+1 for what it's worth. Although, it would be nice to also officially make fixed promotable strings as well. Allowing old UUID string schema to read newer UUID encodings.
Best, Jack On Mon, Jan 22, 2024 at 1:40 PM Ryan Blue <[email protected]> wrote: > +1 for storing as 16-byte fixed. > > On Mon, Jan 22, 2024 at 1:44 AM Martin Grigorov <[email protected]> > wrote: > > > I just want to add that it would be nice to add an interop test, e.g. an > > Avro file written by one SDK should be readable by the other SDKs. > > > > On Fri, Jan 19, 2024 at 11:58 AM Driesprong, Fokko <[email protected] > > > > wrote: > > > > > Hey everyone, > > > > > > Last December I proposed support for encoding UUIDs as fixed[16] > > > <https://lists.apache.org/thread/h4lvc2lk194om6cjm76wzosm22cop2xw>, as > > > this > > > has several advantages > > > < > > > > > > https://docs.google.com/document/d/16_oSWrEM7AFUCTe0uuraAEHxywezLfoEz5ahzwvhGUk/edit#heading=h.43xuauwfk7ow > > > >. > > > I know that this was over the holidays, so I would like to revisit > this. > > > Let me know if there are any questions or concerns, or feel free to > jump > > in > > > on the PR <https://github.com/apache/avro/pull/2672>. There are also > > > implementations already for Java > > > <https://github.com/apache/avro/pull/2652> (thanks > > > Christophe!), and one for Rust (thanks Martin!). If there are no > > concerns, > > > I would like to move this forward next week, and also see if we can > start > > > the release process. > > > > > > Kind regards, > > > Fokko > > > > > > > > -- > Ryan Blue > Tabular >
