Handling complex formulas with := (data.table assignment by reference)

Question

To calculate some aggregates of x over label and add it to data I could use following code, for mean it is :

library(data.table)  
setDT(data)[, y := mean(x), label]

but how to calculate means only when size of group given by label is over 5 and input 0 otherwise. I was first trying to calculate size of groups using length,nrow instead of mean keyword, but it is not the right way and doesn't work. Sample dataset I work with :

set.seed(123)

data<-data.frame(label=sample(c("A","B"),10,replace=TRUE),x=rnorm(10))
data
#   label          x
#1      A  1.7150650
#2      B  0.4609162
#3      A -1.2650612
#4      B -0.6868529
#5      B -0.4456620
#6      A  1.2240818
#7      B  0.3598138
#8      B  0.4007715
#9      B  0.1106827
#10     A -0.5558411

I see that trying code like :

setDT(data)[, y := ifelse(nrow(x)>10,mean(x),0), label] # don't run

is wrong direction.

Answer 1

I'd suggest you avoid ifesle all together both because efficiency and because it's just wrong to put 0 when you don't want to calculate the mean, what will happen if one of the groups also will have a zero mean, how would you distinguish between them? I'd just do

setDT(data)[, y := mean(x)[.N > 4] , label][]
#     label          x          y
#  1:     A  1.7150650         NA
#  2:     B  0.4609162 0.03327823
#  3:     A -1.2650612         NA
#  4:     B -0.6868529 0.03327823
#  5:     B -0.4456620 0.03327823
#  6:     A  1.2240818         NA
#  7:     B  0.3598138 0.03327823
#  8:     B  0.4007715 0.03327823
#  9:     B  0.1106827 0.03327823
# 10:     A -0.5558411         NA