bestfit nonlinear to a list of dataframes - r

Question

I have a list of data frames and I would like to calculate the nonlinear bestfit to each dataframe in the list and to get a list with the best fit objects. I am trying to use lapply but I am having problems with the parameters.
# define a function for D
fncTtr <- function(n,d) (d/n)*((sqrt(1+2*(n/d))-1))
# define a function for best fit
bestFitD <- function(dat,fun) {
res <- nls(dat$ttr~fun(n,d),data=dat,start=list(d=25),trace=T)
return(res)
}
resL <- lapply(dData2,bestFitD,dat=dData2,fun=fncTtr)
When I execute this code I am getting the following error:
Error in FUN(X[[1L]], ...) : unused argument(s) (X[[1]])
I want the lapply to go thru each dataframe in dData2 and to execute the bestFitD function. How should I define the parameters for the function in lapply. The dData2 object is a list. I am using it as a parameter for bestFitD but this function expects one dataframe as a parameter. How can I define this parameter? When I execute the function bestFitD by itself with one dataframe, it is running correctly. example of a dData2 list with components that are dataframes:
$`1`
n ttr d id
1 35 0.6951 27.739 1
2 36 0.6925 28.072 1
3 37 0.6905 28.507 1
4 38 0.6887 28.946 1
5 39 0.6790 28.003 1
6 40 0.6703 27.247 1
7 41 0.6566 25.735 1
8 42 0.6605 26.981 1
9 43 0.6567 27.016 1
10 44 0.6466 26.026 1
11 45 0.6531 27.667 1
12 46 0.6461 27.128 1
13 47 0.6336 25.751 1
14 48 0.6225 24.636 1
15 49 0.6214 24.992 1
16 50 0.6248 26.011 1
$`2`
n ttr d id
17 35 0.6951 27.739 2
18 36 0.6925 28.072 2
19 37 0.6905 28.507 2
20 42 0.6605 26.981 2
The following code seems to be ok:
res <- bestFitD(dData2[[1]],fncTtr)
but when I execute the following:
res <- bestFitD(dData2[[2]],fncTtr)
I am getting the followin error:
Error in model.frame.default(formula = ~dat + ttr + n, data = dat) :
invalid type (list) for variable 'dat'
Why? Both are dataframes!
But it seems that There is something strange with the second component!

Answer 1

just get rid of the dat$ in your nls function call. i believe it's looking for dat$dat$ttr which obviously will break. That is, your bestFitD function should be:
bestFitD <- function(dat,fun) {
res <- nls(ttr~fun(n,d),data=dat,start=list(d=25),trace=T)
return(res)
}
Now, call using lapply as:
resL <- lapply(dData2, bestFitD, fun = fncTtr)

Answer 2

This should work:
resL <- lapply(dData2, function(x,fun){
bestFitD(x,fun)
},fun='fncTtr')
Where I rewrite,bestFitD using do.call
bestFitD <- function(dat,fun){
nls(ttr~do.call(fun,list(n,d)), data=dat,
start=list(d=25),trace=T)
res
}
0.003237457 : 25
0.0009393089 : 26.77943
0.0009362902 : 26.84895
0.0009362902 : 26.84898
0.001282807 : 25
4.771935e-05 : 27.64267
4.389588e-05 : 27.80729
4.389584e-05 : 27.80781
EDIT
my solution can be simplified to (similar but not exactly to Anthony solution)
lapply(dData2, bestFitD, fun = 'fncTtr')