toString of grouped data-frames should include variables, number of groups, and number of records

[holgerbrandl]

In R's toString implementation of grouped data-frames, the grouping attributes are displayed along with the complete table structure providing a much more informative view:

> iris %>% group_by(Species)
# A tibble: 150 x 5
# Groups:   Species [3]
   Sepal.Length Sepal.Width Petal.Length Petal.Width Species
          <dbl>       <dbl>        <dbl>       <dbl> <fct>
 1          5.1         3.5          1.4         0.2 setosa
 2          4.9         3            1.4         0.2 setosa

In comparison, with kdf we can hardly imagine what's behind a grouped data-frame (making subsequent aggregations and transformation harder to work out interactively)

> irisData.groupBy("Species").print()
      Species    group
 0     setosa [50 x 5]
 1 versicolor [50 x 5]
 2  virginica [50 x 5]

Expected: A grouped-df should render the same as an ungrouped table, but with additional information about the grouping attributes and the number of groups.

See https://dplyr.tidyverse.org/reference/group_by.html#ref-examples

[nikitinas]

I suggest to implement it in a more general way.

There is a similar problem with printing ColumnGroups: currently only top-level columns of DataFrame are printed. So, let's add an argument depth into print() function with default value 1. It will control the level of ColumnGroup and FrameColumn expansion in print. And group column of grouped dataframe will be expanded by default.

We can later provide a more precise API to specify particular columns to be expanded in print, but in simple cases depth argument will be enough.