library(data.table)
clicks_train <- read.csv("../input/clicks_train.csv")

# set as data table
dt2 = setDT(clicks_train, key = "ad_id")

# calculate summary variables for each group
t = proc.time()
dt3 = dt2[, .(N = .N,
              N2 = sum(clicked = 1)), by = ad_id]
proc.time() - t

nrow(t)
t