R devel - Oct 2016 - Inconsistent behavior of summary

Here is the code:

summary(c(1:5, NA))
summary(c("a", NA))

It seems in the firs case the number of NAs is reported, but not in the
second.

Tested with R 3.3.1 on a mac.


Best regards,

Kaiyin ZHONG

	[[alternative HTML version deleted]]

On 15/10/2016 5:48 AM, Kaiyin Zhong wrote:
> Here is the code:
>
> summary(c(1:5, NA))
> summary(c("a", NA))
>
> It seems in the firs case the number of NAs is reported, but not in the
> second.
>
> Tested with R 3.3.1 on a mac.
summary() is a generic function, so the output depends on the class of 
the input.  There's no special method for character vectors, so you get 
the default output.  Numeric vectors show more detail.

If you want a particular behaviour for character objects, just write a 
summary.character method.

Here's the default:

 > summary(c("a", NA))
    Length     Class      Mode
         2 character character

Here's a method and example using it:

 > summary.character <- function (x, ...) { cat("this character
vector
has", sum(is.na(x)), " NA value(s)\n")}

 > summary(c("a", NA))
this character vector has 1  NA value(s)

Duncan Murdoch