Reshape

Reshapes an array to a new shape without changing the underlying data. Returns the input unchanged if it already has the target shape.

Usage

nv_reshape(x, shape)

Arguments

x

(arrayish)
Input array.

shape

(integer())
Target shape. Must have the same number of elements as x. At most one entry may be -1, in which case its extent is inferred from the remaining entries and the number of elements of x.

Value

arrayish
Has the given shape and the same data type as x.

Details

Note that row-major order is used, which differs from R's column-major order.

See also

prim_reshape() for the underlying primitive.

Examples

x <- nv_array(1:6)
nv_reshape(x, c(2, 3))
#> AnvlArray
#>  1 2 3
#>  4 5 6
#> [ CPUi32{2,3} ] 
nv_reshape(x, c(2, -1)) # infer the second dimension
#> AnvlArray
#>  1 2 3
#>  4 5 6
#> [ CPUi32{2,3} ] 
nv_reshape(x, -1) # flatten
#> AnvlArray
#>  1
#>  2
#>  3
#>  4
#>  5
#>  6
#> [ CPUi32{6} ]