There are a few small pyarrow regressions related to fsspec filesystems (breaking some dask usage), so I would also be very much in favor of having a 2.0.1 patch release.
All relevant JIRAs should be tagged with a 2.0.1 version. The one other issue that I remember but which hasn't this 2.0.1 label is https://issues.apache.org/jira/browse/PARQUET-1935, in case we want to include it in the patch release, it might make sense to move it back to ARROW to be able to label it (or keep track of it in another way, I don't know how this is done for normal releases) Joris On Thu, 12 Nov 2020 at 19:20, Antoine Pitrou <anto...@python.org> wrote: > > > ARROW-10519 would also be nice to have in a bugfix release. > > Regards > > Antoine. > > > Le 11/11/2020 à 21:21, Micah Kornfield a écrit : > > There are a couple of Parquet bugs that I think might warrant a patch > > release. The most pressing I think is: ARROW-10493 which can potentially > > lose data silently depending on how batch parameters are used with nullable > > structs. > > > > I think this is serious enough that we should consider a patch release. > > > > Thoughts? > > > > -Micah > >
