hop() is the lower level engine that powers slide() (at least in theory). It has slightly different invariants than slide(), and is useful when you either need to hand craft boundary locations, or want to compute a result with a size that is different from .x.

hop(.x, .starts, .stops, .f, ...)

hop_vec(.x, .starts, .stops, .f, ..., .ptype = NULL)

## Arguments

.x [vector] The vector to iterate over and apply .f to. [integer] Vectors of boundary locations that make up the windows to bucket .x with. Both .starts and .stops will be recycled to their common size, and that common size will be the size of the result. Both vectors should be integer locations along .x, but out-of-bounds values are allowed. [function / formula] If a function, it is used as is. If a formula, e.g. ~ .x + 2, it is converted to a function. There are three ways to refer to the arguments: For a single argument function, use . For a two argument function, use .x and .y For more arguments, use ..1, ..2, ..3 etc This syntax allows you to create very compact anonymous functions. Additional arguments passed on to the mapped function. [vector(0) / NULL] A prototype corresponding to the type of the output. If NULL, the default, the output type is determined by computing the common type across the results of the calls to .f. If supplied, the result of each call to .f will be cast to that type, and the final output will have that type. If getOption("vctrs.no_guessing") is TRUE, the .ptype must be supplied. This is a way to make production code demand fixed types.

## Value

A vector fulfilling the following invariants:

### hop()

• vec_size(hop(.x, .starts, .stops)) == vec_size_common(.starts, .stops)

• vec_ptype(hop(.x, .starts, .stops)) == list()

### hop_vec()

• vec_size(hop_vec(.x, .starts, .stops)) == vec_size_common(.starts, .stops)

• vec_size(hop_vec(.x, .starts, .stops)[[1]]) == 1L

• vec_ptype(hop_vec(.x, .starts, .stops, .ptype = ptype)) == ptype

## Details

hop() is very close to being a faster version of:

map2(
.starts,
.stops,
function(start, stop) {
x_slice <- vec_slice(.x, start:stop)
.f(x_slice, ...)
}
)

Because of this, hop_index() is often the more useful function. hop() mainly exists for API completeness.

The main difference is that the start and stop values make up ranges of possible locations along .x, and it is not enforced that these locations actually exist along .x. As an example, with hop() you can do the following, which would be an error with vec_slice() because 0L is out of bounds.

hop(c("a", "b"), .starts = 0L, .stops = 1L, ~.x)
#> [[1]]
#> [1] "a"

hop() allows these out of bounds values to be fully compatible with slide(). It is always possible to construct a hop() call from a slide() call. For example, the following are equivalent:

slide(1:2, ~.x, .before = 1)

hop(1:2, .starts = c(0, 1), .stops = c(1, 2), ~.x)

#> [[1]]
#> [1] 1
#>
#> [[2]]
#> [1] 1 2

# hop() let's you manually specify locations to apply .f at. hop(1:3, .starts = c(1, 3), .stops = 3, ~.x)
# hop()'s start/stop locations are allowed to be out of bounds relative # to the size of .x. hop( mtcars, .starts = c(-1, 3), .stops = c(2, 6), ~.x )