_current_price AS DOUBLE)) AS rounded_up,
       FLOOR(CAST(i_current_price AS DOUBLE)) AS rounded_down,
       ROUND(CAST(i_current_price AS DOUBLE), 1) AS one_decimal
FROM demo.tpcds.item
WHERE i_current_price IS NOT NULL
ORDER BY i_item_sk
LIMIT 5;

Powers, roots, and logarithms

Function	Returns	Meaning
`power(base, exponent)`, `pow(base, exponent)`	Numeric	Exponentiation
`sqrt(number)`	Numeric	Square root
`exp(number)`	Floating point	Exponential
`ln(number)`	Floating point	Natural logarithm
`log10(number)`	Floating point	Base-10 logarithm
`sin(number)`, `cos(number)`, `tan(number)`	Floating point	Trigonometric functions with radians

SELECT i_item_sk,
       SQRT(CAST(i_current_price AS DOUBLE)) AS square_root,
       POWER(CAST(i_current_price AS DOUBLE), 2.0) AS squared_price,
       LN(CAST(i_current_price AS DOUBLE)) AS log_price
FROM demo.tpcds.item
WHERE i_current_price > 0
ORDER BY i_item_sk
LIMIT 5;

Random values and exact values

random() and alias rand() produce a non-deterministic floating-point value. Do not use random output as an expected-result fixture. Decimal calculations preserve declared precision differently from DOUBLE; choose the type intentionally for money and exact comparisons.

SELECT i_item_sk, RANDOM() AS sample_value
FROM demo.tpcds.item
ORDER BY i_item_sk
LIMIT 3;

A numeric function resolving successfully does not guarantee that an unsupported arithmetic expression around it will compile. See operators for integer arithmetic and cast restrictions.

Last modified on November 4, 2026