CODEWITHSUNDEEP
rdplyr
silver arrow
silver arrow

Reputation: 117

dplyr: Unquoting in count fails to identify column

test <- data.frame(name = c(1, 2, 3, 4, 3, 3), value = c(1222, 11122, 111, 122, 318431, 29424))

I am writing a function with tidyverse 1.3.0 in R version 3.6.2 that extracts duplicated ids. So for the above example, it would return 3. The following works as I desire:

 getDupIds <- function(df, id_type) {
  list <- df %>%
    count_(id_type) %>%
    filter(n > 1) %>%
    pull(id_type)
  
  list <- list[!is.na(list)]
  
  if (length(list) != 0) {
    return(list)
  } else {
    return("No duplicate ids.")
  }
}

getDupIds(test, "name")

But I understand that underscored functions (count_) are now to be replaced with unquoting. However, the following throws the error: Error: Can't extract columns that don't exist. x The column `name` doesn't exist.

getDupIds <- function(df, id_type) {
  list <- df %>%
    count(!!id_type) %>%
    filter(n > 1) %>%
    pull(id_type)
  
  list <- list[!is.na(list)]
  
  if (length(list) != 0) {
    return(list)
  } else {
    return("No duplicate ids.")
  }
}

Any ideas why this doesn't work? It seems to unquote name fine, it puzzles me why it cannot then find the column with that name.

Upvotes: 1

Views: 90

Answers (1)

Ronak Shah
Ronak Shah

Reputation: 389155

Convert to symbol first before using !! :

library(dplyr)
library(rlang)

getDupIds <- function(df, id_type) {
   list <- df %>%
           count(!!sym(id_type)) %>%
           #You can also use
           #count(.data[[id_type]]) %>%
           filter(n > 1) %>%
           pull(id_type)

   list <- list[!is.na(list)]

   if (length(list) != 0)  return(list)
   else return("No duplicate ids.")
}

getDupIds(test, "name")
#[1] 3

To pass unquoted column name you can use {{}}, replace the part in function as :

 list <- df %>%
          count({{id_type}}) %>%
          filter(n > 1) %>%
           pull({{id_type}})

and call it like :

getDupIds(test, name)

Upvotes: 2

Related Questions