Hello, There has been some discussion in Arrow DataFusion Python about changing our release versioning to more closely match Arrow DataFusion itself [1]. The reasoning behind this suggestion is two fold.
1. Since the versions do not align it is unclear to the end user what version of datafusion they are actually using under the covers. 2. Most, if not all, of the code in datafusion-python is tightly coupled to datafusion itself and having those versions in a more lock and step approach makes development more straight forward. However changing versioning is something I felt merited feedback from the community so hence this email. Does anyone have any concerns with the versioning approach laid out in the github issue [1]? Thanks, Jeremy Dyer [1] - https://github.com/apache/arrow-datafusion-python/issues/123
