This is really cool. I like the no dependency approach. Also having more clean room implementations will harden the spec.
On Sat, Feb 28, 2026 at 2:09 AM Julien Le Dem <[email protected]> wrote: > Nice, > I think a lot of people will be happy to have an implementation that is not > dependent on Hadoop. > It is easier to have a separate implementation rather than extracting > hadoop from parquet-java while trying to maintain backward compatibility. > The Hadoop Config, Codec and FileSystem get a bit in the way. > This issue should probably also try to share code for parts that are fairly > self contained? (like encodings, ...) > https://github.com/hardwood-hq/hardwood/issues/62 > We should invite Gunnar to come talk about it in the sync. > > On Fri, Feb 27, 2026 at 5:57 AM Rok Mihevc <[email protected]> wrote: > > > Hi all, > > > > I just noticed an announcement [1] of a new Parquet parser project [2] > and > > thought it might be useful to share it here. > > > > Best, > > Rok > > > > [1] https://www.morling.dev/blog/hardwood-new-parser-for-apache-parquet/ > > [2] https://github.com/hardwood-hq/hardwood > > >
