R help - Mar 2020 - Some seemingly odd behavior of survfit in library survival

I ran across an issue that looks like variable scoping in survfit is not acting
as I would expect.  Here's a minimal example:

new<-as.data.frame(list(t=1:4,d=rep(1,4),s=c(1,0,0,1)))
library(survival)
mine<-function(ff,df){
        fit<-coxph(ff,data=df)
        out<-survfit(fit,df)
}
mine(as.formula("Surv(t,d)~s"),new)

I would expect this to fit the proportional hazards regression model using
formula Surv(t,d)~s, using data set new, and then calculate a separate fitted
survival curve for each member of the data set.  Instead I get an error

Error in eval(predvars, data, env) :
  invalid 'envir' argument of type 'closure'

The code runs without error if I modify it by copying the data set new to the
local variable within the function mine before running:

new<-as.data.frame(list(t=1:4,d=rep(1,4),s=c(1,0,0,1)))
library(survival)
mine<-function(ff,df){
        fit<-coxph(ff,data=df)
        out<-survfit(fit,df)
}
df<-new
mine(as.formula("Surv(t,d)~s"),new)

which leads me to believe that there's some variable scoping error.  Can
anyone point out what I'm doing wrong?  Thanks, John

	[[alternative HTML version deleted]]

The coxph function appears to rely on finding the name of the data argument in
the environment in which the formula was created. The lm function does not have
this problem.

Oh, and df is the name of the F distribution density function, which explains
why the error complained about a "closure".

new<-as.data.frame(list(t=1:4,d=rep(1,4),s=c(1,0,0,1)))
library(survival)
mine<-function(ff,df){
        fit<-coxph(as.formula(ff),data=df)
         survfit(fit,df)
}
mine("Surv(t,d)~s",new)


On March 16, 2020 7:23:26 PM PDT, John Kolassa <kolassa at
stat.rutgers.edu> wrote:
>I ran across an issue that looks like variable scoping in survfit is
>not acting as I would expect.  Here's a minimal example:
>
>new<-as.data.frame(list(t=1:4,d=rep(1,4),s=c(1,0,0,1)))
>library(survival)
>mine<-function(ff,df){
>        fit<-coxph(ff,data=df)
>        out<-survfit(fit,df)
>}
>mine(as.formula("Surv(t,d)~s"),new)
>
>I would expect this to fit the proportional hazards regression model
>using formula Surv(t,d)~s, using data set new, and then calculate a
>separate fitted survival curve for each member of the data set. 
>Instead I get an error
>
>Error in eval(predvars, data, env) :
>  invalid 'envir' argument of type 'closure'
>
>The code runs without error if I modify it by copying the data set new
>to the local variable within the function mine before running:
>
>new<-as.data.frame(list(t=1:4,d=rep(1,4),s=c(1,0,0,1)))
>library(survival)
>mine<-function(ff,df){
>        fit<-coxph(ff,data=df)
>        out<-survfit(fit,df)
>}
>df<-new
>mine(as.formula("Surv(t,d)~s"),new)
>
>which leads me to believe that there's some variable scoping error. 
>Can anyone point out what I'm doing wrong?  Thanks, John
>
>	[[alternative HTML version deleted]]
>
>______________________________________________
>R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
>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.
-- 
Sent from my phone. Please excuse my brevity.

Jeff Newmiller wrote/hat geschrieben on/am 17.03.2020
05:39:
> The coxph function appears to rely on finding the name of the data argument
in the environment in which the formula was created. The lm function does not
have this problem.
>
> Oh, and df is the name of the F distribution density function, which
explains why the error complained about a "closure".
>
> new<-as.data.frame(list(t=1:4,d=rep(1,4),s=c(1,0,0,1)))
> library(survival)
> mine<-function(ff,df){
>         fit<-coxph(as.formula(ff),data=df)
>          survfit(fit,df)
> }
> mine("Surv(t,d)~s",new)
>
Therefore a workaround could be to use the character string for the
formula as argument and apply as.formula() within the function.

mine2 <-function(fstr,df){
         fit<-coxph(as.formula(fstr),data=df)
         out<-survfit(fit,df)
         out
}

mine2("Surv(t,d)~s",new)

Heinz
>
> On March 16, 2020 7:23:26 PM PDT, John Kolassa <kolassa at
stat.rutgers.edu> wrote:
>> I ran across an issue that looks like variable scoping in survfit is
>> not acting as I would expect.  Here's a minimal example:
>>
>> new<-as.data.frame(list(t=1:4,d=rep(1,4),s=c(1,0,0,1)))
>> library(survival)
>> mine<-function(ff,df){
>>        fit<-coxph(ff,data=df)
>>        out<-survfit(fit,df)
>> }
>> mine(as.formula("Surv(t,d)~s"),new)
>>
>> I would expect this to fit the proportional hazards regression model
>> using formula Surv(t,d)~s, using data set new, and then calculate a
>> separate fitted survival curve for each member of the data set.
>> Instead I get an error
>>
>> Error in eval(predvars, data, env) :
>>  invalid 'envir' argument of type 'closure'
>>
>> The code runs without error if I modify it by copying the data set new
>> to the local variable within the function mine before running:
>>
>> new<-as.data.frame(list(t=1:4,d=rep(1,4),s=c(1,0,0,1)))
>> library(survival)
>> mine<-function(ff,df){
>>        fit<-coxph(ff,data=df)
>>        out<-survfit(fit,df)
>> }
>> df<-new
>> mine(as.formula("Surv(t,d)~s"),new)
>>
>> which leads me to believe that there's some variable scoping error.
>> Can anyone point out what I'm doing wrong?  Thanks, John
>>
>> 	[[alternative HTML version deleted]]
>>
>> ______________________________________________
>> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
>> 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.
>