If error, next iteration in a for loop in R

Question

I'm looking for a simple way to move on to the next iteration in a for loop in R if the operation inside the for loop errors.

I've recreated a simple case below:

for(i in c(1, 3)) {
  test <- try(i+1, silent=TRUE)
  calc <- if(class(test) %in% 'try-error') {next} else {i+1}
  print(calc)
}

This correctly gives me the following calc values.

[1] 2
[1] 4

However once I change the vector in i to include a non-numeric value:

for(i in c(1, "a", 3)) {
  test <- try(i+1, silent=TRUE)
  calc <- if(class(test) %in% 'try-error') {next} else {i+1}
  print(calc)
}

This for loop doesn't work. I was hoping for the same calc values as above with the vector excluding the non-numeric value in i.

I tried using tryCatch as the following:

for(i in c(1, "a", 3)) {
  calc <- tryCatch({i+1}, error = function(e) {next})
  print(calc)
}

However, I get the following error:

Error in value[[3L]](cond) : no loop for break/next, jumping to top level

Could someone please help me understand how I could achieve this using a for loop in R?

You might want to examine what c(1, "a", 3) actually is. I think you believe that only the middle element is character but that is incorrect. — Dason
– Dason, Commented Sep 7, 2016 at 14:36
I'm not sure why you're trying to do this though. Is this a simplified example? If it isn't then just using as.numeric on the input vector will make the characters into NA. Play around with it as.numeric(c(1, "k", "3")) + 1 — Dason
– Dason, Commented Sep 7, 2016 at 14:38
@Dason I'm actually more interested in error handling of the control flow within the for loop example as this is just a simplified case of a larger script I have. I only violated the definition of vector in i by putting a character element to recreate a scenario that throws out an 'Error'. Hope this clarifies my question. — ed0reddie
– ed0reddie, Commented Sep 7, 2016 at 14:50
The issue you created with your example though is that you created all errors. You didn't leave a good case. — Dason
– Dason, Commented Sep 7, 2016 at 14:59
What output do you want from this? Do you want a vector of where the errors are, or a vector of the valid results? — Jonno Bourne
– Jonno Bourne, Commented Sep 7, 2016 at 17:52

jakub · Accepted Answer · 2016-09-07 14:55:02Z

1

As Dason noted, an atomic vector really is not the best way of storing mixed data types. Lists are for that. Consider the following:

l = list(1, "sunflower", 3)

for(i in seq_along(l)) {
   this.e = l[[i]]
   test <- try(this.e + 1, silent=TRUE)
   calc <- if(class(test) %in% 'try-error') {next} else {this.e + 1}
   print(calc)
}

[1] 2
[1] 4

In other words, your former loop "worked". It just always failed and went to next iteration.

answered Sep 7, 2016 at 14:55

jakub

5,1744 gold badges33 silver badges49 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Jonno Bourne · Accepted Answer · 2016-09-08 09:34:14Z

Here is a solution using the "purr" package that might be helpful. It goes through your list or vector and returns the elements that will cause errors

#Wrap the function you want to use in the adverb "safely" 
safetest <- safely(function(x){ifelse(is.na(as.numeric(x)),
                                  x+1,
                                  as.numeric(x)+1)})

myvect<-c(1,"crumbs",3) #change to list if you want a list

#Use the safe version to find where the errors occur
check <- myvect  %>% 
  map(safetest) %>%
  transpose %>% .$result %>% 
  map_lgl(is_null)

myvect[check]

#This returns the results that did not through an error
#first remove NULL elements then flatten to double.
#The two flatten expresiion can be replaced by a single unlist
myvect  %>% 
  map(safetest) %>%
  transpose %>% .$result %>% 
  flatten()%>%flatten_dbl()

see https://blog.rstudio.org/2016/01/06/purrr-0-2-0/ for the original example.

Collectives™ on Stack Overflow

If error, next iteration in a for loop in R

2 Answers 2

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related