FR: lapply + other columns: improving automatic handling of column names

[MichaelChirico]

DT <- data.table(x = 1:5, y = 1:5, z = 6:10)

DT[ , c(list(x), lapply(.SD, `^`, 2)), .SDcols = c("y", "z")]
#    V1  y   z
#1:  1  1  36
#2:  2  4  49
#3:  3  9  64
#4:  4 16  81
#5:  5 25 100

Is conjugate to the approach without .SD:

DT[ , .(x, y^2, z^2)]
#    x V2  V3
#1: 1  1  36
#2: 2  4  49
#3: 3  9  64
#4: 4 16  81
#5: 5 25 100

Seems like it's clear in the first case that V1 should be named x (fixed with c(list(x=x, ...)))

[arunsrinivasan]

possible dup of #618

[MichaelChirico]

@arunsrinivasan certainly related, and the solution may be the same. But distinct from the user's side.

[jangorecki]

I would say the example in first post is fine, if one wants a named column then named list should be provided, not just list(x).
lapply is another thing that we catch and optimize before evaluation, thus dynamic naming make sense.
So the preferred way to achieve what Michael asks for in first post should be

DT <- data.table(x = 1:5, y = 1:5, z = 6:10)
DT[ , c(list(x = x), lapply(.SD, `^`, 2)), .SDcols = c("y", "z")]

I understand this issue is about naming of column c(list(x), ...). Naming of lapply is in #2311. Thus closing, re-open if you strongly disagree about list(x = x) requirement in such cases.