How to loop lapply to create LAG terms over multiple variables in R

Question

For a model I\'m building, I want to create multiple lag terms for every field/vector in my data table:

For example, with the following data table:

Answer 1

data.table and Map to handle the looping:

vars <- c("b","c")
rpv  <- rep(1:2, each=length(vars))
df[, paste(vars, "lag", rpv, sep="_") := Map(shift, .SD, rpv), by=a, .SDcols=vars]

#   a         b        c   b_lag_1  c_lag_1  b_lag_2  c_lag_2
#1: x 10.863180 393.9568        NA       NA       NA       NA
#2: x  6.139258 537.9199 10.863180 393.9568       NA       NA
#3: x 11.896448 483.8036  6.139258 537.9199 10.86318 393.9568
#4: y 18.079188 509.6136        NA       NA       NA       NA
#5: y  5.463224 233.6991 18.079188 509.6136       NA       NA
#6: y  6.363724 869.8406  5.463224 233.6991 18.07919 509.6136

Answer 2

Here's a way to do it with dplyr:

df %>%
    group_by(a) %>%
    mutate_all(funs(lag1 = lag(., 1), lag2 = lag(., 2)))

Output:

       a         b        c    b_lag1   c_lag1   b_lag2   c_lag2
  <fctr>     <dbl>    <dbl>     <dbl>    <dbl>    <dbl>    <dbl>
1      x  6.663691 689.2483        NA       NA       NA       NA
2      x 11.759130 397.8902  6.663691 689.2483       NA       NA
3      x  3.888010 467.9758 11.759130 397.8902 6.663691 689.2483
4      y  6.221436 355.5437        NA       NA       NA       NA
5      y  2.390940 701.2719  6.221436 355.5437       NA       NA
6      y 17.141815 175.4642  2.390940 701.2719 6.221436 355.5437