lapply function with arguments for dataframe and variable

Question

I have a custom function which summarises a variable. I simplified the function to illustrate my problem, i.e. it is more complex than shown below. Note that the general structure of the function should remain the same: It takes an argument for specifying which dataframe to work on (df), and an argument which variable to summarise (variable_to_test).

my_fun <- function(df, variable_to_test) {

  variable_to_test <- enquo(variable_to_test)
  new_var_name <- paste0(quo_name(variable_to_test), "_new_name")

  df %>% 
    summarise(
      !!new_var_name := sum(!!variable_to_test, na.rm = TRUE)
    ) 
}

Using an example, I can apply the function on each variable in my dataframe:

library(tidyverse)
dat <- tibble(
  variable_1 = c(1:5, NA, NA, NA, NA, NA),
  variable_2 = c(NA, NA, NA, NA, NA, 11:15)
)


> my_fun(dat, variable_1)
# A tibble: 1 x 1
   variable_1_new_name
                 <int>
1                  15


> my_fun(dat, variable_2)
# A tibble: 1 x 1
  variable_2_new_name
                <int>
1                  65

But: how can I list apply the function on all columns in the dataframe? I tried

> dat %>%
+ lapply(., my_fun)
Error in duplicate(quo) : argument "quo" is missing, with no default
Called from: duplicate(quo)

but this returns an error. I'm struggling with the fact that the function takes an argument for both the dataframe to work on, and the variable to summarise. Note that I'd like to keep this structure - I find it more elegant to pass the name of the dataframe into the function instead of just giving the function the variable name and "hard-code" the data frame into the function body. Does anybody have a good idea how to lapply() the function?

Do you need a dplyr solution, or does base R fit your needs? Usually you would solve this by giving the function one static and one variable input, e.g. lapply(dat, function(x) myfun(dat, x)). I'm not versed in dplyr, but maybe try lapply(., function(x) myfun(., x))? — LAP
– LAP, Commented Aug 14, 2017 at 11:54
I already have a base R solution. I tried to rewrite the function the tidyeval-way, since it enhances readability of the function body. So yes, I need atidyeval-solution :) — piptoma
– piptoma, Commented Aug 14, 2017 at 11:58

piptoma · Accepted Answer · 2017-08-15 06:40:41Z

2

Oh, I think you're just mapping over the wrong thing. For the tidyverse solution I would try:

map(dat, ~my_fun(dat, .))

What this does is map over the column names and plug in the column to the ..

edited Aug 15, 2017 at 6:40

piptoma

8161 gold badge10 silver badges22 bronze badges

answered Aug 14, 2017 at 12:25

Shorpy

1,69415 silver badges28 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Lionel Henry · Accepted Answer · 2017-08-14 17:12:55Z

1

You're working at the wrong level. If you map a function over a data frame, then this function should take a column. The problem here is that the function my_fun() expects a data frame rather than a column.

You need to find some other way of solving the problem. One solution is to use the mappers provided by dplyr:

dat %>%
  summarise_all(sum, na.rm = TRUE) %>%
  rename_all(paste0, "_new_name")

You could equivalently use a combination of map() and set_names() from purrr.

dat %>%
  map_df(sum, na.rm = TRUE) %>%
  set_names(paste0, "_new_name")

answered Aug 14, 2017 at 17:12

Lionel Henry

6,83130 silver badges36 bronze badges

Collectives™ on Stack Overflow

lapply function with arguments for dataframe and variable

2 Answers 2

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related