C-Loftus opened a new pull request, #960:
URL: https://github.com/apache/arrow-go/pull/960

   ### Rationale for this change
   
   Any downstream arrow clients like iceberg-go which write geo data to parquet 
end up being labeled in the underlying parquet as bytes / binary blobs and not 
the proper geometry type. This means they are not able to be queried by tools 
like duckdb, etc. and will result in issues like this 
https://github.com/duckdb/duckdb-iceberg/issues/1175
   
   arrow-go needs to implement the logical type interface for both geometry and 
geography to ensure the parquet write path gets the proper metadata.
   
   ### What changes are included in this PR?
   
   Implementations of the logical type interface for both Geometry and Geography
   
   ### Are these changes tested?
   
   Yes
   
   ### Are there any user-facing changes?
   
   Yes. Geometry should now be labelled as such in written parquet.
   
   Note, downstream users like `iceberg-go` are using 
[geoarrow-go](https://github.com/geoarrow/geoarrow-go) for some of the 
arrow-related geospatial metadata imports (i.e. edge interpolation algorithms, 
type constructors, etc.) Should someone in that repo simply implement the 
proper interfaces and then this package would import it directly as a new 
dependency? Since there are geo types in parquet natively now I was wondering 
if it maybe makes sense to pull into the main repo. ccing geoarrow-go creator 
@dwilson1988 on this
   
   Happy to defer to maintainers' preference. 


-- 
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]

Reply via email to