Skip to contents

to_fct() checks whether an argument can be coerced to a factor without losing information, returning it silently if so. Otherwise an informative error message is signaled.

stabilize_fct() can check more details about the argument, but is slower than to_fct().

stabilize_fct_scalar() and to_fct_scalar() are optimized to check for length-1 factors.

Usage

stabilize_fct(
  x,
  ...,
  allow_null = TRUE,
  allow_na = TRUE,
  min_size = NULL,
  max_size = NULL,
  levels = NULL,
  to_na = character(),
  x_arg = caller_arg(x),
  call = caller_env(),
  x_class = object_type(x)
)

stabilize_fct_scalar(
  x,
  ...,
  allow_null = TRUE,
  allow_zero_length = TRUE,
  allow_na = TRUE,
  levels = NULL,
  to_na = character(),
  x_arg = caller_arg(x),
  call = caller_env(),
  x_class = object_type(x)
)

to_fct(
  x,
  allow_null = TRUE,
  levels = NULL,
  to_na = character(),
  x_arg = caller_arg(x),
  call = caller_env(),
  x_class = object_type(x)
)

to_fct_scalar(
  x,
  allow_null = TRUE,
  allow_zero_length = TRUE,
  levels = NULL,
  to_na = character(),
  x_arg = caller_arg(x),
  call = caller_env(),
  x_class = object_type(x)
)

Arguments

x

The argument to stabilize.

...

These dots are for future extensions and should be empty.

allow_null

Logical. Is NULL an acceptable value?

allow_na

Logical. Are NA values ok?

min_size

Integer. The minimum size of the object. Object size will be tested using vctrs::vec_size().

max_size

Integer. The maximum size of the object. Object size will be tested using vctrs::vec_size().

levels

Character. Expected levels. If NULL (default), the levels will be computed by base::factor().

to_na

Character. Values to coerce to NA.

x_arg

Character. An argument name for x. The automatic value will work in most cases, or pass it through from higher-level functions to make error messages clearer in unexported functions.

call

The execution environment of the call. See the call argument of rlang::abort() for more information.

x_class

Character. The class name of x to use in error messages. Use this if you remove a special class from x before checking its coercion, but want the error message to match the original class.

allow_zero_length

Logical. Are zero-length vectors acceptable?

Value

The argument as a factor.

Details

These functions have important differences from base::as.factor() and base::factor():

  • Values are never silently coerced to NA unless they are explicitly supplied in the to_na argument.

  • NULL values can be rejected as part of the call to this function (with allow_null = FALSE).

Examples

to_fct("a")
#> [1] a
#> Levels: a
to_fct(1:10)
#>  [1] 1  2  3  4  5  6  7  8  9  10
#> Levels: 1 2 3 4 5 6 7 8 9 10
to_fct(NULL)
#> NULL
try(to_fct(letters[1:5], levels = c("a", "c"), to_na = "b"))
#> Error in eval(expr, envir, enclos) : 
#>   All values of `letters[1:5]` must be present in `levels` or `to_na`.
#>  Bad values: d and e.

to_fct_scalar("a")
#> [1] a
#> Levels: a
try(to_fct_scalar(letters))
#> Error in eval(expr, envir, enclos) : 
#>   `letters` must be a single <factor>.
#>  `letters` has 26 values.

stabilize_fct(letters)
#>  [1] a b c d e f g h i j k l m n o p q r s t u v w x y z
#> Levels: a b c d e f g h i j k l m n o p q r s t u v w x y z
try(stabilize_fct(NULL, allow_null = FALSE))
#> Error in eval(expr, envir, enclos) : 
#>   `NULL` must not be <NULL>.
try(stabilize_fct(c("a", NA), allow_na = FALSE))
#> Error in eval(expr, envir, enclos) : 
#>   `c("a", NA)` must not contain NA values.
#>  NA locations: 2
try(stabilize_fct(c("a", "b", "c"), min_size = 5))
#> Error in eval(expr, envir, enclos) : 
#>   `c("a", "b", "c")` must have size >= 5.
#>  3 is too small.
try(stabilize_fct(c("a", "b", "c"), max_size = 2))
#> Error in eval(expr, envir, enclos) : 
#>   `c("a", "b", "c")` must have size <= 2.
#>  3 is too big.

stabilize_fct_scalar("a")
#> [1] a
#> Levels: a
try(stabilize_fct_scalar(letters))
#> Error in eval(expr, envir, enclos) : 
#>   `letters` must be a single <factor>.
#>  `letters` has 26 values.
try(stabilize_fct_scalar("c", levels = c("a", "b")))
#> Error in eval(expr, envir, enclos) : 
#>   All values of `"c"` must be present in `levels` or `to_na`.
#>  Bad values: c.