将LOCF填充或携带至一定值

Question

我有数据框 df_have。我想创建数据框 df_want，在其中保留 flag 的值来填充缺失值，直到 flag 将其值更改为每个患者 ID 的非缺失值。我尝试了函数 fill( ,.direction=down) ，但是，这会导致在所有观察中延续先前的值。一旦 FLAG=="Y"，我想停止填充 FLAG 列。

USUBJID <- c(1,1,1,1,1,1, 2,2,2,2,2,2 ,3,3,3,3,3,3)
FLAG <- c("N", NA, NA, "Y", NA, NA,    "N", NA, NA, "Y", NA, NA,   "N", NA, "Y", NA, NA, NA)

df_have<-data.frame(USUBJID, FLAG)
df_have

USUBJID <- c(1,1,1,1,1,1, 2,2,2,2,2,2 ,3,3,3,3,3,3)
FLAG <- c("N", "N", "N", "Y", NA, NA,    "N", "N", "N", "Y", NA, NA,   "N", "N", "Y", NA, NA, NA)

df_want<-data.frame(USUBJID, FLAG)
df_want

Answer 1

foo <- function(x) {
  first_value <- NA_character_
  for (i in seq_along(x)) {
    if (is.na(first_value) && !is.na(x[i])) {
      first_value <- x[i]
    } else if (!is.na(first_value) && is.na(x[i])) {
      x[i] <- first_value
    } else {
      break
    }
  }
  x
}

df_have |>
  group_by(USUBJID) |>
  mutate(FLAG2 = foo(FLAG))
   USUBJID FLAG  FLAG2
#      <dbl> <chr> <chr>
#  1       1 N     N    
#  2       1 NA    N    
#  3       1 NA    N    
#  4       1 Y     Y    
#  5       1 NA    NA   
#  6       1 NA    NA   
#  7       2 N     N    
#  8       2 NA    N    
#  9       2 NA    N    
# 10       2 Y     Y    
# 11       2 NA    NA   
# 12       2 NA    NA   
# 13       3 N     N    
# 14       3 NA    N    
# 15       3 Y     Y    
# 16       3 NA    NA   
# 17       3 NA    NA   
# 18       3 NA    NA