Saving na.rm=TRUE for each function in dplyr

Question

I am using dplyr summarise function. My data contain NAs so I need to include na.rm=TRUE for each call. for example:

group <- rep(c('a', 'b'), 3)
value <- c(1:4, NA, NA)
df = data.frame(group, value)

library(dplyr)
group_by(df, group) %>% summarise(

          mean = mean(value, na.rm=TRUE),

          sd = sd(value, na.rm=TRUE),

          min = min(value, na.rm=TRUE))

Is there a way to write the argument na.rm=TRUE only one time, and not on each row?

https://stackoverflow.com/questions/17418640/is-it-possible-to-set-na-rm-to-true-globally — J_F, Dec 11 '17 at 10:17
You can use `na.omit`: `df %>% group_by(group) %>% na.omit() %>% summarise()` — pogibas, Dec 11 '17 at 10:17
na.omit() will delete the entire row with al least one NA. I don't want that. — Rtist, Dec 11 '17 at 10:26

score 4 · Accepted Answer · answered Dec 11 '17 at 10:26

4

You should use summarise_at, which lets you compute multiple functions for the supplied columns and set arguments that are shared among them:

df %>% group_by(group) %>% 
  summarise_at("value", 
               funs(mean = mean, sd = sd, min = min), 
               na.rm = TRUE)

answered Dec 11 '17 at 10:26

mtoto

23,919
4
58
71

now that across() is the preferred approach in dplyr 1.0.0, how would you do this? – Dominik Jun 01 '20 at 16:54

score 1 · Answer 2 · answered Dec 11 '17 at 10:33

If you're planning to apply your functions to one column only, you can use filter(!is.na()) in order to filter out any NA values of this variable only (i.e. NA in other variables won't affect the process).

group <- rep(c('a', 'b'), 3)
value <- c(1:4, NA, NA)
df = data.frame(group, value)

library(dplyr)

group_by(df, group) %>% 
  filter(!is.na(value)) %>%
  summarise(mean = mean(value),
            sd = sd(value),
            min = min(value))

# # A tibble: 2 x 4
#    group  mean       sd   min
#   <fctr> <dbl>    <dbl> <dbl>
# 1      a     2 1.414214     1
# 2      b     3 1.414214     2

Saving na.rm=TRUE for each function in dplyr

2 Answers2