wave <-c(1, 2, 3, 1, 2, 3, 1, 2, 1, 2)

id   <-c(1, 1, 1, 2, 2, 2, 3, 3, 4, 4)

var1 <-c(6.341464, 13.206567, NA, 5.272296,2.746650, NA, 1.684472,1.874200, 15.861959, 3.938463)

var2 <-c(6.341464, 13.206567, NA, 5.272296,2.746650, NA, 1.684472,1.874200, 15.861959, 3.938463)

var3 <-c(6.341464, 13.206567, NA, 5.272296,2.746650, NA, 1.684472,1.874200, 15.861959, 3.938463)


df<- data.frame(wave,id,var1,var2,var3)

Output:df

How can I replace NA's in wave 3 with values from wave 2? it should copy all the values from second wave (NA too)

I tried for loop but without success!

Another option with Base R, could be using which to find the row indexes of wave 3 and use -1 index to replace the values like this:

df[which(wave == '3'),c("var1","var2","var3")] = df[which(wave == '3')-1,c("var1","var2","var3")]
df
#>    wave id      var1      var2      var3
#> 1     1  1  6.341464  6.341464  6.341464
#> 2     2  2 13.206567 13.206567        NA
#> 3     3  3 13.206567 13.206567        NA
#> 4     1  4  5.272296  5.272296  5.272296
#> 5     2  5  2.746650  2.746650  2.746650
#> 6     3  6  2.746650  2.746650  2.746650
#> 7     1  7  1.684472  1.684472  1.684472
#> 8     2  8  1.874200  1.874200  1.874200
#> 9     1  9 15.861959 15.861959 15.861959
#> 10    2 10  3.938463  3.938463  3.938463

^{Created on 2023-03-18 with reprex v2.0.2}

Another option from @TarJae, thanks!

library(dplyr)
library(tidyr)
df %>% 
  filter(wave == 2 | wave == 3) %>%    
  fill(var1:var3, .direction = "down") %>%    
  bind_rows(df) %>%    
  distinct(id, .keep_all = TRUE) %>%    
  arrange(id)
#>    wave id      var1      var2      var3
#> 1     1  1  6.341464  6.341464  6.341464
#> 2     2  2 13.206567 13.206567        NA
#> 3     3  3 13.206567 13.206567        NA
#> 4     1  4  5.272296  5.272296  5.272296
#> 5     2  5  2.746650  2.746650  2.746650
#> 6     3  6  2.746650  2.746650  2.746650
#> 7     1  7  1.684472  1.684472  1.684472
#> 8     2  8  1.874200  1.874200  1.874200
#> 9     1  9 15.861959 15.861959 15.861959
#> 10    2 10  3.938463  3.938463  3.938463

Old answer, different data OP

You could use fill with a down direction like this:

library(dplyr)
library(tidyr)
df %>%
  fill(var1:var3, .direction = 'down')
#>    wave id      var1      var2      var3
#> 1     1  1  6.341464  6.341464  6.341464
#> 2     2  2 13.206567 13.206567 13.206567
#> 3     3  3 13.206567 13.206567 13.206567
#> 4     1  4  5.272296  5.272296  5.272296
#> 5     2  5  2.746650  2.746650  2.746650
#> 6     3  6  2.746650  2.746650  2.746650
#> 7     1  7  1.684472  1.684472  1.684472
#> 8     2  8  1.874200  1.874200  1.874200
#> 9     1  9 15.861959 15.861959 15.861959
#> 10    2 10  3.938463  3.938463  3.938463

^{Created on 2023-03-18 with reprex v2.0.2}