Supplying several methods in a loop and saving the output - r

Question

EDIT:Actually looking at the method shows it only uses one method, how can I write a loop to select a method given a vector of methods?!
I've looked at a couple of questions(Loop in R: how to save the outputs?) but can't seem to figure out how to save the output of these models. Here is my "function":
library(caret)
Control<-trainControl(method="cv",number=5)
metric<-"Accuracy"
modeler<-function(df,yname,xname,method,metric,control){
df<-df
methods1<-method
f1<-as.formula(paste0(yname,"~",xname))
for(method in methods1){
fit<-do.call("train",list(data=quote(df),f1,metric=metric,trControl=control,
method=method))
}
fit
}
An implementation of the "function":
methods1<-c("rf","rpart")
modeler(iris,yname="Species",xname=".",method = methods1,control=Control,
metric = metric)
Now I could save the above like:
mod1<-modeler(iris,yname="Species",xname=".",method = methods1,control=Control,
metric = metric)
And then:
sapply(mod1,"[",1)
This works but returns only the last model and in not the best of displays.
How can I optimise this process?

Answer 1

Here is an abridge answer for the key parts.
With methods1 as is vector of methods one can use the lapply function:
methods1<-c("rf","rpart")
#use lapply to loop through the methods
#fit will be a list of the results from modeler
fit<-lapply(methods1, function(met) {
modeler(iris,yname="Species",xname=".",method = met,control=Control, metric = metric)
})
If the calculations are time consuming the lapply function can easily made to run in parallel with the parallel package.

Answer 2

Based on #Dave2e 's comment: Here is how I've managed to "solve" it:
library(caret)
Control<-trainControl(method="cv",number=5)
metric<-"Accuracy"
modeler<-function(df,yname,xname,method,metric,control){
df<-df
methods1<-method
f1<-as.formula(paste0(yname,"~",xname))
for(method in methods1){
fit<-do.call("train",list(data=quote(df),f1,metric=metric,trControl=control,
method=methods1))
}
data.frame(fit$results,Type=fit$modelInfo$tags[1])
}
Implement:
methods1<-c("rf","rpart")
lapply(methods1, function(met) {modeler(iris,yname="Species",xname=".",
method = met,control=Control, metric = metric)})
Result with no seed: This is still lacking but it works so;
[[1]]
mtry Accuracy Kappa AccuracySD KappaSD Type
1 2 0.9533333 0.93 0.03800585 0.05700877 Random Forest
2 3 0.9533333 0.93 0.03800585 0.05700877 Random Forest
3 4 0.9533333 0.93 0.03800585 0.05700877 Random Forest
[[2]]
cp Accuracy Kappa AccuracySD KappaSD Type
1 0.00 0.9533333 0.93 0.02981424 0.04472136 Tree-Based Model
2 0.44 0.7733333 0.66 0.14605935 0.21908902 Tree-Based Model
3 0.50 0.3333333 0.00 0.00000000 0.00000000 Tree-Based Model