Scalar functions

Casts, conditionals, maths and type checks — everything that works on one value.

Scalar functions take values and return values. They appear inside where, extend, project and aggregation arguments.

Casts

Function	Returns	Note
`tostring(v)`	string	Never fails
`toint(v)`	integer or null	Null when unparseable — pair with `isnotnull()`
`todouble(v)`	float or null	—
`tobool(v)`	boolean or null	Accepts `true`/`false`/`1`/`0`
`todatetime(v)`	datetime or null	RFC 3339 or unix

Conditionals

Function	Does
`iff(cond, a, b)`	Ternary
`case(c1, v1, c2, v2, …, default)`	First matching condition wins
`coalesce(a, b, …)`	First non-null
`isnull(v)` / `isnotnull(v)`	Null test
`isempty(v)`	Null or empty string

Bucketing with case

read top to bottom

1['api-gateway-prod']
2| extend tier = case(
3    latency < 100, "fast",
4    latency < 1000, "acceptable",
5    "slow")
6| summarize count() by tier, service
7| order by count_ desc

Maths

Function	Does
`round(v, digits)`	Rounds to N decimal places
`floor(v)` / `ceiling(v)`	Rounds down / up
`abs(v)`	Absolute value
`log(v)` / `log10(v)` / `exp(v)`	Logs and exponent
`sqrt(v)` / `pow(v, n)`	Root and power
`bin(v, step)`	Rounds down to a multiple of `step` — works on numbers and datetimes

Integer division truncates

errors / requests on two integers gives 0. Force a float: 100.0 * errors / requests. This is the single most common wrong-looking result in LSQL.

PreviousTabular operators NextAggregations