R help - May 2003 - Creating Dummy Variables with if else phrase

Dear All,

I want to do 52 state dummy variables. In order to make it easier for me, I 
thought of using the below commands. (x is the vector with state variables, 
matrix will correspond to the dummy variables)


x <- c(1,2,NA,4)
matrix <- matrix(0,nrow=4,ncol=4)

for (i in 1:4) {
if (is.real(x[i])) {
matrix[i,x[i]] <- 1 } else {
matrix[i,] <- rep(NA,4)
}
}

This gives me the following matrix:

1  0  0  0
0  1  0  0
0  0  0  0
0  0  0  1

But I want it to look as follows:

1   0  0  0
0   1  0  0
NA NA NA NA
0   0  0  1


I want the forth row to be NA, but it is unchanged, i.e., 0. The second
"else"
phrase does not seem to work, and I cannot see what is wrong.

I would be very grateful for your help.

Regards, Sophie

On Friday 02 May 2003 21:41, langensk at fas.harvard.edu wrote:
> Dear All,
>
> I want to do 52 state dummy variables. In order to make it easier
> for me, I thought of using the below commands. (x is the vector with
> state variables, matrix will correspond to the dummy variables)
>
>
> x <- c(1,2,NA,4)
> matrix <- matrix(0,nrow=4,ncol=4)
>
> for (i in 1:4) {
> if (is.real(x[i])) {
> matrix[i,x[i]] <- 1 } else {
> matrix[i,] <- rep(NA,4)
> }
> }
>
> This gives me the following matrix:
>
> 1  0  0  0
> 0  1  0  0
> 0  0  0  0
> 0  0  0  1
>
> But I want it to look as follows:
>
> 1   0  0  0
> 0   1  0  0
> NA NA NA NA
> 0   0  0  1
>
I don't really understand why you want to create a matrix that way, 
but if you want to do it the way described above
  is.real(x[i])
seems to be the wrong function (because x is a real vector in the 
example above). If you want to test for NA you should use
  !is.na(x[i])
instead.
Z
> I want the forth row to be NA, but it is unchanged, i.e., 0. The
> second "else" phrase does not seem to work, and I cannot see what
is
> wrong.
>
> I would be very grateful for your help.
>
> Regards, Sophie
>
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://www.stat.math.ethz.ch/mailman/listinfo/r-help

langensk at fas.harvard.edu writes:
> Dear All,
> 
> I want to do 52 state dummy variables. In order to make it easier for me, I
> thought of using the below commands. (x is the vector with state variables,
> matrix will correspond to the dummy variables)
> 
> 
> x <- c(1,2,NA,4)
> matrix <- matrix(0,nrow=4,ncol=4)
> 
> for (i in 1:4) {
> if (is.real(x[i])) {
> matrix[i,x[i]] <- 1 } else {
> matrix[i,] <- rep(NA,4)
> }
> }
> 
> This gives me the following matrix:
> 
> 1  0  0  0
> 0  1  0  0
> 0  0  0  0
> 0  0  0  1
> 
> But I want it to look as follows:
> 
> 1   0  0  0
> 0   1  0  0
> NA NA NA NA
> 0   0  0  1
> 
> 
> I want the forth row to be NA, but it is unchanged, i.e., 0. The second
"else"
> phrase does not seem to work, and I cannot see what is wrong.
> 
> I would be very grateful for your help.
is.real is true for all elements in x. You may have intended
!is.na(x[i]).

However, there are simpler ways, e.g.

0+outer(x,1:4,"==")
     [,1] [,2] [,3] [,4]
[1,]    1    0    0    0
[2,]    0    1    0    0
[3,]   NA   NA   NA   NA
[4,]    0    0    0    1

or simply use model formulas and code x as a factor.

-- 
   O__  ---- Peter Dalgaard             Blegdamsvej 3  
  c/ /'_ --- Dept. of Biostatistics     2200 Cph. N   
 (*) \(*) -- University of Copenhagen   Denmark      Ph: (+45) 35327918
~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk)             FAX: (+45) 35327907

There is a bit of a trap for young players involved, but:

	> x <- NA
	> is.real(x)
	[1] FALSE
	> x <- c(1,NA)
	> is.real(x[2])
	[1] TRUE
	> x <- as.real(NA)
	> is.real(x)
	[1] TRUE
	> typeof(x)
	[1] "double"
	> x <- NA
	> typeof(x)
	[1] "logical"

It makes sense when you think it through:  An ``NA'' could be
anything on its own.  When it's an entry of a vector with real
entries then it oughta be a ``not available real''.  It's a bit
subtle when you first bump into it, but it all clicks into place.

Be that as it may --- to avoid the trap, use is.na() if you want
to check whether a value is missing.

				cheers,

					Rolf Turner