Fix issues with different nlevels in groups, enable "n" argument

[strengejacke]

library(report)
wilson2017 <- readxl::read_excel("c:/Users/mail/Downloads/pmed.1002479.s001.xls", sheet = "data")

wilson2017 |>
  report::report_sample(
    group_by = "ethnicgrp",
    select = c("gender", "partners", "msm")
  )
#> # Descriptive Statistics
#> 
#> Variable                | Asian/ Asian British (n=127) | Black/ Black British (n=191) | Mixed/ Multiple ethnicity (n=188) | Other (n=29) | White/ White British (n=1528) | Total (n=2063)
#> -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
#> gender [Female], %      |                         55.1 |                         57.6 |                              63.3 |         48.3 |                          58.9 |           58.8
#> gender [Male], %        |                         44.9 |                         42.4 |                              36.2 |         48.3 |                          41.0 |           41.0
#> gender [Transgender], % |                          0.0 |                          0.0 |                               0.5 |          3.4 |                           0.1 |            0.2
#> partners [1], %         |                         41.7 |                         36.1 |                              33.5 |         37.9 |                          26.8 |           29.4
#> partners [10+], %       |                          6.3 |                          5.2 |                               6.9 |         10.3 |                           9.6 |            8.8
#> partners [2], %         |                         17.3 |                         24.6 |                              16.5 |         13.8 |                          17.1 |           17.7
#> partners [3], %         |                          9.4 |                         11.0 |                              17.0 |          6.9 |                          14.3 |           13.8
#> partners [4], %         |                         11.0 |                          8.9 |                               9.0 |          3.4 |                          10.9 |           10.4
#> partners [5], %         |                          5.5 |                          8.4 |                               8.0 |         17.2 |                           8.7 |            8.5
#> partners [6], %         |                          5.5 |                          3.1 |                               3.7 |          3.4 |                           6.0 |            5.5
#> partners [7], %         |                          0.0 |                          1.6 |                               2.7 |          6.9 |                           3.3 |            2.9
#> partners [8], %         |                          2.4 |                          1.0 |                               1.6 |          0.0 |                           2.1 |            1.9
#> partners [9], %         |                          0.8 |                          0.0 |                               1.1 |          0.0 |                           1.2 |            1.1
#> msm [99], %             |                          1.6 |                          1.6 |                               0.0 |          3.4 |                           1.1 |            1.1
#> msm [msm], %            |                         14.2 |                          5.8 |                               8.5 |         17.2 |                          13.9 |           12.7
#> msm [other], %          |                         84.3 |                         92.7 |                              91.5 |         79.3 |                          85.0 |           86.2

wilson2017 |>
  report::report_sample(
    select = c("gender", "partners", "msm", "ethnicgrp"),
    n = TRUE
  )
#> # Descriptive Statistics
#> 
#> Variable                                    |    Summary
#> --------------------------------------------------------
#> gender [Female], %, n                       | 58.8, 1213
#> gender [Male], %, n                         |  41.0, 846
#> gender [Transgender], %, n                  |     0.2, 4
#> partners [1], %, n                          |  29.4, 606
#> partners [10+], %, n                        |   8.8, 181
#> partners [2], %, n                          |  17.7, 365
#> partners [3], %, n                          |  13.8, 285
#> partners [4], %, n                          |  10.4, 215
#> partners [5], %, n                          |   8.5, 176
#> partners [6], %, n                          |   5.5, 113
#> partners [7], %, n                          |    2.9, 60
#> partners [8], %, n                          |    1.9, 40
#> partners [9], %, n                          |    1.1, 22
#> msm [99], %, n                              |    1.1, 23
#> msm [msm], %, n                             |  12.7, 262
#> msm [other], %, n                           | 86.2, 1778
#> ethnicgrp [Asian/ Asian British], %, n      |   6.2, 127
#> ethnicgrp [Black/ Black British], %, n      |   9.3, 191
#> ethnicgrp [Mixed/ Multiple ethnicity], %, n |   9.1, 188
#> ethnicgrp [Other], %, n                     |    1.4, 29
#> ethnicgrp [White/ White British], %, n      | 74.1, 1528

report::report_sample(iris, n = TRUE)
#> # Descriptive Statistics
#> 
#> Variable                   |          Summary
#> ---------------------------------------------
#> Mean Sepal.Length (SD), n  | 5.84 (0.83), 150
#> Mean Sepal.Width (SD), n   | 3.06 (0.44), 150
#> Mean Petal.Length (SD), n  | 3.76 (1.77), 150
#> Mean Petal.Width (SD), n   | 1.20 (0.76), 150
#> Species [setosa], %, n     |         33.3, 50
#> Species [versicolor], %, n |         33.3, 50
#> Species [virginica], %, n  |         33.3, 50

^{Created on 2023-04-28 with reprex v2.0.2}

[strengejacke]

We may think about a different layout of n (extra column?), but I think none of the above examples were previously working, so this PR is actually fixing some bugs.

[strengejacke]

1. Issue: When grouping, and a variable had no values in one group (like "gender" -> "transgender" above), we had different numbers of rows as zero-rows were omitted.

2. Issue: n did not work for factors.

[rempsyc]

Cool, thank you @strengejacke!

I updated the snapshots but now am getting the following on the tests:

── Error ('test-report_performance.R:105:3'): report_performance Bayesian) ─────
Error in `eval(code, test_env)`: cannot change value of locked binding for 'stan_glmer'

That looks like a false positive because it didn't do that when you ran your tests and I didn't change anything but the snapshot, and it has nothing to do with the current changes. Should I just squash and merge the change and hope this error goes away by itself?

[rempsyc]

Is this the same as #367?

[codecov-commenter]

Codecov Report

Merging #368 (74d92fb) into main (43faa6c) will increase coverage by 4.41%.
The diff coverage is 73.68%.

❗ Current head 74d92fb differs from pull request most recent head 1470444. Consider uploading reports for the commit 1470444 to get more accurate results

❗ Your organization is not using the GitHub App Integration. As a result you may experience degraded service beginning May 15th. Please install the Github App Integration for your organization. Read more.

@@            Coverage Diff             @@
##             main     #368      +/-   ##
==========================================
+ Coverage   67.31%   71.72%   +4.41%     
==========================================
  Files          46       46              
  Lines        3246     3254       +8     
==========================================
+ Hits         2185     2334     +149     
+ Misses       1061      920     -141     

Impacted Files	Coverage Δ
R/report_sample.R	88.34% <73.68%> (-0.04%)	⬇️

... and 7 files with indirect coverage changes

[rempsyc]

Thanks for the PR @strengejacke ! :)