Skip to content

is.na()<- on a character vector

7 messages · Göran Broström, Eric Berger, Jeff Newmiller +2 more

Original
Göran Broström
# 
I'm confused:
[1] NA  2

OK, but
A 
"A" "B"  NA  

What happens?

G_ran
Eric Berger
# 
maybe this will make it clear

x <- c(a=1, b=5)
is.na(x) <- "b"

i.e. your second case is dealing with a named vector

HTH,
Eric


On Fri, Dec 16, 2022 at 8:29 PM G?ran Brostr?m <goran.brostrom at umu.se>
wrote:

  
  


  


      

    

    

      
      

        


  

        

      Jeff Newmiller
    

    

      
      #
    

  


  

      
I don't find _either_ of these acceptable.

On the other hand,

    x[ is.na( x ) ] <- 1

should have no effect on x.

        
On December 16, 2022 10:28:52 AM PST, "G?ran Brostr?m" <goran.brostrom at umu.se> wrote:

            

      
      
      
    

        

      
      
    

        

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Bert Gunter
    

    

      
      #
    

  


  

      
To do what the OP wanted, one can use:

x <- c('A', 'B')
is.na(x) <- x=='A'  ## rhs is a logical index vector

## yielding

            

      
      
      
    

        
[1] NA  "B"

Cheers,
Bert

        
On Fri, Dec 16, 2022 at 10:43 AM Eric Berger <ericjberger at gmail.com> wrote:

        

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

        

  
  


  


      

    

    

      
      

        


  

        

      Göran Broström
    

    

      
      #
    

  


  

      
Thanks to all. I was cofused and forgot that in is.na(x) <- value,
value is an index vector. Clearly stated on the help page. So Bert's 
suggestion is the right one.


Den 2022-12-16 kl. 19:44, skrev Jeff Newmiller:

            

      
      
      
    

        

      
      
    

            

      
      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Bill Dunlap
    

    

      
      #
    

  


  

      
I think that
   is.na(x) <- i
generally does the same to x as does
   x[i] <- NA

I say 'generally' because some classes (e.g., numeric_version) do not allow
x[i]<-NA but do allow is.na(x)<-i.  It is possible that some classes mess
up this equivalence, but I think that would be considered a bug.

-Bill

On Fri, Dec 16, 2022 at 10:29 AM G?ran Brostr?m <goran.brostrom at umu.se>
wrote:

            

      
      
      
    

        

      
      
    

        

  
  


  


      

    

    

      
      

        


  

        

      Göran Broström
    

    

      
      #
    

  


  

      
Yes, I was confused: In is.na(x) <-value, value is supposed to be an index vector, clearly stated on the help page. So Bert?s suggestion is the way to go.

Thanks to all, G?ran

-----Ursprungligt meddelande-----
Fr?n: Bill Dunlap <williamwdunlap at gmail.com>
Datum: fredag, 16 december 2022 20:15
Till: G?ran Brostr?m <goran.brostrom at umu.se>
Kopia: r-help at r-project.org <r-help at r-project.org>
?mne: Re: [R] is.na()<- on a character vector
I think that is.na <http://is.na/><http://is.na/%3e>(x) <- i
generally does the same to x as does
x[i] <- NA


I say 'generally' because some classes (e.g., numeric_version) do not allow x[i]<-NA but do allow is.na <http://is.na/><http://is.na/%3e>(x)<-i. It is possible that some classes mess up this equivalence, but I think that would be considered a bug.


-Bill

        
On Fri, Dec 16, 2022 at 10:29 AM G?ran Brostr?m <goran.brostrom at umu.se<mailto:goran.brostrom at umu.se> <mailto:goran.brostrom at umu.se<mailto:goran.brostrom at umu.se>>> wrote:

        
I'm confused:

            

      
      
      
    

        
[1] NA 2

OK, but

            

      
      
      
    

        
A
"A" "B" NA

What happens?

G_ran

______________________________________________
R-help at r-project.org<mailto:R-help at r-project.org> <_blank> mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help <https://stat.ethz.ch/mailman/listinfo/r-help><https://stat.ethz.ch/mailman/listinfo/r-help%3e>
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html <http://www.r-project.org/posting-guide.html><http://www.r-project.org/posting-guide.html%3e>
and provide commented, minimal, self-contained, reproducible code.