Problem with assigning variables in a for-loop with lares:: corr_var function

Question

I want to test the first columns of a data frame for correlation with their last columns. I'm implementing this to a bigger function, so I never know the variable names. Therefore I need an automated solution, which I'm not able to find. I tried the following:

library(lares)

data("iris")

corlist <- list()
for(i in 1:2){
  corlist[i] <- corr_var(df = iris,
                         var = names(iris)[i],
                         ignore = names(iris)[c(1:2, 5)[-i]],
                         max_pvalue = 0.05,
                         top = 50)
}

Which returns:

Error in corr_var(df = iris, var = names(iris)[i], ignore = names(iris)[c(1:2)[-i]],  : 
  Not a valid input: names(iris)[i] was transformed or does not exist.

I tried using get(), paste(), or noquote(), but neither worked. I also tried to insert the column itself like iris[, i], but I always receive the same error. The only thing that seems to work is e.g. var = Sepal.Length and so on, but these can't be looped. \

Answer 1

While we wait for an update on lares::corr_var(), this should work for you:

corlist <- list()
for(i in 1:2) {
  var_name <- names(iris)[i]
  var_sym <- rlang::sym(var_name)  # Construct symbol from string
  corlist[[i]] <- corr_var(
    df = iris,
    var = !!var_sym,  # Unquote and evaluate symbol
    ignore = names(iris)[c(1:2, 5)[-i]],
    max_pvalue = 0.05,
    top = 50
  )
}