dot alias not understood as expected in by/keyby

[MichaelChirico]

I think there are a couple of issues on similar fronts addressing that we should be more consistent about how/when . is understood in data.table and more specifically within [.data.table calls. Here's a new one I came across today:

DT = data.table(a = 1:10, b = 10:1)
DT[ , mean(b), by = {
  a2 = a^2
  .(a2mod5 = a2 %% 5)
}]

This returns an error since evaluation looks for the . function (i.e., the call has not replaced . with list) (replacing . with list, this evaluates as expected):

Error in .(b2mod5 = b2%%5) : could not find function "."

Contrast this to something similar in j:

DT = data.table(a = 1:10, b = 10:1)
DT[ , {
  b2 = b^2
  .(b2mod5 = sum(b2 %% 5))
}, by = a %% 3]
#    a b2mod5
# 1: 1      6
# 2: 2      6
# 3: 0      8

Some related issues: #2391, #1912, #852, a few others I can't find just now

[Henrik-P]

Perhaps among the 'few others': dot list alias doesn't work with between

[MichaelChirico]

yes, exactly the thread I was looking for, thanks!

…

On Oct 28, 2017 12:12 AM, "Henrik-P" ***@***.***> wrote: Perhaps among the 'few others': dot list alias doesn't work with between <#2315> — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#2447 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AHQQdcYDRAOJZwcJkZ98SxPy3QvZ-Lsbks5swgDmgaJpZM4QIzog> .