[Rd] bug in function rpart (PR#9245)

[yma at biostat.bsc.gwu.edu]

Full_Name: Yong Ma
Version: R 2.3.1
OS: Windows 2000
Submission from: (NULL) (128.164.162.1)


I have a data set containing 3000 observations. while I was trying to build a
survival tree using the function rpart, I kept getting the following error
message.
>tree<-rpart(coxph(Surv(diabv, diabf) ~ weight, data = sgotdat)) 
 Error in "$<-.data.frame"(`*tmp*`, "yval2", value = c(1, 638)) : 
        replacement has 2 rows, data has 1
It was confusing to me because I was able to use the coxph function on the same
data set with no problem at all. So I googled the internet and found some help
from the following link.

http://tolstoy.newcastle.edu.au/R/devel/05/04/0598.html

It suggested that the problem has to do with the length of the data. So I tried
a subset of my original data containing 2500 observations and BINGO it worked! 

However, I need to run the analysis on the complete data set and the patch
proposed by Andy is not clear to me how to implement. Please fix the bug
systematically.