You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Is your feature request related to a problem or challenge?
I would like to create the first release soon so that we can test out the release process and scripts.
I am not sure if we should stick with version 0.1.0 for this first release, or use the same version as the DataFusion version that we build against (like DataFusion Python and Ballista subprojects).
I would softly vote to stick with 0.1.0 initially to convey that significant public API changes are still possible, and increment the minor version until that is the case. But I struggled with this question myself and see it both ways.
I think it's good to match the python strategy eventually; my understanding of that is that it matches upstream datafusion's major/minor version but not necessarily the patch version, incrementing that as needed if there is a bindings-specific change.
Activity
andygrove commentedon May 25, 2026
@pgwhalen @LantaoJin any thoughts on version number?
pgwhalen commentedon May 25, 2026
I would softly vote to stick with 0.1.0 initially to convey that significant public API changes are still possible, and increment the minor version until that is the case. But I struggled with this question myself and see it both ways.
I think it's good to match the python strategy eventually; my understanding of that is that it matches upstream datafusion's major/minor version but not necessarily the patch version, incrementing that as needed if there is a bindings-specific change.
LantaoJin commentedon May 26, 2026
+1 for 0.1.0 now, and use the same version as the DataFusion version as long-term.