R-saemixData函数在函数环境中找不到数据“文件错误（文件，“ rt”）：无法打开连接”

Question

In R (version 3.3.2) I get the error "Error in file(file, "rt") : cannot open the connection" when I use the function saemixData inside a function. It is as if saemixData could not find the dataframe sim when sim is in the function environment. The code works well when it is run outside of the function f, or when sim is in the global environment... It seems like a bug of the function saemixData, do you have a solution ?

Thanks !

rm(list=ls())
library(saemix)  # saemix_2.1.tar.gz


f=function(){

    sim = data.frame(patient=c(1,1,1,2,2,2),
                          time=c(1,4,8,1,4,8),
                          HBA1C_obs=c(9,8,7,8,7.5,6))   

    saemix.data <- saemixData(name.data       = sim,
                              name.group      = "patient",
                              name.predictors =  c("time"),
                              name.response   = "HBA1C_obs")  

    saemix.data
}


f()

# Reading data from file sim 
# Error in file(file, "rt") : cannot open the connection
# Error in read(x) : 
#   The file sim does not exist. Please check the name and path. 

Answer 1

I also have this same problem. As a workaround you can assign sim to the global environment from within the function by using a double arrow:

sim <<- data.frame(patient=c(1,1,1,2,2,2),
                   time=c(1,4,8,1,4,8),
                   HBA1C_obs=c(9,8,7,8,7.5,6))

But it's only a poor fix as it will overwrite any sim object in the global environment each time the function is called.

Answer 2

yes I'm aware of this issue which you can chalk down to an insufficient understanding on my part of the environment issues in R. I wanted a function that could accept both objects in the environment and files on disk. I'm very sorry about this and if anyone has a better way of coding the function saemix.data please feel free to send it !

Another workaround is to write data to a temporary file although this is clearly inefficient if you have a large file:

f=function(){
    sim = data.frame(patient=c(1,1,1,2,2,2),
                      time=c(1,4,8,1,4,8),
                      HBA1C_obs=c(9,8,7,8,7.5,6))
    tempname<-tempfile()
    write.table(sim,tempname,quote=F,col.names=T)
    saemix.data <- saemixData(name.data       = tempname,
                          name.group      = "patient",
                          name.predictors =  c("time"),
                          name.response   = "HBA1C_obs")
    saemix.data
}

f()

Emmanuelle