Loop to create crosstabs of columns using tidyr-CodePudding

I would like to use a loop to create crosstabs of one column with every other column in a df. I started with this code (substituting in the iris df), which works nicely for two variables:

iris <- iris
tbl <- iris %>% 
  tabyl(Species, Sepal.Length, show_missing_levels = FALSE, show_na = FALSE) %>%
  adorn_percentages("row") %>%
  adorn_pct_formatting(digits = 0) %>%
  adorn_ns() %>%
  adorn_title("combined") %>%
  knitr::kable()
print(tbl)

My df contains ~200 columns. I thought I would write a for loop to print a crosstab for one variable with each of the other variables. Here's what I tried:

cols <- c('Sepal.Length', 'Sepal.Width')
for (c in cols){
  tbl <- iris %>% 
    tabyl(Species, c, show_missing_levels = FALSE, show_na = FALSE) %>%
    adorn_percentages("row") %>%
    adorn_pct_formatting(digits = 0) %>%
    adorn_ns() %>%
    adorn_title("combined") %>%
    knitr::kable()
  print(tbl)
}

This returns Column `c` is not found.

This seems like it should be simple, but I can't figure it out. Thanks for any help.

CodePudding user response：

Change the c in your code to !!sym(c). I can't explain this non-standard tidyverse evaluation thingy, but in layman's terms, you want to access an object (i.e. "c") outside of your pipe (iris). That's why you need !!sym.

CodePudding user response：

You can use the .data pronoun when passing columns names as strings.

cols <- c('Sepal.Length', 'Sepal.Width')

for (col in cols){
  tbl <- iris %>% 
    tabyl(Species, .data[[col]],show_missing_levels = FALSE,show_na = FALSE) %>%
    adorn_percentages("row") %>%
    adorn_pct_formatting(digits = 0) %>%
    adorn_ns() %>%
    adorn_title("combined") %>%
    knitr::kable()
  print(tbl)
}