detect() predicate

[lionel-]

Parameterise detect() with a predicate:

detect <- function(.x, .f, ..., .right = FALSE, .p = is_true) {
  for (i in index(.x, .right)) {
    if (.p(.f(.x[[i]], ...))) {
      return(.x[[i]])
    }
  }
  NULL
}

Useful for detecting non-null values for instance.

[lionel-]

Along with something like:

detect_value <- function(.x, .f, .p, ..., .right = FALSE) {
  for (i in index(.x, .right)) {
    value <- .f(.x[[i]], ...)
    if (.p(value)) {
      return(value)
    }
  }
  NULL
}

detect_value(x, some_fn, negate(is_null))

[lionel-]

Or should we test with is_trueish() instead?

is_trueish <- function(x) {
  if (is_true(x)) {
    return(TRUE)
  }
  if (is_integerish(x, n = 1L, finite = TRUE)) {
    return(x != 0)
  }
  FALSE
}

Then we can use it with e.g. anyDuplicated()

  detect(list(1:3, 4:6, c(7L, 8L, 7L)), anyDuplicated)
  #> [1] 7 8 7

Or should we keep is_true() by default, parameterise .p, and add is_trueish() to rlang?

[lionel-]

Would get in the way of #470, and can be solved by wrapping .f with .p manually.