How create series of indexing matrix with R? - r

Question

I would like to create a for-loop in order to generate several tables called S1, S2...
For the moment, this is what I have:
S1<-int[int$IdVar == "X1",1:n]
S2<-int[int$IdVar == "X2",1:n]
S3<-int[int$IdVar == "X3",1:n]
S4<-int[int$IdVar == "X4",1:n]
S5<-int[int$IdVar == "X5",1:n]
S6<-int[int$IdVar == "X6",1:n]
S7<-int[int$IdVar == "X7",1:n]
But I can have more or less factors for IdVar variable. I have to add or suppress lines...which is not very efficient!
Could you help me please to find the best way to create my loop ?
I hope I have made this sufficiently clear.
Thank you very much for your help,

Answer 1

We could loop the "X1:X7" values using lapply and subset the 'int' dataset. Change the names of the list elements as S1:S7
lst <- setNames(lapply(paste0('X', 1:7), function(x)
int[int$IdVar==x, 1:n]), paste0("S", 1:7))
It would be better to keep the datasets within the list rather than creating individual objects in the global environment. If you want to do this, then list2env can be used.
list2env(lst, envir=.GlobalEnv)
Another option is using assign with a for loop.

Answer 2

A sillier but nonetheless effective solution:
Get R to automatically write a script to assign the variables and then source that script.
for(s in 1:7){
write(paste0("S", s, "<-int[int$IdVar == \"X", s, "\",1:n])", "tmp.R", append = T)
}
source("tmp.R")
file.remove("tmp.R")
Such an approach I also find useful for referring back to variables that you've created in an automated manner (as you are doing). You can use assign to create variables, but not to refer back to them afterwards. (If anyone knows another way to do that I'd be interested).

Answer 3

Consider using a list instead of creating global variables. 'split' will do what you want, and if you don't like the names "X1", you can change then to "S1" easily. This will also handle all the values in 'IdVar' without you having to specify them:
> n <- 100
> int <- data.frame(IdVar = sample(paste0("X", 1:7), n, TRUE)
+ , val = runif(n)
+ , stringsAsFactors = FALSE
+ )
> # split into a list
> S <- split(int, int$IdVar)
>
> str(S)
List of 7
$ X1:'data.frame': 13 obs. of 2 variables:
..$ IdVar: chr [1:13] "X1" "X1" "X1" "X1" ...
..$ val : num [1:13] 0.104 0.515 0.135 0.501 0.94 ...
$ X2:'data.frame': 18 obs. of 2 variables:
..$ IdVar: chr [1:18] "X2" "X2" "X2" "X2" ...
..$ val : num [1:18] 0.7697 0.6108 0.9354 0.2199 0.0235 ...
$ X3:'data.frame': 16 obs. of 2 variables:
..$ IdVar: chr [1:16] "X3" "X3" "X3" "X3" ...
..$ val : num [1:16] 0.758 0.347 0.48 0.781 0.157 ...
$ X4:'data.frame': 11 obs. of 2 variables:
..$ IdVar: chr [1:11] "X4" "X4" "X4" "X4" ...
..$ val : num [1:11] 0.658 0.247 0.515 0.731 0.114 ...
$ X5:'data.frame': 15 obs. of 2 variables:
..$ IdVar: chr [1:15] "X5" "X5" "X5" "X5" ...
..$ val : num [1:15] 0.502 0.71 0.394 0.738 0.147 ...
$ X6:'data.frame': 14 obs. of 2 variables:
..$ IdVar: chr [1:14] "X6" "X6" "X6" "X6" ...
..$ val : num [1:14] 0.687 0.625 0.705 0.468 0.382 ...
$ X7:'data.frame': 13 obs. of 2 variables:
..$ IdVar: chr [1:13] "X7" "X7" "X7" "X7" ...
..$ val : num [1:13] 0.903 0.466 0.558 0.799 0.527 ...