paleolimbot commented on pull request #11133:
URL: https://github.com/apache/arrow/pull/11133#issuecomment-917306807
Fixed the base with length != 1 issue, although I think my error message
could be better since base with length != 1 *is* supported, just not as a
non-expression. Reprex summary of new functionality:
``` r
library(arrow)
library(dplyr, warn.conflicts = FALSE)
# with base as a column
RecordBatch$create(a = 2:5, b = 2:5) %>%
mutate(log(a, base = b)) %>%
collect()
#> # A tibble: 4 × 3
#> a b `log(a, base = b)`
#> <int> <int> <dbl>
#> 1 2 2 1
#> 2 3 3 1
#> 3 4 4 1
#> 4 5 5 1
# with base as a scalar
RecordBatch$create(a = 2:5) %>%
mutate(log(a, base = 3)) %>%
collect()
#> # A tibble: 4 × 2
#> a `log(a, base = 3)`
#> <int> <dbl>
#> 1 2 0.631
#> 2 3 1
#> 3 4 1.26
#> 4 5 1.46
# errors for base with length != 1
RecordBatch$create(a = 2:5) %>%
mutate(log(a, base = 3:4)) %>%
collect()
#> Warning: In log(a, base = 3:4), base with length != 1 not supported by
Arrow;
#> pulling data into R
#> # A tibble: 4 × 2
#> a `log(a, base = 3:4)`
#> <int> <dbl>
#> 1 2 0.631
#> 2 3 0.792
#> 3 4 1.26
#> 4 5 1.16
```
Should `x` as a scalar be supported? It currently triggers bringing the data
in to R but could theoretically not:
``` r
# errors
RecordBatch$create(a = 2:5) %>%
mutate(a * log(2)) %>%
collect()
#> Warning: Expression a * log(2) not supported in Arrow; pulling data into R
#> # A tibble: 4 × 2
#> a `a * log(2)`
#> <int> <dbl>
#> 1 2 1.39
#> 2 3 2.08
#> 3 4 2.77
#> 4 5 3.47
# works!
RecordBatch$create(a = 2:5) %>%
mutate(a * log(Expression$scalar(2))) %>%
collect()
#> # A tibble: 4 × 2
#> a `a * log(Expression$scalar(2))`
#> <int> <dbl>
#> 1 2 1.39
#> 2 3 2.08
#> 3 4 2.77
#> 4 5 3.47
```
--
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]
