R devel - Apr 2007 - ifelse behaviour

Hi!

I'm puzzled by the return value of ifelse

consider

x<-integer(0)
ifelse(is(x, "character"), paste(x), x)
[1] NA

whereas
if (is(x, "character")) return(paste(x)) else x
[1] integer(0)

or
x<-integer(1)
ifelse(is(x, "character"), paste(x), x)
[1] 0

work as I had anticipated. Is this correct behaviour?

Regards,

   Matthias


 >sessionInfo()
R version 2.5.0 (2007-04-23)
i686-pc-linux-gnu

locale:
C

attached base packages:
[1] "stats"     "graphics"  "grDevices"
"datasets"  "utils"     "methods"
[7] "base"

other attached packages:
rcompletion    rcompgen
     "0.1-2"    "0.1-10"


-- 
Matthias Burger                     Project Manager/ Biostatistician
Epigenomics AG    Kleine Praesidentenstr. 1    10178 Berlin, Germany
phone:+49-30-24345-371                          fax:+49-30-24345-555
http://www.epigenomics.com           matthias.burger at epigenomics.com
--
Epigenomics AG Berlin           Amtsgericht Charlottenburg HRB 75861
Vorstand:   Geert Nygaard (CEO/Vorsitzender),  Dr. Kurt Berlin (CSO)
               Oliver Schacht PhD (CFO),  Christian Piepenbrock (COO)
Aufsichtsrat:   Prof. Dr. Dr. hc. Rolf Krebs (Chairman/Vorsitzender)

On 4/26/2007 9:53 AM, ml-it-r-devel at epigenomics.com
wrote:
> Hi!
> 
> I'm puzzled by the return value of ifelse
> 
> consider
> 
> x<-integer(0)
> ifelse(is(x, "character"), paste(x), x)
> [1] NA
The test evaluates to a length 1 logical vector containing FALSE.  So 
ifelse() tries to return the first entry of x.  But x[1] doesn't exist, 
so it evaluates to NA, and that's what ifelse()
returns.
> 
> whereas
> if (is(x, "character")) return(paste(x)) else x
> [1] integer(0)
This is quite different. It says to return the expression in the else 
clause, which is x.
> 
> or
> x<-integer(1)
> ifelse(is(x, "character"), paste(x), x)
> [1] 0
> 
> work as I had anticipated. Is this correct behaviour?
Yes, this is correct.

Duncan Murdoch

A simpler version of your "puzzling call to ifelse" is

ifelse(FALSE, character(0), integer(0))

The most obvious way to satisfy the requirements stated in the 
documentation is to extend integer(0) to length 1 by creating an NA 
value, and that's what you get as a return value (here the 'test' 
argument has length 1, and the 'no' argument has length 0).

-- Tony Plate


ml-it-r-devel at epigenomics.com wrote:
> Hi!
>
> I'm puzzled by the return value of ifelse
>
> consider
>
> x<-integer(0)
> ifelse(is(x, "character"), paste(x), x)
> [1] NA
>
> whereas
> if (is(x, "character")) return(paste(x)) else x
> [1] integer(0)
>
> or
> x<-integer(1)
> ifelse(is(x, "character"), paste(x), x)
> [1] 0
>
> work as I had anticipated. Is this correct behaviour?
>
> Regards,
>
>    Matthias
>
>
>  >sessionInfo()
> R version 2.5.0 (2007-04-23)
> i686-pc-linux-gnu
>
> locale:
> C
>
> attached base packages:
> [1] "stats"     "graphics"  "grDevices"
"datasets"  "utils"     "methods"
> [7] "base"
>
> other attached packages:
> rcompletion    rcompgen
>      "0.1-2"    "0.1-10"
>
>
>