[Feature Request]: order splits alphabetically in qtable for character vars

[anajens]

Feature description

When using qtable with character variables it would be a nice feature if the row and column splits are ordered alphabetically. This would help improve readability of the table with multiple categories. Also this would follow the behaviour of dplyr::count.

> library(rtables)
Loading required package: formatters
Loading required package: magrittr
> library(dplyr)

Attaching package: ‘dplyr’

The following objects are masked from ‘package:stats’:

    filter, lag

The following objects are masked from ‘package:base’:

    intersect, setdiff, setequal, union

> 
> set.seed(123)
> tmp <- data.frame(
+   x = rep("x", 50),
+   y = sample(letters[1:5], size = 50, replace = TRUE, prob = c(0.1, 0.2, 0.3, 0.35, 0.05)),
+   stringsAsFactors = FALSE
+ )
> 
> # Categories are not sorted:
> qtable(tmp, "y")
count   all obs
        (N=50) 
———————————————
d         17   
b          8   
c         14   
a          7   
e          4   
> 
> # dplyr sorts:
> tmp %>% count(y)
  y  n
1 a  7
2 b  8
3 c 14
4 d 17
5 e  4

Code of Conduct

• I agree to follow this project's Code of Conduct.

Contribution Guidelines

• I agree to follow this project's Contribution Guidelines.

Security Policy

• I agree to follow this project's Security Policy.

[gmbecker]

This makes sense in isolation but I'm not sure it does within the larger scheme of things, because the rtables framework as a whole matches what happens now (implicit levels of character vectors are ordered by their first appearance in the full dataset, and then fixed in that order regardless of faceting hierarchy).

So while i think this would be nice for qtable and match it's purpose well, I'm not sure it's behavior should diverge from the core tabulation engine, and changing that that might be desirable but would be a fairly large breaking change we would need to think hard before implementing.

Lets discuss this more before we implement it