CODEWITHSUNDEEP
rtidyr
dfrankow
dfrankow

Reputation: 21387

How to get tidyr's complete to work with a variable specifying columns?

Here is some code using complete:

dat <- data.frame(a=c(1,NA), b=c(3,NA), val=c(20,30))
dat %>% complete(a, b, fill=list(val=0))

# A tibble: 4 x 3
      a     b   val
  <dbl> <dbl> <dbl>
1     1     3    20
2     1    NA     0
3    NA     3     0
4    NA    NA    30

How do I make a function that takes the columns to complete? Here is a failed attempt:

foo_func <- function(dat, the_cols) {
  dat %>% complete(all_of(the_cols), fill=list(val=0))
}
foo_func(dat, c('a', 'b'))

Error: Join columns must be present in data.
x Problem with `all_of(the_cols)`.

Here's another:

foo_func <- function(dat, the_cols) {
  dat %>% complete(!!the_cols, fill=list(val=0))
}
foo_func(dat, c('a', 'b'))

Error: Join columns must be present in data.
x Problem with `<chr>`

I want the_cols to be a character vector, because this is in an existing body of code that passes things around that way.

Upvotes: 4

Views: 242

Answers (1)

akrun
akrun

Reputation: 887108

We could convert to symbols and use !!!

foo_func <- function(dat, the_cols) {
  dat %>% complete(!!! rlang::syms(the_cols), fill=list(val=0))
}

-checking

foo_func(dat, c('a', 'b'))
# A tibble: 4 x 3
#      a     b   val
#  <dbl> <dbl> <dbl>
#1     1     3    20
#2     1    NA     0
#3    NA     3     0
#4    NA    NA    30

Upvotes: 2

Related Questions