How to create nested list from data frame with two factors?

Question

I'd like to create a nested list from this data frame:

dat <- data.frame(var1 = c("A", "A", "B", "B"),
                  var2 = c("A_1", "A_2", "B_1", "B_2"),
                  val = 1:4)

> dat
  var1 var2 val
1    A  A_1   1
2    A  A_2   2
3    B  B_1   3
4    B  B_2   4

I split the data frame by var1 first:

mylist <- split(dat, dat$var1)
> mylist
$A
  var1 var2 val
1    A  A_1   1
2    A  A_2   2

$B
  var1 var2 val
3    B  B_1   3
4    B  B_2   4

Now I want to create nested lists for var2, I tried:

mylist <- lapply(mylist, function(x) split(x, x$var2))

> mylist
$A
$A$A_1
  var1 var2 val
1    A  A_1   1

$A$A_2
  var1 var2 val
2    A  A_2   2

$A$B_1
[1] var1 var2 val 
<0 Zeilen> (oder row.names mit Länge 0)

$A$B_2
[1] var1 var2 val 
<0 Zeilen> (oder row.names mit Länge 0)


$B
$B$A_1
[1] var1 var2 val 
<0 Zeilen> (oder row.names mit Länge 0)

$B$A_2
[1] var1 var2 val 
<0 Zeilen> (oder row.names mit Länge 0)

$B$B_1
  var1 var2 val
3    B  B_1   3

$B$B_2
  var1 var2 val
4    B  B_2   4

But how can I avoid the empty data frames of inexistent combinations of var1 and var2 being created?

Answer 1

Just wrap the second argument to split in droplevels . This will get rid of the now extraneous factor levels that were constructed in the original data.frame.

lapply(mylist, function(x) split(x, droplevels(x$var2)))
$A
$A$A_1
  var1 var2 val
1    A  A_1   1

$A$A_2
  var1 var2 val
2    A  A_2   2


$B
$B$B_1
  var1 var2 val
3    B  B_1   3

$B$B_2
  var1 var2 val
4    B  B_2   4