[Feature request] as.data.frame() method for Categories objects

[rossellhayes]

Including an as.data.frame() method for Categories objects would make it easier to use join functions to map between the names and numeric codes associated with a variable.

Here are the functions that I'm currently using in my scripts to achieve this:

as.data.frame.Category <- function(category) {
  data <- category@.Data
  names(data) <- category@names
  data[lengths(data) == 0] <- NA
  as.data.frame(data)
}

as.data.frame.Categories <- function(categories) {
  do.call(rbind, lapply(categories@.Data, as.data.frame))
}

[gergness]

Yeah, seems like an oversight because there is one for variables.

A complication I can foresee would be that I don't think the category metadata schema is consistent between variables. Do you think numeric_value or datetime column should always be present, or only when relevant?

[rossellhayes]

I think my expectation would be that an as.data.frame() method would produce a data frame that looks like the print method for Categories objects, i.e. only including the columns that are relevant for that variable type, but I would be happy with either implementation!

[1beb]

I think I have a method for this already in crunchtabs. I think it's in codebook. https://github.com/Crunch-io/crunchtabs/blob/main/R/codeBookSummary.R#L137