Compute Ordered Indices — idx_ord

Computes ordered indices. Similar to order, except the user must supply a vector, a list of equal-length vectors, a data.frame or a matrix (row-wise and column-wise are both supported), as the input.

For a vector x,
idx_ord_v(x) is equivalent to
order(x).

For a data.frame or a list of equal-length vectors x, with p columns/elements,
idx_ord_df(x) is equivalent to
order(x[[1]], ..., x[[p]]).

For a matrix (or array) x with p rows,
idx_ord_m(x, margin = 1) is equivalent to
order(x[1, ], ..., x[p, ], ...).

For a matrix (or array) x with p columns,
idx_ord_m(x, margin = 2) is equivalent to
order(x[, 1], ..., x[, p], ...).

Note that these are merely convenience functions, and that these are actually slightly slower than order (except for idx_ord_v()), due to the additional functionality.

Usage

idx_ord_v(
  x,
  na.last = TRUE,
  decr = FALSE,
  method = c("auto", "shell", "radix")
)

idx_ord_m(
  x,
  margin,
  na.last = TRUE,
  decr = FALSE,
  method = c("auto", "shell", "radix")
)

idx_ord_df(
  x,
  na.last = TRUE,
  decr = FALSE,
  method = c("auto", "shell", "radix")
)

Arguments

x: a vector, data.frame, or array
na.last, method: see order and sort.
decr: see argument decreasing in order
margin: the margin over which to cut the matrix/array into vectors.
I.e. margin = 1L will cut x into individual rows, and apply the order on those rows.
And margin = 2L will cut x into columns, etc.

Value

See order.

Examples


x <- sample(1:10)
order(x)
#>  [1]  7  6  1  2 10  9  5  8  4  3
idx_ord_v(x)
#>  [1]  7  6  1  2 10  9  5  8  4  3
idx_ord_m(rbind(x, x), 1)
#>  [1]  7  6  1  2 10  9  5  8  4  3
idx_ord_m(cbind(x, x), 2)
#>  [1]  7  6  1  2 10  9  5  8  4  3
idx_ord_df(data.frame(x, x))
#>  [1]  7  6  1  2 10  9  5  8  4  3