Closures & Functions

Define a closure, call it

A closure is a parenthesized parameter list followed by a block, assigned to a variable with let. Calling the variable runs the body; when the call is the final statement of a request, its result comes back as a one-value frame:

Closures capture their scope

The body can reference variables from the enclosing scope - that is what makes it a closure rather than just a function. Here $base is not a parameter; it is captured:

Parameters win over captures of the same name. Inside the body, $x is the argument, not the outer variable:

Bodies are blocks

A closure body is a full block: it can declare its own variables, and its last expression is the return value:

Closures compose - a closure can call another closure it captured:

Using closures in queries

To feed a closure's result into a pipeline, bind it with let and use the variable:

What does not work is calling the closure variable inline inside a pipeline expression - map and filter do not resolve $f(4) as a callable:

Named functions with udf

When the whole point is to call the function from inside a pipeline, declare it with udf instead. A named udf is callable anywhere a built-in function is:

udf bodies use explicit return, and return exits immediately - combined with if that gives guard-clause style logic:

A udf returning a boolean works as a reusable predicate:

And a udf can wrap statements, not just expressions - including writes. This is the pattern for "do this parameterized thing N times" inside one request:

One request, then gone

Closures and udfs are script constructs, not schema objects. They exist from their definition to the end of the request - the next request starts clean: