Following ARROW-13344, we should handle the case in which the user calls summarise() but does not call any aggregate functions in it.

Currently this is unhandled and results in errors (which are then are converted to warnings) like:

Warning: Error in x[c("fun", "options")] : object of type 'environment' is not subsettable ; pulling data into R

or when the expression in summarise() has only scalar literals:

Warning: Error : $ operator is invalid for atomic vectors
; pulling data into R 

or when there are zero arguments passed to summarise() and there is no group_by() preceding it:

Warning: Error : Invalid input type, expected 'character' actual 'NULL'
; pulling data into R 

We should aim for consistency with dplyr::summarise() in these cases, but if that's not practical then we should detect this case and throw a more informative error.

Reporter: Ian Cook / @ianmcook
Assignee: Neal Richardson / @nealrichardson

Related issues:

• [R] Use Arrow engine for summarize() by default (blocks)
• [R] Initial bindings for ExecPlan/ExecNode (relates to)
• [R] Support for dplyr::distinct() (is depended upon by)

PRs and other links:

• GitHub Pull Request #11078

_{Note: This issue was originally created as ARROW-13543. Please see the migration documentation for further details.}