Skip to contents

Variables: scalar functions of an individual and time

Source: R/variable.R
variable.Rd

A variable is a scalar, real-valued function \(f(i, t)\) written with the data pronouns .i$field, .t, .b and .x. A static_variable is a variable that does not depend on time – a function of the individual's facts alone, \(f(i)\).

Construct one from a pronoun expression or a ~ formula. The constructors return the most specific type they can prove: a time-invariant expression becomes a static_variable, and one that is also structurally logical becomes an indicator. static_variable() additionally asserts that the expression does not use .t (and fails if it does); indicator() asserts a \(\{0,1\}\) value. So variable(.i$pension) is a static_variable and variable(.i$pension > 0) is an indicator.

value() evaluates the variable against a single individual's facts .i (a named list) and a time vector .t. It is the plain-R reference path for testing and understanding – not the performance path. For a static_variable, .t may be omitted.

Usage

variable(expr)

static_variable(expr)

is_variable(x)

is_static_variable(x)

value(x, .i, .t = NULL, ...)

# Default S3 method
value(x, .i, .t = NULL, ...)

# S3 method for class 'variable'
value(x, .i, .t = NULL, ...)

Arguments

expr

A pronoun expression, optionally written as a ~ formula.

x

A variable.

.i

A named list of scalar facts.

.t

A vector of datey (omit for a time-invariant variable).

...

Unused.

Value

variable() and static_variable() return a variable (the latter also classed static_variable).

is_variable() and is_static_variable() return a scalar logical.

value() returns a vector aligned to .t.

Examples

amounts <- static_variable(.i$pension)
value(amounts, .i = list(pension = 1000))
#> [1] 1000

v <- variable(.i$pension * 2)
value(v, .i = list(pension = 1000))
#> [1] 2000