基于 R 中变量值的相对行索引

Question

想知道是否有有效且 terse/concise data.table 解决以下问题的方法。

请假设我有以下 data.table:

library(data.table)

DT <- data.table(store = c("A", "A", "A", "A", "B", "B", "B", "B"),
                 time = c(1,2,3,4,1,2,3,4),
                 treat_time = c(0,0,1,0, 0,1,0,0))

这里，treat_time是店铺接受治疗的时间段。请注意，A店和B店的治疗时间不同。我想创建一个列time_rel，它描述了相对于treat_time = 1时的时间段。即data.table 应该是这样的：


DT_outcome <- data.table(store = c("A", "A", "A", "A", "B", "B", "B", "B"),
                 time = c(1,2,3,4,1,2,3,4),
                 treat_time = c(0,0,1,0, 0,1,0,0),
                 time_rel = c(-1,0,1,2, 0,1,2,3))


    store  time treat_time time_rel
   <char> <num>      <num>    <num>
1:      A     1          0       -1
2:      A     2          0        0
3:      A     3          1        1
4:      A     4          0        2
5:      B     1          0        0
6:      B     2          1        1
7:      B     3          0        2
8:      B     4          0        3

谢谢！

Answer 1

我们可以从序列中减去

library(data.table)
DT[,   time_rel := seq_len(.N) - seq_len(.N)[treat_time == 1] + 1 ,store]

-输出

> DT
    store  time treat_time time_rel
   <char> <num>      <num>    <num>
1:      A     1          0       -1
2:      A     2          0        0
3:      A     3          1        1
4:      A     4          0        2
5:      B     1          0        0
6:      B     2          1        1
7:      B     3          0        2
8:      B     4          0        3

或dplyr

中的相同逻辑

library(dplyr)
DT %>% 
  group_by(store) %>%
  mutate(time_rel = row_number() - which(treat_time == 1) + 1) %>% 
  ungroup

基于 R 中变量值的相对行索引

Relative row index based on a variable value in R

r

data.table