Argsort

Returns the indices that would sort the array: over every element by default, or along one axis. It is the index twin of nv_sort() and takes the same axis, so the two always describe the same ordering.

Usage

nv_argsort(x, axis = NULL, decreasing = FALSE, stable = FALSE)

Arguments

x

(arrayish)
One input. Can be any data type. An R value materializes at its default data type.

axis

(integer(1) | NULL)
Axis along which to compute the sort permutation. Negative values count from the end, i.e. -1 refers to the last axis. If NULL (default), the input is first flattened to a 1-D array, like nv_sort(), and the indices refer to that flattening.

decreasing

(logical(1))
If TRUE, returns indices that produce a decreasing sort. Default FALSE.

stable

(logical(1))
If TRUE, the sort is stable: indices for equal values keep their original relative order. Default FALSE.

Value

(arrayish)
Has the default integer data type (see default_dtypes()) regardless of the input's, and the input's shape – or 1-D holding every element's index when axis is NULL. For a size-0 axis, the output is an empty array of the same shape (a valid empty permutation). Indexing the flattened input by the result reproduces nv_sort()'s output.

NaN handling

NaN values sort to the end (ascending) or beginning (descending), regardless of sign. +0 and -0 compare equal.

See also

nv_sort(), prim_sort().

Examples

# the indices come out at the default integer data type
x <- nv_array(c(3, 1, 4, 1, 5))
nv_argsort(x)
#> AnvlArray
#>  2
#>  4
#>  1
#>  3
#>  5
#> [ CPUi32{5} ] 

m <- nv_matrix(c(3, 1, 5, 2, 4, 0), nrow = 2, byrow = TRUE)
nv_argsort(m) # indexes the flattened matrix
#> AnvlArray
#>  6
#>  3
#>  2
#>  1
#>  4
#>  5
#> [ CPUi32{6} ] 
nv_argsort(m, axis = 2L) # a permutation per row
#> AnvlArray
#>  2 1 3
#>  3 1 2
#> [ CPUi32{2,3} ]