In a package I'm developing, I'm allowing the user to select column names, in a way that is functionally similar to the example included in the documentation on data mask programming.
Let's say that I have a function such as:
my_summarise <- function(data, var) {
data |> dplyr::summarise({{ var }})
}
In the real world, the function is obviously more complex, so I troubleshoot it in an interactive session.
Is there some easy way to work with this in a usual interactive R session, passing to var the desired value? E.g. , testing this function as if the user war running the following function?
my_summarise(data, var = string)
and run the code:
data |> dplyr::summarise({{ var }})
without getting any error?
I appreciate that more formal debugging approaches, breakpoints, etc., would probably be more robust, but for some quick checks more informal solutions and testing in a plain R session still feel easier to me.
Any suggestions welcome? If you feel I shouldn't really be trying to work with it this way, or you have suggestions for a better workflows, I'm all happy to read your thoughts.