NRQL supports the use of basic and advanced mathematical operators within a `SELECT`

clause. You can apply mathematical calculations on both individual attributes and also the results of aggregator functions.

## Use basic math operators with `SELECT`

To use basic math functions in NRQL, include operators within the `SELECT`

clause:

- Addition:
`+`

- Subtraction:
`-`

- Multiplication:
`*`

- Division:
`/`

Here are some examples:

`SELECT duration-databaseDuration FROM Transaction`

`SELECT count(*)/uniqueCount(session) FROM PageView`

`SELECT average(duration-databaseDuration) FROM Transaction`

## Use advanced math operators with `SELECT`

NRQL also includes some advanced mathematical capabilities that can be used for complex calculations. This is helpful if you want to process data to display it more effectively in the UI, or make statistical or psychometric calculations on queried results in a single step.

`abs`

`abs(n)`

returns the absolute value of *n*. For non-negative *n* it returns *n*, and for negative *n* it returns the positive number -*n*. For example `abs(2) = 2`

, and `abs(-4) = 4`

.

`clamp_max`

, `clamp_min`

The clamping functions impose an upper or lower bound on values.

For example, `clamp_max(duration, 10)`

returns the duration, unless it exceeds 10, in which case 10 is returned.

Similarly `clamp_min(duration, 1)`

will not return any value lower than 1.

The following chart shows the result of clamping both min and max to keep the value in the range 70-90.

Sample graph showing raw data with clamp function applied.

`exp`

Computes the natural exponential function of the argument: `exp(n) = pow(e, n)`

.

### Logarithmic functions: `ln`

, `log`

, `log2`

, `log10`

These functions compute the logarithm of the argument for various bases.

`ln(n)`

computes the natural logarithm: the logarithm base e.`log2(n)`

computes the logarithm base 2.`log10(n)`

computes the logarithm base 10.`log(n, b)`

allows logarithms to be computed with an arbitrary base b.- All logarithms satisfy the identity:
`log(pow(b, n), b) = n`

.

Note that `log(0)`

is undefined, for all bases. Be aware that if you take the logarithm of something that might be zero, you may end up getting `no value`

back from your query.

`pow`

`pow(n, m)`

computes *n* raised to the power *m*. (I.e. `n * n * ... * n`

, with *m* copies of *n*)

`abs(n)`

returns the absolute value of *n*. For non-negative *n* it returns *n*, and for negative *n* it returns the positive number -*n*. For example `abs(2) = 2`

and `abs(-4) = 4`

.

### Rounding functions: `round`

, `floor`

, `ceil`

These three functions force decimal numbers to one of the neighboring integers.

`floor(n)`

returns the closest integer less than or equal to*n*.`ceil(n)`

(short for "ceiling") returns the closest integer greater than or equal to*n*.`round(n)`

returns the closest integer to*n*in either direction.

Sample graph showing raw data, with floor, round, and ceiling functions applied.

`sqrt`

`sqrt(n)`

returns the square root of *n*, that is, the number such that `sqrt(n) * sqrt(n) = n`

.

## Results with STRING or FLOAT

Here is how NRQL handles strings present in math calculations:

`pow(n, m)`

computes *n* raised to the power *m*. (for example, `n * n * ... * n`

, with *m* copies of *n*)

Examples:

`sum(1+STRING)`

= 0`sum(1+MIXED)`

= skips records where`MIXED`

is a string`average(1+STRING)`

= 0`average(1+MIXED)`

= skips records where`MIXED`

is a string

NULL and zero both appear as 0 in the dashboard.

To override NULL values with another numeric value, use the syntax:

`SELECT average(purchasePrice OR 0)`

This will replace NULL values with 0 or any number specified.

#### Tip

This can also be used to test whether something returns NULL or zero. `(zero) OR 1`

returns 0. `(NULL) OR 1`

returns 1.