paleolimbot commented on issue #45438: URL: https://github.com/apache/arrow/issues/45438#issuecomment-2675298553
> I think one of the challenges with this is that the schema isn't necessarily fixed for each geometry extension type. There's a place for this in Acero, but it looks like it's not implemented for extension types: https://github.com/apache/arrow/blob/12cdaaafb7a5ba39db81ba6d7e565b8e7a24117e/cpp/src/arrow/compute/kernel.h#L103-L107 . (This is also a problem in DataFusion, where metadata is considered when matching things and extension types are hard to incorporate. I am not quite sure where the arrow package fits in with what you're trying to do. I wonder if you can "just" have R functions that are liberal about what they accept (arrow, sf, whatever) and output something you have control over (S3 wrapper around a Rust pointer). You can implement `as_arrow_array()` and/or `as_nanoarrow_array()` on those. (This is vaguely how Kyle structured geoarrow for Python: use `__arrow_c_array/stream__` for inputs, output something he owns that implements the requisite protocols to be useful. -- 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]
