Range Reduction

The smallest and the largest element along the specified axes, stacked along a new first axis. You can also use the range() generic.

Usage

nv_range(x, axes = NULL, nan_rm = FALSE)

# S3 method for class 'AnvlArray'
range(..., na.rm = FALSE)

Arguments

x

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

axes

(integer() | NULL)
Axes to reduce over. NULL (default) reduces over all of them, which makes the result a length-2 array like base::range(). Negative values count from the end.

nan_rm

(logical(1))
How to handle NaN values in float inputs. If FALSE (default), NaN propagates. If TRUE, NaN values are skipped.

...

(arrayish)
Arrays to reduce, plus named arguments for nv_reduce_min() and nv_reduce_max() (e.g. axes), which are only accepted when there is a single array to reduce.

na.rm

(logical(1))
Forwarded to the nan_rm argument of nv_reduce_min() and nv_reduce_max().

Value

(arrayish)
Has the same data type as x and the shape of the reduced array with a leading axis of size 2 added: element 1 is the minimum, element 2 the maximum.

The range() generic

range() reduces over all axes and, like base::range(), takes several data arguments: range(x, y) is the range of both arrays. Beyond base R, named arguments are passed on, so range(x, axes = 1L) reduces a single axis – but only when x is the only data argument.

See also

nv_reduce_min(), nv_reduce_max()

Examples

nv_range(nv_array(c(3, 1, 4)))
#> AnvlArray
#>  1
#>  4
#> [ CPUf32{2} ] 
nv_range(nv_matrix(1:6, nrow = 2), axes = 1L)
#> AnvlArray
#>  1 3 5
#>  2 4 6
#> [ CPUi32{2,3} ]