[ARROW-13499] [R] Aggregation on expression doesn't NSE correctly

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 6.0.0
• Component/s: R
• Labels: pull-request-available query-engine
• External issue URL: https://github.com/apache/arrow/issues/29158
• Assignee: Neal Richardson
• Reporter: Neal Richardson
• Created: 29/Jul/21 20:00
• Updated: 11/Jan/23 08:33
• Resolved: 24/Aug/21 16:36

Description

  expect_dplyr_equal(
    input %>%
      mutate(has_words = nchar(verses) < 0) %>%
      group_by(some_grouping) %>%
      summarize(all(has_words)) %>%
      arrange(some_grouping) %>%
      collect(),
    tbl
  )

works, but if you define the expression inline, it doesn't:

  expect_dplyr_equal(
    input %>%
      group_by(some_grouping) %>%
      summarize(has_words = all(nchar(verses) < 0)) %>%
      arrange(some_grouping) %>%
      collect(),
    tbl
  )

The error you get I think means that base::nchar is being called. I'm not sure if this is a general issue with our nse_funcs calling other nse_funcs, or if this is about the way the aggregation function environment is constructed separately, but it's not good.

Issue Links

• blocks: ARROW-13618 [R] Use Arrow engine for summarize() by default
• relates to: ARROW-13344 [R] Initial bindings for ExecPlan/ExecNode
• links to: GitHub Pull Request #10938