How to filter out where specific columns are all na [duplicate]

Question

Here is a sample dataset:

id <- c("Item1","Item2","Item3","Item4","Item5","Item6")
var1 <- c(2,3,NA,NA,5,6)
var2 <- c(NA,3,5,NA,5,NA)
var3 <- c(NA,3,4,NA,NA,6)
test <- data.frame(id, var1, var2, var3)

I want to filter out where var1, var2 and var3 are all na. I know it can be done like this:

test1 <- test %>% filter(!(is.na(var1) & is.na(var2) & is.na(var3)))
test1

     id var1 var2 var3
1 Item1    2   NA   NA
2 Item2    3    3    3
3 Item3   NA    5    4
4 Item5    5    5   NA
5 Item6    6   NA    6

Is there a better way of doing this?

Answer 1

If the filtering is focused on certain columns, e.g. var1:var3, you can use

library(dplyr)

option 1

test %>%
  filter(rowSums(across(var1:var3, ~ !is.na(.))) > 0)

option 2

test %>%
  filter_at(vars(var1:var3), any_vars(!is.na(.)))

option 3

test %>%
  rowwise() %>% 
  filter(sum(!is.na(c_across(var1:var3))) > 0) %>%
  ungroup()

output

# # A tibble: 5 x 4
#   id     var1  var2  var3
#   <chr> <dbl> <dbl> <dbl>
# 1 Item1     2    NA    NA
# 2 Item2     3     3     3
# 3 Item3    NA     5     4
# 4 Item5     5     5    NA
# 5 Item6     6    NA     6