Dewey Dunnington created ARROW-16517:
----------------------------------------
Summary: [R] as_arrow_table() and as_record_batch() do not
preserve row count for empty data.frame
Key: ARROW-16517
URL: https://issues.apache.org/jira/browse/ARROW-16517
Project: Apache Arrow
Issue Type: Improvement
Components: R
Reporter: Dewey Dunnington
It's definitely an edge case, but preserving the row-count on zero-column
tables/record batches is helpful when programming on those objects (I ran into
this when working with GeoParquet tables, where I separate the geometry and
non-geometry columns and then recombine. The functionality is there but it
looks like it's not implemented for the data.frame case.
{code:R}
arrow::as_arrow_table(tibble::tibble(.rows = 5))
#> Table
#> 0 rows x 0 columns
arrow::record_batch(tibble::tibble(.rows = 5))
#> RecordBatch
#> 0 rows x 0 columns
# workaround
arrow::arrow_table(x = 1:5)[integer()]
#> Table
#> 5 rows x 0 columns
{code}
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
