Iteratively fill missing values

Question

I have a data frame with one column completely filled and the other one partially filled. I would like like to fill up the na s in the second column with this formula: col3 = col1*lag(col2)

When I'm using mutate, it's iterating once. I prefer using a dplyr function since I will need to apply this function over a larger sample.

Here is the data frame

dd <- tibble::tribble(~col1, ~col2,
  1.1127493, 34533.98,
  0.9432176, 32573.06,
  0.9130100, 29758.64,
  0.6593648, 19621.80,
  0.9222222, 18095.66,
  0.7349398, 13299.22,
  0.9836066, 13081.20,
  0.8518519,       NA,
  1.0000000,       NA,
  1.0000000,       NA,
  1.0000000,       NA,
  1.0000000,       NA)

Answer 1

I find it hard to understand the procedure you have in mind and why is it guaranteed to terminate. May be you can specify your procedure in some more detail. In the meantime, I created a procedure close the one you are asking for, to serve you as a guide, so that you can write your own solution.

iter_func <- function(c1, c2){
  c3 <- c1 + dplyr::lag(c2)
  na_pos <- which(is.na(c3))
  non_na_pos <- which(!is.na(c3))

  while(length(na_pos)){
    c3[na_pos[1]] <- c3[non_na_pos[length(non_na_pos)]]
    c3 <- c1 + c3

    na_pos <- which(is.na(c3))
    non_na_pos <- which(!is.na(c3))
  }

  c3
}

dd %>% dplyr::mutate(col3 = iter_func(col1, col2))

HTH