Bug in as.data.frame.Node: NA for certain lists slots

[thothal]

There are some cases of list fields where print (respectively the underlying as.data.frame.Node) does not work as intended:

library(data.tree)
data(acme)
## Problem: elements which are length 1 _lists_ (Case #1)
acme$Set(bad1 = list(list(list(a = 1)), 
                     list(list(a = 1, b = 2))), 
         filterFun = isNotRoot)
## Problem: elements will be transformed to matrix (Case #2)
acme$Set(bad2 = list(list(a = 1, b = 2)))

print(acme, "bad1") ## prints but shows NA and warning
#                           levelName bad1
# 1  Acme Inc.                          NA
# 2   ¦--Accounting                     NA
# 3   ¦   ¦--New Software               NA
# 4   ¦   °--New Accounting Standards   NA
# 5   ¦--Research                       NA
# 6   ¦   ¦--New Product Line           NA
# 7   ¦   °--New Labs                   NA
# 8   °--IT                             NA
# 9       ¦--Outsource                  NA
# 10      ¦--Go agile                   NA
# 11      °--Switch to R                NA
# Warning message:
# In `[<-.data.frame`(`*tmp*`, colName, value = list(`Acme Inc.` = NA,  :
#   provided 11 variables to replace 1 variables

print(acme, "bad2") ## doesn't even print
# Error in `[<-.data.frame`(`*tmp*`, colName, value = c(1, 2, 1, 2, 1, 2,  : 
#   replacement has 22 rows, data has 11

The error comes from these lines.

• Case #1: stems from the fact that sapply creates a list, because for a length(.) == 1 object it simply returns the object, which is in this case a length 1 list, thus the result must eb a list as well
• Case #2: comes from teh fact that Get does some simplification by default, and transforms the output to a 2 x # of nodes matrix, which is then treated as a vector by sapply.

Solution to this problem is to use Get with simplify = FALSE, retort to the more predictable lapply and doing an unlist in the end.