How do I use approx() inside mutate_at() with a conditional statement in dplyr?

Question

I want to interpolate missing values using dplyr, piping, and approx().

Data:

test <- structure(list(site = structure(c(3L, 3L, 3L, 3L, 1L, 1L, 1         


        

Answer 1

This will do what you are looking for...

test_int2 <- test %>% 
             group_by(site) %>% 
             mutate_at(vars(c(var1, var2)),
                       funs("i"=if(sum(!is.na(.))>1) 
                                  approx(depth, ., depth, rule=1, method="linear")[["y"]] 
                                else 
                                  NA))

test_int2
# A tibble: 12 x 6
# Groups:   site [3]
      site depth  var1  var2 var1_i var2_i
    <fctr> <int> <int> <int>  <dbl>  <dbl>
 1 wetland     0     1     1    1.0    1.0
 2 wetland    -3    NA    NA    2.5    2.5
 3 wetland    -4     3     3    3.0    3.0
 4 wetland    -8     4     4    4.0    4.0
 5    lake     0     1    NA    1.0     NA
 6    lake    -1     2    NA    2.0     NA
 7    lake    -3    NA    NA    3.0     NA
 8    lake    -5     4    NA    4.0     NA
 9  stream     0     1    NA    1.0     NA
10  stream    -2    NA     2    2.0     NA
11  stream    -4    NA    NA    3.0     NA
12  stream    -6     4    NA    4.0     NA