Thanks for the excitement and reviews! I think the current proof-of-concept implementations demonstrate that the concepts we've included in the spec don't introduce undue dependencies or complexity into existing implementations (i.e., they are enough for me personally to be comfortable with the language in the format PR); however, I'm new here and it is a great point that the POC PRs need to be polished.
As a basis for adding files to parquet-testing, I would propose using a subset of files from the geoarrow-data repository [1], which contains examples of files of all geometry types, CRSes, and a few other concepts that we've found important when testing GeoParquet and GeoArrow implementations. I'm happy to add those. [1] https://github.com/geoarrow/geoarrow-data?tab=readme-ov-file#geoarrow-data On Thu, Feb 6, 2025 at 2:15 PM Micah Kornfield <[email protected]> wrote: > I'm really excited to see this come together but I think this vote might be > a little bit premature. > > 1. It looks like both sample implementations all have requested changes > (maybe reviewers just forgot to do LGTM but at least C++ there are at least > some style issues which probably aren't blocking per se). > 2. Following up on a different thread, I'm not clear if we've added parquet > files to parquet-testing to verify interop? > > Thanks, > Micah > > On Thu, Feb 6, 2025 at 11:40 AM Daniel Weeks <[email protected]> wrote: > > > +1 (binding) > > > > > > > > On Thu, Feb 6, 2025 at 11:15 AM Russell Spitzer < > [email protected] > > > > > wrote: > > > > > Ha sorry, wrong ML (non binding) > > > > > > On Thu, Feb 6, 2025 at 8:14 PM Russell Spitzer < > > [email protected]> > > > wrote: > > > > > > > +1 (binding) > > > > > > > > On Thu, Feb 6, 2025 at 6:52 PM Szehon Ho <[email protected]> > > > wrote: > > > > > > > >> +1 (non-binding). > > > >> > > > >> Thanks for driving this Gang! This has been a huge collaboration, > > > across > > > >> Parquet, Iceberg, Geo-Parquet, Sedona communities and many > companies, > > > and > > > >> the Iceberg proposal would have not been able to make much progress > > > >> without > > > >> your help! I am excited to see this get closer to delivering great > > > value > > > >> for geospatial use cases. > > > >> > > > >> Szehon > > > >> > > > >> On Thu, Feb 6, 2025 at 8:50 AM [email protected] <[email protected]> > > > wrote: > > > >> > > > >> > +1 (binding) > > > >> > > > > >> > Thanks to everyone that worked on getting this update done! It's > > been > > > an > > > >> > amazing amount of discussion and I'm excited to see it ready to > go. > > > >> > > > > >> > On Thu, Feb 6, 2025 at 8:11 AM Jia Yu <[email protected]> wrote: > > > >> > > > > >> > > +1 (non-binding) > > > >> > > > > > >> > > I’m really looking forward to this! It’s going to be a fantastic > > > >> addition > > > >> > > to the community. > > > >> > > > > > >> > > On 2025/02/06 15:54:49 Gang Wu wrote: > > > >> > > > Hi, > > > >> > > > > > > >> > > > We would like to add geometry and geography logical types to > > > >> Parquet: > > > >> > > > > > > >> > > > https://github.com/apache/parquet-format/pull/240 > > > >> > > > > > > >> > > > This is proposed together with Apache Iceberg to support > > > geospatial > > > >> > data: > > > >> > > > > > > >> > > > https://github.com/apache/iceberg/pull/10981 > > > >> > > > > > > >> > > > Two prototype implementations are available: > > > >> > > > > > > >> > > > 1. Parquet Java: > > https://github.com/apache/parquet-java/pull/2971 > > > >> > > > 2. Parquet C++: https://github.com/apache/arrow/pull/43977 > > > >> > > > > > > >> > > > This vote will be open for at least 72 hours. > > > >> > > > > > > >> > > > [ ] +1 Add these types to the format specification > > > >> > > > [ ] +0 > > > >> > > > [ ] -1 Do not add these types to the format specification > > > because... > > > >> > > > > > > >> > > > Thanks, > > > >> > > > Gang > > > >> > > > > > > >> > > > > > >> > > > > >> > > > > > > > > > >
