Conditional Element Selection

Selects elements from yes or no based on test, mirroring R's ifelse().

nv_if() is the other conditional: it mirrors R's if construct and branches between two functions, evaluating only the selected one.

Usage

nv_ifelse(test, yes, no)

Arguments

test

(arrayish)
Predicate array. Must be a boolean or an R logical, and scalar or the same shape as the non-scalar arguments.

yes, no

(arrayish)
Values to return where test is TRUE / FALSE. Can be of any data type; the two are promoted to a common data type. Scalars (including test) are broadcast to the shape of the non-scalar arguments.

Value

(arrayish)
Has the common data type of yes and no, and their broadcast shape.

See also

prim_ifelse() for the underlying primitive, nv_if().

Examples

test <- nv_array(c(TRUE, FALSE, TRUE))
nv_ifelse(test, nv_array(c(1, 2, 3)), nv_array(c(4, 5, 6)))
#> AnvlArray
#>  1
#>  5
#>  3
#> [ CPUf32{3} ] 
# scalar branches are broadcast and promoted to a common data type
nv_ifelse(test, nv_scalar(1L), nv_scalar(0.5))
#> AnvlArray
#>  1.0000
#>  0.5000
#>  1.0000
#> [ CPUf32{3} ]