Sort

Sorts an array along an axis.

You can also use sort() directly.

Usage

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

# S3 method for class 'AnvlArray'
sort(x, decreasing = FALSE, ..., axis = NULL)

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 sort. 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 base::sort().

decreasing

(logical(1))
If TRUE, sort in decreasing order.

stable

(logical(1))
If TRUE, the sort is stable: equal values keep their original relative order along axis. Default FALSE. Stability is only observable for floats when -0 / +0 or -NaN / +NaN are mixed (they compare equal under the total order used here); for distinct values the result is identical either way.

...

No additional arguments.

Value

(arrayish)
Has the input's shape and data type.

NaN handling

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

The sort() generic

Like base::sort(), nv_sort() with axis = NULL flattens a multi-axis array into one sorted vector, so sort() on an anvl array agrees with base R (the flatten order does not matter once the elements are sorted). It differs in one respect: base R drops NA by default, whereas NaN is kept and sorted to the end. Pass axis to sort each slice along one axis instead, which keeps the shape.

See also

prim_sort() for the underlying primitive, nv_argsort(), nv_top_k(), nv_median(), nv_argmax(), nv_argmin().

Examples

# sorting moves elements, so the data type and shape stay
x <- nv_array(c(3, 1, 4, 1, 5, 9, 2, 6))
nv_sort(x)
#> AnvlArray
#>  1
#>  1
#>  2
#>  3
#>  4
#>  5
#>  6
#>  9
#> [ CPUf32{8} ] 
sort(x) # via the S3 generic
#> AnvlArray
#>  1
#>  1
#>  2
#>  3
#>  4
#>  5
#>  6
#>  9
#> [ CPUf32{8} ] 
nv_sort(x, decreasing = TRUE)
#> AnvlArray
#>  9
#>  6
#>  5
#>  4
#>  3
#>  2
#>  1
#>  1
#> [ CPUf32{8} ] 

m <- nv_matrix(c(3, 1, 5, 2, 4, 0), nrow = 2, byrow = TRUE)
nv_sort(m) # one sorted vector, like base R
#> AnvlArray
#>  0
#>  1
#>  2
#>  3
#>  4
#>  5
#> [ CPUf32{6} ] 
nv_sort(m, axis = 2L) # each row sorted, shape kept
#> AnvlArray
#>  1 3 5
#>  0 2 4
#> [ CPUf32{2,3} ]