[Rd] Why is any() only defined for a numeric and not logical data.frame?

[Rui Barradas]

Hello,

As it turns out, this is valid for all generics of the ?Summary group. 
 From help("Summary"):

Group "Summary":

all, any

sum, prod

min, max

range

methods("Summary") shows that there is a method for df's.
And the code of Summary.data.frame has an explicit test

        if (!is.numeric(x) && !is.complex(x))
             stop("only defined on a data frame with all numeric variables")


So now the question is even more pertinent (?)(I think):

If sum(<logical vector>) is a valid instruction, why only

sum(<numeric df>)
sum(<complex df>)

are valid?


Rui Barradas

Às 19:23 de 16/02/20, Karolis Koncevičius escreveu:
> Hello,
> 
> I recently stumbled on an unusual behaviour of any() and all() and have 
> been adviced from StackOverflow to share it here [1].
> 
>      df1 <- data.frame(A=TRUE, B=FALSE)
>      df2 <- data.frame(A=1, B=0)
> 
>      > any(df1)
>      Error in FUN(X[[i]], ...):
>        only defined on a data frame  with all numeric variables
> 
>      > any(df2)
>      [1] TRUE
>      Warning message: In any(c(1, 2), na.rm = FALSE):
>        coercing argument of type 'double' to logical
> 
> Same results happen when using all()
> 
> So the any() and all() do not work on data frames with logical values, 
> but work if the values are numeric.
> 
> This doesn't seem to be a bug because error correctly states that any() 
> will only work on an all-numeric data.frame. However the behaviour 
> doesn't seem intentional and I cannot come up with any reason for it 
> behaving this way.
> 
> Maybe any() and all() need to be modified to not work on data.frames() 
> at all, which would also be consistent with is.nan() ?
> 
> [1]: 
> https://stackoverflow.com/questions/60251847/why-is-any-only-defined-for-a-numeric-and-not-logical-data-frame 
> 
> 
> ______________________________________________
> R-devel using r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel