sum over past window-size dates per group

Question

The problem is similar to How do I do a conditional sum which only looks between certain date criteria but slightly different and the answer from that does not fit into curr

Answer 1

Here's a (supposedly efficient) solution using the new non-equi joins and the by = .EACHI features in data.table (v1.9.8+)

library(data.table) #v1.10.4

## Convert to a proper date class, and add another column in order to define the range
setDT(input)[, c("Date", "Date2") := {
  Date = as.IDate(Date)
  Date2 = Date - 2L
  .(Date, Date2)
}]

## Run a non-equi join against the unique Date/Group combination in input
## Sum the Scores on the fly
## You can ignore the second Date column 

input[unique(input, by = c("Date", "Group")), ## This removes the dupes
      on = .(Group, Date <= Date, Date >= Date2), ## The join condition
      .(Score = sum(Score)), ## sum the scores
      keyby = .EACHI] ## Run the sum by each row in unique(input, by = c("Date", "Group"))

#     Group       Date       Date Score
#  1:     A 2017-04-01 2017-03-30     1
#  2:     A 2017-04-02 2017-03-31     3
#  3:     A 2017-04-04 2017-04-02     6
#  4:     A 2017-04-07 2017-04-05     7
#  5:     B 2017-04-02 2017-03-31     4
#  6:     B 2017-04-05 2017-04-03     5
#  7:     B 2017-04-08 2017-04-06     8
#  8:     C 2017-04-02 2017-03-31     2
#  9:     C 2017-04-03 2017-04-01     5
# 10:     C 2017-04-06 2017-04-04     6
# 11:     C 2017-04-09 2017-04-07     9