[ENH] From "with_groups()" to "by" within verbs....

[coforfe]

Feature Type

• Adding new functionality to datar

• Changing existing functionality in datar

• Removing existing functionality in datar

Problem Description

Hi,

The latest versions of dplyr includes an important design change when grouping.

• Instead of using a group_by() before an operation (mutate, summarize, etc) and using an ungroup() after completing it.
• Now the grouping will be another parameter in the verbs like mutate(dataframe, newvar_definition, by = group).
• This approach simplifies the code significatively.

For your reference Hadley Wikham has recognized recently that using group_by() was a design mistake:
https://news.ycombinator.com/item?id=30067406

Feature Description

So the proposal would be to align datar functionality to dplyr by modifying the verbs (mutate, slice, summarize) in order to include a new parameter by where one can specify the variable/s to use for the grouping.

This new functionality would substitute the current with_groups() and ungroup() functions.

Thank!
Carlos.

Additional Context

No response

[pwwang]

This feature is still experimental with dplyr.

See:

https://dplyr.tidyverse.org/reference/mutate.html
https://dplyr.tidyverse.org/reference/summarise.html
https://dplyr.tidyverse.org/reference/slice.html

We will introduce it until it's stabilized.

[coforfe]

OK. Good!.