dragosmg opened a new pull request #11417:
URL: https://github.com/apache/arrow/pull/11417
The approach I propose is simple - add an `abort()` if there is a
`try-error` in the evaluated RHS of the `case_when()`. This error is then
handled my `mutate()` which captures the call, forces a `collect()` and
re-evaluates the `case_when()` call as a regular {dplyr} call.
There are a couple of potential implications for this ☝🏻 approach:
* we are working with the assumption that `case_when()` is only used inside
`mutate()` and relying on `mutate()` to handle the error (via `abandon_ship()`).
* the error is not captured at `case_when()` level (which would be a bit
more informative), but rather at `mutate()` level (when it becomes a bit less
specific, i.e. one-step removed). The error message at `case_when()` level is
not surfaced.
``` r
record_batch(tbl) %>%
mutate(
cw = case_when(!(!(!(lgl))) ~ factor(chr),
TRUE ~ fct)
) %>%
collect()
Warning: Expression case_when(!(!(!(lgl))) ~ factor(chr), TRUE ~ fct) not
supported in Arrow; pulling data into R
# A tibble: 10 × 8
int dbl dbl2 lgl false chr fct cw
<int> <dbl> <dbl> <lgl> <lgl> <chr> <fct> <fct>
1 1 1.1 5 TRUE FALSE a a a
2 2 2.1 5 NA FALSE b b b
3 3 3.1 5 TRUE FALSE c c c
4 NA 4.1 5 FALSE FALSE d d d
5 5 5.1 5 TRUE FALSE e NA NA
6 6 6.1 5 NA FALSE NA NA NA
7 7 7.1 5 NA FALSE g g g
8 8 8.1 5 FALSE FALSE h h h
9 9 NA 5 FALSE FALSE i i i
10 10 10.1 5 NA FALSE j j j
```
<sup>Created on 2021-10-14 by the [reprex
package](https://reprex.tidyverse.org) (v2.0.1)</sup>
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
