Specific Aggregate Type for Each Column - na.rm

[michaelgloven]

Continuing the discussion from Specific Aggregate Type for Each Column:

Hi Nir,

One detail I forget to ask about this post, the aggregation functions should have an argument for na.rm = T or F. I generally default to T but would like to know where I could insert this argument in the code you provided?

Mike

[nirgrahamuk]

I'll leave you with the same exercise to convert hand written code to computer written code; but the general answer is to either use an inline anonymous function call to set additional parameters, or explicitly make your own function with the desired defaults.

library(tidyverse)

# example data preperation
(hiris <- head(iris))
hiris[1,1]<-NA
hiris

# the defaults with NA :(
hiris  |>
  summarise(across(starts_with("Sepal"),
                   list(mean=mean)))
# Sepal.Length_mean Sepal.Width_mean
# 1                NA         3.383333

# inline with anonymouse function
hiris  |>
  summarise(across(starts_with("Sepal"),
                   list(mean=\(x)mean(x,
                                      na.rm=TRUE))))
# Sepal.Length_mean Sepal.Width_mean
# 1              4.92         3.383333


#use purrr's partial to make your own function

mymean <- purrr::partial(mean,na.rm=TRUE)
hiris  |>
  summarise(across(starts_with("Sepal"),
                   list(mean=mymean)))
# Sepal.Length_mean Sepal.Width_mean
# 1              4.92         3.383333