R Fill NA With Adding 5 to Previous Value-CodePudding

HAVE = data.frame(STUDENT = c(1,1,1,2,2,2,3,3),
                  TIME = c(1,2,3,1,2,3,1,2),
                  SCORE = c(7, NA, NA, 5, NA, 19, NA, 2))

WANT = data.frame(STUDENT = c(1,1,1,2,2,2,3,3),
                  TIME = c(1,2,3,1,2,3,1,2),
                  SCORE = c(7, 12, 17, 5, 10, 19, NA, 2))

I wish to modify SCORE by doing the following step: If is.na(SCORE) then take previous value for STUDENT and add 5 to it. But only do this downwards not upwards.

I try this with no success, it doesn't work when there is two NA in a row for the STUDENT

HAVE %>% group_by(STUDENT) %>% mutate(WANT = ifelse(is.na(SCORE), lag(SCORE)   5, SCORE))

CodePudding user response：

You can use purrr::accumulate:

library(dplyr)
library(purrr)
HAVE |> 
  mutate(SCORE = accumulate(SCORE, ~ if(is.na(.y)) .x   5 else .y),
         .by = STUDENT)

#   STUDENT TIME SCORE
# 1       1    1     7
# 2       1    2    12
# 3       1    3    17
# 4       2    1     5
# 5       2    2    10
# 6       2    3    19
# 7       3    1    NA
# 8       3    2     2

Or, in base R, with Reduce and accumulate = TRUE:

na_add <- function(col, n = 5){
  Reduce(function(x, y) if(is.na(y)) x   n else y, col, accumulate = TRUE)
}
with(HAVE, ave(SCORE, STUDENT, FUN = na_add))
#[1]  7 12 17  5 10 19 NA  2