Create and handle an exception in R - r

Question

I want to have a function where for a particular exceptions it throws and return a message, and then simply check whether what is returned is one of my defined 'exceptions'. For example say I have the function:
divideByX <- function(x){
# If x is NA throws exception
if(is.na(x)){
return(exception('x is NA'))
}
# If x is 0 throws exception
else if(x == 0){
return(exception('Cannot divide by zero'))
}
else{
return(10/x)
}
}
So if x is 0 it returns the expception 'Cannot divide by zero', if x is NA it returns the exception 'x is NA' and for all other values of x it tries to evaluate the expression 10/x.
Then I would want to run something like this:
tempList <- list('a' = 2, 'b' = 0, 'c' = 5, 'd' = NA)
lapply(tempList, function(x){
if(is.exception(x)){
return(x)
}
else{
y <- divideByX(x)
return(y^2)
}
})
So it first checks if the value is one of my defined exceptions and if so returns the message, else it squares my value, so the above should return
$a
[1] 25
$b
[1] 'Cannot divide by zero'
$c
[1] 4
$d
[1] 'x is NA'
Does anyone know the best way to do this? Please let me know if anything is unclear.
Thanks in advance

Answer 1

Create a function to generate exceptions. The exceptions can be a linear hierarchy that extends the simple error class
exception <-
function(class, msg)
{
cond <- simpleError(msg)
class(cond) <- c(class, "MyException", class(cond))
stop(cond)
}
Here's your function
divideByX <- function(x){
# If x is 0 throws exception
if (length(x) != 1) {
exception("NonScalar", "x is not length 1")
} else if (is.na(x)) {
exception("IsNA", "x is NA")
} else if (x == 0) {
exception("DivByZero", "divide by zero")
}
10 / x
}
and use to generate the output you asked for
lapply(tempList, function(x) tryCatch({
divideByX(x)
}, MyException=function(err) {
conditionMessage(err)
}))
or to treat some exceptions differently from others
> lapply(list(NA, 3:5), function(x) tryCatch({
+ divideByX(x)
+ }, IsNA=function(err) {
+ warning(err) # signal a warning, return NA
+ NA
+ }, NonScalar=function(err) {
+ stop(err) # fail
+ }))
Error: x is not length 1
In addition: Warning message:
x is NA

Answer 2

Unless you are always dividing 10, you would want to include the numerator in yoru function.
divideByX <- function(X, num=10) {
if(is.na(X))
return('X is NA')
if(X == 0)
return('Cannot divide by zero')
return(num / X)
}
usage:
y <- 3
lapply(tempList, divideByX, num=y^2)
# $a
# [1] 4.5
#
# $b
# [1] "Cannot divide by zero"
#
# $c
# [1] 1.8
#
# $d
# [1] "X is NA"