remove group if any member containes NA in R

user_14514085 0
Viewed 266

How can I remove entire group if one of its values is NA. For ex - remove category B because it contains NA.

library(dplyr)

tbl = tibble(category = c("A", "A", "B", "B"),
             values = c(2, 3, 1, NA))
dplyr r
4 Answers

We can use filter after grouping by 'category'

library(dplyr)
tbl %>% 
    group_by(category) %>% 
    filter(!any(is.na(values))) %>%
    ungroup

-output

# A tibble: 2 x 2
  category values
  <chr>     <dbl>
1 A             2
2 A             3
user_37322710 
tbl %>% 
  filter(!category %in% category[is.na(values)])

Output

  category values
  <chr>     <dbl>
1 A             2
2 A             3
user_63824340 
tbl %>% 
  group_by(category) %>% 
  filter(all(!is.na(values)))

  category values
  <chr>     <dbl>
1 A             2
2 A             3
user_133216470

You can get the categories which has at least one NA value and exclude them.

subset(tbl, !category %in% unique(category[is.na(values)]))

#  category values
#  <chr>     <dbl>
#1 A             2
#2 A             3

If you prefer dplyr::filter.

library(dplyr)

tbl %>% filter(!category %in% unique(category[is.na(values)]))
user_39629140
Related