Documentation needs data examples and not just code

[holgerbrandl]

It's hard to understand the docs because they present how to program the API but do not illustrate the result.

E.g. comparing https://dplyr.tidyverse.org/reference/summarise.html with https://kotlin.github.io/dataframe/groupby.html#aggregation it becomes clear (to me at least) that data examples are very helpful to understand a table processing API.

Expected: The docs should inline the results of the various API examples.

[koperagen]

Have a look at the updated https://kotlin.github.io/dataframe/groupby.html :)

[pacher]

I can finally see it after trying in Chrome, not visible in Firefox

[holgerbrandl]

Thank yoy, it certainly has improved.

However, it still falls short compared to https://dplyr.tidyverse.org/reference/group_by.html because of multiple reasons

• output and commands are not aligned (first commands are listed, then all results)
• output is hidden in hierarchical structure, forcing the user to click and click to get to the result
• examples are not reproducible: e.g. df is declared as val df = dataFrameOf(key, data) but a line later is used to group by city. Imho examples should use bundled example data sets with clear names (cars, flights, ... and never df) for clarity and reproducibility
• It's unclear to me why it's listing input, step 1 and output. No other data-science prompt would work in that way.