基于条件 R 的列值顺序增加

Question

我有一个 R 数据框，它有一个 ID 列，其中包含一个 ID 的多个记录。当 ID 的标志设置为 1 时，我想创建一个从 1 开始并以 6（1,6,12 ...）为增量顺序增加的新时间轴列。如何使用 dplyr 在 R 中实现此目的？

下面是一个示例数据框

ID	Timepoint	Flag
A	0	0
A	6	0
A	12	0
A	18	1
A	24	0
A	30	0
A	36	0

预期数据帧

ID	Timepoint	Flag	New_Timepoint
A	0	0
A	6	0
A	12	0
A	18	1	1
A	24	0	6
A	30	0	12
A	36	0	18

Answer 1

一个选项是按 'ID' 分组，创建 'Timepoint' 的 lag 并将 n 指定为 'Flag' 的位置，其中值是 1 (-1)

library(dplyr)
df1 %>% 
   group_by(ID) %>% 
   mutate(New_Timepoint = dplyr::lag(replace(Timepoint, !Timepoint, 1),
           n = which(Flag == 1)-1)) %>%
   ungroup

-输出

# A tibble: 7 x 4
#  ID    Timepoint  Flag New_Timepoint
#  <chr>     <int> <int>         <dbl>
#1 A             0     0            NA
#2 A             6     0            NA
#3 A            12     0            NA
#4 A            18     1             1
#5 A            24     0             6
#6 A            30     0            12
#7 A            36     0            18

或者使用双cumsum创建索引

df1 %>% 
   group_by(ID) %>% 
   mutate(New_Timepoint = Timepoint[na_if(cumsum(cumsum(Flag)), 0)]) %>%
   ungroup

数据

df1 <- structure(list(ID = c("A", "A", "A", "A", "A", "A", "A"), 
    Timepoint = c(0L, 
6L, 12L, 18L, 24L, 30L, 36L), 
     Flag = c(0L, 0L, 0L, 1L, 0L, 0L, 
0L)), class = "data.frame", row.names = c(NA, -7L))

Answer 2

另一个dplyr选项

df %>%
  group_by(ID) %>%
  mutate(New_Timepoint = pmax(1, Timepoint - c(NA, Timepoint[Flag == 1])[cumsum(Flag) + 1])) %>%
  ungroup()

给予

  ID    Timepoint  Flag New_Timepoint
  <chr>     <int> <int>         <dbl>
1 A             0     0            NA
2 A             6     0            NA
3 A            12     0            NA
4 A            18     1             1
5 A            24     0             6
6 A            30     0            12
7 A            36     0            18

基于条件 R 的列值顺序增加

Sequential Increase in Column value based on a condition R

r

dplyr

conditional-statements

data-transform

sequential-number

数据