Re: [R] Looping survdiff

[Terry Therneau]

The number of recent questions from umn.edu makes me wonder if there's homework 
involved....
Simpler for your example is to use get and subset.
dat <- structure(.....    as found below
var.to.test <- names(dat)[4:6]   #variables of interest
nvar <- length(var.to.test)
chisq <- double(nvar)
for (i in 1:nvar) {
    tfit <- survdiff(Surv(time, completion==2) ~ get(var.to.test[i]), data=dat, 
subset=(group==3))
    chisq[i] <- tfit$chisq
    }
write.csv(data.frame(var.to.test, chisq))

On 10/19/2012 05:00 AM, r-help-requ...@r-project.org wrote:
Hello,

I am trying to set up a loop that can run the survdiff function with the
ultimate goal to generate a csv file with the p-values reported.  However,
whenever I try a loop I get an error such as "invalid type (list) for
variable 'survival_data_variables[i]".

This is a subset of my data:

structure(list(time = c(1.51666666666667, 72, 72, 25.7833333333333,
72, 72, 72, 72, 72, 72, 1.18333333333333, 0.883333333333333,
1.15, 0.866666666666667, 72, 1.03333333333333, 72, 1.05, 72,
22.7666666666667), group = c(2L, 1L, 3L, 3L, 3L, 4L, 4L,
1L, 3L, 3L, 3L, 3L, 4L, 3L, 3L, 4L, 3L, 4L, 3L, 4L), completion =
structure(c(2L,
1L, 1L, 2L, 1L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 1L, 2L, 1L,
2L, 1L, 2L), .Label = c("1", "2"), class = "factor"), var1 =
structure(c(2L,
2L, 3L, 1L, 1L, 3L, 1L, 1L, 1L, 3L, 2L, 2L, 4L, 3L, 2L, 4L, 2L,
4L, 2L, 4L), .Label = c("1", "2", "3", "4"), class = "factor"),
     var2 = structure(c(3L, 3L, 1L, 1L, 2L, 4L, 3L,
     3L, 2L, 4L, 2L, 1L, 2L, 1L, 2L, 2L, 4L, 4L, 2L, 3L), .Label = c("1",
     "2", "3", "4"), class = "factor"), var3 = structure(c(4L,
     2L, 3L, 1L, 3L, 4L, 4L, 2L, 2L, 4L, 2L, 2L, 1L, 2L, 2L, 2L,
     1L, 3L, 4L, 1L), .Label = c("1", "2", "3", "4"), class = "factor")),
.Names = c("time",
"group", "completion", "var1", "var2",
"var3"), row.names = c(NA, 20L), class = "data.frame")


The loop I have been trying for just group 3 is:

d=data.frame()
for(i in 4:6){
     a=assign(paste("p-value",i,sep=""),
         survdiff(Surv(time, completion=="2")~dat[i],
         data=dat[group=="3",],
         rho=0))
     b=as.matrix(a$chisq)
     d=rbind(d, b)
write.csv(d, file="C:/.../junk.csv", quote=FALSE)}

Perhaps I am making this more difficult than it needs to be.  Thanks for
any help,

Charles

______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.