What makes lapply interpret a data frame as atomic or recursive?

Question

I am curious to understand why there is a difference in the following code. Where mylist is a list of data frames and a single entry in the list of the "try-error" class. This problem does not happen if the list is only filled with data frames.

mylist = split(iris, iris$Species)
mylist[['error-entry']] = try(errorplease)
Error in try(erlarorplease) : object 'errorplease' not found

Then when I run lapply

lapply(mylist, FUN=function(x){summary(x$Species)}

Which returns the following error:

Error in x$score : $ operator is invalid for atomic vectors

After reading a bit I've found that I can circumvent this issue by using

lapply(mylist, FUN=function(x){summary(x["Species"])})

But real question here is that if I enter to the debugging browser while on the function and the ask for the summary of x$Species, it works!

lapply(mylist, FUN=function(x){browser();summary(x$Species)})
Called from: FUN(X[[1L]], ...)
Browse[1]> summary(x$Species) 
[1] setosa setosa setosa setosa setosa setosa setosa setosa setosa setosa
[11] setosa setosa setosa setosa setosa setosa setosa setosa setosa setosa
[21] setosa setosa setosa setosa setosa setosa setosa setosa setosa setosa
[31] setosa setosa setosa setosa setosa setosa setosa setosa setosa setosa
[41] setosa setosa setosa setosa setosa setosa setosa setosa setosa setosa

What fundamentals am I missing? Is it that when I access the browser I traverse the list in a different order? Guessing this after @Roland 's help

You should provide a [reproducible example](http://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example). This sounds more like you are not using the variable you think you are or a typo rather than a fundamental R misunderstanding. — MrFlick, Dec 09 '15 at 16:19
Sorry guys I made the typical mistake of copying some code and typing the other. Fixing it now. — pedrosaurio, Dec 09 '15 at 16:21
Yes @Roland, it is a list of data.frames. Edited question for clarity. — pedrosaurio, Dec 09 '15 at 16:27
Well, then we have to wait until you produce the reproducible example. — Roland, Dec 09 '15 at 16:37
Hmmm the toy example doesn't show the error - classic! Should I close the question while I get this working? — pedrosaurio, Dec 09 '15 at 16:40
Are you really, really sure that you have a list of data.frames and not matrices? Please check `sapply(mylist, class)`. — Roland, Dec 09 '15 at 16:45
All but one which is a "try-error" class. I do this on purpose for other reasons. — pedrosaurio, Dec 09 '15 at 16:49
Ok, That's the problem. I added a broken entry to the toy example list by using `try()` so now not all of them are data.frames and this reproduces the error. — pedrosaurio, Dec 09 '15 at 16:52
Finished formatting the question for future reference of other users. — pedrosaurio, Dec 09 '15 at 17:03

What makes lapply interpret a data frame as atomic or recursive?

0 Answers0