[R] crash when reading with duckdb
- Dominant language
- C++
- Stars
- 17.1k
- Forks
- 4.3k
- Avg merge
- 3d 13h
- Merged PRs (30d)
- 88
Description
### Describe the bug, including details regarding any error messages, version, and platform.
Using arrow 15 in R through dplyr, I got a crash (R session is crashed) when reading a dataset with to_duckdb(). On my dataset, pretty large composed of some parquet files -- I can send a dropbox link if needed, there are 2 columns that provoque the crash. There are encoded as factors (dictionary) and I suspect that different dictionaries for the different parquet are the cause of the crash.
so `open_dataset(mydts) |> select(field1) |> to_duckdb() |> collect()` crash everything
`open_dataset(mydts) |> select(field1) |> mutate(field1 = as.character(field1) |> to_duckdb() |> collect()` works
and
`open_dataset(mydts) |> select(field1) |> collect()` works too.
I tried to rewrite the dataset loading it first (without duckdb) and then using write_dataset with no succes.
using arrow 15.0.1 and duckdb 0.10
### Component(s)
R
Contributor guide
Assessment
This issue has not been assessed yet.