paleolimbot commented on PR #41823: URL: https://github.com/apache/arrow/pull/41823#issuecomment-2157009669
Just a note that I was reminded of a possible use-case for this recently (reading Parquet files with a logical type that was defined in a later version of Parquet: https://github.com/apache/arrow/pull/41765 > Without some attempt to canonicalize well-known pairs of type and vendor, we could lose opportunities for interoperability. Perhaps another wording is that the vendor name is informational purposes or as an escape hatch that is more likely to enable a workaround (rather than some value an application should be relying on). For example, duckdb currently returns its extension "geometry" type as Arrow binary that many people mistakenly try to interpret as WKB. The real solution is for DuckDB to just return geoarrow extension types in this case; however, this proposal would let DuckDB return an `Other` extension type, which would enable a workaround (an R package could spot it and issue an informative error or do the required conversion until DuckDB can export its extension types to Arrow in this way). -- 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: github-unsubscr...@arrow.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
