data.table : column name not working with j = c() and by

Question

Difficult to formulate my question (hence difficult to search efficiently in the archives...). The code below should be clear.

Why in the second command the last column name of the output is not "nb_ob" ? c(lapply(.SD, mean), nb_obs = .N) should provide a named list with 4 elements that should become a column in the final result.

Curiously c(lapply(.SD[,1:4], mean), nb_obs = .N) (third command) provides the intended result. If I remove the by argument (last command) I also obtain the expected column name (with a warning for the character column "Species").

Code run with data.table_1.10.4, R version 3.4.1 for Ubuntu 16.04.3 LTS (I can provide more if needed)

iris <- data.table(iris)
iris[, c(lapply(.SD, mean), nb_obs = .N), by = Species] # 2nd command
#       Species Sepal.Length Sepal.Width Petal.Length Petal.Width  N
# 1:     setosa        5.006       3.428        1.462       0.246 50
# 2: versicolor        5.936       2.770        4.260       1.326 50
# 3:  virginica        6.588       2.974        5.552       2.026 50

iris[, c(lapply(.SD[,1:4], mean), nb_obs = .N), by = Species] # 3rd command
#       Species Sepal.Length Sepal.Width Petal.Length Petal.Width nb_obs
# 1:     setosa        5.006       3.428        1.462       0.246     50
# 2: versicolor        5.936       2.770        4.260       1.326     50
# 3:  virginica        6.588       2.974        5.552       2.026     50

iris[, c(lapply(.SD, mean), nb_obs = .N)] # Fourth command
#    Sepal.Length Sepal.Width Petal.Length Petal.Width Species nb_obs
# 1:     5.843333    3.057333        3.758    1.199333      NA    150

Answer 1

This is an issue in data.table project backlog raised based on this question:

• https://github.com/Rdatatable/data.table/issues/2311

---

This answer is based on @Frank's comment above.