R：反算补缺值

Question

I have received a data set in which all values below 10 have been replaced by a *. However, the data set also contains row and column totals, which partially make the back calculation possible.

For the rows I have already managed this, but for the columns I lack the inspiration how this could work.

When reading the CSV files, the * are converted to NAs, so a sample dataset looks like this:

ID	V1	V2	V3	VS
A1	11	12	13	36
A2	NA	11	12	32
A3	NA	12	NA	24
AS	27	35	32	92

In this example the NA for ID A2 should be replaced by 9 [ 32 - (11 + 12) ]. The next step is to calculate the NAs for ID A3. V1 should be replaced by 7 [ 27 - (11 + 9) ] and V3 by 5 [ 32 - (13 + 12) ]

I feel like this is actually one of the simplest problems, but I just can't come up with the solution. Can anyone help me out with this?

Thanks a lot Benne

Answer 1

A solution with dplyr : I do it in two steps, 1) calculate the imputation that is performed rowwise, and 2) the column-specific imputation.

df <- read.table(text= "ID  V1  V2  V3  VS
A1  11  12  13  46
A2  NA  11  12  32
A3  NA  12  NA  24
AS  27  35  32  102", h = T) 


library(dplyr)
df %>%
  slice(1:2) %>% 
  mutate(across(V1:VS, ~ifelse(is.na(.x), VS - V3 - V2, .x))) %>% 
  rbind(df[3:4,1:5]) %>% 
  mutate(across(V1:VS, ~ifelse(is.na(.x), dplyr::lead(.x) - dplyr::lag(.x, 2) - dplyr::lag(.x,1), .x)))

Result :

  ID V1 V2 V3  VS
1 A1 11 12 13  46
2 A2  9 11 12  32
3 A3  7 12  7  24
4 AS 27 35 32 102