variable number of variables in data frames

-----Original Message-----
From: murphyk at relay.eecs.berkeley.edu
[mailto:murphyk at relay.eecs.berkeley.edu]On Behalf Of Kevin Murphy
Sent: Monday, July 16, 2001 6:19 PM
To: r-help at hypatia.math.ethz.ch; Henrik Bengtsson
Cc: murphyk at cs.berkeley.edu
Subject: variable number of variables in data frames


I would like to specify that certain columns in a data frame should be
treated as ordered factors. I know what numbers these columns are, but
not their names.
How do I do this?

For example, if I know columns 1:4 are to be treated as factors, I can
write

dat <- matrix(c(2,1,1,1, 1,1,1,1), 2, 4)
D <- as.data.frame(dat)
# force all variables to be treated as binary
# regardless of the small data set
D$V1 <- factor(D$V1, 1:2)
D$V2 <- factor(D$V2, 1:2)
D$V3 <- factor(D$V3, 1:2)
D$V4 <- factor(D$V4, 1:2)

But how do I do this in general? What I would like to say is something
like

for (i in my.factor.columns) {
  D$Vi <- factor(D$Vi, 1:my.nlevels[i])
}

Presumably I could do something tricky using eval, but I don't know how.
Besides, I'd prefer to avoid eval, since it is slow.
(Also, I don't want to rely on the fact that the columns are named "Vnn"
by default.)


Kevin

I would like to specify that certain columns in a data frame should be
treated as ordered factors. I know what numbers these columns are, but
not their names.
How do I do this?

For example, if I know columns 1:4 are to be treated as factors, I can
write

dat <- matrix(c(2,1,1,1, 1,1,1,1), 2, 4)
D <- as.data.frame(dat)
# force all variables to be treated as binary
# regardless of the small data set
D$V1 <- factor(D$V1, 1:2)
D$V2 <- factor(D$V2, 1:2)
D$V3 <- factor(D$V3, 1:2)
D$V4 <- factor(D$V4, 1:2)

But how do I do this in general? What I would like to say is something
like

for (i in my.factor.columns) {
  D$Vi <- factor(D$Vi, 1:my.nlevels[i])
}

Presumably I could do something tricky using eval, but I don't know how.
Besides, I'd prefer to avoid eval, since it is slow.
(Also, I don't want to rely on the fact that the columns are named "Vnn"
by default.)

Kevin

dat <- matrix(c(2,1,1,1, 1,1,1,1), 2, 4)
D <- as.data.frame(dat)
makefactor <- function(d,index) { d[index] <- sapply(d[index],factor); return(d)

# say one wants the two first columns to be factors
index <- c(1,2)
str(makefactor(D,index))

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

Kevin, you can treat data frames as matrices, i.e. you can do

  dat <- matrix(c(2,1,1,1, 1,1,1,1), 2, 4)
  D <- as.data.frame(dat)
  for (i in 1:ncol(D)) 
    D[,i] <- factor(D[,i], 1:2)

for (i in my.factor.columns) {
  D$Vi <- factor(D$Vi, 1:my.nlevels[i])
}

D[1:2]<-lapply(D[1:2],factor,levels=1:2)
summary(D)