Jefffrey commented on issue #7303: URL: https://github.com/apache/arrow-datafusion/issues/7303#issuecomment-1786026813
> FWIW option 2 is what we are doing now and it works easily enough. But I think option 3 sounds pretty good to me. But that seems like it would basically just boil down to adding methods to `PhysicalExtensionCodec`? Yeah haven't given too much thought to the interop code yet, just mainly been considering the proto side, so unsure how the code would turn out. If you've already been implementing option 2, how have you been handling encoding/decoding `ParquetSink`s and `CsvSink`s? Since they contain `FileSinkConfig`s which contain `FileTypeWriterOptions` which then if you keep going, will inevitably contain `parquet::file::properties::WriterProperties` for parquet and `arrow::csv::WriterBuilder` for csv, both from `arrow-rs`. Do you just encode the options as a map of string to string? -- 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]
