我们需要为每个患者创建唯一的治疗组，因此我们需要去掉列中的“第一天”和“最后一天”字符串。然后group_byPatient和Treatment列以创建唯一组，并使用reframe基于Day的min和max生成新的Day序列。

library(tidyverse)

a %>% 
  mutate(Treatment = sub(" first day| last day", "", Treatment),
         Day = as.integer(Day)) %>% 
  group_by(Patient, Treatment) %>% 
  reframe(Day = min(Day):max(Day))

# A tibble: 12 × 3
   Patient Treatment   Day
   <chr>   <chr>     <int>
 1 1       A             3
 2 1       A             4
 3 1       A             5
 4 1       B            10
 5 1       B            11
 6 1       B            12
 7 1       B            13
 8 1       B            14
 9 2       A             2
10 2       A             3
11 2       A             4
12 2       A             5

首先使用sub删除多余的文本。然后将by拆分为ID变量（在本例中为1:2）和cbind1 Dataframe 的第一行减去日期变量-3，然后在do.call中使用:来获得序列，最后rbind所有内容。我添加了一些额外的列来证明它们是保留的。

a$Treatment <- sub('\\s.*', '', a$Treatment)
by(a, a[1:2], \(x) cbind(x[1, -3], Day=do.call(`:`, as.list(as.integer(sort(x[[3]])))), row.names=NULL)) |> 
  do.call(what=rbind)
#    Patient Treatment Sex   X Day
# 1        1         A   m 0.5   3
# 2        1         A   m 0.5   4
# 3        1         A   m 0.5   5
# 4        2         A   f 0.4   2
# 5        2         A   f 0.4   3
# 6        2         A   f 0.4   4
# 7        2         A   f 0.4   5
# 8        1         B   f 0.3  10
# 9        1         B   f 0.3  11
# 10       1         B   f 0.3  12
# 11       1         B   f 0.3  13
# 12       1         B   f 0.3  14

• 数据：*

a <- structure(list(Patient = c(1L, 1L, 1L, 1L, 2L, 2L), Treatment = c("A first day", 
"A last day", "B first day", "B last day", "A first day", "A last day"
), Day = c(3L, 5L, 10L, 14L, 2L, 5L), Sex = c("m", "m", "f", 
"f", "f", "f"), X = c(0.5, 0.5, 0.3, 0.3, 0.4, 0.4)), class = "data.frame", row.names = c("1", 
"2", "3", "4", "5", "6"))

以下是complete与fill组合的解决方案：

library(dplyr)
library(tidyr)

a %>%
  separate(Treatment, into = c("Treatment", "first_last"), sep = " ", extra = "drop") %>% 
  group_by(Patient, Treatment) %>%
  mutate(Day = as.integer(Day)) %>% 
  complete(Day = seq(min(Day), max(Day))) %>%
  fill(Treatment, .direction = "updown") %>% 
  select(-first_last) %>% 
  ungroup()

# A tibble: 12 × 3
   Patient Treatment   Day
   <chr>   <chr>     <int>
 1 1       A             3
 2 1       A             4
 3 1       A             5
 4 1       B            10
 5 1       B            11
 6 1       B            12
 7 1       B            13
 8 1       B            14
 9 2       A             2
10 2       A             3
11 2       A             4
12 2       A             5