Index of the Minimum

Returns the index of the minimum value over one or more axes. Ties are broken by returning the smallest index.

Usage

nv_argmin(x, axes = NULL, drop = TRUE, nan_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. Negative values count from the end, i.e. -1 refers to the last axis. If NULL (default), reduces over all axes.
drop: (logical(1))
Whether to drop the reduced axes: removed from the output shape if TRUE, set to 1 if FALSE.
nan_rm: (logical(1))
How to handle NaN values in float inputs. If FALSE (default), NaN propagates. If TRUE, NaN values are skipped.

Value

(arrayish)
Has the default integer data type (see default_dtypes()) regardless of the input's, and the input's shape with axes removed (drop = TRUE) or set to 1 (drop = FALSE).

Reducing several axes

nv_argmin() is the index to nv_reduce_min()'s value: called with the same axes and drop, it points at the element whose value nv_reduce_min() returns. Reducing several axes ranks their elements together, and the result indexes the column-major flattening of those axes – the order nv_flatten() produces and base::which.min() reports – which is also the order ties are broken in.

NaN handling

With nan_rm = FALSE (default), if any entry being reduced is NaN, the returned index points at the first such NaN. With nan_rm = TRUE, NaN entries are skipped.

Examples

# the index comes out at the default integer data type
nv_argmin(nv_array(c(3, 1, 4, 1, 5, 9, 2, 6)))
#> AnvlArray
#>  2
#> [ CPUi32{} ] 
m <- nv_matrix(c(3, 1, 5, 2, 4, 0), nrow = 2, byrow = TRUE)
nv_argmin(m) # indexes the flattened matrix
#> AnvlArray
#>  6
#> [ CPUi32{} ] 
nv_argmin(m, axes = 2L) # one index per row
#> AnvlArray
#>  2
#>  3
#> [ CPUi32{2} ] 
nv_argmin(nv_array(c(2, NaN, 1, 3)))
#> AnvlArray
#>  2
#> [ CPUi32{} ] 
nv_argmin(nv_array(c(2, NaN, 1, 3)), nan_rm = TRUE)
#> AnvlArray
#>  3
#> [ CPUi32{} ]