[ARROW-13615] [R] Bindings for stringr::str_to_sentence

• Type: Sub-task
• Status: Closed
• Priority: Major
• Resolution: Won't Fix
• Fix Version/s: 7.0.0
• Component/s: R
• Labels: good-first-issue kernel pull-request-available
• External issue URL: https://github.com/apache/arrow/issues/29256
• Assignee: Dragoș Moldovan-Grünfeld
• Reporter: Nicola Crane
• Created: 12/Aug/21 13:52
• Updated: 11/Jan/23 08:34
• Resolved: 21/Oct/21 12:39

Description

There is more to this issue than meets the eye. The stringr::str_to_sentence() does 2 things:

• capitalise the first word
• if there are multiple sentences provided as a single string, attempts to find sentence breaks and capitalise the first word of each sentence.

The stringr implementation wraps stringi::str_trans_totitle(), which in turns uses ICU’s BreakIterator to locate specific text boundaries. As a consequence stringr::str_to_title() is not able to identify a full stop / period (".") as a sentence end and does not capitalise words following it. Thus, there is a discrepancy between behaviour of the utf8_capitalize kernel (which capitalises the first word of a string without making any attempt to break into sentences) and the behaviour of stringr::str_to_sentence().

For more extensive discussions around the stringi / stringr implementation see stringr issues 202 and 231.

Due to the complexity of this issue and the relatively niche use cases, the recommendation is to postpone implementation.

Issue Links

• depends upon: ARROW-12944 [C++] String capitalize kernel
• links to: GitHub Pull Request #11484