How can i get my function arguments to be of the corrct type for datatables in r?-CodePudding

I have a data table called data:

[![This is a screengrab of the data table][1]][1]

The goal is to write a function that mimics the following code:

data[Region == "Northeast",mean(Awareness, na.rm = TRUE), by = Product][order(-rank(V1))][1:5,.(Product)]

So far I have:

topx_engagement = function(state_of_engagement, respondent_variable, respondent_variable_sub, rank_length = 3){
  
  respondent_variable_sub = as.character(respondent_variable_sub)
  
  data[eval(respondent_variable == respondent_variable_sub), mean(get(state_of_engagement), na.rm = TRUE), by = Product][order(-rank(V1))][1:rank_length,.(Product)]
}

The function topx_engagement should allow a user to enter respondent_variable which coincides with Region, a respondent_variable_sub which coincides with "Northeast" and subsequent arguments. I would like to focus on the preceding arguments as all the other work fine.

Currently, when I call:

topx_engagement(state_of_engagement = Awareness, respondent_variable = Region, respondent_variable_sub = Northeast, rank_length = 3)

I get an error:

Error in topx_engagement(state_of_engagement = Awareness, respondent_variable = Region, : object 'Northeast' not found

Alternately, running

topx_engagement = function(state_of_engagement, respondent_variable, respondent_variable_sub, rank_length = 3){
  
  #respondent_variable_sub = as.character(respondent_variable_sub)
  
  data[eval(respondent_variable == respondent_variable_sub), mean(get(state_of_engagement), na.rm = TRUE), by = Product][order(-rank(V1))][1:rank_length,.(Product)]
}

topx_engagement(state_of_engagement = Awareness, respondent_variable = Region, respondent_variable_sub = Northeast, rank_length = 3)

throws an error of Error in eval(.massagei(isub[[2L]]), parent.frame(), parent.frame()) : object 'Region' not found

I need assistance with getting the function inputs in the right format. [1]: https://i.stack.imgur.com/hseWS.png

CodePudding user response：

We may use deparse/substitute to convert to character string if we are passing unquote arguments

topx_engagement = function(state_of_engagement, 
     respondent_variable, respondent_variable_sub, rank_length = 3){
  
  state_of_engagement <- deparse(substitute(state_of_engagement))
  respondent_variable <- deparse(substitute(respondent_variable))
  respondent_variable_sub <- deparse(substitute(respondent_variable_sub))
 
 
  
  
  data[eval(as.name(respondent_variable)) == respondent_variable_sub, 
        mean(get(state_of_engagement), na.rm = TRUE), 
           by = Product][order(-rank(V1))][1:rank_length,.(Product)]
}

then call it as

topx_engagement(state_of_engagement = Awareness, 
   respondent_variable = Region, 
       respondent_variable_sub = Northeast, rank_length = 3)