How To Count The Number Of Elements In A Column Other Than 'X' In R

Question

[Don]

I have a column that contains several "NA" while the other are alphanumeric. I want to count the elements that are NOT "NA". I used

M1<-length(M11$DAID[M11$DAID =="NA"])

to count the NA but

NM1<-length(M11$DAID[M11$DAID !="NA"])

wouldn't work for non-NA. Thanks for your help

Answer 1

[Stefano Berri]

The problem with your code is that NA are not character, but NA.

> a <- c("Kat", "cat", "Kat", NA) 
> length(which(a == "Kat")) 
[1] 2 
>lengthwhichis.na(a)))
[1] 1 
> length(which(a != "Kat")) 
[1] 1 
> length(which(!is.na(a)))
[1] 3

Above, is how you should do it, however

> length(which(a != "NA")) 
[1] 3
> length(which(a == "NA")) 
[1] 0

which might look strange. The first seem to work, the second doesn't. But they are both "wrong".

Basically this happens because in the first case there are three elements that are not character "NA". NA is not a character, so it does not count In the second case, there are no elements that match with character "NA". Once again, NA is not a character, so it does not count.

Further reads: http://cran.r-project.org/doc/manuals/R-intro.html#Missing-values (the whole "book" is worth having in the bookmarks when working with R)

Answer 2

[Jeremy Leipzig]

> a<-c(1,2,NA,4)
> lengthwhichis.na(a)))
[1] 1
> length(which(!is.na(a)))
[1] 3

[Don]

hi Jeremy, what if it was "Kat" rather than NA. is there an opposite for =="kat" (ie count every other element but kat)? Thanks

[Stefano Berri]

> a <- c("Kat", "cat", "Kat")
length(which(a == "Kat"))
[1] 2
length(which(a != "Kat"))
[1] 1

NA are not character, but NA, that is, I think, why your code does not work.