Check if transformation/coercing of a vector is good enough

This function is primarily aimed to check if the transformation of a vector was successful enough to return the transformed value instead of the original.

exceed_threshold(
  original,
  transformed,
  threshold = 0.9,
  force = FALSE,
  ask = FALSE,
  var_name = "the input vector"
)

Arguments

original	the original vector
transformed	the transformed vector with NA-values for non transformed values
threshold	is a numeric value in [0,1] specifying the proportion of cells in transformed that should be recognized as correctly coerced to accept the new class. This does not effect the function output (except when force = TRUE) but will have some diagnostic benefits.
force	Should a candidate vector (candidate according to threshold) be forced to its suggested class (with non-coercible elements set to NA). FALSE by default but if the function is called interactively, the user will also have the option to set force = TRUE on the fly.
ask	this argument gives you the chance to interactively inspect your data and specify if a column is a date or not, on the fly. This is FALSE by default for as.Dates.default but TRUE for as.Dates.dataframe. It only applies when the function is run interactively and only when force == FALSE.
var_name	a name for the object to be used in messages (you could probably just leave this as default, NULL; it is mostly used for internal purposes!).

Value

Either original or transformed.

Examples

x <- c(rep("2012-01-01", 9), "foo")
exceed_threshold(x, as.Date(x))
#> More than 90 % of the cells in "the input vector" can be coercd to "Date" but the following can not: "foo"
#>  [1] "2012-01-01" "2012-01-01" "2012-01-01" "2012-01-01" "2012-01-01"
#>  [6] "2012-01-01" "2012-01-01" "2012-01-01" "2012-01-01" "foo"       
exceed_threshold(x, as.Date(x), force = TRUE)
#> Warning: the input vector coerced to Date with foo set to NA!
#>  [1] "2012-01-01" "2012-01-01" "2012-01-01" "2012-01-01" "2012-01-01"
#>  [6] "2012-01-01" "2012-01-01" "2012-01-01" "2012-01-01" NA          
exceed_threshold(x, as.Date(x), ask = TRUE)
#> More than 90 % of the cells in "the input vector" can be coercd to "Date" but the following can not: "foo"
#>  [1] "2012-01-01" "2012-01-01" "2012-01-01" "2012-01-01" "2012-01-01"
#>  [6] "2012-01-01" "2012-01-01" "2012-01-01" "2012-01-01" "foo"       
exceed_threshold(x, as.Date(x), threshold = 1)
#>  [1] "2012-01-01" "2012-01-01" "2012-01-01" "2012-01-01" "2012-01-01"
#>  [6] "2012-01-01" "2012-01-01" "2012-01-01" "2012-01-01" "foo"       
exceed_threshold(x, as.Date(x), var_name = "bar", force = TRUE)
#> Warning: bar coerced to Date with foo set to NA!
#>  [1] "2012-01-01" "2012-01-01" "2012-01-01" "2012-01-01" "2012-01-01"
#>  [6] "2012-01-01" "2012-01-01" "2012-01-01" "2012-01-01" NA          

x <- c(1:9, "baz")
exceed_threshold(x, suppressWarnings(as.numeric(x)))
#> More than 90 % of the cells in "the input vector" can be coercd to "numeric" but the following can not: "baz"
#>  [1] "1"   "2"   "3"   "4"   "5"   "6"   "7"   "8"   "9"   "baz"