从 data.table 子集中重复采样选择有替换的行，而替换被关闭

Question

I have a data.table and one column is initially empty (NA values). I would like to select all rows with NA values in the NA column and then select two random samples and replace the NA value for them with an index variable coming from a loop. This step shall be repeated 3 times.

My code somehow seems not to produce correct subsets as already assigned non-NA values are overwritten.

Desired possible output:

1   3
2   2
3   NA
4   2
5   NA
6   3
7   1
8   NA
9   1

Real possible output (2x3 values should have no NA-values):

1   3
2   2
3   NA
4   NA
5   NA
6   3
7   NA
8   NA
9   1

MWE:

  d <- data.table(a=c(1,2,3,4,5,6,7,8,9), c=numeric())
  col_name <- "c" # 
  for(chunk in seq(1,3)) {
    d[d[is.na(get(col_name)), .I[sample(.N, 2, replace = FALSE)]], toString(col_name) := chunk]
  }

Why is this not working?

Answer 1

Here is one possible solution:

library(data.table)
set.seed(123)

for(chunk in seq(1,3)) {
  d[sample(which(is.na(get(col_name))), 2), (col_name) := chunk]
}
d

#   a  c
#1: 1 NA
#2: 2  2
#3: 3  1
#4: 4  2
#5: 5  3
#6: 6  1
#7: 7 NA
#8: 8 NA
#9: 9  3

---

This can also be done without a loop:

n <- 3
d[sample(.N, n * 2), (col_name) := rep(seq_len(n), 2)]
d