Kirill Müller

Thanks for the benchmark! This looks like a case where we don't support particular semantics but fall back to R instead. What happens if you change the code to `sum()`...

Thanks for the follow-up! Supporting `sum(na.rm = TRUE)` is definitely on our agenda.

@jrosell: Please open a new issue for questions like this.

Thanks. Code like `file_paths %>% map(duckplyr_df_from_csv) %>% bind_rows()` has worked for me in practice, but I agree that this should be streamlined. Would you like to contribute a PR?

`bind_rows()` reads into memory, `%>% reduce(union_all)` is better but will also read into memory in duckplyr 0.4.0 (works better in duckplyr 0.3.0): https://github.com/tidyverse/dplyr/pull/7049 . What *should* work is `duckplyr_df_from_csv("file_*.csv")`, but...

The action items here are a subset of those in https://github.com/duckdblabs/duckplyr/issues/181#issuecomment-2215275517, let's move the discussion there.

Thanks, interesting idea. The translations currently aren't very well organized (a lot happens in `R/relational.R`). Let me think about how to open this for extension.

Until then, would you like to share examples of fallback situations that would help you most (perhaps one issue per example)?

We now at least have a list of functions that _are_ translated: https://github.com/duckdblabs/duckplyr/pull/179/files#diff-a202cfba76540d6822868ac7755edd4945b6344057d78e0092f4836e33c0d4eaR11 .

Because we will never be able to translate everything, a way to provide custom translations can help bridge the gap in some cases. For now, clone the repository and follow...