I have a data frame in R with country name, year and several variables as columns. For each country, I have data for my variables in the time span 1970-2020. The problem is that I have missing values on both ends of my dataset for most countries, so for the years 1970 and 1971 for country A or for 2019 and 2020 for country B.
So, for example:

Country-Year-column1-column2-column3

USA     1970- NA-    0,54 -   NA
USA     1971  NA -   0,58  -  0,9
USA     1972  0,23 - 0,6  -   0,85

etc.

I tried to solve the problem using na.approx, which I had previously used for interpolating missing values. It had worked well when I had values before and after the missing value, but not for the marginal values now. I was able to run the code but nothing changed in my dataset, the missing values remained. Maybe this is possible with na_spline, but I didn’t know how to do it. So I opted for another approach: I want those missing values to be based on the next / last 6 values now.

My code looks like this:

columns_interpoliert<-c("life_expectancy", "economic_growth", "prison_rates"
function_rand<-function(x){
if(
is.na
      (x[1])) {
    first_nona<-min(which(!
is.na
(x)))
    differential <- (x[first_nona + 5] - x[first_nona]) / 5
    for(i in 1:first_nona - 1) {
        x[i] <- x[first_nona] - differential * (first_nona - i)
    }
}
if(
is.na
(x[length(x)])) {
    last_nona<-max(which(!
is.na
(x)))
       differential <- x[last_nona - 5] - x[last_nona] / 5
    for(i in last_nona + 1:length(x)) {
        x[i] <- x[last_nona] + differential * (i - last_nona)
    }
}
return(x)
} 
data_new <- wvs_filled %>%
group_by(country) %>%
mutate(across(columns_interpoliert, function_rand)) 

I get an error code here:

"Error in `mutate()`:
ℹ In argument:

across(columns_interpoliert, function_rand).
ℹ In group 1: country = 2.
Caused by error in across():
! Can’t compute column life_expectancy.
Caused by error in dplyr_internal_error():”

Does anybody know why this might happen? Any help is appreciated!