[
https://issues.apache.org/jira/browse/ARROW-12711?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17585019#comment-17585019
]
Travis Lim commented on ARROW-12711:
------------------------------------
[~icook] Any updates on bindings for dplyr summarise with paste(collapse) or
str_c(collapse) in upcoming releases?
A potential workaround was floated for Python here
https://issues.apache.org/jira/browse/ARROW-12710 but having this in R would be
a game changer, especially for NLP applications :pray: :pray: :pray:
> [R] Bindings for paste(collapse), str_c(collapse), and str_flatten()
> --------------------------------------------------------------------
>
> Key: ARROW-12711
> URL: https://issues.apache.org/jira/browse/ARROW-12711
> Project: Apache Arrow
> Issue Type: New Feature
> Components: R
> Reporter: Ian Cook
> Priority: Major
> Labels: query-engine
>
> These are the aggregating versions of string concatenation—they combine
> values from a set of rows into a single value.
> The bindings for {{paste()}} and {{str_c()}} might be tricky to implement
> because when these functions are called with the {{coallapse}} argument
> unset, they do _not_ aggregate.
> In {{summarise()}} we need to be able to use scalar concatenation within
> aggregate concatenation, like this:
> {code:java}
> starwars %>%
> filter(!is.na(hair_color) & !is.na(eye_color)) %>%
> group_by(homeworld) %>%
> summarise(hair_and_eyes = paste0(paste0(hair_color, "-haired and ",
> eye_color, "-eyed"), collapse = ", ")){code}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
