Thanks David. Indeed it seems that exposing IpcWriteOptions is going to be critical here. I'd like to avoid an "environment variable" workaround at the C++ level instead only providing such things in e.g. Python like we did for the alignment patch
On Mon, Jun 29, 2020 at 9:30 AM David Li <li.david...@gmail.com> wrote: > > This would cause compatibility issues for Flight servers/clients > between versions as well. The situation is a little worse since > IpcWriteOptions isn't exposed and so you can't control what version > you write. But just exposing them in lieu of a full negotiation (which > we should start thinking about) should be enough to work through this. > > I see there's https://issues.apache.org/jira/browse/ARROW-8190 so I'll > try to tackle this soon (and do the same for Java) since it should be > independent of whether the format change goes through. > > Best, > David > > On 6/28/20, Wes McKinney <wesmck...@gmail.com> wrote: > > I opened a PR https://github.com/apache/arrow/pull/7566 > > > > We should prioritize getting through the other format changes, but we > > can vote on this in the meantime if there is consensus > > > > On Fri, Jun 26, 2020 at 2:58 PM Micah Kornfield <emkornfi...@gmail.com> > > wrote: > >> > >> I agree I think we have to do this given the number of changes in flight > >> (especially union types). > >> > >> On Fri, Jun 26, 2020 at 7:29 AM Wes McKinney <wesmck...@gmail.com> wrote: > >> > >> > I created a JIRA about this > >> > > >> > https://issues.apache.org/jira/browse/ARROW-9231 > >> > > >> > This issue is quite important so please take a look. > >> > > >> > On Thu, Jun 25, 2020 at 8:53 AM Wes McKinney <wesmck...@gmail.com> > >> > wrote: > >> > > > >> > > On Thu, Jun 25, 2020 at 5:31 AM Antoine Pitrou <anto...@python.org> > >> > wrote: > >> > > > > >> > > > > >> > > > Le 25/06/2020 à 12:18, Antoine Pitrou a écrit : > >> > > > > > >> > > > > Le 25/06/2020 à 00:40, Wes McKinney a écrit : > >> > > > >> hi folks, > >> > > > >> > >> > > > >> This has come up in some other contexts, but I believe it would > >> > > > >> be a > >> > > > >> good idea to increment the version number in Schema.fbs starting > >> > with > >> > > > >> 1.0.0 to separate the pre-1.0 and post-1.0 worlds > >> > > > >> > >> > > > >> https://github.com/apache/arrow/blob/master/format/Schema.fbs#L22 > >> > > > >> > >> > > > >> Given that we are contemplating a number of changes to assist > >> > > > >> with > >> > > > >> forward compatibility and a breaking serialization change for > >> > unions, > >> > > > >> this would seem prudent so that we do not risk breaking > >> > compatibility > >> > > > >> with 0.17.1 and prior. > >> > > > >> > >> > > > >> Given that there are no major backwards incompatibilities, there > >> > > > >> should be no problem with 1.0.0 readers reading data generated > >> > > > >> by > >> > > > >> libraries <= 0.17.1. > >> > > > > > >> > > > > Actually, it seems that a dense array with top-level null values > >> > > > > (represented in 0.17.1 fashion) would need non-trivial rewriting > >> > > > > of > >> > its > >> > > > > offsets and child arrays (at least one child array) to represent > >> > > > > the > >> > > > > nulls at the child level. > >> > > > > > >> > > > > This is unless we keep the top-level union null bitmap in C++ and > >> > only > >> > > > > avoid emitting it on the IPC side. Which would be a slightly > >> > > > > weird > >> > > > > arrangement, but would limit incompatibilites on the C++ API > >> > > > > side. > >> > > > > >> > > > Actually, if we do this, the same problem will appear on the IPC > >> > > > write > >> > > > side (C++-created dense union arrays with a top-level null bitmap > >> > > > will > >> > > > need regenerating some of the child buffers). > >> > > > >> > > I see. Well I think we can shut down this issue by giving up on Union > >> > > forward compatibility V4 / pre-1.0 libraries. > >> > > > >> > > > Regards > >> > > > > >> > > > Antoine. > >> > > >
